This is the 5th ODE (Open Dynamics Engine) tutorial.
This time, let’s study a body (rigid body) which is a target for dynamics, and a geometry which is a target for collision detection.
ODE implemented the dynamics core and the collision detection library separately. The advantage of it is easy to use other collision libraries instead of ODE’s. ODE, currently, uses the OPCODE library and the Gimpact library. OPCODE is the default collision detection library and it is faster than Gimpact. Gimpact, however, is more accurate than OPCODE.
OpenHRP3, Open-architecture Human-centered Robotics Platform version 3, also uses OPCODE for collision detection.
Another advantage of the implementation is that only collision detection can be applied to objects. For example, The Tsukuba challenge, Real World Robot Challenge, you do not have to calculate dynamics about buildings, trees, and fences. Only calculating dynamics is sufficient to simulate the environment.
Due to such advantages, ODE creates a World for the dynamics calculation and create a Space for the collision detection . An object has two attributes, that is, a body (rigid body) for dynamics and a geometry for collision detection.
This time, there is no sample code. It is, however, a very important point to use ODE. Next time, a sample code will describe how to use the collision detection function.