Collision detection is to use collision detection algorithm to detect two rigid bodies at a specific moment to judge whether the current position can move.

FCL open source detection library provides object collision detection and object approach distance calculation [51]. FCL can detect

Collision or distance detection of traditional triangular planes and basic shapes including spheres, squares and cylinders. In addition the

Besides, collision detection can also be carried out on point cloud. FCL can also handle deformable models without doing anything about them

It is assumed that the boundary of the object at each point in time is represented as a triangular grid. FCL believes that all geometries that can be detected can be represented as two interfaces: 1) overlap, which checks the overlap between the boundary volume of the geometry and that corresponding to some other object. 2) Intersection, check the exact intersection between geometric objects. This is primarily used to perform precise neighborhood queries between low-level primitives.

FCL collision detection library detects the collision between two objects using the object’s position, attitude,

Bounding box and surface mesh. Bounding box collision detection and grid collision detection are two commonly used collision detection methods

Train of thought. Bounding box detection is firstly to add bounding bodies on the surface of complex objects, in order to use the bounding bodies for fast collision

Bump detection. The bounding body types include sphere, axially symmetric bounding box (AABB), directed bounding box (OBB),

Enveloping ball collision detection is to surround the whole geometry with a sphere, the implementation of intersection test is simple, but tight

Bad sex. When objects are evenly distributed in each coordinate system, a large gap will be left and the pre-processing time will be taken

Expensive. After the bounding boxes are detected, even if the two boxes do overlap, it doesn’t mean two objects have happened

Collision, because the shape of the object is irregular. The bounding box detection can only be used as a rough detection method

The method is very fast, and when there is no overlap between two bounding boxes, you can be sure that there is no collision between objects.

Mesh collision detection uses a mesh that forms the surface of an obstacle. FCL is applied to both surfaces

Some grids detect whether there is overlap between any two grids. In contrast to bounding box collision detection, this

This method is more time consuming. Therefore, the whole collision detection scheme is to execute bounding box detection first, and then use grid collision detection as the fine collision detection process if there is a collision. Conversely, if no collision occurs in the bounding box, then

It is judged that there is no collision, and the collision detection process is stopped.

In the system platform to achieve the collision detection function, you can use rVIZ visual environment for testing. First of all, we

Use code to import obstacles in the visual environment, and then edit the obstacle object by adjusting the obstacles

The size and position allow the robot to collide with obstacles. After the obstacle is placed, the end of the manipulator is driven

———————————————— Copyright notice: This article is originally published BY CSDN blogger “Kobesdu” under CC 4.0 BY-SA copyright agreement. Please attach the link to the original source and this statement. The original link: blog.csdn.net/kobesdu/art…