Network design is a complex and time consuming affair with many steps and processes. However from a high level it could be considered that there are 4 main steps in the planning cycle.
The process begins with information gathering and objective setting. Information gathered at this stage will include both marketing and technical data. The marketing information is important so that realistic objectives can be set. Technical data will include information about the technology to be used, spectrum related data and possibly equipment performance data from a vendor. Phase 1
Information Gathering + Initial Objective Setting
Site Selection + Backhaul Planning
RF Predictions + Confirm Assumptions
Build Plan + Drive Test Optimisation
Figure 1 – High Level Design Life Cycle
Information gather during this rst phase is used to test the objectives and determine the viability of the business case. Since there are no major investments at this stage it is also a good time to analyse the risks involved using known information. The assumptions and objectives can be tested iteratively until some initial design is decided.
The second phase used the outputs of phase one to determine the best location for the base sites and to determine the back haul requirements. Issues of co-location and new site builds would be addressed at this stage.
Once all the site locations have been determined the initial assumptions regarding coverage will need to be validated. This is possible through the use of software RF planning tools. Some design optimisations can be determined during this stage. Choice of software tools and models will have to be made, this is often a matter of scale and budget.
Phase 4 is the build out of the system. Some starting point must be determined, possibly from the demographic information from the marketing team or from site availability. At some point during this stage drive tests should be carried out to conrm the accuracy of the software planning models used in stage three and if necessary some redesign and optimisations can be made. The use of additional software tools to plan the deployment may be used at this stage.
Phase 1 Detailed Procedure
As stated above phase 1 is the information gathering and objective setting stage. The more information that can be gathered and tested at this stage, the better the understanding of the design and the behaviour of the system when variables are included. Some of the additional steps that need to be considered in the early stages of planning are;
Gather relevant technical and marketing information
Set primary objectives based in some initial assumptions, type of service, coverage, capacity etc
Draft initial plan based on objectives and other assumptions, equipment selection, technology selection
Determine the number of base station required, through simple modelling techniques to full the initial objectives
Test the performance of the initial design based on market assumption variability
Test the business case based on market variability and equipment performance
Iterate the results and make necessary changes to basic plan.
Figure 2 – Information Required for Phase 1 Planning
Phase 1 Information
Phase 1 of planning is primarily about information gathering and initial system modelling, the more information that can be gathered at his stage will allow for more detailed and accurate modelling. More time spent at this at this stage understanding how the system responds to changes in design inputs should result in more solid and reliable design in the later stages. The basic premise of phase one design is to determine the optimum number of base stations to meet the required objectives of coverage and capacity.
A planning process can also be considered at this time taking into account what tools are available to the designer, RF planning tools, spreadsheets used to determine system operating criteria etc.
Phase 2 Detailed Procedures
The output of phase 1 is, amongst others, is the number of base stations required to meet the objectives, however the location of the base stations is yet to be determined. Phase 2 is about site selection and conrming the assumptions from the rst stage holds true against the real location of sites.
Many operators will have existing sites on which they may co-locate the new LTE equipment., however one of the implications of mobile broadband is the number of new sites that may have to be deployed (depending on the spectrum used). This will involved detailed site planning and acquisition to be carried out.
In addition the backhaul requirements for both the co-located sites and new sites will have to be calculated and planned.
Introduce real site location including existing and new sites
Test system performance using real location against initial objectives
Site Acquisition - Planning processes - Site Availability - Owned or Leased - Cost
Figure 3 – Phase 2 Information Required
Phase 3 Detailed Procedures
Once the site locations have been established, software tools can be used to conrm the coverage and capacity assumptions made in the rst stage. Changes can be made to the initial design at this stage as well the selection of ideal locations for new sites. It is important at this stage to develop a build out plan that will quickly establish the required coverage and capacity in the least amount of time with the least amount of cost, there are software tools available that can develop this plan.
Use software tools to conrm initial assumptions for coverage and capacity
Before a major build is undertaken the accuracy of the software tools must be determined, therefore it is not uncommon to run drive test against a test site, this can be used to conrm the coverage predicted by the RF tools and if the site is fully functional some estimate of cell capacity can also be determined. Any major discrepancy between the RF prediction and the actual measurements can be used to tune the prediction models. Tuning of the software models is important in order to reduce the amount of retro planning/ site building further in to the build process.
Drive test to conrm the software planning models used
Optimise radio plan if necessary
Phase 3 - 4 Information
Phase 3 and 4 are primarily about site selection and building, where the use of RF software planning, capacity planning tools and optimisation tools are heavily used. The selection of tools is based on the type of system that is being planned and the budget given to the planning department. There are many different stand-alone tools that ca be used in the process and an increasing number of integrated tools that will allow the planner to manage the design process from start to nish.
Typical tools required during the third and fourth stages are:
Whilst LTE technology is new and complex some of the basic rules of system planning do not change. Much of the complexity of LTE is designed to make the best use of the available spectrum, better spectral efciency, in other words. Achieving better efciency means that higher data rates can be achieved in systems that are spectrum limited. Indeed LTE is design to support a single channel reuse pattern with out resorting to tricks like spread spectrum.
When considering capacity planning, or general system planning, these are some of the factors that should be taken in to account.
Each of the factors mentioned above will have some impact on the overall system design and the ultimate capacity in each cell and across the system as a whole. Frequency Band
There are many frequency bands potentially available for the deployment of LTE, the bands listed opposite have been identied through work done by the ITU and the WRCs. The bands are part of the IMT spectrum and many are in use already with cellular technologies like GSM, UMTS and WiMAX.
It is not expected for a UE to support all of the bands shown here, but is highly likely that UE will support a sunset of the bands depending on the intended are of deployment, allowing national and international roaming as cost effectively as possible.
Figure 5 – FDD IMT Frequency Bands
The chosen spectrum will have a very large impact on the planning process since the nominal radius of the LTE radio cell is dependant on the frequency of operation.
Generally speaking the lower the frequency the larger the radio cell, the better the building penetration, the less sensitive to atmospheric issues the system becomes. This is of great interest to operators since the cost of deploying LTE networks is likely to be very high, lower frequency allocations can save many millions of dollars in CAPEX, i.e. there will be less eNBs to buy.
e.g. The US operator Verizon is deploying its LTE network in the 700MHz band (band 13) whilst DoCoMo in Japan have won spectrum in the 1500MHz band. A band of interest for many European operators is the 2.6GHz band.
Figure 6 – TDD IMT Frequency Bands
Allocated Spectrum and Channel Bandwidth
The bands are regulated in terms of the allowed operating bandwidth. This is driven largely by the amount of available spectrum in each of the bands. Some of the bands do not allow the use of the narrow channels, whilst others prohibit the use of the larger bandwidths.
The amount of allocated spectrum will impact the overall network capacity and the individual sector capacity. As with many aspects of system planning more is better. Planning a system with 1 or 2 channels is very challenging, even when the technology provides some complex mechanisms to allow for reuse factors of 1, there will still be a negative impact on capacity.
In some cases the operator may have the exibility to choose the channel bandwidth depending on the total amount of spectrum they have. Some analysis may have to be done on the advantages and disadvantages of a few large bandwidth channels (e.g. 2x10MHz) versus more, lower bandwidth channels (e.g. 4x5MHz)
Figure 7 – Available Capacity and Channel Bandwidths for LTE
LTE Channel Parameters
Once the individual channel bandwidths are know, it is possible to work out what the likely capacity of the channel will be. This is less straight forward in LTE for many reasons, not least of which is the nature of the OFDM technique employed on the radio interface.
The table opposite shows the main attributes of the various channel bandwidths. It can be
seen that the entire channel is not occupied due to the FFT sampling of the channel, this will yield a lower than expected capacity using the Nyquist and Shannon assumptions
Figure 8 – LTE Channel Parameters
Maximum Bit Rate per Channel
Based on a simple Nyquist calculation and an assumption of the overall efciency (80%) of the radio, the table opposite shows the maximum data rates that could be expected from the various channel bandwidths.
Figure 9 – Maximum Downlink Capacity per Radio Channel
However the actual cell capacity in LTE may vary due to considerations of serving cell load and adjacent cell load and also the interference coordination feature of LTE.
Figure 10 – Maximum Uplink Capacity per Radio Channel
System performance will be affected by many factor related to the equipment used in the network. The fundamental aspects of the link budget rely entirely on the performance of the equipment. In many case the vendor spec sheet will provide the majority of the information required to perform basic ink budgets. This may be enough during the initial phase of planning to establish a baseline for capacity and performance. Once the basic performance parameters have been worked out and certain levels of performance have been determined, it is then possible to include the more complex features of the equipment to determine the additional gains possible. For example MIMO, beamforming antennas, vendor specic algorithms for interference management.
