Indoor localization using infrared global vision system and LRF for twin brushes floor polishing robots

We have been developing floor polishing robots, which have two rotary brush polishers. These omnidirectional mobile robots require no driving for locomotion and steering because the friction between the floor and rotating brushes is used as the driving force. The localization for these robots cannot be used dead reckoning because water and washing liquid make wheels slip. This paper proposed the localization using the infrared global vision system and a laser range finder (LRF) without touching the floor. In the infrared global vision system, assuming the model that the infrared LED is a point light source, we calculated the center of a LED from a gray scale image. From the experimental result that the processing time was approximately 40(ms) and the accuracy was approximately 16(mm), a high-speed and accurate localization was achieved. And the method using a LRF was also achieved a high speed and precise localization by the comparison with the infrared global vision system. Furthermore, we proposed the method of the precise localization and the obstacle avoidance in all of a room by using these two methods at the same time.