Metric state space reinforcement learning for a vision-capable mobile robot

Zhumatiy, ViktorGomez, FaustinoHutter, MarcusSchmidhuber, Jürgen2015-12-089781586035952http://hdl.handle.net/1885/34046We address the problem of autonomously learning controllers for vision-capable mobile robots. We extend McCallum's (1995) Nearest-Sequence Memory algorithm to allow for general metrics over state-action trajectories. We demonstrate the feasibility of ourCopyright Information: © 2006 The Author(s).Keywords: Computer models; Discretizations; Learning controllers; Partial observability; Piecewise-continuous; Sensor inputs; State space; Algorithms; Reinforcement learning; Mobile robots Mobile robots; Reinforcement learningMetric state space reinforcement learning for a vision-capable mobile robot20062016-02-24