Real Time Robot Control Using Micro-Genetic Algorithm

In this project the image of the arena was captured using an overhead camera. The obstacles were in white colour the bot in red and the rest of the arena in black. Using certain simple image processing techniques the bot image was separated and the obstacle information from the image was fed into an integer matrix with value '1' for an obstacle and '0' for a obstacle free pixel. The microgenetic algorithm code was then used to find an optimum ,obstacle free path avoiding rapid curvature changes ,from an initial location to a final location. This path was represented using a bezier curve with 8 control points. The next part of the project included breaking up this bezier curve into straight lines. The bot was then made to run through serial communication on these straight lines taking feedback from the camera. The bot stopped when it reached a certain threshhold area around the destination pixel.