• BS/UE Power Output • BS/UE Antenna Gains • Receiver sensitivity • Link Budget Gains and Losses • MIMO Gains • Vendor Specic Requirements
Figure 11 – Equipment Parameters Considered for Capacity
Coverage limited design Coverage limited systems are those whose performance is limited by the coverage possible from a given set of performance attributes. The system design for coverage will maximise the range from the base station at the expense of capacity. Coverage limited systems will likely have a few widely spaced base stations.
Capacity Limited Design A system that is limited by its capacity will deliver maximum capacity for a given set of conditions. Capacity will be delivered at the expense of coverage. Systems designed for capacity will have many closely spaced base stations.
Having established the performance capabilities of LTE and the vendor specic equipment the job of planning must then determine the capacity or coverage objectives. The objectives will of course vary from area to area depending on the planning criteria.
Alternating electrical current passing through a conductor causes an electromagnetic eld to be produced in the air around the conductor. This electromagnetic eld will also alternate in turn with the current that generates it. If the frequency of the current is sufciently high the electric and magnet elds will propagate away from the conductor at the speed of light.
Figure 14 – A Representation of Electric and Magnetic Fields
It should be noted the electric (E) and magnetic (H) elds are perpendicular to each other. The orientation of the electric eld is used to determine the polarisation of the transmitted energy, it is also used to describe the orientation of the antenna that transmits the signal, a vertically oriented antenna will transmit a vertically polarised electromagnetic signal.
Figure 15 – Polarisation is determined by the Angle of the Electric Field
The radio frequency signal has that property that it will propagate away from the transmitting element making it suitable to act as a carrier of information.
Early systems of radio transmission made used of very simple information systems, simply switching the transmitter on and off the send information, Morse code maybe the best know example of this kind of transmission system.
However today we have much more complex signals that we wish to transmit, voice, video, high speed broadband information, the information that represents the data that we wish to transmit is known as the baseband information.
The diagram below shows an analogue representation of the speech band, human speech happens to be very wide, up to 20KHz, however we choose not to transmit all of the information since our brains are able to understand what is being said with much less information in the signal. This is also convenient for transmission systems since the amount of information they can typically carry is limited. In voice based transmission systems, wired or wireless the amount of speech information that is transmitted is normally limited to only 3.1KHz of the total amount of information.
Figure 16 – A Comparison of Audio Signal Bandwidths
The diagram below represents the speech information in the time domain, showing how the amplitude of the information varies with time.
Figure 17 – An Analogue Signal Shown in the Time Domain
This diagram shows the same information but now the amplitude is shown against the frequency domain, it is possible to see from this kind of spectral analysis the bandwidth of the voice signal and the nature of the individual frequency components.
Figure 18 – An Analogue Signal Shown in the Frequency Domain
In today’s communication systems it is more common to convert the analogue information (shown above) into digital signals. The diagram below shows the time domain representation of a digital signal. This signal is simply an ON/OFF wave form, real digital systems would have much more complex waves, however it is a good starting point to describe the way in which digital system attributes can be described.
Figure 19 – A Time Domain Representation of a Square Wave
The same information from above can be shown in the frequency domain. From the signals shown below it is possible to see that the simple square waveform has signal components at the fundamental frequency of the wave form and then odd harmonic components. This is a simplied description of a much more complex theory in communication known as the Fourier Transform.
Figure 20 – A Frequency Domain Representation of a Square Wave
In fact Fourier stated that any complex wave form can be described by the sum of a series of sinusoidal components. The diagram below again illustrates the principle of the simple square wave built from sinusoidal wave forms.
Figure 21 – Showing the Addition of Fundamental and Harmonic Components
In general it can be said that the Decibel (dB) is another way of representing factors or absolute values, it turns out to be a very convenient way to represent very small or very large numbers, and consider them on a reasonable scale.
To dene the decibel we should rst look at the way in we represent the numbers associated with the logarithms
Figure 22 – Dening the Base and Index of a Number
When considering the product of two number that are raised to the power of some index, m and n in this case, the indexes can be added or subtracted as shown below.
Figure 23 – Showing the Addition and Subtraction of Number Indexes
From the statement below it can be understood that the value 10 raised to the index x will yield the value N, and that the logarithm to the base 10 of N will yield the value x .
This can be seen in the following numerical example.
The vaules above however are simply logarithms, the decibel refers much more specicaly to factors and absolute values.
The example below shows the ratio of two values P1 and P2. If P1 = 10 and P2 = 5 then the linear value would be 5 , the logartihm i.e. log10 (P1/P2) would be 0.7.
However the answer in dB requires a mutilication by 10 there for the ratio of 10 and 2 is 7dB. The answer in this case is a simple factor, and can be used to describe the gain or loss of ampliers, components, pathloss etc.
Figure 24 – Finding the Total Gain of a System
In some case it is necessary to describe absolute values in dB therefore the value in question must be referenced against some know value. For measurements of power the reference value of 1mW is often used. The following expression can be used to convert from linear Watts to dBm.
In the example system shown below, each component has a value of performance expressed as a gure of gain in dB, to establish the total performance of the combined components we can simply add the gures together.
Figure 26 – Gain and Loss Expressed as dB can be Added and Subtracted
It should be noted that dB values that expressed absolute level of power or ratios cannot be added in this way, the gures must converted back in to linear values before the addition is made.
The table below shows some commonly used dB values and their linear conversions.
Figure 27 – Table of Typical Values and their Conversions
Calculating Noise in RF systems
Thermal noise is the wideband electromagnetic radiation that is emitted from all objects, the cosmos, the stars, the earth and the conducting components that comprise a radio system. Noise is something that is inevitable in radio systems and cannot be completely eliminated. However its possible quantify the noise and to design system that will still work satisfactorily despite the noise.
The expression below determines the amount of noise present in a radio channel of a dened bandwidth. The constant k and temperature T are often taken together to be a constant value of -174dBm/Hz, this amounts to -174dBm of noise power present in one hertz of radio bandwidth, it follows therefore that the total amount of noise present will be proportional to the to actual bandwidth of the channel. (This is covered in more detail in Section 3)
System components congured in series or cascade will contribute to the overall noise present in any radio system. The diagram below illustrates the principle. If we could measure the signal to noise ratio (SNR) at the input and output of a system, represented by the box in the middle, then the total noise contribution is the difference of the SNR dB at the input and output. This gure is often expressed as the Noise Figure (NFdB) of the system.
Figure 28 – Calculating Noise Figure from Signal to Noise Ratios
Where there are multiple components in the receiver system, such as feeders, lters, ampliers, each component will contribute noise to the total NF of the system. However the noise gure of the total system cannot be better than the noise gure of the rst component. Also the gain of the rst stage will impact the noise seen in the subsequent stages of the system, thus a cascade calculation must be carried out to determine the total noise in the system, this concept is outlined in the diagram below and is covered in more detail in section 3.
Figure 29 – Noise in Cascaded Systems
Noise in radio systems will also be affected by the ambient noise level generated from man made sources, such as street lighting, car ignition systems, electricity distribution. It follows that urban areas will exhibit more noise than rural areas given the greater density of electrical systems. This noise may need to be considered as a margin when planning mobile radio systems, however radio systems operating above 1GHz or so are less affected by this source of noise.
As suggested earlier in this section there are two types of signal in radio systems, the carrier and the baseband information. The process of modifying the radio frequency carrier to represent or carry the baseband data is known as modulation.
The diagram below shows the 3 principle methods used by digital modulation schemes.
Amplitude Shift Keying (ASK) the amplitude or power of the radio carrier is varied to represent the baseband information, in this example a low power represents a digital 0 and a high power represents a digital 1. Such systems are simple in there concept but rather more difcult to implement with good performance in practice, since any variation in the radio signal during propagation will also distort the baseband information leading errors in the receiver.
Frequency Shift Keying (FSK) systems keep the power constant and vary the transmitted frequency to represent the baseband information. In this example a higher frequency represents the 0 whilst a lower frequency represents the 1. This is a more practical system and is used in mobile technologies such as GSM, it also has the advantage of being rather power efcient since the constant envelope of the modulated signal can be amplied easily. It could be said that FSK systems are not as spectrally efcient since they occupy a wide radio channel compared to the amount of data that can be sent over the channel.
Phase Shift Keying (PSK) are arguable the most spectrally efcient of modulation schemes allowing a large amount of data to be sent relative to the amount of radio spectrum occupied. However these systems tend to be rather complex and less power efcient than FSK systems. The baseband information is no encoded in to the angle or phase of the transmitted radio carrier. PSK system can be absolute, in that the angle of the carrier directly represents the baseband information, or they can be differential where the information is encoded in to the direction and magnitude of the phase change.
BPSK modulation is the simplest of the PSK family, the transmitted radio signal has only two possible angle, typically 0o and 180o. the angles can represent the 1 or the 0 of the baseband data. The diagram below shows the phase change occurring during the change of the baseband data from a 0 to 1 or 1 to 0.
The time domain representation of the BPSK modulated signal is sometimes a little complex to study there fore the diagram below is a vector representation of the same signal. In fact most PSK based modulation schemes are shown using this representation.
