Path planning algorithms for mesh networked robots based on WiFi geo-location

摘要

Robot formation and motion planning is a challenging area in mesh networked robots (MNRs). The accuracy and speed with which two or more robots can align them so as to serve a purpose is very important. This paper reports results from our study of the characteristics of robots operating in a real world wireless network environment where robots plan their motion based on the position determined using a Wi-Fi based positioning algorithm. Further we propose and study three methods of formation planning for a leader follower system and evaluate their accuracy and efficiency. First, the followers decide their own path based on the leader's position. In the second algorithm, the motion command for the follower was decided by the leader based on the follower's position and in the third, the followers predict the leader's next position using its current position and direction of motion and moved accordingly. These methods were studied in different working environments and the effect of disturbances on the accuracy of formation was determined using a prototype MNR we built. RSSI fingerprinting was used to calculate the position of the robots and based on which, the path was planned by the leader as well as the followers. Results from the comparison of path planning algorithms are provided.