Loading System Collision Detection with Safety-Certified Controllers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing loading systems, particularly cranes, face challenges in collision avoidance due to high data processing demands that exceed the certification criteria for functional safety, necessitating the use of uncertified high-performance hardware.
Innovation Solution
Shift the collision calculation to a certified control node by providing a digital load image file and a protective distance, converting sensor data into reduced depth images, and performing differential calculations on a safety-certified control unit to ensure rapid collision detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Power
If high-performance industry PC with high computing power is used for processing sensor data and creating environment maps, then collision detection capability is improved, but the system cannot be certified according to functional safety criteria
Solution Approach 1:
The system is divided into two separate nodes: a sensor node that performs high-performance sensor data processing and environment mapping, and a control node that performs collision detection using certified algorithms. This segmentation allows each component to be optimized for its specific function while maintaining overall system safety certification.
Solution Approach 2:
The patent introduces an intermediary approach where the sensor node processes raw sensor data and generates environment maps, which are then used by the control node for collision detection. The control node receives pre-processed data and performs differential calculations to detect collisions, acting as an intermediary between high-performance sensing and safety-critical control.
2Measurement precision
If high-frequency three-dimensional sensor capture is performed, then collision detection accuracy is improved, but storage and computing demands increase beyond certified controller capabilities
Solution Approach 1:
The patent extracts only the essential collision-relevant information from the full three-dimensional sensor data. The sensor node creates environment maps from high-frequency sensor capture, but only the differential changes in these maps are transmitted to the control node for collision detection, removing unnecessary data while preserving safety-critical information.
Solution Approach 2:
The system performs partial action by focusing computational resources only on detecting changes in the environment that could indicate collisions. Rather than processing all sensor data in full detail, the control node performs differential calculations that identify only the relevant changes, reducing computational demands while maintaining detection accuracy.
3Reliability
If a certified controller is used for collision avoidance, then functional safety compliance is improved, but processing speed and computational capacity are reduced
Solution Approach 1:
The sensor node performs preliminary processing of sensor data by creating environment maps before the control node performs collision detection. This preliminary action prepares the data in advance, allowing the certified controller to focus only on differential calculations and collision assessment, thereby improving processing speed without compromising safety certification.
Data Source
AI summary
An automation system for controlling a loading system and operating method for the loading system which loads or moves a load along a route, wherein a collision between the load and objects in the environment during the loading process is avoided, where a digital load image file describing a spatial overall extent of the load is provided, an environment image file is recorded cyclically via an imaging sensor system, a protective distance is added to data of the load image file to provide a collision model image file, a cutout file is provided as a virtually recorded image file with a viewing angle from the load as a function of the route, a differential image file is provided from the collision model image file and the cutout file via differentiation, and an evaluation step is performed via which the differential image file is examined for a possible collision.