Improved Occupancy-grid Map Building and Updating for Robot Navigation

초록

In the map building, the robot navigates and collects environment information into a database automatically. Then, an off-line procedure is performed to build an initial environment model. Also, this model is improved by a modified Bayesian probability updating rule. In the travelling stage, the robot moves along the learned environment automatically, locates itself by map matching, and preserves necessary informatioin for map updating. These procedures are implemented and tested using an Pioneer 2-DXE and primary results are presented in this paper.