Intercepting algorithm for differential robots

In a highly dynamical environment, such as robotic soccer, robots must in real-time make its next position to intercept the fast moving target as quickly as possible, and with obstacle-avoidance. This paper describes a computationally inexpensive, yet look-ahead intercepting method for mobile robots. The linear trajectory of the target's motion was achieved by least squares fit, the ellipse was established as the motion model for the target based on the linear trajectory, and the tangential point in the ellipse was accepted as the predicted position navigating robot to go in every grabbing cycle. This method changes the mode of robot directly moving to the target into that of robot moving to the motion zone of the target, and guarantees that the robot both reaches the predicted destination with high speed and avoids moving obstacle smoothly along continuous trajectory. The algorithm was successfully probed in the FIRA robotic soccer system.