AI based steering of Robot with Zero Collision by Using Image Segmentation J.Jasper Gnana Chandran 1 , Tharani Vimal 2 Professor& Head 1 , Department of Electrical and Electronics Engineering 1 Associate Professor 2 , Department of Information Technology 2 St.Peter’s College of Engineering and Technology 1,2 ,Chennai,Tamilnadu jasper_fx@yahoo.co.uk 1 ,tharanimadhan@yahoo.com 2 Abstract The robot in the working environment should be provided with good vision for detection and avoidance of the obstacle present in the path. The proposed method called Pixel Based Image Segmenter (PBAS) is an image segmentation technique that follows a nonparametric background modeling paradigm which segment and extract the details about the obstacles from the static background. In the PBAS, the background update is based on a learning parameter and the foreground decision depends on a decision threshold. The Fire Bird V robot is used here to implement the proposed approach i.e. the PBAS segmentation algorithm. The camera mounted on the robot senses the obstacles present in the vicinity. The video sequence is transmitted to the PC in which the segmentation process is carried out. The segmentation algorithm is applied on the image frames of the video sequence for segmenting the dynamic foreground from the static background. The decision making is performed over time to choose an alternative path for avoiding the obstacle. Thus the proposed method can effectively segment the image frames for detecting the obstacles in the vicinity of robot and the decision making can be carried and finally it makes the direction better for avoiding the obstacles by providing the details about the unknown surface. Keywords: Pixel Based Image Segmenter, Fire Bird Robot, Segmentation Process, Decision Thershold 1.INTRODUCTION The robot varies widely with researchers, engineers, and robot manufacturers. However, robot is accepted and used in many industries in its working environment. Robots are expected to have higher mobility and dexterity than the traditional machine tools. They must have the ability to work in a large reachable range, to access crowded places, handle a variety of work pieces, and perform complex and flexible tasks. The origin of industrial robots is in two preceding technologies: numerical control for machine tools, and remote manipulation which generate control actions based on stored data and performs a task at a distance respectively. It is used in environments where human workers cannot access easily or safely, e.g. for handling radio-active materials, some deep sea and space applications. Here operator's motion is transformed into electrical signals, which are then transmitted to the mechanical arm to cause the same motion as the one that the human operator performs. The research [10], [12] in the area of robot navigation involved only the single features such as edge/texture to detect objects which suffers from poor accuracy while detecting objects, since it employs single feature. As reported in [10], edge detection based path planning using SOBEL operator assumes the edges formed in the images are due to Vol 40, 2020 2307 Tierärztliche Praxis ISSN: 0303-6286