OPPORTUNISTIC COMMUNICATIONS AT THE DOWNLINK OF MOBILE SYSTEMS IN THE PRESENCE OF INACCURATE CHANNEL INFORMATION Banafsheh Lashkari 1 , Mehrzad Biguesh 1 , Saeed Gazor 2 1 Wireless Comm. Lab, Elect. Eng. Dept., Shiraz University, Shiraz, Iran 2 Dept. of ECE., Queen’s University, Kingston, Ontario, K7L 3N6 Canada ABSTRACT The throughput of a multi-user broadcast opportunistic com- munications system is highly affected by the accuracy of the available channel state information (CSI) at the transmitter side. It is a questionable assumption to have almost perfect knowledge of the user SNRs at the transmitter due to the error in the channel measurement, or due to reporting feedback errors from users to the transmitter. As a result such errors, the system can not gain its maximum throughput. In this paper, we investigate the performance of opportunistic communications system considering imperfect SNR reports. We show that in order to maximize the throughput, the transmitter needs to estimate the channel state information. The imposition of a correlation constraint will prevent the estimate of channel from fluctuating too widely. Employ Kalman filter, we estimate the channel state information and show that the throughput is significantly improved. I. INTRODUCTION The quality of wireless communication is impacted by fad- ing which is the result of multi-path propagation of signals. Various diversity techniques are proposed to combat fading. The fundamental idea behind the diversity is to transmit signals via different and independent dimensions, such as time, frequency or space [1]. A type of diversity, so-called multiuser diversity is used in multi-user systems. The multi- user diversity is best motivated from an information theory result of Knopp and Humblet [11]. Knopp and Humblet in [11] focused on the uplink in a single cell system with multiple users, and showed that the uplink capacity could be maximized by allowing only one user with the best channel to transmit the data at each given time. In [6] similar results are reported for the downlink from base station to the mobile users. In practical communication systems, the traffic demand is generally higher in the downlink communication than that of uplink. Thus, it is very desirable to maximize the throughput of the downlink [3]. To achieve this goal, many algorithms have been proposed which require some accurate feedback about channel state information from receiver to the transmitter. Such an accurate feedback is not always possible in practice, especially when the number of users is large. As a further consequence of such imperfect feedback, a significant loss can be imposed on the performance of the opportunistic communications system. In this manuscript, we study the robustness of the oppor- tunistic communication system against the erroneous SNR feedback reports. We also employ a channel state estimator in order to enhance the inaccurate reports. Our results show that the capacity of the system can be significantly improved. II. SYSTEM MODEL We consider a multi-user downlink with K single antenna users and assume that the channels are narrow-band and that the channel gains are independent and have identical Rayleigh distribution. The transmitter has to send a private data to each user in this system. The decision to transmit data for a user in this system is made based on the channel state information (CSI) of all users. To measure the channel quality, base station broadcasts some foreknown training signal to all users. The base-band received signal by the kth user at time instant n is written as, y k (n)= h k (n)s(n)+ v k (n) (1) where h k (n) is the complex Gaussian channel between transmitter and the kth user, s(n) is the transmitted symbol selected from constellation C , and v k (n) denotes the white zero-mean complex noise process with variance σ 2 k . For the sake of simplicity, we assume that the noise power is the same for all users, that is σ 2 k = σ 2 for k =1,...,K. The average power of transmit symbols is also denoted by E{|s(n)| 2 } = σ 2 s . We assume that h k (n) is smoothly time varying and is almost constant over P successive time instances, i.e., h k (n) h k [m] for all n ∈{mP, mP +1, ··· (m + 1)P }, where P is a given integer known as the channel coherence time. Thus, the SNR for the kth user during such time interval is expressed by SNR k (n) = E{|h k (n)s(n)| 2 } E{|v(n)| 2 } = h k [m] 2 σ 2 s σ 2 . (2) 978-1-4244-1946-3/08/$25.00 ©2008 IEEE 356 24th Biennial Symposium on Communications