Position Estimation and Autonomous Travel by Mobile Robots in Natural Terrain

In this chapter, we will address two senses of the term “mobile robot navigation:” (1) finding a position or a course; and (2) travelling from place to place. In simplistic language, these two meanings correspond to answers to the questions “Where am I?’ and “How do I get there?’ Although the robotics community has devoted significant effort to unattended air and water vehicles, in this chapter we restrict attention to ground robots. One reason for this restriction is that the majority of research and development efforts have concentrated on land vehicles, so we have learned more about ground-travelling robots than about flying, gliding, swimming, or floating robots. Another reason for the restriction is that, in general, the density of obstacle fields is greater on the ground than in either air or water, and as a consequence, mobile robots on the ground face the more challenging navigation problems. Mobile robotics is a young field, having started, by most accounts, in the late 1960s with work on Shakey at SRI and the Cart at Stanford. In the three decades since the field’s inception, an estimated 100 to 1,OOO mobile robots have been built and operated. In the process, many lessons have been learned and many problems have been solved, enabling mobile robots to achieve such feats as driving 140 km/h on the Autobahn, rappelling into an active volcano, and driving on highways from Pittsburgh to San Diego.