Navigation system based on ceiling landmark recognition for autonomous mobile robot -position/orientation control by landmark recognition with plus and minus primitives

摘要

A template matching is a popular method for visual recognition. It costs much time to detect the target object. We have applied fuzzy and neural network algorithm to solve this problem. However, the problem to mistake recognizing the target object is not solved. Hence, in this paper, we propose robust visual recognition algorithm for the autonomous robot navigation. The proposing algorithm employs a fuzzy template matching with the target object as well as the similar objects in the environment. As a result, the robot verifies the tentative recognition by comparing the likelihood of plus elements (target objects) with that of minus elements (similar objects). To integrate this algorithm into the robot navigation system, we propose a hierarchical control architecture. The validity of the proposing navigation system is shown with experimental results.