Power control for a mobile satellite system

This thesis describes the design and performance of closed loop power control algorithms for a mobile satellite system. A two-loop control strategy is considered. It consists of a fast-acting 'inner' loop which controls received signal strength to a desired setpoint Pdes (dB) and a slow-acting 'outer' loop which adjusts the setpoint in order to maintain a desired code-word error rate at the receiver, WERdes.A variety of 'inner' loop controllers were designed. The Predictive 'inner' loop controller forward predicts the channel gain disturbance and then issues appropriate power requests to the transmitter. The linear phase-lag controller attempts to compensate the channel gain disturbance by adopting a suitably large open-loop gain. Phase-lag compensators based on pure Integral, Ragazzini and Smith's design methods were considered. The Variable Step Integral (VSI) controller is shown to be the best choice phase-lag compensator in terms of quality of compensation and control filter complexity. The Predictive controller yields superior compensation at LEO and MEO delays whilst the VSI controller is better suited at GEP delay. The Integro-Predictive (IP) controller is a novel arrangement of a VSI filter followed by a forward predictor. The predictor allows the VSI filter to adopt a comparatively larger open-loop gain without affecting closed loop stability. The performance of the IP controller is akin to that of the Predictive controller at LEO and MEO delays whilst it is similar to the VSI controller to GEO delay.In addition, two 'outer' loop control algorithms were designed to maintain WERdes at the receiver. The adaptive step (AS) algorithm updates the setpoint at fixed intervals but by a variable amount. In contrast, the fast response (FR) algorithm updates the setpoint by a fixed amount but at variable intervals. Simulation results show that power control with FR 'outer' loop algorithm is best suited to yield WERdes. It is established that for a given WERdes, the power costs increase with roundtrip delay. Hence, transmit power consumption at LEO delay is notably smaller than that at MEO or GEO delays. The uncorrelated diffuse multipath component of the fading channel limits the performance of the control scheme. So the transmit power requirements increase as the diffuse multipath component of the fading channel becomes increasingly dominant. As a result, the transmit power consumption increases as a mobile user travels from a rural to an urban surrounding.