Motion-Planning Using RRTs for a Swarm of Robots Controlled by Global Inputs
Small-scale robots have great potential to bring transformation in the field of medical applications, defense systems, security, micro-assembly and many other areas. Imagine a group of milli, micro or nano-robots that can navigate inside the body to solve medical problems, or a swarm of robots that can paint a beautiful painting. Robots containing iron can be steered using a magnetic field generated by an MRI scanner, but all the robots will move in the same direction because the magnetic effect is global and not local. This thesis uses a customized version of the motion-planning technique called Rapidly Exploring Random Tree (RRT) to solve this problem for multiple robots by using obstacles. This thesis project solves instances of this motion-planning problem for more than one robot when steered using a magnetic field, and develops a motion planner that searches for sequences of control inputs to simulations steer the robots to desired goal positions.