Cobot Joint Torque Estimation Using AI and Angular Position Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing collaborative robot systems face challenges in achieving precise and safe torque measurement, particularly in shared work environments with humans, due to the high cost, weight, and imprecision of conventional torque sensors, and the need for redundant sensors, which affect flexibility and dynamics.
Innovation Solution
A method using artificial intelligence to determine output torque based on input and output angular positions, supplemented by angular position sensors, allowing for precise and redundant torque determination without direct torque measurement, and enabling continuous training to adapt to environmental changes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a torque sensor is used to measure output torque, then measurement precision is improved, but weight and cost increase
Solution Approach 1:
The patent replaces the mechanical torque sensor with an AI-based computational model that calculates torque from encoder measurements. The neural network model processes input variables (joint angles, angular velocities, motor currents) to predict output torque, eliminating the need for physical torque sensing hardware and its associated weight.
Solution Approach 2:
The patent introduces encoder sensors as intermediary measurement devices that indirectly provide torque information through AI processing. Instead of directly measuring torque with a sensor, the system uses angular position and velocity measurements from encoders as intermediate variables that the AI model transforms into torque estimates.
2Measurement precision
If a torque sensor is used to measure output torque, then measurement precision is improved, but cost increases
Solution Approach 1:
The patent replaces expensive torque sensors with software-based AI processing. The neural network model runs on existing robot controllers, transforming hardware cost into computational algorithms that leverage already-present sensors (encoders and current sensors), significantly reducing overall system cost.
Solution Approach 2:
The patent creates a virtual copy of torque measurement through AI simulation. Instead of purchasing physical torque sensors, the system replicates torque measurement functionality through computational modeling that mimics what a torque sensor would measure, using data from existing sensors.
3Reliability
If redundant torque sensors are used to ensure safety, then reliability is improved, but device complexity and weight increase
Solution Approach 1:
The patent replaces redundant physical torque sensors with redundant computational checks within the AI model. The neural network can be configured with multiple output predictions or validation mechanisms that provide reliability without adding physical sensors, thereby reducing system complexity.
Solution Approach 2:
The patent implements feedback mechanisms where the AI model continuously monitors its own predictions and compares them against expected physical constraints and multiple sensor inputs (encoders, current sensors). This internal feedback loop ensures reliability by detecting inconsistencies without requiring additional external sensors.
Data Source
AI summary
The invention relates to a method for precisely determining an output-side torque, in particular an output-side torque of an actuator gearing mechanism of a joint of a collaborative robot, by means of an artificial intelligence which is designed to output one or more output variables on the basis of input variables, wherein the input variables of the artificial intelligence comprise: a first angular specification which corresponds to an input-side angular position and a second angular specification which corresponds to an output-side angular position, and additionally the output variables comprise the output torque determined by the artificial intelligence. (FIG. 1)