Research On Robot Learning System Based On Artificial Immune Mechanism

As we know, computation theories inspired from natural phenomena are being paid more and more attention in the domestic research field. Different from traditional methods, and with the condition of admitting the existent means the logical, naturally heuristic methods always try to find the order of nature discovered by now, and then regard them as the rule or principles of constructing novel computer structures or algorithms with high performance. In which, a typical example is the study and the development on the research of artificial immune system (namely, AIS for short).The nature immune system, as a parallel, distributed and self-adaptive information processing system with high intelligence performance, provide a new way to deal with real time problem. Study on how to perfect artificial immune models reported, establish new ones, and investigate their theory and applications through sufficiently excavating, and using abundant resources of the immune system for reference, has become important research contents and main development trend of artificial immune systems in artificial intelligence.Under this background, the paper firstly uses the cooperated function principle of B cell, kinds of T cells and related assistant construction in the natural immune system for reference, puts forwards a new artificial immune network model, and have a study for the dynamics feature of its three-rank and four-rank system based on that model, separately discussing the influence of the feedback feature of transconductance matrix TH, leukocyte feature and assistant TS cell feature for network dynamics feature behavior. The result of simulation study illustrates that, in the network of interconnected four units, system can have the phenomenon of limited cycle and chaos in the same time. Meanwhile, it shows that there are close connections between the dynamics feature of network and the intension, type and transconduntance of assistant Tcell.Through the analysis of network, we can see that network will achieve to stabilization or balance through the function of antibody's multiplication weakening antigen gradually. That is, system's final balanced state is gradually stable, so the paper analyzes the stabilization of non-feedback network using Lyapunov method. Then using Lyapunov method, the paper discusses the stable problem of new network when the counteractive feature of assistant cell groups belongs to non-linear function.