Robot localisation using interval analysis

This paper describes a deterministic approach for the sensor-based localisation and navigation of a mobile robot. This approach is based on interval analysis and the robot equipped with ultrasonic sensors. For the localisation, it is assumed that the map is two-dimensional and also it is assumed to be known a-priory to the robot. It has already been shown by Jaulin et.al. that mobile robot localisation and tracking using interval analysis and an interval model of the robot, with ultrasonic sensors only can be achieved. Here we use the same algorithm for robot localisation but without using an interval model of the robot. Instead the physical limitations of the robot is used to predict and track the robots position. In classical methods such as Kalman filters for robot localisation, the data association step is very complex and they are based on linearisation. Where as the method proposed here using interval analysis bypasses the data association step and deals with the problem as nonlinear and in a global way.