ECE416 Progress Report A software-controlled fading channel simulator

Transcription

1 ECE416 Progress Report A software-controlled fading channel simulator Chris Snow Faculty Advisor: Dr. S. Primak Electrical/Computer Engineering Project Report (ECE 416) submitted in partial fulfillment of the requirements for the degree of Bachelor of Engineering Science Department of Electrical and Computer Engineering The University of Western Ontario London, Ontario, Canada November 8, 2002

2 1 Introduction Some progress has been made on the 416 project since the initial project proposal of September 23. Specifically, a literature review has been conducted, some initial design plans have been made, and some investigation of the mathematical methods of channel modeling has been done. A summary of the work and results thus far is presented in the sections below. 2 A Brief Overview of the Project A fading channel can be described as a communication channel in which the instantaneous received power, and hence the signal-to-noise ratio (SNR), fluctuates greatly from moment to moment. In wireless communications, fading often occurs because the receiver is in a non-line-of-sight (NLOS) position with respect to the tranmitting antenna. For this reason, only diffuse signals to arrive at the receiving antenna. Because the reflection and diffraction of radio waves is dependent on the position and geometry of a large number of objects between the transmitter and receiver, the movement of both the objects and the receiving antenna causes the received SNR to fluctuate with time. The central focus of this project is to somehow be able to cause a wireless receiver to see a fading signal power without having to set up the transmitter and receiver in a fading situation. For reasons explained below, this will be done by modifying the hardware of the transmitting antenna. A software package will allow control of the parameters of the simulated fading. 3 Literature Review A review of the literature has been conducted, focusing on two important areas: the mathematical modeling of the fading channel, and methods of simulating the channel. Because many channels 1

3 (e.g. the 800, 900, 1800, and 1900 MHz bands used in digital cellular communication) suffer from fading, there has been a great deal of interest in the mathematical modeling as well as the simulation of the fading channel [1, 2, 3, 4]. 3.1 Statistics of the fading channel Envelope distributions The probability density function (PDF) of the received signal envelope is an important part of fading channel modeling. This distribution is required for the calculation of the average probability of error and, along with the joint PDF, can be used to calculate the average duration of fades as well as other important statistics. Several classical PDFs are often applied in the study of the distribution of the fading channel envelope: the Rayleigh, Rician, and Nakagami-m PDFs. For lack of space, the actual mathematical formulae are left to be found in [1]. Briefly, the Rayleigh PDF for the envelope applies in a situation where many diffuse waves (and no direct ray) arrive at the receiver. In the case where there is a strong direct ray, the envelope PDF is given by the Rician distribution. The Nakagami-m PDF is a generalized probability density function with parameter m which can be varied to represent different amounts of contribution from direct rays and from diffuse components. Newer research has examined analytical models for received signals composed of two strong (direct) rays and diffuse components [5] Correlation function The classical work of Jakes [6] established that the correlation function of a mobile receiver is given by R(τ) = J 0 (2πf D τ), where J 0 ( ) is the Bessel function of the first kind [7], τ is the bit duration (or block duration if reception of blocks is being considered), and f D is the maximum Doppler frequency of the mobile, given by f D = vm λ c with v m being the velocity of the mobile and λ c being 2

4 the carrier frequency. Typical Doppler frequencies may be on the order of f D = 10Hz for pedestrian users, and f D = 100Hz for vehicular (cars, trains) users. This result is important because any fading channel simulator should attempt to match the statistics of the real fading channel as closely as possible. Many of the more mathematically tractable models do not replicate Jake s correlation function [8], and some academic debate has been waged over the issue of how much accuracy is necessary in modeling the correlation function [9, 10]. 3.2 Characterizing channel error sequences An overview of models for channels with memory is given by the classic paper of Kanal and Sastry [11]. The models given are useful for the characterization of the error process of the channel. These models can be used to generate bit error patterns which resemble those of a real channel. More recent work includes the use of Finite-State Markov Modeling [8, 12] and Hidden Markov Modeling [13]. One example of the application of a Markov model to performance studies is [14]. 3.3 Channel simulators Because of the ever-present development and deployment of new communication technologies, there is great interest in the simulation of mobile radio channels in software [15, 16] as well as hardware [17]. Generally speaking, some type of algorithm is devised for approximating the error conditions on the fading channel. The simulator can operate at the signal level (describing the instantaneous SNR), the bit level (making an error/no error decision for each bit), or the block level (making a decision about blocks at a time). The appropriate level of modeling depends on the type of study being performed. From the point of view of this project, signal and bit level channel models are the most relevant. 3

