Design of Multi-Sensor Switching Path-Planning Control System for Target Tracking and Obstacle Avoidance of Mobile Robot

• Main Authors: You-Wei Lin, 林有為
• Other Authors: Rong-Jong Wai
• Format: Others
• Language: en_US
• Online Access: http://ndltd.ncl.edu.tw/handle/w5s7ya

博士 === 元智大學 === 電機工程學系 === 103 === Abstract The purpose of this dissertation is to design a multi-sensor switching path-planning control system for target tracking and obstacle avoidance of a mobile robot. This dissertation focuses on the combination of multi sensors and a robust dynamic petri recurrent-fuzzy-neural-network (DPRFNN) framework to achieve superior trajectory tracking performance and adaptive path planning for obstacle avoidance. The first part of this dissertation derives the dynamic model of a vision-based mobile robot via a tilt camera for capturing the real position information to take the place of wheels sensors. The mobile robot with the feedback signals from the camera can track moving objects and solve the conventional slip problem to significantly enhance the practical application level. The second part of this dissertation designs a robust DPRFNN scheme to manipulate the mobile robot for the target tracking via a tilt camera. In the robust DPRFNN, the concept of a petri net and the recurrent frame of internal feedback loops are incorporated into a traditional fuzzy neural network (FNN) to alleviate the computation burden of parameter learning and to enhance the dynamic mapping of network ability. The third part of this dissertation investigates a multi-sensors scheme including a single camera with its ability to track moving objects and sonar sensors with their ability to measure the distance between the obstacle and the mobile robot. Thus, the robot can gradually approach its object without detailed environmental information, large memory size and heavy computation burden. This dissertation is aimed to achieve the following three objectives: (1) The development of a DPRFNN framework for the trajectory tracking to alleviate the computation burden of parameters learning without the requirement of detailed system information and auxiliary compensation controllers. (2) The application of a robust DPRFNN scheme with camera feedback signals to a mobile robot for tracking moving objects and solving the conventional slip problem. (3) The integration of camera and sonar sensors into a mobile robot to achieve the goals of moving object tracking and dynamic obstacles avoidance. Numerical simulations and experimental results are used to verify the effectiveness of the proposed intelligent trajectory tracking and path planning scheme for a mobile robot. Keywords－Trajectory tracking, path planning, obstacles avoidance, dynamic petri recurrent fuzzy neural network, robust control.