High Level Modulation Schemes, QPSK, 8PSK
Using this vector based approach it is easier to show the high order modulation schemes. Below is the QPSK (used in LTE) modulation constellation where each point or angle can represent 2 bits of information and 8 PSK where each angle represents 3 bit of information (EDGE uses 8PSK)
When the number of angle is more than 8 the receivers become more sensitive to noise and interference and it becomes more efcient to use the angle domain and the amplitude domain together, these system are known as Quadrature Amplitude Modulation (QAM) schemes. The constellation shown below is 16QAM and each point on the constellation now represents 4 bits of information. Such systems are highly spectrally efcient, however there is a requirement for low noise in the radio link in order that the receiver can correctly determine the point on the constellation. LTE also uses the 16QAM scheme.
Below is the 64QAM modulation scheme, each point on the constellation now represents 6 bits of information. This is a very efcient scheme however it can only be used successfully in the best signal areas. 64QAM is used by LTE.
In the diagram below we can see the impact of noise and interference on the 16QAM modulation system. Instead of the information being perfectly aligned with each target point, the noise in the radio channel causes the information to arrive in a less than perfect location, thus the information appears “spread” out over the angle and amplitude domains. Some distortion is allowed in the channel however the more complex the scheme the less distortion can be tolerated before the receiver begins to make errors.
There is more detail about the maximum distortion allowed in section 3 where we discuss more the required SNR for each of the modulation schemes of LTE
In today’s advanced mobile radio systems multiple modulation and error coding schemes are used and the link can dynamically adapt to the current radio conditions. This will ensure that the link can trade throughput or capacity for reliability for any given UE across the cell. What this means in practice is that many users in the radio cell will be using different modulation and coding schemes depending on their location. The diagram below shows the probable situation where 4 modulation and coding schemes are available. This also means that is becomes very difcult to dimension the raio cell for capacity since a user communicating using the QPSK modulation scheme will use 3 times more cell resources than a user that is situated closer to the base station using 64QAM.
In cases like this a base station function known generally as the scheduler is highly important to the efcient use of system resources.
Given the limited resources of the radio spectrum it is important that these communication systems off the highest possible capacity i.e. large number if users able to communicate apparently simultaneously. These systems are known as Multiple Access systems. in radio systems there is generally only two domains that can be shared to achieve multiple access, the frequency and time domains, other systems such as those based on spread spectrum techniques exploit information theory to allow user to communicate at the same time.
. Figure 39 – The Multiple Access Concept
Frequency Division Multiple Access
FDMA (Frequency Division Multiple Access) schemes divide a spectrum allocation into smaller frequency segments, allocating each signal a different frequency. Simple 1st Generation systems used this method.
Figure 40 – Separate Radio Channels in FDMA Multiple Access
Time Division Multiple Access
TDMA (Time Division Multiple Access) allows signals to be transmitted on the same frequencies, but not at the same time – each signal is given its own time slot within this frequency band. Note that GSM uses a combination of both of these schemes. Network Operators are allocated a portion of spectrum which is divided into radio carrier frequencies spaced 200kHz apart (FDMA). Each carrier frequency band is then divided into eight separate timeslots (TDMA).
Figure 41 – Individual User Time Slots in TDMA Multiple Access
Systems like GSM use both the time and frequency domains to create multiple sperate radio channels each divided in the time domain into timeslots. Thus a channel allocation will include both a frequency domain and time domain description.
Figure 42 – Radio Channels and Time Slots in Hybrid TDMA/FDMA
Code Division Multiple Access
The third type of access scheme, CDMA (Code Division Multiple Access), allows all signals to share the same frequency and time domains. In order to distinguish signals at the receiver, unique codes are attached to each signal. A common analogy which is made between the TDMA and CDMA schemes which are the basis of 2G cellular systems is as follows
Figure 43 – User Information Spread in the Time and Frequency Domains
Imagine a crowded room. In a TDMA system, everyone in the room is speaking the same language. Therefore in order to hear someone speaking on the other side of the room, it is necessary for everyone else to stop speaking. Each person could therefore be allocated a recurring timeslot during which they could speak, with multiple conversations supported by allocating a different timeslot to each. In CDMA, everyone in the room is speaking a different language. Therefore even when other people in the room are speaking at the same time, it is still possible to pick out what the person on the other side of the room is saying, so long as they are speaking the language that you understand.
Multi Carrier Transmission (OFDM)
Multi-carrier systems split the high speed stream of serial baseband data in to lower speed parallel streams. The lower bit rate on each sub-carrier results in a narrower radio channel that is resistant to the frequency selective fade.
Figure 44 – Single Carrier and Multiple Carrier Comparison
OFDM (Orthogonal Frequency Division Multiplexing)
However, these multi-carrier systems need to exhibit good spectral efciency, each sub carrier must be placed close to its adjacent carrier without causing interference. The channel spacing is 1/Ts where Ts is the symbol time of information modulated onto the carrier. Spacing the channels in this manner ensures that the centre of each carrier corresponds with a zero crossing point for each of the neighbouring sub-carriers. This means that the centre of the sub-carriers can be sampled, free from interference of the adjacent sub-carriers.
Figure 45 – Data is Sent in Parallel Radio Channels
Orthogonal Frequency Division and Multiple Access
Whilst the concept of multi-channel systems have many performance benets in the multipath environment, there is still a requirement to allow multiple access, that is allow many people at one time to access the services of the system.
LTE uses Orthogonal Frequncy Division Multiple Access (OFDMA) to organsise and schedule data transmission to the users in the cell. Simple OFDM systems on exploit the time domain to allow multiple access however OFDMA also allows multiple access to extend to the frequency domain. This yeilds a system that is very exible and efceint but at the same time fairly complicated to manage, hence the importance, again, of the scheduler funciton within the base station.
Section 1 Assignment Questions Q1 For your own company discover what the radio planning practices are, and comment on the differences between your own practice and those described in lesson 1. _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________
Q2 Look at a typical link budget for your own system and comment on where the main differences would be when considering an LTE link budget. Where possible include details of the vendors you may choose for the LTE network. _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________
Q3 LTE supports upto 16 different modulation and coding schemes, list the schemes supported and the SNR required for good performance. _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________
It is generally assumed that the radio wave will travel in straight lines, however this is not the case. The radio wave will follow a curved trajectory determined by the properties of the medium though which it travels. This means that the radio horizon is further that the optical or geometric horizon, the diagram below illustrates this.
Figure 49 – the Geometric and Radio Horizon
The radio wave can be assumed to have a vertical dimension which increases as the wave front travels further from the transmission source, this means that the top and bottom of the wave front will be travelling through a transmission medium which as different properties. The air in this case is the transmission medium, and the air has a certain refractive index which is determined by the air pressure, temperature, and water vapour pressure. It can be generally stated that the refractive index is less as height increases.
The variation in refractive index will vary the speed at which the radio wave travels, effectively moving faster at the top of the wave front, thus causing the entire wave front to follow the curved path.
Figure 51 – Refractive Index Reduces with Altitude
The gure below shows an alternative view where the radio wave is shown as a straight line and the geometric line is drawn as a curved line. This is referred to as the 4/3 model, where the relative size of the earths radius would have to be increase to 4/3s of it actual radius to cause the radio wave to be drawn as a straight line. The 4/3 rule applies to normal refractive and propagation conditions, however there are extreme conditions where the 4/3 does not apply
The 4/3 earth radius scan be calculated based on the following expression
Figure 53 – Calculating Earth Radius
The refractive index is given the value N, which is a unitless value. Under normal refractive conditions this value can be seen to change by 40 units for every 1000m gained in altitude. It is normally shown in a graphical format as seen below.
Figure 56 – The Effect of Sub Conditions On the Radio Path
When the refractive index falls more rapidly than standard it is referred to as superrefractive conditions and is illustrated below.
Figure 57 – Super-Refraction
When this condition occurs the radio wave will follow a more curved trajectory causing it to be bent more toward the earth than under standard conditions. The impact in this case maybe reduced radio range.
Figure 58 – The Effect of Super Conditions On the Radio Path
Extreme Cases, Ducting
Where there are extreme variations in temperature, air pressure or water vapour pressure a phenomenon known as ducting can occur. In the case below the refractive index falls with altitude but then reverses and begins to increase, this sharp change in refractive index will cause the radio wave to be reected from the boundary and become “trapped” in the duct. The duct can exhibit a very low propagation loss and the signal may travel for many miles before becoming very weak. Ducts like this may be the cause of de-coupled point to point link an interference.
Areas around the Middle East and other regions where there is extreme temperature and humidity, particularly in coastal areas, would tend to suffer from the ducting effects.
Figure 59 – Surface Duct
The two diagrams below illustrate other forms of ducting that may occur, areas where cool thermal layers sit over warm surface air (or vice versa) will cause these elevated ducts.
