Recognition Of 3-dimensional Environment For A Mobile Robot Using Structured Light

Abstract

In this paper, we develop a robust and simple structured light sensory system which is composed of a CCD camera and a laser scanner. The laser scanner projects a horizontal stripe and vertically rotates the stripe by a fixed step to cover thc whole scene. At each vertical angle of the stripe, the stereoscopically located camera views the stripe pattem, and the 3-D position values of the stripe are computed. Since mobile robots usually handlc hugc amount 01 range data, we propose a data manipulation tcchniquc and an effective modeling method to reduce the time required for path planning and necessary memory capacity. We implement these methods to a personal computer to show the feasibility of real-time navigation. Experiments in diverse situations show the effectiveness of the structured light sensor and relevant data manipulation methods for a mobile robot.