Multi-Sensor Robot Interaction Prediction for Safety Stop Avoidance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Collaborative robots (COBOTs) face inefficiencies due to safety system limitations, which reduce their productivity by restricting speed and load capacity and requiring frequent stops or slowdowns when interacting with humans or other robots, despite advanced safety measures.
Innovation Solution
A predictive system using a distributed vision system with multiple sensors and machine learning models to detect and classify potential interactions, allowing for proactive control adjustments and alerts to avoid risky situations, thereby enhancing safety and efficiency without replacing traditional safety systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If safety systems are programmed to be very sensitive and drastic in responding to potential collisions, then safety is improved, but productivity is reduced due to frequent stops and slowdowns
Solution Approach 1:
The system performs preliminary classification of detected objects to distinguish between safe and unsafe interactions before safety systems activate. By pre-analyzing the nature of interactions using machine learning models, the system allows COBOTs to continue operating at full speed when interactions are classified as safe, while only slowing down or stopping when unsafe interactions are predicted, thus maintaining productivity while ensuring safety.
Solution Approach 2:
The patent introduces an intermediary classification system between the sensors and the safety system. This intermediary layer uses machine learning models to interpret sensor data and determine whether detected interactions require safety intervention. This mediator prevents unnecessary safety system activations for benign interactions while ensuring proper response to dangerous situations, resolving the contradiction between safety sensitivity and operational efficiency.
2Reliability
If COBOTs move slowly and carry small masses to guarantee safety, then safety is improved, but efficiency of movement is reduced
Solution Approach 1:
The system dynamically adjusts COBOT operational parameters based on real-time interaction classification. When the environment is determined to be safe, COBOTs operate at full speed and full load capacity. When potential unsafe interactions are detected and classified, the system dynamically reduces speed or stops only for those specific situations. This dynamic adaptation eliminates the need for consistently reduced operational parameters, thereby maintaining efficiency while ensuring safety.
Solution Approach 2:
The patent changes operational parameters (speed, load) conditionally based on classified interaction types rather than maintaining fixed conservative parameters. The machine learning model determines appropriate parameter adjustments based on the specific interaction context, allowing COBOTs to operate at optimal parameters for each situation rather than being constrained by universally reduced parameters, thus resolving the efficiency-safety tradeoff.
3Reliability
If traditional safety systems completely stop the device upon detecting potential collision, then safety is improved, but efficiency of operation is dramatically reduced
Solution Approach 1:
The system applies different safety response strategies to different types of interactions rather than using a uniform stop-for-all approach. By locally classifying each interaction type (e.g., human contact, robot-to-robot contact, object contact) and applying appropriate responses, the system avoids unnecessary stops for safe interactions while ensuring proper safety responses for dangerous ones, thereby maintaining operational efficiency while preserving safety.
Solution Approach 2:
The classification system performs preliminary analysis of detected interactions to determine whether they require safety intervention before the safety system activates. This pre-screening prevents unnecessary complete stops for benign interactions that would otherwise trigger safety systems, allowing continuous operation for safe interactions while maintaining safety responses for dangerous ones, thus resolving the contradiction between safety activation and operational efficiency.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The system increases the safety and operational efficiency of COBOTs by predicting and mitigating potential interactions, allowing them to operate more effectively while maintaining human safety.
Implementation Method 1
a distributed vision system capable of modeling and interpreting, in a real time fashion (e.g., by means of 2D and 3D camera), the interactions and potential interactions between robots and between robots and humans
Implementation Method 2
The vision system may include non-optical components, including LIDAR
Implementation Method 3
The vision system may include non-optical components, including LIDAR, radar
Implementation Method 4
The vision system may include non-optical components, including LIDAR, radar, ultrasonic sensors
Data Source
AI summary
A predictive system and process that predicts safety system activation in industrial environments when collaborative robots (COBOTs), automated guidance vehicles (AGVs), and other robots (individual or collectively “robots”) are interacting (i) between one another or (ii) between a robot and human. As provided herein, the predictive system is not meant to substitute traditional safety systems, but rather to detect and classify robot-to-robot and robot-to-human interactions and potential interactions thereof so as to limit or avoid those interactions altogether, thereby increasing safety and efficiency of the robots.