Research Of The Adaptive Control Based Of Rough Reasoning And Its Robustness

Rough reasoning can be extracted for more complex data rules, and adaptive controloptimization rule extraction is very difficult. This paper studied the rough set improvedthe robustness of adaptive control and numerical simulations are conducted.Firstly, this paper studied the Rough reasoning reduction control rules. The maintask of Rough reasoning is approximate classification, knowledge reduction, attributedependency analysis, according to the decision table to produce optimal or suboptimaldecision control algorithm. The simulation results show that the Rough reasoning is veryeffective for reduction of decision rules. Rough reasoning provides the basis for thisstudy improving rule adaptive control algorithm.Secondly, this paper studied the robust adaptive control algorithm. Compared withother correction method it does not make the system performance degradation. Theadaptive law can guarantee the asymptotic stability of the tracking error and the signal inthe circumstance of no interference. Simulation results verify the robustness andeffectiveness of the proposed control strategy.Finally, this paper study presents a new algorithm of robust adaptive neural networkbased on Rough reasoning. The new algorithm uses the Rough reasoning for data analysisand rule extraction method, extraction rules from data is mapped to a robust adaptiveneural network control input output subspace. And then using the robust adaptive neuralnetwork is trained to approximate the subspace.Simulation shows that the robust adaptive neural network Rough inference controlpattern recognition effectively reduce the training samples of the neural network basedrules, reduces the training steps and time. Because the Rough reasoning can filter theredundant data information. The algorithm has better capability of generalization of thenoise data and stronger robustness.