AUTONOMOUS ROBOT NAVIGATION IN BUILDINGS

The aim of the project was to program an autonomous robot known as the E R l to move around without human interference avoiding obstacles and recognising and staying out of the way of humans. The point of the exercise was to see if it was possible to achieve these goals using every day items as opposed to special expensive equipment. Two programs were written. The first was aimed at obstacle avoidance and worked by using the machine vision technique of edge detection. The obstacle was counted as the edge and depending on the position of this obstacle depended oh where the robot would turn. The second program was designed to recognise humans by identifying the shoes worn by them. The program worked by using a series of fihers to generalise the shape of the shoe and then using the machine vision technique known as SIFT (Scale Invariant Feature Transform) would process this image and then compare it to a library of stored images to find a match. When a match was found, directions were given to the robot to move away from the human along with an audio warning. The overall result was that both programs worked to an extent though future work would be required to improve the performance of both programs.School of Computing, Communications and Electronic