Rapid modelling of the target tracking loop of a passive IR guided SAM, using artificial neural systems

[摘要] ENGLISH ABSTRACT: This thesis deals with the modelling of the target tracking system of a specific passive infrared guidedsurface to air missile (SAM). Modelling is accomplished by making use of rapid modelling principlesand procedures in order to establish whether an accurate target tracking model may be constructedwhilst at the same time'keeping construction time to an absolute minimum. The specific rapid modellingtechniques used during the course of the study centre around the application of hybrid artificial neuralstructures in an attempt to adequately represent the missile's target tracking control system.Various criteria are developed, whereby the target tracking control system features and structures ofthis type of missile may be extracted through minimal analysis and disassembly of the missile. Timerequirements are thereby minimised. The tracking loop is characterised by making use of signalinjection into the missile electronics. The eventual construction of a model is discussed as aprogression, beginning with the development of a fully artificial neural structure, and developingtowards a fully optimised hybrid neural model.The final model's eventual output results are analysed according to error energy, statistical principlesand frequency content. Finally, a comprehensive hardware in loop (HIL) simulation is completed,whereby the model replaces the missile's target tracking electronics under real-time physical testconditions, while tracking a true infrared target. The model's response compares favourably to the truemissile's behavior during HIL proceedings. Together with the favourable theoretical analysis of themodel, this illustrates that rapid modelling of the tracking system may be successfully completed whenapplying the techniques developed in this thesis.

[发布机构] Stellenbosch University