Occupancy grid mapping using artificial neural networks

Robot environment map is used for the path planning. Sonar Sensor measurement is converted to the probabilistic representation of the environment (occupancy grid). In this paper we used sonar sensor model to identify the different regions of the robot on the floor (cells). We used the grid mapping to describe the emptiness and occupancy and the unknown areas of the robot environment and the nearby region (around the robot). But the environment changes affect these conversions. Artificial neural network (ANN) is introduced which learns these conversions. The conversion of sensor data remains adaptive to changes in either the sensor or its environment.