Multi-Axis Machining Collision Avoidance Using Depth Data and AI

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for avoiding collisions during three-dimensional machining, such as laser cutting or welding, are complex or require long computation times, making them inefficient.

Innovation Solution

A method utilizing a machine learning algorithm, specifically a neural network, to process input data on the geometry and distance between a tool head and a workpiece, predicting potential collisions and adjusting machining paths accordingly, aided by a sensor for real-time data acquisition.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If manual checking or brute-force algorithms are used to avoid collisions, then collision avoidance is achieved, but the method becomes complex or requires very long computation times

Engineering Contradiction:
Improvecollision avoidanceVSAvoidcomputation time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent transforms the collision detection problem from analyzing continuous 3D geometry to processing 2D depth image data. By changing the parameter representation from detailed geometric models to compressed depth maps, the computational complexity is reduced while maintaining collision detection accuracy. The depth images provide sufficient information about tool-head proximity to the workpiece without requiring full 3D geometric analysis.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent creates a simplified digital representation (depth image) that copies only the essential collision-relevant information from the complete 3D scene. Instead of processing the entire geometric model, the system uses depth images that capture only the distance information needed for collision detection, significantly reducing computation time while preserving the critical safety data.

Inventive Principle:
Principle #26Copying

2Measurement precision

If detailed geometric analysis is performed to ensure collision avoidance, then accuracy is improved, but computational complexity increases

Engineering Contradiction:
Improvecollision detection accuracyVSAvoidcomputational complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent extracts only the essential collision-relevant information from the complete geometric scene. By using depth images that contain only distance data from the tool head to the workpiece surface, the system discards unnecessary geometric details while retaining the critical information needed for collision detection. This extraction approach reduces computational complexity without sacrificing detection accuracy.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

Instead of analyzing whether tool head geometry intersects with workpiece geometry through complex 3D modeling, the patent inverts the approach by measuring the distance from the tool head to the workpiece surface and checking if this distance falls within a collision threshold. This inverse method using depth information simplifies the computational problem while maintaining precision.

Inventive Principle:
Principle #13The other way round (Inversion)

Data Source

PatentEP4177688B1Method and device for preventing collisions in workpiece machining by a multi-axis machining machine
Publication Date: 2025.12.03 TRUMPF WERKZEUGMASCHINEN GMBH & CO KG
  • EP4177688B1 patent drawingFigure 1~2
  • EP4177688B1 patent drawingFigure 3~4
  • EP4177688B1 patent drawingFigure 5

AI summary

The invention relates to a method (26) and a device (10) for preventing collisions between a part of a machine tool (12) and a workpiece (14) being machined by the machine tool (12). For this purpose, distances between at least a part of the machine tool (12) and the workpiece (14) are simulated and/or measured by a sensor (56) at values ​​(40a, b) of at least one first axis and one second axis of the machine tool (12) during workpiece machining. These input data (42) are preferably processed into a feature vector (48) and fed to a trained machine learning algorithm (28). Depending on the feature vector (48), the machine learning algorithm (28) determines collision-free values ​​(40c) for the axes. After verification, these values ​​(40c) can be fed back to the machine learning algorithm (28) as feedback, either directly or indirectly, for training purposes.