UBC Theses and Dissertations
Cross-layer resource scheduling for wireless systems over correlated fading channels Karmokar, Ashok Kumar
Packet scheduling is very important in future wireless networks due to their limited resources and increasing demand for high data rates. Time-varying incoming traffic and channel gains make the scheduling decision very challenging. Due to the inherent dynamic nature of packet scheduling, we formulate the scheduling problem as a Markov decision process. We consider a single user communicating over a correlated fading channel. The incoming traffic is randomly varying and stored in the finite buffer before transmission. We formulate the scheduling problem from a cross-layer viewpoint, by considering both the physical and the data link layer optimization objectives. Our objective is to maximize throughput, and minimize power, buffering delay, packet overflow and bit error rate. First, we present optimal and suboptimal packet scheduling over Rayleigh fading channels. We analyze the problem using both the information-theoretic and the multilevel modulation transmission models. Two different ways of computing optimal policies are given and their benefits and drawbacks are discussed. The performance of the suboptimal scheduler is compared with that of the optimal scheduler. On the top of employing a diversity technique at the receiver, we also show that by adapting packet transmission across different layers the scheduler can save a significant amount of power. The rate adaptation problem is then extended for schemes with a multiple-input multiple-output channel and a selective-repeat automatic repeat request protocol. Second, we develop several adaptation techniques for type-I hybrid automatic repeat request schemes. We investigate the rate and power adaptation problem for two cases: when both the perfect channel state information and observation feedback are known, and when only latter is known. We analyze the adaptation problem for both flat-fading and frequency-selective channels. The policy heuristic-based solution for coding rate adaptation is later given for these schemes in the case of no perfect channel state information, and then extended for joint coding rate and modulation order adaptation. Finally, cross-layer scheduling for the incremental redundancy hybrid automatic repeat request system is given using rate compatible punctured code. We propose three different adaptation models and compare their performances.
Item Citations and Data