Title :
Fuzzy target acquired by reinforcement learning for parking control
Author :
Yasunobu, Seiji ; Matsubara, Tomoya
Author_Institution :
Inst. of Eng. Mech. & Syst., Tsukuba Univ., Ibaraki, Japan
Abstract :
In this paper, a flexible action deciding method under various situations like a human being based on the concept of "fuzzy target" is proposed, and it applied to parking control. The "fuzzy target" is acquired by reinforcement learning, and it contains the grade of value of various targets. Fuzzy target is applied to intelligent parking control system for nonholonomic vehicle. The simulation results show the effectiveness of the proposed method.
Keywords :
digital simulation; fuzzy control; intelligent control; knowledge acquisition; learning (artificial intelligence); predictive control; traffic control; vehicles; digital simulation; flexible action deciding method; fuzzy target; intelligent parking control; knowledge acquisition; nonholonomic vehicle; predictive control; reinforcement learning;
Conference_Titel :
SICE 2003 Annual Conference
Conference_Location :
Fukui, Japan
Print_ISBN :
0-7803-8352-4
