An Algorithm on Collision Detection by Computing the Minimum Distance between Two Convex Polyhedra

The problem of tracking the distance between two convex polyhedra is finding applications in many areas of robotics, including intersection detection, collision detection, and path planning. An algorithm for computing the minimum distance between two convex polyhedra is presented. The algorithm applies to polyhedral objects that can be represented as convex hulls of its vertex value in three-dimensional space. Nonlinear programming techniques are then employed to compute the minimum distance between two convex polyhedra, and according to the minimum distance, it can conclude whether or not the objects has collided at some time. The simulation results demonstrate the performance of this new method.