2P1-B04 Collision Estimation from Information of an Omni-directional Camera

Abstract

This paper presents a collision estimation method from information of the RoboCup Middle-Size League (MSL) robot equipped with an omni-directional camera. This applies a method of a collision-avoidance calculation with radar information of a ship to an omni-directional camera image. The Closest Point of Approach (CPA) and Time to CPA (TCPA) are calculated from courses of a robot and a high-speed moving object, as a result, the predicted position in collision detection can be estimated. We demonstrate effectiveness by testing that a RoboCup MSL goalkeeper robot intercepts a ball.