There are many mechanisms by which radio energy propagates around the environment, the actual effect of these mechanisms depend largely on the wavelength of the radio signal
Radio energy which arrives at a surface will be reected or scattered. The amount energy reected depends on the wavelength and the nature of the material itself and the angle of incidence. Smooth, conducting surfaces such as metal or sea water will tend to reect the signal. A reected signal will carry most of the energy of the incident wave, some of the energy will be absorbed or transmitted through the surface.
Figure 62 – Radio Wave Refection
Scattering of the radio wave would tend to occur when the height of the surface features is large relative to the wave length of the signal. The incident wave would be dispersed in multiple directions each of the new signal components having a low energy compared to the incidence wave.
When planning macro or micro level cells diffraction of radio energy around objects in the radio path is one of the main mechanisms that is analysed when making signal predictions. A radio wave that strikes an object would tend to be “bent” around the object yielding a “soft” shadow behind the object.
Figure 64 – Radio Wave Diffraction
The amount of energy diffracted is dependant on the wave length an shape of the object, basic mathematical analysis of diffraction would model spherical and knife edge objects. The path between transmitter and receiver may of course have multiple objects therefore more advance analysis will calculate multiple edge diffraction in order to predict the signal strength. Software planning tools do this as a matter of course and use both terrain and building features in their predictions.
Through this analysis it is possible to determine curves such as the one shown below for the amount of signal energy behind the object, the shape of the curve being dened by the wave length, the shape of the object and the percentage of obstruction of the radio signal.
Figure 66 – The Effect of Diffraction on the Radio Wave
Attenuation through Penetration
Another major mechanism of interest when making propagation predictions is the amount
of energy that will pass though objects, this is of particular importance when predicting in-building coverage in macro and micro cellular systems. The radio frequency, building material, thickness and the angle of incidence will all determine the amount of energy transmitted thought the object. These penetration loss values are often built empirically from tests on different types of building using different frequency bands. There is no single reference table that can be consulted during the planning stages since local variations play a large part in the nal attenuation value.
Figure 67 – Loss of Energy Through Penetration
In point to point or LOS systems it is expected the radio path can be designed largely free from mid path objects however the denition of path clearance must be determined with respect to the 1st Fresnel zone.
Fresnel zones are described by path lengths that are ½, 1, 1½ …wavelengths longer than a direct bore sight path between the transmitter and receiver antennas. When determining clearance it is only the 1st Fresnel zone that is of interest. The 1st Fresnel zone is all paths between the transmitter and receiver that are ½ wavelength longer than the bore sight path
The radius of this zone can be calculated using the expression shown below
The 1st Fresnel zone is shown in cross section below, in point to point links about 9% of the transmitted power is delivered in this zone, therefore clearance of the zone is important.
Figure 69 – The Fresnel Zones in Cross Section
The zone however does not need to be 100% clear. It is sufcient to have 60% of the 1st Fresnel zone clear for maximum power over the link. Engineers who plan these links will establish a path prole and determine the height of the transmitting and receiving antenna based on a 60% clearance.
In non line of sight system all of the above explained propagation mechanisms will be present to ensure that there is some level of coverage in all locations in the required cell area. The mechanisms described however create an environment where there is no single line of sight path between the transmitter and receiver, there will be instead many paths of radio energy, this is referred to as the multipath environment. One of the many issues in these kind of environments is the problem of fading.
Where there is many radio paths and each of the radio paths has a roughly equal power distribution the multipaths cause deep fading of the received signal. As much as 30 – 40 dB less than the expected mean signal. These environments are known as Rayleigh fading.
Multipath can exist where one of the signal paths has a much higher energy than the other paths, fading will still occur however the magnitude of the fading is much less than that experienced in the Rayleigh case, fades of up to 10-20 dB less than the expected mean can be seen.
Radio systems that use large radio cells (traditional PMR) may not use very many base stations but they are unable to offer very high capacity (number of simultaneous call, Mbps). Since the 1940s it has been known that using smaller radio cells and reusing the same bock of frequencies over and over again will yield much higher network capacities. However the regulatory regime and the technology were unavailable at that time to allow such systems to be built.
The diagram below illustrates the main concept of frequency reuse, where cell A though G will use the same radio channel or set of radio channels. The trick in these types of systems is to manage the amount of co-channel interference across the system. The more capacity require the greater the number of time the same radio channel will be used over the same area, unfortunately this also means that the level of interference will also be higher. It is a ne balance in designing high capacity networks.
The parameter that affects the amount of interference is the distance between the cell centres of the reuse cells, this is illustrated in the diagram below. Whilst the reuse distance is of some importance, the ratio of cell radius to reuse distance has more of an impact on the amount of interference.
Figure 74 – Calculating the Re-Use Distance
The expression above for the reuse distance can be transposed to ;
D/R = √3N Where N is the number of cells in the reuse pattern. A value of N = 7 will yield a particular capacity and interference value, where N=4 the capacity will be higher and the interference will also be higher.
The diagram below describes the interference concept. At the cell edge the mobile device will receive a wanted signal C but will also receive unwanted power from the interferer I. The amount interference is expressed as a ratio of these two values, C/I. C/I is also a factor when calculating the total SNR experience by the device and will determine the capacity available to the user in that location. This is particularly important in systems like LTE since the selection of modulation and coding scheme is driven largely by the SNR.
For LTE networks the challenge of frequency reuse is very high since it is very unlikely that operators will have more than 3-6 channels. Verizon in the USA, for example has deployed the rst phase of its LTE system using only a single 10MHz radio channel. This means that every radio cell will be using the same radio channel, potentially leading to very high co-channel interference.
LTE uses a mechanism called interference coordination where each base station is network to its neighbour cells and will negotiate the use of time and frequency resources. In some cases where there will be very high use of the radio channel a base station can announce what amounts to an interference warning to all the adjacent sites, thus allowing them to avoid resource collisions and therefore high interference. This coordination mechanism is crucial to the successful operation of LTE networks.
Practically, a network will not have cells of only one size, the cell sizes will depend on factors such as the type of area to be covered and the capacity required in those locations. The diagram below show the progressive splitting of cells to meet the local capacity requirements of an urban area.
Figure 76 – Cell Splitting of Capacity Increase
Cell Deployment in LTE
LTE is designed to work using radio cells from just a few meters wide to 100Km. Depending on the frequency band used it seems that the initial deployments will be microcells and smaller, certainly in Europe and the Middle East where the most likely frequency band to become available will be the 2.6GHz band. The gure below shows the basic concept and names given to radio cells of different sizes.
Figure 77 – Typical Cell Sizes for Cellular Systems
Systems that support mobility often have multiple layers of cells to increase network reliability and capacity. It is possible in these systems to services mobiles with different levels of mobility i.e. speed, with the different layers of radio cell. Smaller radio cells can be overlaid on the larger macro cells and will have antenna heights of lower altitude. Of course these days it is very common to have base stations inside public buildings to increase the reliability of the network.
Large buildings such as shopping centres’ and airports may use distributed antenna systems and remote radio heads to provide coverage in a cost effective manner.
All antenna theory stems from the basic concept of the isotropic radiator. The isotropic radiator is a theoretical point source of energy that radiates equally in all directions. The concept is show in the diagram below. From this concept the gain of real antennas can be dened as well as the basics of radio propagation and pathloss.
Figure 79 – The Isotropic Radiator
The Dipole Antenna
The simplest antenna that can be practically constructed is the ½ wave dipole. A have wave dipole is a self resonating antenna which is normally fed from the centre. The vertical dimension of the antenna is determined from the wavelength of the radio signal that is being transmitted, maximum power transfer is achieved if the antenna is ½ the wavelength and the feeder in impedance match to the antenna.
Figure 81 – The Radiation Patterns for a Dipole Antenna
When compared to the radiation from the isotopic antenna the dipole effectively focus the energy in a more specic direction. In the diagram below the edges of the radiation elds are effectively equal power contours, therefore the dipole appears to push the energy eld further from the point of radiation. This can be described as the gain of the antenna.
Figure 82 – Dipole Radiation Compared to the Isotropic
To determine the gain of the antenna if the power could be measured at the same physical point from both the isotropic and dipole antennas the formula below would allow us to determine the actual gain of the dipole over the isotropic. When the isotropic is used as the reference antenna is turns out the gain of the dipole is 2.15dB.
The isotropic antenna is normally used as the reference to describe the gain of practical antennas. The diagram below show the isotropic, the dipole and a practical directional antenna in comparison. It can be deduced that the gain of the practical antenna has a signicant performance increase over the isotropic in a specic direction, this gain is described with reference to the isotropic and expressed in dBi.
The gain of the antenna is an important factor when performing link budgets in radio planning, having a positive impact (generally) on the performance of the radio link. The radiation patterns are not considered for link budgeting purposes but are important when predicting coverage when using software planning tools. The software tool will take the vertical and horizontal radiation eld and predict the shape of the radiated energy from the site, this is particularly important when predicting the behaviour of the radio system when performing antenna tilting.
