Robot Sensor Consistency Monitoring via Physical Model Constraints
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting sensor failures in industrial robots are limited in detecting all possible failure conditions, often resulting in false positives or false negatives, and are costly due to the need for redundant sensors, which may not always be feasible or effective.
Innovation Solution
Monitoring robot sensor readings for mutual consistency based on constraints derived from a physical model of the system, allowing detection of sensor inconsistencies and system issues without the need for sensor redundancy, using constraints such as kinematics, torque-generation, voltage-velocity, thermal-response, and load-dynamics constraints to initiate fail-safe procedures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If redundant sensors are used to detect sensor failures, then reliability of sensor failure detection is improved, but device complexity and cost increase
Solution Approach 1:
The patent creates a virtual copy of sensor functionality through mathematical models rather than physical sensor duplication. The system uses a dynamic model of the robot system to generate expected sensor readings and compares these with actual sensor measurements, eliminating the need for redundant physical sensors while maintaining detection capability.
Solution Approach 2:
The patent replaces the mechanical approach of redundant physical sensors with a computational/model-based approach. Instead of using additional physical sensing elements, the system uses mathematical models and algorithms to predict sensor behavior and detect anomalies, substituting mechanical redundancy with intellectual redundancy.
2Measurement precision
If redundant sensors are deployed to cover all failure modes, then measurement precision of sensor failure detection is improved, but manufacturing cost increases
Solution Approach 1:
The patent makes the dynamic model serve multiple functions: it is used for robot control, prediction of sensor behavior, and failure detection. This multi-functionality allows the same computational infrastructure to provide precise failure detection without requiring additional specialized hardware, thereby reducing manufacturing costs while maintaining high detection precision.
Solution Approach 2:
The system creates a virtual replica of the sensor measurement process through mathematical modeling. This virtual copy allows for precise comparison between expected and actual sensor readings, enabling accurate failure detection without the cost of purchasing and installing redundant physical sensors.
3Ease of operation
If monitoring for implausible sensor values is used, then ease of operation is improved, but reliability of failure detection deteriorates due to false positives and negatives
Solution Approach 1:
The patent implements a feedback mechanism where the dynamic model continuously predicts sensor readings based on the current system state, and these predictions are fed back to compare with actual sensor measurements. This closed-loop feedback system provides reliable failure detection while maintaining ease of operation, as the comparison process is automated and does not require complex manual monitoring.
Solution Approach 2:
The system performs preliminary calculation of expected sensor values using the dynamic model before comparing them with actual measurements. This preliminary action allows the system to identify deviations and potential failures proactively, improving detection reliability while keeping the monitoring process simple and automated.
Data Source
Figure 1A
Figure 1B
Figure 2
AI summary
Sensors associated with a robot or sub-system thereof (e.g., a series elastic actuator associated with a robot joint) may be monitored for mutual consistency using one or more constraints that relate physical quantities measured by the sensors to each other. The constraints may be based on a physical model of the robot or sub-system.