Kernel-Based Approximate Dynamic Programming for Real-Time Online Learning Control: An Experimental Study
Kernel-Based Approximate Dynamic Programming for Real-Time Online Learning Control: An Experimental Study
Xin Xu,Chuanqiang Lian,L. Zuo,Haibo He
2014 · DOI: 10.1109/TCST.2013.2246866
IEEE Transactions on Control Systems Technology · 引用数 70
TLDR
It is shown that both online learning control schemes are effective for approximating near-optimal control policies of nonlinear dynamical systems with model uncertainties and it is demonstrated that KDHP can achieve better performance than conventional DHP, which uses multilayer perceptron neural networks.