5 4 Progress to Date In order to simulate fading, either the output power of the transmitting antenna or the amount of power delivered from the receiving antenna to the receiver must be controlled. This project is intended as a proof-of-concept for a testing system for new mobile devices. Generally speaking, the most interesting measures of performance for such devices are in the downlink from the base station to the mobile station (i.e. with the mobile device is acting as the receiver). To allow for the study of a wide variety of mobile devices, the receiving side of the radio link (the mobile device under test) should have as few modifications as possible. For this reason, a method of controlling the output power of the transmitter is required. 4.1 Controlling output power In a practical application, the output power of the transmitter (e.g. a cellular base station) is controllable, and thus controlling the antenna output power requires no special attention. However, for our proof-of-concept design, the transmitter/receiver pair being used does not have a controllable output power setting, and thus another method is needed to vary the output power of the transmitter. While the scheme developed is somewhat basic, it should suffice as a proof of concept. The idea, as illustrated in Fig. 1, is to switch on or off a series of resistors connected in parallel with the antenna, thus causing some of the output power of the transmitter to be dissipated in the resistors, lowering the transmitted power. After having constructed the device, the power delivered to the antenna can be measured for the different combinations of resistors, and these values can be used by the software controller in order to select the proper combination of resistors to switch on for any desired antenna output power. This scheme is of course limited by the number of combinations of resistances available, and 4

6 Transmitter Transmit Antenna... FET Switches... Figure 1: Scheme for modifying the instantaneous output power of the antenna some distortion will obviously be introduced because of the mismatch in impedance between the transmitter output and the antenna. However, as long as the power delivered to the antenna can be measured for each combination, the software controlling the switching will be able to choose the correct subset of resistors to switch on in order to achieve the desired output power. 5 Future Work 5.1 Prototyping, Construction and Testing The next major task to be undertaken is the prototyping, construction, and testing of the transmitter and receiver. The transmitter, receiver and antenna have been acquired, and initial circuit designs are underway. The first step is to prototype the circuit on a breadboard. Once the prototype is working, a PCB can be designed and fabricated, and the final construction and testing can be done. 5

7 5.2 Software The software control will be done through Matlab. Once the transmitter and receiver hardware are ready, the characterization of output power described above must be done. With this complete, all required information for software development will be available. The software package will have a graphical front-end which will allow the user to select the average severity and duration of fades. From this, the control software must decide at what levels to set the output power of the transmitter, and send the appropriate signals to the hardware. At the receiving end, the software will read the data from the receiver hardware, and compare the received data with what was expected. The receiver will then display statistics about the quality of the link. 6 Conclusions The project is currently progressing well. No problems have arisen. Signature of student:

8 References [1] M. Simon and M.-S. Alouini, Digital Communication over Fading Channels: A Unified Approach to Performance Analysis. New York: Wiley-Interscience, [2] M. Zorzi, R. Rao, and L. Milstein, Error statistics in data transmission over fading channels, IEEE Trans. Comm., vol. 46, pp , Nov [3] R. H. Clarke, A statistical theory of mobile radio reception, Bell Syst. Tech. J., vol. 47, pp , [4] H. Suzuki, A statistical model for urban radio propagation, IEEE Trans. Comm., vol. COM- 25, pp , July [5] G. D. Durgin, T. S. Rappaport, and D. A. de Wolf, New analytical models and probability density functions for fading in wireless communications, IEEE Trans. Comm., vol. 50, pp , June [6] W. C. Jakes, ed., Microwave mobile communications. New York: Wiley, [7] M. Abramowitz and I. A. Stegun, eds., Handbook of mathematical functions with formulas, graphs, and mathematical tables. New York: Dover, [8] H. S. Wang and N. Moayeri, Finite-State Markov Channel a useful model for radio communication channels, IEEE Trans. Vehic. Tech., vol. 44, pp , Feb [9] C. Tan and N. Beaulieu, First-order Markov modeling for the Rayleigh fading channel, in GLOBECOM 1998, vol. 6, (Sydney, NSW, Australia), pp , Nov [10] C. Tan and N. Beaulieu, On first-order Markov modeling for the Rayleigh fading channel, IEEE Trans. Comm., vol. 48, pp , Dec [11] L. Kanal and A. Sastry, Models for channels with memory and their applications to error control, Proc. IEEE, vol. 66, pp , July [12] F. Babich and G. Lombardi, A Markov model for the mobile propagation channel, IEEE Trans. Vehic. Tech., vol. 49, pp , Jan [13] W. Turin and R. van Nobelen, Hidden Markov modeling of flat fading channels, IEEE J. Select. Areas Commun., vol. 16, pp , Dec [14] C. Snow and S. L. Primak, Performance evaluation of tcp/ip in bluetooth based systems, in Proc. VTC Spring 2002, vol. 1, (Birmingham, AL), pp , May [15] U. Dersch, R. Ruegg, H. Kaufmann, and R. Rufener, Modelling and simulation of indoor radio channels, in Proc. IEEE ICC 93, vol. 3, (Geneva), pp , May [16] C. Snow and S. L. Primak, On a filter based simulator of Isotropic Sattering Omnidirectional Receiving Antenna (ISORA) fading channel. Internal Note I, October [17] E. Casas and C. Leung, A simple digital fading simulator for mobile radio, IEEE Trans. Vehic. Tech., vol. 39, pp , Aug