Collision Detection Configuration Guide

Overview

Configure collision detection to identify potential collisions during robot grasping operations.

Collision detection is primarily achieved by setting collision detection pairs and their corresponding collision thresholds.

  • Collision detection pairs refer to objects that may collide with each other.

  • Collision threshold represents the tolerance level for collisions. A higher value indicates greater tolerance for collisions.

Collision Detection Configuration

Follow these steps to configure collision detection:

  1. Open Configuration Matrix

    Click Open Configuration Matrix to open the collision detection matrix configuration window.

  2. Select Collision Detection Pairs

    You can configure collisions between robot joints, tools, scene objects, and dynamic obstacles (scene point clouds, matched result models). Select appropriate collision detection pairs based on your needs.

    Click the corresponding cell in the matrix based on row and column information to open the collision detection pair configuration window.

  3. Set Collision Detection Threshold

    1. In the collision detection pair configuration window, select the collision detection unit and set the threshold under the collision threshold section.

    2. Click Add Configuration. The configured collision detection pair cell will turn blue and display the threshold parameter.

    3. Click Save Configuration.

  4. Set Collision Grid Size

    During collision detection, point clouds do not directly participate in collision detection. Instead, the 3D space containing the point cloud is recursively divided into eight sub-cubes until the cube edge length reaches the set collision grid size. If these cubes collide with other objects, it is considered that the point cloud has collided. Adjusting the collision grid size affects the accuracy and efficiency of collision detection.

    In collision detection, the collision grid size directly affects detection accuracy: smaller grid sizes result in more precise collision detection but longer computation times.

    For the same point cloud:

    Collision Grid Size

    Computation Time

    Illustration

    1mm

    Longer

    pick collision voxel small

    3mm

    Shorter

    pick collision voxel medium

    5mm

    Shorter

    pick collision voxel large

  5. Set Detection Mode

    • Quick Mode: During robot path planning, collision detection for the current path point and remaining path points is immediately terminated once any collision is detected between collision pairs.

    • Complete Mode: During robot path planning, all collision pairs for all path points are fully traversed and detected, regardless of whether collisions are found. Specifically:

      • All collision detection pairs for a single point are detected

      • All path points on a single path are completely detected

  6. Perform Collision Detection

    1. Click Take Photo and wait for the photo to complete.

    2. Click Detect and wait for detection to complete.

    3. Click Collision Detection and wait for collision detection to complete.

    4. Click the detection details entry on the right parameter panel to view collision detection results.

View Detection Details

detection details

Click the details icon icon to view the grasp point’s name, ID, priority, corresponding grasp object ID, and grasp point pose information.

Click the move to point icon icon to move the robot to that path point.