A Collision Detection Algorithm For Virtual Robot-Centered Flexible Manufacturing Cell

Collision detection is crucial in virtual manufacturing applications such as virtual prototyping, virtual assembly and virtual robot path planning. For accurate simulation of manufacturing systems and processes in virtual environment, physical interaction with the objects in the scene are triggered by collision detection. This thesis presents a collision detection algorithm for accurate simulation of a virtual flexible manufacturing cell. The technique utilizes the narrow phase approach in detecting collision detection of non-convex object by testing collision between basic primitive and polygon. This algorithm is implemented in a virtual flexible manufacturing cell for the loading and unloading process performed by the robot. The robot’s gripper is treated as non-convex object and the exact point of collision is represented with a virtual sphere and collision is tested between the virtual sphere and the polygon. To verify the collision detection algorithm, it is tested with different positions and heights of the storage system during simulation of the virtual flexible manufacturing cell. The results showed that the collision detection algorithm can be used to support the concept of hardware reconfigurablility of FMC which can be achieved by changing, removing, recombining or rearranging its manufacturing elements in order to meet new demands such as introduction of new product or change.

Preview PDF FK_2008_44.pdf Download (260kB)

Actions (login required)

View Item