Global Tracking Control of Strict-Feedback Systems Using Neural Networks

摘要: Most existing adaptive neural controllers ensure semiglobally uniform ultimately bounded stability on the condition that approximation remains valid for all time. However, such a is difficult to verify beforehand. As result, deterioration of tracking performance or even instability may occur in real applications. A common recourse activate an extra robust controller outside active region pull back transient. Such approach, however, has been restricted dynamic systems with matched uncertainty. We extend it strict-feedback mismatched uncertainties via multiswitching-based backstepping methodology. Each virtual and actual proposed design switches between controller, switching algorithm being sufficiently smooth and, hence, able be incorporated tool. The overall ensures globally ultimate boundedness while simultaneously avoiding possible control singularity. Simulation results demonstrate validity designs.