Catalogue of Artificial Intelligence Techniques
Author(s): Ashley Walker
Robot navigation is a multi-faceted subject for which little general theory has been formulated, but many empirically derived solutions exist. Navigation systems can be built with elaborate model-based controllers or simple reactive behaviour based controllers (i.e., controllers using distributed, non-symbolic representations). The term is currently applied to mobile robots as well as assembly arms and usually involves a number of sensors (and sensor fusion techniques) and one or more actuating mechanisms. Traditionally, navigation systems were responsible for enabling the following behaviours: (i) obstacle avoidance, (ii) localisation (i.e., position determination), and (iii) path planning. Early research robots and many present-day industrial robots navigated with reference to an a priori specified map or a system of artificial beacons and signposts distributed through the environment. Recent interest in navigation systems for use in dynamic and unknown environments has caused the task of map building and maintenance to be considered as an essential navigation behaviour.
- Mataric, M.J. , Integration of representation into goal-driven behaviourbased robots
IEEE Transactions on Robotics and Automation, 8 (1992), 304--312.