Dynamic collision avoidance of multiple moving objects

• Main Author: Hoe, Chin Lun.
• Other Authors: Pang Wee Ching
• Format: Final Year Project
• Language: English
• Published: 2012
• Subjects: DRNTU::Engineering::Mechanical engineering::Mechatronics
• Online Access: http://hdl.handle.net/10356/50428
• Institution: Nanyang Technological University

Over the past decades, the demand for autonomous robots had increased in different areas to replace humans for difficult and tedious tasks that required high precision and high speed in harsh conditions and harmful environments. Autonomous robots desired for the ability to navigate in a dynamic environment without colliding with obstacles. Minimum Potential Field method and A-star Search method were developed to enable an autonomous robot to avoid collision with moving obstacles and stationary obstacles. Besides, limitations and restrictions for Minimum Potential Field method and A-star Search method which were applied in dynamic obstacle avoidance were resolved. By using GLUT, virtual work space and virtual entities were established. Efficiency of Minimum Potential Field method and A-star Search method was evaluated through five sets of virtual simulation. Average path length, average number of retraction and average execution time were used to analyze the efficiency of Minimum Potential Field method and A-star Search method. Minimum Potential Field method and A-star Search method are efficient in different aspects. Minimum Potential Field method is efficient for a moving entity to avoid obstacles and reach its destination in short execution time. A-star Search method is efficient for a moving entity to avoid obstacles and reach its destination in a short path generated. Mix reality simulation was suggested to examine performance of Minimum Potential Field method and A-star Search method in a real robot. Application of Minimum Potential Field method and A-star Search method in a three dimensional dynamic obstacle avoidance problem was highlighted.