Research On Error Exponents And Transmission Strategies In Multiuser Information Theory

With the development of society, the demand for high data rates and different quality-of-service is growing rapidly. Improving the throughput and reliability have been the evolutional directions of wireless communication systems. As the theoretical foundation of broadband wireless communication networks, multi-user information theory can capture the aspects of information flow in real-world multipoint-to-multipoint communication networks (Multiple sources, multiple destinations, multi-access, broadcasting, interference, cooperation, feedback, etc.). Multi-user information theory is devoted to investigate the information theoretic capacity limits in networks where multiple sources and destinations interactive with each other. Therefore, the multi-user information theory has become an important research topic in recent years.This dissertation mainly focuses on cognitive radio channels (also called cognitive interference channels) and multi-source relay broadcast channels. The main contribution of our work lies in deriving several achieve rate regions under the corresponding transmission strategies, and showing the relationship among the communication reliability, transmission rate and the code length from an error exponent perspective.Firstly, the achievable rate regions of the four-user discrete memoryless cognitive radio channel are obtianed by rate splitting, superposition coding, Gel’ fand-Pinsker coding and simultaneous decoding strategies. According to the different encoding and decoding strategies, the Gel’fand-Pinsker coding is only applied to cancel the private messages of the non-cognitive users at the cognitive transmitters in the case that public messages can be decoded by all the receivers, and the Gel’ and-Pinsker coding is applied to cancel all the messages of the non-cognitive users at the cognitive transmitters in the case that public messages can not be decoded between cognitive receivers and non-cognitive receivers. Also, we show that the region is identical to the existing results via fixing some auxiliary random variables for the the four-user discrete memoryless cognitive radio channel in which the public messages can be decoded by all the receivers.Secondly, the error exponent for the cognitive radio channels is investigated, in which the superposition coding, Gel’ fand-Pinsker coding and maximum likelihood decoding are considered. Under this model, we derive the error exponent region for the discrete memoryless cognitive radio channel with the inequality lemma and Gallager’s bounding technique. Also, the error exponent region for the Gaussian cognitive radio channels under choosing the special parameter case is analyzed.Thirdly, this dissertation investigates the capacity of the degraded multi-source relay broadcast channels, where each source wants to communicate individual messages with its intended destination via both direct and relay links. An achievable rate region of this channel is obtained, in which the block Markov regular encoding and the sliding-window decoding transmission strategies are exploited. Also, the outer bound is derived via using the Fano’s inequality, the max-flow min-cut theorem and condition entropy lemma.Finally, the error exponents of the two-hop multiple source-destination relay channels are obtained. Two relay modes are mentioned:amplified and forward (AF) and decode and forward (DF). An achievable rate region can be obtained by the rate splitting strategies which each source can split its message into private and public parts. We determine the system error probability via integrated exponent function under AF relay strategies. The most significant difference between AF and DF system error probability evaluations lies in a minimized cut-set bound of transmission rate under the DF strategy. There are many cases among transmission rate intervals for different system parameters (e.g. transmit power) and we derive the system error probability for the DF strategy. Moreover, in order to draw deep insight into the reliability requirement of each source node in this network, we provide the error exponent region for different source nodes to show the trade-off of error probability among source nodes.