MODELLING AND SIMULATION OF LOCAL AREA WIRELESS CHANNELS FOR WLAN PERFORMANCE ANALYSIS

Transcription

1 MODELLING AND SIMULATION OF LOCAL AREA WIRELESS CHANNELS FOR WLAN PERFORMANCE ANALYSIS Simmi Dutta, Assistant Professor Computer Engineering Deptt., Govt. College of Engg. & Tech., Jammu. Ph , Jyoti Mahajan Lecturer Computer Engineering Deptt., Govt. College of Engg. & Tech., Jammu. Abstract The wide-spread adoption of IEEE wireless local area networks (WLANs) operating at the unlicensed ISM frequency bands is one of the few highlights of communication technologies in recent years. Several WLAN configurations have been implemented in the operational environment with each one providing different Quality of Service. There is requirement to measure and relatively grade the performance of a WLAN connect. This paper adapts and implements a WLAN system channel model so as to obtain simulation results about Quality of Service with varying channel conditions and also to analyze the performance of one of the most widely deployed WLAN based on IEEE a standard. The channel adaptation process essentially captures the propagation characteristics of a typical WLAN enabled office ranging from a small room to a medium size shop floor. Matlab & Simulink has been chosen as the modelling tool because of its flexibility and mathematical simulation capabilities. The results obtained with varying channel conditions represented by typical models are indicative of performance of OFDM in different environments and clearly show that certain channels are better than the others and will help in evolving signal power benchmark for effective communication. Keywords Indoor Channel Modeling, Rayleigh multipath, Power delays Profiles, a simulation. Introduction In the recent years, IEEE Wireless Local Area Networks (WLANS) have increasingly been deployed in a variety of situations, such as homes, business enterprises, academic campuses, and public places such as airports, hotels, and shopping centers [1]. It has therefore become very important to understand the performance of such networks as well as the effective network design, deployment and management. The IEEE standard [2] specifies the medium access control (MAC) and physical (PHY) layers for so called WLANs.WLAN performance is an indicator of how productive a wireless user s connectivity will be, and thus is a key metric when a consumer or an IT organization makes a purchasing decision. This paper is intended to adapt a model for the WLAN system channel so as to obtain simulation results about Quality of Service (QoS) with varying channel conditions. The paper documents the implementation of a proposed channel model and the configuration of the complete system to provide as a workbenchto evaluate the effect of channel conditions on WLAN realistically. Matlab & Simulink has been chosen as the modeling tool because of its flexibility and mathematical simulation capabilities The SISO WLAN Models and the Proposed Channel Model The indoor communication environment is the subject of interest in this study. This environment is typical of the WLAN systems deployed for communication in a small office or house. The propagation in these conditions is characterized by a possibly direct LOS and multipath 15

2 propagation from the transmitter to the receiver. A channel model will generally comprise of three distinct terms namely Path Loss, Path Delay and Doppler Spread. Since every wireless user s environment will likely differ, there is no way to predict exactly what performance can be obtained in any one environment or at any instance. However, there can be an industry-standard benchmark to gauge relative performance. A set of WLAN channel models was developed by Medbo et al. [3,4]. In this, five delay profile models were proposed for different environments (Models A-E).The multipath channel model adapts and implements the channel characteristics proposed by the IEEE task group on n channel models. It is proposed to use the three models developed by Medbo [3] with three additional models proposed by the IEEE task group on n channel models [5]. These channel models are representative of smaller environments, such as residential homes and small offices, for our modelling purposes. The resulting models that are proposed are as follows: a) Model A (optional, should not be used for system performance comparisons), flat fading model with 0 ns rms delay spread (one tap at 0 ns delay model). This model can be used for stressing system performance, occurs small percentage of time (locations). b) Model B with 15 ns rms delay spread. c) Model C with 30 ns rms delay spread. d) Model D with 50 ns rms delay spread. e) Model E with 100 ns rms delay spread. f) Model F with 150 ns rms delay spread. Model A defines an artificially small room with no multipath reflections and is included to help calibrate the channel emulator. Model B emulates reflections of a typical small room. Models C through F represent increasingly larger office spaces and outdoor settings. The most commonly deployed WLAN operates on the IEEE b/g standard because of its commercial viability and backward compatibility with IEEE b clients. The higher speeds in IEEE802.11g are still achieved using the OFDM modulation. The comparison between the primary specifications of the various IEEE standards is shown in table 1. 16