Figure 83 – Practical Antenna Radation Compared to the Isotropic and Dipole References
It is common to use phrases such as “ the radiation pattern” when describing antenna performance and the gain of the antenna is normally attributed to the radiation performance of the antenna, however the gain in the forward direction (transmitting) can be assumed to be the same in the reverse (receiving) direction. This is referred to as reciprocity.
One of the other important antenna performance attributes considered when planning is the horizontal and vertical beamwidth of the antennas. The beamwidth is normally determined from a point at the edges of the radiated eld where the power is 3dB below the bore sight (main lobe). Antennas used in sectored sites would normally use antennas who’s beamwidth was between 90o and 60o for a geometric 120 o sector.
Other important antenna attributes include the Front to Back ratio, and the Cross Polar Discrimination performance.
Front to Back ratio is a measure of how well the antenna discriminates between signals entering the front lobe (bore sight) and the rear of the antenna and is an important factor in reducing interference.
Antennas are arranged such that they operate in one plane of polarisation either vertically or horizontally, this is particularly useful when mitigating co-channel interference in cellular frequency reuse systems. Thus the ability of the antenna to discriminate between horizontal and vertical signals is important for interference reduction.
Increasing Antenna Gain
When considering an omni-directional dipole the gain of the antenna may be increased by increasing the number of radiating elements in the antenna. The individual dipole elements are normally placed in a vertical array with a phasing coil connecting each
element together, the radiated energy from each element will interfere with the others to change the shape of the total radiated eld. And increase in gain is typically associated with a lower vertical or horizontal beamwidth depending on the mechanism used.
The diagram below shows the typical arrangement of stacked antennas and their respective beamwidth and radiation pattern along with the theoretical gain. These patterns show the vertical radiation pattern from an omni-directional antenna system, further gains can be obtained by placing a reector behind the array increasing the gain in a specic direction.
Figure 85 – Increasing Antenna Gain by Stacking Elements
Antenna tilt is an important tool in the design and optimisation of mobile networks. Tilt can be achieved through a mechanical process where the antenna is physically tilted manually by and engineer, or the antenna may have the capability to be tilted electrically by remote control.
The diagram below shows the various options available with mechanical and electrical tilt and combinations of both.
Figure 86 – Mechanical and Electrical Antenna Tilt
Tilt is used to limit the range of the radiated signal there by reducing cell range and interference where required.
Antenna Diversity Confgurations
It is very common to use multiple antennas in a diversity conguration to mitigate the effects of the multipath fading environment. When spatial diversity is used, arrangements must be made to coordinate the transmit and receive functions of the antenna array.
In diagram below a separate antenna is used for the transmit, this necessitates the use of 3 antenna panels but eliminates the need for a duplex lter. The two receive antenna are arranged in a spatial array and feed the received signals to a diversity combiner.
Figure 87 – Spatial Diversity with Separate Transmit
The arrangement shown below makes use of a duplex lter which allows a single antenna panel to be used for both transmit and receive functions.
This is the “traditional” method of accessing the radio channel. Each transmitter has a single antenna, as does each receiver. This method is used as the baseline against which the performance of all multiple antenna techniques is compared.
Figure 90 – Single In Single Out
Multiple Input Single Output (MISO) – Transmit Diversity
MISO is also known as transmit diversity. Each transmit antenna transmits essentially the same stream of data. The multipath environment impacts upon the transmitted signal resulting in the arrival of time displaced replicas of the same signal at the receiver. This is used to improve the signal to noise ratio at the receiver and thus the reliability of data transmission. It is usual to apply antenna-specic coding to the signals prior to transmission to increase the diversity effect. Transmit diversity does not increase data rates as such, but rather supports the same data rates using less power or, allows a higher order modulation scheme to be used if sufcient improvement in SNR is experienced at the receiver. The performance of transmit diversity can be enhanced if the receiver is able to feedback parameters to be used by the transmitter to adjust the balance of phase and power used for each antenna.
Figure 91 – Multiple In Single Out, or Transmit Diversity
Single Input Multiple Output (SIMO)
SIMO uses one transmitter and two or more receivers and is usually referred to as receive diversity. It is particularly well suited for low SNR conditions. There is no improvement in the data rate as only one data stream is transmitted, but coverage at the cell edge is improved due to the lowering of the usable SNR.
Figure 92 – Single In Multiple Out, or Receive Diversity
Multiple Input Multiple Output (MIMO)
MIMO requires two or more transmitters and two or more receivers. Multiple data streams are transmitted simultaneously in the same frequency and time, taking full advantage of the multiple paths in the radio channel. For a system to be described as MIMO, it must have at least as many receivers as there are transmit streams.
Figure 93 – Multiple In Multiple Out, Spatial Multiplexing
Multiple Input Multiple Output (MIMO)
Adding receive diversity (SIMO) to Tx diversity (MISO) does not create MIMO, even though there are now two Tx and two Rx antennas involved. If N data streams are transmitted from fewer than N antennas, the data cannot be fully descrambled by any number of receivers since overlapping streams results in interference. However, by spatially separating N streams across at least N antennas, N receivers will be able to fully reconstruct the original data streams provided the crosstalk and noise in the radio channel are low enough. One other crucial factor for MIMO operation is that the transmissions from each antenna must be uniquely identiable so that each receiver can determine what combination of transmissions has been received. This identication is usually done with pilot or reference signals.
The spatial diversity of the radio channel means that MIMO has the potential to increase the data rate. The gure below shows a simplied illustration of spatial multiplexing. In this example, each transmit antenna transmits a different data stream. One data stream is uniquely assigned to one antenna. The multipath characteristics of the channel should ensure that each receiver antenna sees a combination of each stream. The receivers decode the received signals by analyzing the patterns that uniquely identify each transmitter and then determine what combination of each transmit stream is present. The application of an inverse lter and summing of the received streams recreates the original data.
Figure 94 – Spatial Multiplexing Requires a Mutlipath Environment
A more advanced form of MIMO includes special pre-coding which results in each stream being spread across more than one transmit antenna. For this technique to work effectively the transmitter must have knowledge of the channel conditions and, in the case of FDD, these conditions must be provided in real time by feedback from the UE. Such optimization signicantly complicates the system but can also provide higher performance. Pre-coding for TDD systems do not require receiver feedback as the transmitter can independently determine the channel conditions by analyzing the received signals that are on the same frequency.
Single User, Multiple User, and Co-operative MIMO
Single User MIMO (SU-MIMO)
This is the most common form of MIMO and can be applied in the uplink or downlink. The primary purpose of SU-MIMO is to increase the data rate to a single user. There is also a corresponding increase in the capacity of the cell. Figure 15 shows the downlink form of 2x2 SU-MIMO in which two data streams are allocated to a single UE. The two data streams (red and blue) are pre-coded in such a way that each stream is represented at a different power and phase on each antenna. The two mixed data streams are then transmitted from each antenna. The transmitted signals are further mixed by the channel. The purpose of the precoding is to optimize the transmissions to the characteristics of the radio channel so that when the signals are received, they can be more easily separated back into the original data streams.
MU-MIMO is used only in the uplink. MU-MIMO does not increase an individual user’s data rate but does offer cell capacity gains. In the gure, the two data streams originate from different UE. The two transmitters are much further apart than in the single user case, and the lack of physical connection means there is no opportunity to optimize the coding by mixing the two data streams. However, the extra spatial separation does increase the chance of the eNB picking up pairs of UE which have uncorrelated paths. This maximizes the potential capacity gain. This contrasts to the pre-coded SU-MIMO case in which the closeness of the antennas could be problematic, especially at frequencies less than 1 GHz. MU-MIMO has an additional important advantage: the UE does not require the expense and power drain of two transmitters, yet the cell still benets from increased capacity.
The essential element of Co-MIMO is that two separate entities are involved at the transmission end. The example in Figure 16 shows two eNB “collaborating” by sharing data streams to pre-code the spatially separate antennas for optimal communication with at least one UE. When this technique is applied in the downlink it is sometimes called network MIMO. The most advantageous use of downlink Co-MIMO occurs when the UE is at the cell edge. Here the SNR will be at its worst but the radio paths will be uncorrelated, which offers signicant potential for increased performance. Co-MIMO is also possible in the uplink but is fundamentally more difcult to implement as no physical connection exists between the UE to share the data streams. Uplink Co-MIMO is also known as virtual MIMO. Co-MIMO is not currently part of the Release 8 LTE specications but is being studied as a possible enhancement to LTE in Release 9 or Release 10 to meet the goals of the ITU’s IMT 4G initiative.
