Multiplexing schemes and ARQ protocols for multiple-receiver systems

The throughput performances of some multiplexing and ARQ schemes are evaluated for a multiple-receiver system in which the quality of the channels varies over time. Packetized data are sent to the receivers from a server (transmitter), which uses a multiplexing scheme to assign bandwidth, as well as an ARQ protocol for error control. Two multiplexing schemes, round-robin multiplexing and an adaptive scheme, are considered. In round-robin multiplexing, each receiver is served periodically. With the adaptive scheme, the transmitter selects at each slot, the receiver whose channel quality is judged to have the highest probability of successful data packet reception. These multiplexing schemes are evaluated in conjunction with the three standard ARQ schemes, stop-and-wait, go-back-N, and ideal selective repeat ARQ. The throughput analysis takes into account the effects of feedback errors and imperfect channel state estimates. A scheme for reducing the unfavorable effects of feedback errors on the performance of continuous ARQ protocols is proposed and analyzed for a point-to-point stationary channel. In this scheme, the transmitter may not immediately retransmit a packet that has timed out but whose status is unknown due to lost feedback. Instead, the transmitter may delay retransmission for up to a given number of slots (the “retransmission delay parameter”), sending new packets in the mean time while waiting for successfully received feedback. The analysis considers the case where complete information of the receiver state is fed back to the transmitter. Under this assumption, it is shown that the effects of feedback errors can be greatly reduced when the retransmission delay parameter is optimized. This scheme is extended to the multiple-receiver system as well. Also discussed is a procedure for maximizing the effective data transmission rate of an ARQ system under certain bandwidth and power constraints. The procedure provides a means for jointly optimizing system design parameters such as the packet length, as well as coding and modulation schemes. The principles behind this procedure may also be useful for evaluating and optimizing the data transmission rate performance of the multiple-receiver system.Applied Science, Faculty ofElectrical and Computer Engineering, Department ofGraduat