Robotics based simultaneous localization and mapping of an unknown environment using Kalman Filtering

In this paper, Robotic application using a microcontroller ATmega2560 based kit mounted with low cost IR sensors for mapping and localization of unknown area is accurately carried out by using Kalman Filtering to determine or measure different parameters. This paper highlights the comparison between the accuracies of output data obtained from IR sensor using Average Technique as well as Kalman Filter. Various cases depending on robot movement in forward and backward direction and wheel rotation in clockwise and anticlockwise direction, are considered in this paper to calculate position encoder resolution. Also probabilistic estimation of actual position of robot is carried out using different techniques i.e Probability Density Function (P.D.F) for verifying the uncertainty in its position and found to be very close to actual position. Simultaneous Localization And Mapping (SLAM) is the only way a robot can navigate through an unknown environment with the use of minimum sensors and get reliable output of the same in identifying an unknown area. Using this method the robot not only create a map more accurate than GPS maps but also localize itself to determine its next position according to the map created more accurately at lower cost. Primarily used for creating a map of an unknown location, this concept can also be used to perform Qualitative Analysis of a given area by equipping the robot with appropriate sensors.