Figure 97 – Cooperative MIMO
Beamforming uses the same signal processing and antenna techniques as MIMO but rather than exploit de-correlation in the radio path, beamforming aims to exploit correlation so that the radiation pattern from the transmitter is directed towards the receiver. This is done by applying small time delays to a calibrated phase array of antennas. The effectiveness of beamforming varies with the number of antennas. With just two antennas little gain is seen, but with four antennas the gains are more useful. Obtaining the initial antenna timing calibration and maintaining it in the eld are challenges. Turning a MIMO system into a beamforming system is simply a matter of changing the pre-coding matrices. In practical systems, however, antenna design has to be taken into account and things
are not so simple. It is possible to design antennas to be correlated or uncorrelated; for example, by changing the polarization. However, switching between correlated and uncorrelated patterns can be problematic if the physical design of the antennas has been optimized for one or the other.
Since beamforming is related to the physical position of the UE, the required update rate for the antenna phasing is much lower than the rates needed to support MIMO pre-coding. Thus beamforming has a lower signalling overhead than MIMO.
Figure 98 – Beamforming Operation
LTE Downlink Multiple Antenna Schemes
The following multiple antenna schemes previously described are supported in the LTE downlink: Single-Antenna transmission, no MIMO• Transmit diversity• Open-loop spatial
multiplexing, no UE feedback required• Closed-loop spatial multiplexing, UE feedback required• Multi-user MIMO (more than one UE is assigned to the same resource block)• Beamforming•
Open-loop Tx Diversity
This is the simplest downlink LTE multiple antenna scheme. LTE supports either two or four antennas for Tx diversity. Figure 17 shows a two Tx example in which a single stream of data is assigned to the different layers and coded using space-frequency block coding (SFBC). Since this form of Tx diversity has no data rate gain, the code words CW0 and CW1 are the same. SFBC achieves robustness through frequency diversity by using different subcarriers for the repeated data on each antenna. Receive Diversity
RX diversity is mandatory for the UE. It is the baseline receiver capability for which performance requirements will be dened. A typical use of Rx diversity is maximum ratio combining of the received streams to improve the SNR in poor conditions. Rx diversity provides little gain in good conditions.
Spatial Multiplexing – MIMO
MIMO is supported for two and four antenna congurations. Assuming a two-channel UE receiver, this scheme allows for 2x2 or 4x2 MIMO. A four-channel UE receiver, which is required for a 4x4 conguration, has been dened but is not likely to be implemented in the near future. The most common conguration will be 2x2 SU-MIMO. In this case the payload data will be divided into the two code-word streams CW0 and CW1 and processed according to the steps in the gure below.
Depending on the pre-coding used, each code word is represented at different powers and phases on both antennas. In addition, each antenna is uniquely identied by the position of the reference signals within the frame structure, as illustrated in Figure 18. The UE must obtain accurate picture of channel conditions for each antenna. Therefore, when a reference signal is transmitted from one antenna port, the other antenna ports in the cell are idle.
Closed Loop Spatial Multiplexing
As the streams must be pre-coded, the transmitter must have knowledge of the channel. The UE estimates the radio channel and selects the optimum pre-coding matrix. This channel information is provided by the UE on the uplink control channel. The channel feedback uses a codebook approach to provide an index into a predetermined set of pre-coding matrices. Since the channel is continually changing, this information will be provided for multiple points across the channel bandwidth, at regular intervals, up to several hundred times a second. The exact details are still to be specied. However, the UE that can best estimate the channel conditions and then signal the best coding to use will get the best performance out of the channel. Although the use of a codebook for pre-coding limits the best t to the channel, it signicantly simplies the channel estimation process by the UE and the amount of uplink signalling needed to convey the desired precoding.
If the UE is moving at a high velocity, the quality of the feedback may deteriorate. Thus, an open loop spatial multiplexing mode is also supported which is based on predened settings for spatial multiplexing and pre-coding. The eNodeB will select the optimum MIMO mode and pre-coding conguration. The information is conveyed to the UE as part of the downlink control information (DCI) on PDCCH.
Cyclic Delay Diversity (CDD)
This technique adds antenna-specic cyclic time shifts to articially create multi-path on the received signal and prevents signal cancellation caused by the close spacing of the transmit antennas. Normally multipath would be considered undesirable, but by creating articial multipath in an otherwise at channel, the eNB UE scheduler can choose to transmit on those RBs that have favourable propagation conditions. The CDD system works by adding the delay only to the data subcarriers while leaving the RS subcarriers alone. The UE uses the at RS subcarriers to report the received channel atness and the eNB schedules the UE to use the RB that it knows will benet from the articially induced ”multipath”. By not applying the CDD to the RS, the eNB can choose to apply the CDD on a per-UE basis.
In order for MIMO schemes to work properly, each UE has to report information about the mobile radio channel to the base station. A lot of different reporting modes and formats are available which are selected according to the MIMO mode of operation and network choice.
The reporting may consist of the following elements:
CQI (Channel Quality Indicator) is an indication of the downlink mobile radio channel quality as experienced by this UE. Essentially, the UE is proposing to the eNodeB an optimum modulation scheme and coding rate to use for a given radio link quality, so that the resulting transport block error rate would not exceed 10%. 16 combinations of
modulation scheme and coding rate are specied as possible CQI values. The UE may report different types of CQI.
A so-called “wideband CQI” refers to the complete system bandwidth. Alternatively, the UE may evaluate a “sub-band CQI” value per sub-band of a certain number of resource blocks which is congured by higher layers. The full set of sub-bands would cover the entire system bandwidth. In case of spatial multiplexing, a CQI per code word needs to be reported.
PMI (Pre-coding Matrix Indicator) is an indication of the optimum pre-coding matrix to be used in the base station for a given radio condition. The PMI value refers to the codebook table. The network congures the number of resource blocks that are represented by a PMI report. Thus to cover the full bandwidth, multiple PMI reports may be needed. PMI reports are needed for closed loop spatial multiplexing, multi-user MIMO and closed-loop rank 1 precoding MIMO modes.
RI (Rank Indication) is the number of useful transmission layers when spatial multiplexing is used. For transmit diversity the rank is equal to 1. The reporting may be periodic or aperiodic and is congured by the radio network. Aperiodic reporting is triggered by a CQI request contained in the uplink scheduling grant. The UE would send the report on PUSCH. In the case of periodic reporting, PUCCH is used if no PUSCH is available.
Section 2 Assignment Q1 Super refraction, sub refraction and ducting are examples of extreme propagation conditions, research what causes these conditions and give examples of where these might occur. Also, determine what might be done to mitigate these effects when planning radio systems. _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________
Q2 Building attenuation is one of the most common factors when performing link budgets and it is generally said that building penetration increases with frequency. Do some research and build a table of typical attenuation values for different building materials and frequency bands. _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________
Q3 LTE supports many different frequency bands and it is likely that operators will choose different band depending on local availability and regulatory conditions. Do some research and nd out what frequency bands are being proposed in different parts of the world, give examples of operators and which bands that are proposing to use. Comment on some of the issues regarding the diversity of bands. _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________
Q4 MIMO is an antenna technique that will be widely deployed by LTE operators. Do some research and n out which vendors are supporting the MIMO technology, which operators are planning to use it and where possible include the results of any trails that have taken place and comment on the performance increases. _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________
Every radio system is a series of components and links, from the transmitter to the receiver. Each element of the system will exhibit some attribute of performance that affects the overall performance of the end to end system. A typical link budget exercise will need to quantify each of these performance attributes and understand the impact it may have on the system performance, i.e. the capacity and coverage.
Many factors can be determined from manufacturers data sheets, thing such as the Tx power, feeder losses, antenna gains etc..However some parts of the system, the radio interface, must be modelled in order to determine a satisfactory plan.
Figure 102 – Typical Arrangement of Radio System Components and Variables
The diagram above shows the typical arrangement of components in a radio system, each of the performance values shown, operating frequency, feeder loss, antenna gain and noise gures etc., must be understood and quantied in order to successfully plan a radio system, this module will help to understand some of those values and the impact it has on link planning
One of the main aims of calculating a link budget is to determine the maximum path loss allowed across the radio link for a given performance objective. The link loss will be due in part to the performance of the transmitter and receiver components as well as the impact of the environment through which the signal will propagate.
The goal of link planning is to determine the parameter MAPL (Maximum Allowable Path Loss)
MAPL = System Gain – Margin(fade, body, building, trees)
System Gain is a function of the radiated power from the transmitter system and the minimum signal power that can be presented to the face of the receiving antenna. The value of System gain is an indication of the maximum and minimum values in the link budget.
Link Margins are subtracted from the System Gain to determine the maximum path loss for a given set of assumptions for the transmitting and receiving system. This MAPL can subsequently converted in to a nominal cell range using an appropriate propagation model.
System gain is determined by subtracting the maximum transmit power from the minimum receive power.
The values of feeder and connector (and any other) losses can be determined from manufacturer data sheets as can the Tx_PWR of the eNB and UE. It is likely that the Rx-SENS will also be quoted by the vendor for the eNB and UE however the calculation is rather complex and can involve many parameters that will ultimately have a great impact on the overall system performance, it is worth therefore, a closer examination.
