Position compensation algorithm for omnidirectional mobile robots and its experimental evaluation

Abstract

Though the basic objective of mobile robots is movement between positions, they have a drawback that position accuracy is not completely guaranteed because of the slip between the wheels and ground. Various strategies employing additional sensors have been adopted to decrease the position error. In this research, multiple ultrasonic distance sensors were used to measure the position and orientation of the mobile robot, and a position compensation algorithm was developed to minimize the position error between the current position and the desired position. In contrast to the conventional mobile robots with wheels that require at least a few steps of path planning to approach a designated position, the Mecanum wheel adopted in this research has a unique structural property that imparts omnidirectionality to the mobile robot; hence, the mobile robot can immediately move in arbitrary directions without any rotation of the body. This enables the proposed algorithm to directly eliminate the position and orientation errors without any complicated path planning. The algorithm was experimentally validated. The performance of the proposed algorithm applied to omnidirectional wheels was experimentally compared to that of independent dead-reckoning control with general wheels from the viewpoints of position accuracy and movement time.