No abstract

Traditional bearings-only measurements (BOM) passive target tracking methods have an intrinsic shortage, that is, it depends on the established target models excessively. In order to solve the problem, a novel RL-based BOM tracking method is proposed. First, a reinforcement learning (RL)-based BOM target tracking framework is established, and sensor actions and rewards function are properly defined. Then, based on the typical Q-learning techniques, and Cerebellar model articulation computer (CMAC) method, a novel BOM tracking algorithm is proposed. Finally, a simulation example is provided to show effectiveness and efficiency of the proposed passive target tracking method. Keywords: reinforcement learning, Q-learning, target tracking, bearings only Classification: Navigation, Guidance and Control Systems IEICE Communications Express, Vol.5, No.1,[19][20][21][22][23][24][25][26] nonlinear transformation of means and covariance in filters and estimators," IEEE Trans. Automat. Contr., vol. 45, no. 3, pp. 477-482, 2000

No abstract