3 Table 1: Primary IEEE specification comparison. Data rate, range, throughput and compatibility vary among the various WLAN standards. These variations are caused by differences in the frequency, modulation schemes and number of data rates.the Orthogonal Frequency Division Multiplexing (OFDM) transmission scheme is an implementation of multi-carrier transmission scheme. In the past, as well as in the present, the OFDM is referred in the literature as Multi-carrier, Multi-tone and Fourier Transform [6]. OFDM is the technology of future and will form the base communication strategy for the next generation communication systems. Therefore, the Performance of OFDM based systems against differing real life channels is a necessary benchmark to be established. The Mathematical Rayleigh multipath Channel Model The band-limited discrete multipath channel is modeled based on guidelines discussed above and the analytic treatment given in section in [7]. It is assumed that the delay power profile and the Doppler spectrum of the channel are separable. The multipath fading channel is therefore modeled as a linear finite impulse-response (FIR) filter. Let {s i } denote the set of samples at the input to the channel. Then the samples {y i } at the output of the channel are related to {s i } through: y i = N 2 s i n n= N1 g n (2) where {g n } is the set of tap weights given by: y i (3) = K k= 1 In the equations above, τ k ak sin c n, where N1 Ts n N T s is the input sample period to the channel. {τ k } where 1 k K, is the set of path delays. K is the total number of paths in the multipath fading channel. 2 17

4 {a k } where 1 k K, is the set of complex path gains of the multipath fading channel. These path gains are uncorrelated with each other. N 1 and N 2 are chosen so that {g n } is small when n is less than - N 1 or greater than N 2 Each path gain process {a k } is generated by the following steps: a) A complex uncorrelated (white) Gaussian process with zero mean and unit variance is generated in discrete time. b) The complex Gaussian process is filtered by a Doppler filter with frequency response H ( f ) = S( f ),, where S(f) denotes the desired Doppler power spectrum. d) The resulting complex process z k is scaled to obtain the correct average path gain. In the case of a Rayleigh channel, the fading process is obtained as: a k 2 = Ω k z k, where Ω k = E( a k ) Implementation of the Channel Models and Simulation Framework In order to successfully simulate the digital communication through a realistic channel as proposed by IEEE [5], a host of ancillary sub programs were developed. Matlab program was written to assemble the cluster powers in a unified manner and prepare the channel implementation involving the computation of Power delay Profiles (PDP) of each channel model (Figure 1). c) The filtered complex Gaussian process is interpolated so that its sample period is consistent with that of the input signal. A combination of linear and polyphase interpolation is used. Figure 1: The Power Delay Profiles of each of the Channel Models proposed by IEEE adapted for implementation in a SISO WLAN scenario 18

5 These Power Delay Profiles are representative of multipath effect experienced from small indoor area to a large indoor complex [5]. Once the PDP had been evolved, this consolidated data was used for writing a program to implement the FIR Rayleigh Multipath filter algorithm as described in the previous section. This algorithm can be configured to implement all the five channel models, with appropriate tap calculations and PDP. The rayleigh multipath simulation program requires another parameter quantifying the Doppler components of the multipath clusters for successful implementation. Since the environment assumed in the study for this implementation was limited to indoor WLAN links, a fair estimate of 3 Hz as maximum Doppler was assumed principally, due to movement of people and plants placed in the indoor environment [5]. Once the Rayleigh channel simulation program was developed it was integrated with the Simulink simulation environment. Simulation Test Bench Simulation model, developed and adapted for different indoor channel conditions [2,3] by the author was coded as a Tapped Delay Line based FIR (Finite Impulse response) filter in Simulink software. In order to generate the simulation results a detailed simulation model for the IEEE a Physical layer was used as prefabricated code. This model is available with the communication Blockset of the Simulink software. This code was modified to suite the requirements of simulation for this study. The PHY model was integrated with the channel models developed by the author to form an integrated test-bench. This testbench could be controlled with a MATLAB based script to excite different communication modes, channel models and undertake monte-carlo simulations. A Simulink block diagram of test-bench simulation setup is given in figure 2. Figure 2: IEEE a WLAN Physical Layer with Modified Channel simulation setup. 19