LTE link Budget variables
The basic expression from above can be formulated to include aspects of the LTE link budget. The expressions on the opposite page show how the MAPL for the uplink and downlink may be calculated. It is assumed in these cases that the UE will have no losses due to cables or connectors. Since most cellular systems are limited by the performance of the uplink it is common to being the link budgeting process with the uplink and look for a link balance with the downlink.
The power output of the UE is pretty straight forward since at the present time only a single maximum power output is specied. However it could be possible in future to have different power outputs depending on the power class of the UE in each of the different specied bands.
The following maximum output powers can be assumed: -
23 dBm for the UE,
Figure 103 – Table of Equipment Classes and Power Outputs
The maximum power output of an LTE UE is specied to be 23dBm, however there are other factors that might result in a reduced power out put, this rst is;
Maximum Power Reduction (MPR)
Maximum Power Reduction (MPR) is a reduction in the power output of the UE due to a high order modulation scheme being used, this reduction in power eases some of the problems that occur with high peak values in the power amplier, it is thought that the disadvantages of reduction in power is out-weighed by reduced complexity in the power amplier stages of the transmitter.
Figure 104 – Power Reduction for LTE Modulation Schemes
Additional-Maximum Power Reduction (A-MPR)
It is possible for the network to signal additional power reductions in specic deployments where there are tighter requirements of Adjacent Channel Leakage Ration (ACLR) and other spectrum emission requirements.
According to the 3GPP specications there are 3 classes of base station.
Wide Area Base Stations are characterised by requirements derived from Macro Cell scenarios with a BS to UE minimum coupling loss equal to 70 dB. No upper limit for power output is specied by 3GPP for this class of base station (some regional limits apply, in addition there are CEPT band limits that should also be considered)
Local Area Base Stations are characterised by requirements derived from Pico Cell scenarios with a BS to UE minimum coupling loss equal to 45 dB. The limitations on power output depend on the number of antenna ports used and are shown in the table opposite
Home Base Stations are characterised by requirements derived from Femto Cell scenarios. The limitations on power output depend on the number of antenna ports used and are shown in the table opposite
For link budgets the typical eNB power outputs for macro cell deployments would however be in the range 20 – 60W (43 – 48dBm) depending on channel bandwidth.
Typical power outputs may depend on the bandwidth being used;
43 dBm (5 Mhz, 1.25 MHz)
Figure 106 – Additional Power Reduction Factors
Typical Losses in the eNB
Within the eNB system there will be many components that insert loss in to the transmitted and received signals. It is a general rule that losses should be kept to a minimum. The total amount of loss will determine the radiated power (EiRP) and the received signal. Additionally the losses in the receive path will also add noise which change the SINR requirement on the link.
Other Losses in the transmit/receive system
There are other components in the transmitter and receiver chains that will incur additional insertion losses. Ideally these will be kept to a minimum by choosing high quality components, or keeping feeder runs to a minimum length.
There is much which is yet unknown about the LTE UE antenna systems, given that MIMO is likely to be present in the devices, this places a great deal of challenge in the design and implementation of the UE antenna. However for basic link budgeting purposes it is acceptable to assume a low gain gure for the antenna, typically 0dBi. This of course will depend on the type of LTE device, USB dongles, handheld smartphone devices and even cameras and other consumer devices are likely to have differing antenna performances. It will be largely up to the vendors of the devices to provide the relevant gures.
The eNB in most cases can make use of the familiar cellular antennas that have been used for other mobile broadband systems such as WiMAX and UMTS/HSPA. A typical example of LTE antenna specications is shown on the page opposite.
It is possible of course that the operator will implement spatial multiplexing or transmit diversity, this will have an impact on the link budget calculations. It is expected that the vendors of these systems will provide the appropriate gures of gain to be included in any calculations.
In calculating the required or minimum IRL it is necessary to determine the sensitivity of the receiver. It is highly probable that the vendor of the eNB and mobile devices will quote the sensitivity in the spec sheets for their product. However it is important to be able to derive the sensitivity of the receivers for all cases of modulation/coding schemes and resource block usage.
The expression below shows the calculation and all the parameters required to make the calculation. The following pages will explain each parameter.
RXsens_eNB = -174dBm/Hz + 10log(Nrb x 180KHz) + NFeNB + SNR + IM
RXsens_UE = -174dBm/Hz + 10log(Nrb x 180KHz) + NFUE + SNR + IM – 3dB - D FB
- -174dBm/Hz is k x T (Boltzmann Constant x Temperature) - Nrb is the Number of Radio Blocks Allocated - 180KHz is the bandwidth of 1 RB -
NFeNB is the total noise gure of the eNB system
- SNR is the Signal to Noise Ratio required i.e. for the modulation scheme in use - IM is an Implementation Margin depends Modulation and Coding used - -3dB is the multiple antenna gain for the UE -
DFB is a frequency band specic relaxation factor for the UE
Thermal noise is present in all things, it is a measure of the amount of noise power present due to the random motion of the atoms and molecules excited by temperature. In electronic and radio systems the noise is always present and there is little to be done to eliminate the noise completely.
In radio systems the noise is present in two forms;
Thermal background noise
Noise present in the system components
The thermal background noise is present as a result of the “big bang” (cosmic background radiation), the galaxies, the stars, our own sun and natural radiation from the surface of the earth and the objects upon it. There is no way that we can prevent this kind of noise entering the radio system but there is a way to quantify the amount of noise present. The expression;
N = k TB t
k is Boltzmann’s Constant 1.38 x 10-23
T is temperature (normally 290K)
B is the Bandwidth of the Channel in Hz
shows that noise is proportional to the bandwidth of the radio systems and temperature. The bandwidth of the radio system under investigation is really the only variable since temperature is taken to be that of the “warm earth” or 290K.
The graph below shows the rise of noise with radio channel bandwidth and the range of LTE radio channel bandwidths plotted for comparison.
Figure 109 – Bandwidth and the Impact on Background Noise
Type of Service and Impact on Noise Floor
LTE is very exible, not only in terms of the system bandwidth, but also the amount of bandwidth or Resource Blocks that can be allocated to a singe mobile device. This variable allocation can be demonstrated in the following example.
A typical voice call in LTE may require 64Kbps, for example, given that call reliability will be important across the whole radio cell, robust modulation schemes may be allocated for the voice call events, QPSK 1/3 for example, in this case only two RBs will be required, a total allocated bandwidth of 2x180KHz or 360KHz, this gure can be used to work out the thermal noise oor.
In contrast a device that is attempting to receive 1Mbps will have to be allocated between 2 and 13 RBs, depending on the selected modulation and coding scheme. Thus the noise oor could rise up to 10dB (or more) for high capacity allocations.
The graph below shows the potential noise oor rise for RB allocation between 1 and 25 RB (25 RB corresponds to a channel bandwidth of 5MHz)
Data Service at 1Mbps – depends on modulation scheme
13 RB for QPSK 1/3
2 RB for 64QAM 2/3
Nt/RB = -174dBm/Hz + (NRB x 180KHz)
Figure 110 – Resource Block Allocation Effect on the Noise Floor
Use the spread sheet to examine the effect of the number of radio blocks on the sensitivity of the receiver. You should use the “sensitivity” tab on the spread sheet, the “link budgets” tab will be examined later.
Implementation Margin, UE, eNB
Included in the sensitivity calculation is a margin due to the implementation of the modulation scheme. It is not possible for the receiver to be 100% accurate particularly for the higher order schemes therefore an implementation margin is added. Typical values are given below.
The margin accounts for the difference in the theoretical SINR values and the practical implementation actually possible.
Receiver Noise Figure
Noise will also be present in the receiver it self. The noise performance of the receiver is normally quoted as the NF (Noise Figure). How much noise is present is largely down to the design of the receiver by the vendor of that component however is expected that the noise will be no more then the example gures given below for a typical eNB and UE receiver.
Typical eNB NF
Typical UE NF
9dB* (same as WCDMA)
The noise gure (NF) will have an impact on cell range. The LTE documents specify a gure similar to those for WCDMA devices and it is felt that the gure is a compromise between reasonable cell range and practical receiver design performance. It is range of values also allow some scope for the vendors to improve the performance of the device receivers and therefore improve the sensitivity of the devices, this is also a key differentiator in the device market.
Total Noise Floor
The overall system noise oor is the sum of the external noise present and the total component noise. This is illustrated in the gure opposite.
Where there are multiple components (active and passive) in the receiver system, the total noise can be calculated using the cascade method.
When using the Cascade formula, the noise gure reference point can be assigned at any point before the rst active (amplier) component. The rst system component will have the greatest inuence, meaning that the system NFdB can’t be better than the NF dB of the rst component, on the system NFdB. Stages after an amplier have progressively less impact on total system NFdB.
