期刊名称：Lecture Notes in Engineering and Computer Science
印刷版ISSN：2078-0958
电子版ISSN：2078-0966
出版年度：2018
卷号：2233&2234
页码：29-34
出版社：Newswood and International Association of Engineers
摘要：This paper presents an intelligent navigation
system to generate safe motion for a mobile robot. It consists of
two main modules of recognizing obstacle and making decision
with artificial intelligence, respectively. Firstly, a recognition
algorithm using laser range finder (LRF) and robot odometry is
developed to detect objects near the robot. The proposed
recognition algorithm provides both position and velocity
information of obstacles by using range data accumulated for a
certain time period. Then, the result is used for computing a
safe moving direction of the robot to avoid collision with objects.
For that, an artificial intelligence algorithm of multi-layered
neural network was designed and trained by deep learning
method with many data sets of information including pairs of
sensor data and its solution of motion command for various
situations.
关键词：Artificial Intelligence; Deep learning; Mobile
robot; Collision avoidance
