| Popis: |
An intelligent robotic system must possess the capabilities to sense the environment, perceive the environment, reason on the environment and then act based upon the conclusions of the reasoning process. Autonomous robotic navigation is an intelligent process. Ergo, it must be able to sense, perceive, reason, and act. This thesis concerns the use of a scanning laser range finder as the sensing device for an autonomous navigation system. It concentrates on developing a perception process. Perception is the representation of sensed data that permits cognition. To understand the requirements of perception, a chapter is devoted to analyzing and reviewing contemporary navigation reasoning modules. An analysis of scanning laser range finder characteristics is then presented. With this background, an effective perception process is developed. The perception process is then incorporated with simulated laser scanning data and a neural network navigation reasoning module. The three modules display attractive attributes for autonomous navigation. |
