Fuzzy Logic for Fusion of Tactical Influences on Vehicle Speed Control

Abstract

An important aspect in autonomous vehicle navigation is determining and maintaining proper vehicle speed. For example, planetary rovers need to navigate at the highest velocity possible, yet be able to slow down in order to maneuver safely around obstacles and over various types of terrains. In the case of automated highway driving, vehicle safety can be endangered by going around a curve too fast and crossing into the oncoming lane, traveling too fast to avoid hitting nearby vehicles if they suddenly stop or change lanes, or speeding down hills. Ideally, an autonomous vehicle would be able to sense and adapt to these conditions on-line, just as human drivers do. In addition, the robot vehicle could take advantage of external sources of information broadcasted to its on-board computer, such as terrain data and reports of changes in the local speed limits due to weather, construction, and traffic conditions. Speed control is also an important concern for teleoperation of planetary rovers and for auto-pilot highway guidance systems. These systems assist, rather than totally replace, the driver. Therefore, it is desirable that a solution for speed control in an autonomous vehicle be compatible with the one required for semi-autonomous systems as well.