The performance of a cascaded system of components is based on the conguration and performance parameters of the individual components. The above two systems use the same components in different congurations. The key to performance of these two systems is the placement and performance of the Low Noise Ampliers (LNA). The rst stage in a cascade of stages limits the receiver system NFdB—it can never be better than the NFdb of the rst component! The purpose of the LNA is to increase the noise oor high enough to reduce the impact of loss from successive stages while having a minimum effect of the C/N. A high gain LNA with a low NFdB can provide benet even if it is after a coax loss. Without sufcient gain, benet is minimum. Too much gain can overdrive the receiver in the presence of a strong receive signal.
System 1: A signicant loss in front of the LNA limits the receiver system NFdB. A high gain in the LNA can help minimize the post-LNA losses. This conguration (indoormounted LNA) can be benecial if the coax loss to the LNA is reasonably low and the LNA has sufcient gain relative to the post-LNA losses. A low gain LNA offers little performance benet in this, or any deployment. An LNA with too much gain reduces the dynamic range of the receiver and could overload the receiver, causing other problems.
Figure 112 – Calculating Total Noise in Cascaded Systems
System 2: Theoretically, this can provide the best performance. If there is a signicant amount of gain in the LNA, the post-LNA losses have little impact on the system NFdB. If a small amount of gain is used, the LNA provides little or no benet. In cellular deployments, this is referred to as a TTA (Tower-Top Amplier). Since LNAs are typically rated for their operating NFdB at 23° C ambient temperature, there can be a degradation of performance when the ambient temperature increases above this value. Remember, an LNA with too much gain reduces the dynamic range of the receiver and could overload the receiver, causing other problems.
Convert the values from dB to linear terms and use the cascade formula to determine the total noise contribution from the conguration shown above.
You can attempt this with a calculator or use the provided spreadsheet
Typical SNR for LTE Modulation and Coding Schemes
Given that there are different modulation and coding schemes in use for the LTE radio interface the SINR for each must be determined, this is largely down to the design of the receive and the efciency of the error coding schemes used, the table below shows the expected values of SINR and the respective IM, however the actual number may vary between vendors.
Figure 113 – Typical SNR Requirements for LTE Modulation and Coding Schemes
Duplex Gap and Duplex Distance, Effect on Receiver Sensitivity
The calculation for UE sensitivity includes an extra parameter which is a margin due to the separation between the uplink and downlink radio channels. Where the channel bandwidth is very large and the duplex separation between them is relatively small this causes the UE receiver to fall directly into the shoulders of the transmitter spectral output. This will require better ltering in the UE, lters with the characteristics required to eliminate any signicant receiver desensing have a higher insertion loss which therefore contributes to a higher receiver NF. For the bands affected by this problem a relaxation factor is taken into account when calculating the sensitivity of the receiver, DFB.
Once all the equipment operating parameters have been determined the EiRP and Sensitivity can be calculated. From this the System Gain can be determined. System Gain is a measure of the maximum drop of power from the transmit antenna to the receive antenna, but does not take in to account any additional margin from radio interface effects such as fading and penetration losses.
Figure 116 – Typical Link Budget Profle
Environmental Factors and Noise Rise
Having worked out the System gain it is now possible to determine the MAPL. The Maximum Allowable Path Loss is the system gain less any environmental margins. Typical margins include;
MAPL = System Gain – Margin(fade, body, building, trees)
When deploying NLOS implementations, shadow fading (due to path obstructions) must be considered. Measurements have shown that for any distance from a base station, the path loss at different locations is random and has a log-normal distribution. Over a large number of measurement locations having the same distance between subscriber unit and base station, the random shadowing effects are described by a log-normal distribution. This is often referred to as Log-normal Shadowing (or fading).A common approach is to calculate the lognormal probability of adequate signal strength in a coverage area.
The probability is a function of the path loss exponent and the standard deviation of signal values for a given environment. The amount of margin determined from the environmental values is based on coverage objectives for a given implementation. Mobile radio (cellular) prioritizes the area service objective, while xed wireless services may consider margin for area or edge coverage.
The propagation constant (n), also called the path loss exponent, accounts for the distance dependent mean of the signal level based on the propagation environment.
Figure 118 – Typical Path Loss Exponent Values
The standard deviation (σ) statistically describes the path loss variability for arbitrary locations with the same distance between subscriber unit and cell site. The ratio of σ/n is used to determine the amount of margin required to satisfy an area reliability objective
Figure 119 – Typical Values for Standard Deviation
The following expression can be used to work the percentage of useful service area (assuming a circular cell) where the factors listed below are known. It is more usual to specify the area reliability gure (e.g. 90% area reliability) and then work out what margin is required in the link budget to achieve the required availability.
Foliage loss is a function of absorption and scattering. Building loss is primarily absorption loss.
Wet surfaces will generally increase the amount of energy reected rather than transmitted thus increasing overall penetration loss.
In both foliage and building loss, it is important to establish local parameters to be used during planning processes.
Figure 120 – Building and Foliage Penetrations Losses
In mobile cellular systems, handheld devices will incur an additional loss due to absorption by the human body. The actual gure will depend on the use of the device i.e. held near the head, away from the body holding angle of the device. UE antenna radiation patterns may also affect the amount of energy lost.
The gure normally assumed for radio planning purposes is 3dB.
Noise rise occurs in TDMA/FDMA systems when the same frequency and time resources are used simultaneously in neighbouring cells. This will be a key factor for implementing LTE networks, the eNBs will communicate across the X2 interface regarding resource allocation either warning of potential noise or simply indicating what resources are currently being used. In lightly loaded systems the noise rise should be kept to a minimum by the interference coordination between the base stations, however when the system becomes loaded the noise rise is likely to have a greater impact on overall system performance.
Propagation modelling or prediction is the science of predicting the pathloss of a particular radio frequency when some of the system attributes are know, typically the radio frequency, tower and UE heights and distance are the information required, however more complex models can use the average height of buildings or terrain, relative angle of roads, antenna tilts etc to produce more accurate results.
The model shown opposite is at the heart of this science. This models the theoretical wave front from an isotropic radiator and predicts the eld strength at a given distance.
Figure 123 – Isotropic Radiation and Spreading Loss
If a value for the receive antenna attributes is included it is possible to derive the Free Space Pathloss model. In the free space pathloss model energy radiated from the source decays in proportion to the square of the distance - a doubling of distance will increase the path loss by a factor of 4.
Figure 124 – Converting Spreading Loss into Free Space Loss
Coverage from link budget
Having calculated the MAPL above it is now possible to convert the pathloss into a nominal cell range using an appropriate propagation model. The results will vary according to the model used. There are many different kinds of model, the classical empirical models such as Okamura-Hata , Walsch-Ikegami and those used by RF planning models. It is important to select the correct model and some model tuning is required to obtain theoretical results that reect the actual loss or distance likely to be experienced in the eld.
The following is a list of empirical models can be used in the preliminary stages of planning.
Comparison of models There are of course many different models that can be used under different circumstances, the choice of model will depend on system design parameters such as the frequency band used, LOS or NLOS systems, antennas above or below rooftop height etc.
The table on the page opposite shows some of the standard models in common use and the range of frequencies over which the model will return sensible results.
Some of the models are empirical models which means that they are also dependant on the circumstances under which they were developed. In many cases different models will return different pathloss results for the same set of inputs (frequency, tower height, link distance etc) therefore several models may need to be test to see which model returns the most accurate results for the are being designed.
Many RF planning tools will allow you to select different propagation models in order that comparisons can be made, in addition the RF software development companies will offer their own models that use a combination of empirical and physical models to predicate the pathloss.
The model shown below is the COST 231 model which is an adaptation of the well known Okamura-Hata model. The COST 231 is an empirical model designed to model NLOS radio systems in the frequency range 1.5GHz to 2GHz making it suitable for cellular systems such as GSM1800, UMTS and even Mobile WiMAX technologies.
This is a baseline model which can be used to make comparisons of other empirical and custom designed models.
Figure 128 – COST 231 Propagation Model
The WINNER Model
The WINNER model had been developed by Information Society Technologies (IST) for predication for indoor and outdoor systems.
The novel features of the WINNER models are its parameterisation, using of the same modelling approach for both indoor and outdoor environments, new scenarios like outdoor-to-indoor and indoor-to outdoor, elevation in indoor scenarios, smooth time (and
space) evolution of large-scale and small-scale channel parameters (including crosscorrelations), and scenario-dependent polarisation modelling. The models are scalable from a single single-input-single-output (SISO) or multiple-input-multiple-output (MIMO) link to a multi-link MIMO scenario including polarisation among other radio channel dimensions.
WINNER II channel models can be used in link level and system level performance evaluation of wireless systems, as well as comparison of different algorithms, technologies and products. The models can be applied not only to WINNER II system, but also any other wireless system operating in 2 – 6 GHz frequency range with up to 100 MHz RF bandwidth. The models supports multi-antenna technologies, polarisation, multi-user, multi-cell, and multi-hop networks.
The link budgets calculations done previously can now be used with the propagation models to determine the nominal cell range based on the equipment performance assumptions. The pathloss models require some transposition to derive distance rather than pathloss, this is best done by modelling within spreadsheets or other software models.