Development of an autonomous-drive personal robot "An environment recognition system using image processing and an LRS"

We are developing an autonomous personal robot able to perform practical tasks in a human environment based on information derived from camera images and an LRS (a laser range sensor). It is very important that the robot be able to move autonomously in a human environment, and to select a specific target object from among many objects. For this reason, we developed a system by which these functions would become possible. This environmental recognition system is composed of an autonomous driving system and an object recognition system. First, the autonomous driving system calculates the driving route from the visual information provided by the CCD camera. The robot is driven by this system. The object recognition system proceeds by identifying the specified object using image processing and an LRS. The robot can grasp the object using this system. An environment recognition system is essential to both of these functions. Here we explain the algorithm by which the robot recognizes the surrounding environment. In addition, we apply this system to the robot, evaluate its performance and discuss our experimental results.