6 The test-bench setup was run to generate simulation results that were documented, analyzed and archived. There are five channel models B through F [3] that were simulated with the communication system operating in different native modes of IEEE a specifications. Model with a simple AWGN (Additive White Gaussian Noise) was taken as the baseline (Model A) setup during the simulation. There are eight modes of the IEEE a communication system characterized by speeds, channel coding and Modulation. The eight modes are specified in Table 2. Table 2: IEEE Communication mode Specification. The real life channel simulation involves multipath propagation from the transmitter to the receiver that generates Rayleigh Fading and motion of the transmitter, receiver or the environmental scatterers generating Doppler shifts in the received signal. When a static environment like the WLAN in office, residential or small factory premises is considered multipath is abundant but the Doppler is restricted because of very little movement involved in between the transmitter and the receiver. A ballpark figure of 3Hz Doppler content is considered typical of indoor environment because of foliage and movement of people in the propagation medium. Figure 3 depicts the baseline characteristics of channel communication. It documents the effect of varying SNR (Signal to Noise Ratio) through the communication channel on the BER (Bit Error Rate) the principle figure of Merit in the communication systems. It can be seen that as the SNR increases the BER decreases and also as the channel rate increases the SNR has to increase, if the BER is to remain constant, which a Quality of Service requirement is. The channel considered in this plot is a benign environment that hardly captures the realistic environmental effects. 20

7 Figure 3 : Baseline Characteristics of channel communication The realistic environments involve multipath effects and Doppler effects and the simulation results have been captured and presented in Figure 4 for different environmental conditions represented by Models B through F. The plotted results highlight the following aspects in general: (a) The BER is affected by the Channel speed and the coding rate and SNR. (b) The best throughput speeds will change based on the channel speeds and the prevailing SNR in the channel. (c) There is no single best speed for communication to achieve a best throughput. A complex combination of the speed coding rates, SNR and channel model needs to be evolved to figure out the best communication strategy. 21

8 Figure 4: BER Characteristics for different channel models. A different look at this simulation data involves comparing each OFDM mode performance in different channel model conditions. Detailed graphs of the Mode performance with channel models B through F are shown in figure 5. The graphs indicate that the best performance in most of the OFDM modes is achieved when the communication channel is emulating channel models D and E. The power delay profiles of these the two channel models have been presented in detail by the author earlier [Figure 1]. It can be observed from the comparative plots in figure 4 that to retain the BER at 10-3 different modes require different SNR values in the communication channel. The highest speed mode requires a SNR of at least 30dB with the channel models, emulating the best propagation environment. 22

9 Figure 5 : BER Characteristics for different OFDM modes A closer inspection of figure 5 reveals that mode 3 exhibits the best BER performance with the least SNR. To achieve a BER of 1/1000 a signal to noise ratio of 12.5dB approximately is required for channel models D, E whereas for the highest speed mode this SNR power requirement increases to at least 30dB. The difference of 17.5 db is indicative of an increase in power levels by a ratio of 1:56. Analysis of the Power Delay Profiles driving the channel models reveal that model D and E have a powerful single path combined with a weaker multipath signal level. This phenomenon may result in lower interference and hence a better BER performance. Conclusion This paper has explored the simulation of a standard wireless communication system through a set of Channel Models adapted by the author. The simulation test bench for experimentation has been successfully implemented and results of simulation were analyzed. The results indicate that a complex combination of parameters (SNR, Coding Ratio, Modulation scheme and Bit Rate) govern the BER performance through realistic channel models. Channel Models D and E provide the best performance for a given SNR and can be exploited for effective infrastructural WLAN deployment. The specular component characteristic of Raician channel has not been considered as it was not within the premise of the original channel model development process as proposed be IEEE TGn Channel model group. The baseline results presented in this paper can be used evolve a complex figure of merit that integrates a realistic channel model to experiment with other OFDM based communication systems. The simulation test bench can also be extended to plan WLAN deployment. Acknowledgement The author would like to express sincere thanks to the Principal, Government College of Engineering & Technology, Jammu for financial support. 23

10 References [1] Gast MS Wireless Networks: The Definitive Guides, O Reilly, [2] IEEE Standard for Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) Specifications, ISO/IEC :1999E, August [3] J. Medbo and P. Schramm, Channel models for HIPERLAN/2, ETSI/BRAN document no. 3ERI085B. [4] J. Medbo and J-E. Berg, Measured radiowave propagation characteristics at 5 GHz for typical HIPERLAN/2 scenarios, ETSI/BRAN document no. 3ERI084A. [5] TGn Channel Models, IEEE Std /940r4, May, [6] Bahai, Ahmad R. S., and Burton R. Saltzberg, Multi-Carrier Digital Comunications: Theory and Applications of OFDM, New York: Kluwer Academic/Plenum Publishers, [7] Jeruchim, M. C., Balaban, P., and Shanmugan, K. S., Simulation of Communication Systems, Second Edition, New York, Kluwer Academic/Plenum,