Research: Optimizing the Homogeneity Between Path Planning and Motion Planning Algorithms
Mentor: Sireesh Gururaja
In robotics, the ability to navigate an environment without human input is known as autonomous navigation. This skill is divided into two categories defined by the nature of the navigational techniques. Heuristic algorithms rely on random motion in a pre-mapped environment. The optimal approach is a technique used in an environment that has not been mapped and aims for a robot to minimize uses of the desired resources whilst navigating. However, there is a margin in error of what the robot predicts as a path and what path the robot travels. This research aimed to minimize this margin through the combined research of both path planning and trajectory algorithms, in order to examine the statistical benefits and disadvantages posed by each method. Path planning algorithms were selected at random and paired with a motion planning algorithm to measure the margin of error between the two processes. This step was repeated numerous times to observe if there were any noticeable relationships between the pairings. As a result, this research provides a novel approach to the optimization of the two data sets in order to create a more synchronized execution of a planned route. While there was no strong relation between the random pairing of algorithms, the research was able to pinpoint flaws in algorithms that could be fixed with weights and biases in order to obtain a motion planning algorithm that followed the proposed path with a smaller margin of error than before. The main limitation in this research however is that this has only been tested in a simulation and not on a physical mechanism as of to date.