Environment Mapping Using Infra-Red Sensor Data and Probability Rules

This paper presents a method of improving the accuracy and precision of the IR sensor data in environment mapping. The mobile robot has been tested for its ability to move and navigate in its environment. However, the IR sensors have limited accuracy in distance measurement and therefore the mapping produces inaccurate environment map. To improve the map quality, it is proposed that probability techniques are used. To refine the decision, occupancy grid mapping technique is used for the mapping together with Bayes’ Theorem to decide whether a grid is occupied or not. The accuracy and precision of the map will be verified a using series of iterative experiments. The resultant map should be able to indicate the location of the object according to its correct grid, and to reduce the effects of shadows. The sensor distance measurement readings were tested, and the probability map were generated based on the relative frequency of the measurements which showed that when the object is nearer to the sensor, the occupancy probability is higher compared to when the object is further away from the sensor. An occupancy grid map was generated by using the probability data that was obtained. This occupancy grid map is generated by testing two different prior probabilities, 0.5 and 0.7. When used prior probability of 0.7, the occupancy probability of the object is higher compared to when 0.5 is used.