Path Planning of a Mobile Robot as a Discrete Optimization Problem an Adjustment of Weight Parameters in the Objective Function by Reinforcement Learning

Authors:
Harukazu Igarashi
Affiliations:
-
Venue:
RoboCup 2000: Robot Soccer World Cup IV
Year:
2001

Citing 5
Cited 0

Gross motion planning—a survey

ACM Computing Surveys (CSUR)
Simple Statistical Gradient-Following Algorithms for Connectionist Reinforcement Learning

Machine Learning
Artificial intelligence and mobile robots: case studies of successful robot systems

Artificial intelligence and mobile robots: case studies of successful robot systems
Map learning and high-speed navigation in RHINO

Artificial intelligence and mobile robots
Active mobile robot localization

IJCAI'97 Proceedings of the Fifteenth international joint conference on Artifical intelligence - Volume 2

Quantified Score

Hi-index	0.00

Visualization

Abstract

In a previous paper, we proposed a solution to path planning of a mobile robot. In our approach, we formulated the problem as a discrete optimization problem at each time step. To solve the optimization problem, we used an objective function consisting of a goal term, a smoothness term and a collision term. This paper presents a theoretical method using reinforcement learning for adjusting weight parameters in the objective functions. However, the conventional Q-learing method cannot be applied to a non-Markov decision process. Thus, we applied Williams's learning algorithm. REINFORCE, to derive an updating rule for the weight parameters. This is a stochastic hill-climbing method to maximize a value functions. We verified the updating rule by experiment.