i have a small bot on the screen and he is in a room. he can detect when he is about to hit a wall and take action to avoid it. i have just added an array that stores any points that the bot was about to hit. my idea was that for each move the bot would check its heading , a few moves in front of it and see if it is going to hit any of the points stored in its memory , and if so, it will turn left or right to avoid. its all based on line intersection and cartesian points.im having a few problems though. if the robot gets into a tight corner, and is detecting a few moves in front, and sees it is going to collide with a wall, takes corrective action, but again finds collision is imminent, it kind of gets stuck. can anyone find any other potential flaws, and or advise me on thier views. bear in mind that i am not getting into deep AI, just basic memory and avoiding of collision is all im after. thanks all