Robotic path planning and obstacle avoidance: A neural network approach
Norwood, John David
Cheatham, John B., Jr.
Master of Science
Robotic path planning and obstacle avoidance has been the subject of intensive research in recent years. Most solutions to this problem have been reached through the use of traditional Artificial Intelligence search techniques. However, these methods have proven inadequate when applied to highly unstructured or unknown environments. By using an Artificial Neural Network, one can generate near optimal paths using only low level information about the scene. In this way, it is possible to navigate from a start position to a goal position while avoiding all obstacles. Major advantages of the method presented herein are that the solution is very fast and does not rely on any a priori knowledge of the robot's environment. The system presented herein has proven very effective for path generation when used in conjunction with a simulated Laser Imaging System.
Mechanical engineering; Computer science; Artificial intelligence