Adaptive Sensor Model Updating for Ageing and Drift Compensation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sensor technologies face challenges in accurately detecting and mitigating the effects of ageing and drifting, particularly in low-cost sensors, which affect the reliability and accuracy of sensor signals over time.
Innovation Solution
A calculation unit is integrated into the sensor device to process sensor signals using a variational autoencoder (VAE) architecture, which includes a model processor to detect anomalies indicative of ageing or drifting, and performs model updates to compensate for these issues, utilizing transfer learning and uncertainty evaluation to ensure accurate future readings.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If low-cost sensors are used, then device complexity and cost are reduced, but measurement precision and reliability deteriorate due to ageing and drifting
Solution Approach 1:
The system implements feedback by continuously monitoring sensor signals for anomalies and automatically updating the sensor model based on detected ageing or drifting. The model processor evaluates sensor signals, detects anomalies, and triggers model updates to compensate for degradation, creating a closed-loop system that maintains accuracy without requiring expensive hardware
Solution Approach 2:
The system changes the parameters of the sensor model over time to adapt to sensor ageing and drifting. By updating model parameters based on detected anomalies and using transfer learning to adjust to new sensor characteristics, the system maintains measurement precision despite physical sensor degradation
2Productivity
If sensor operation time is extended, then productivity is improved, but measurement precision deteriorates due to ageing and drifting effects
Solution Approach 1:
The sensor model is made dynamic and adaptive rather than static. The model processor continuously evaluates sensor signals and automatically updates the model parameters in response to detected anomalies, allowing the system to maintain precision throughout extended operation periods by adapting to changing sensor characteristics
Solution Approach 2:
The system performs self-service by automatically detecting its own degradation through anomaly detection and correcting itself through model updates. The model processor monitors sensor performance and autonomously adjusts the sensor model to compensate for ageing effects without external intervention, enabling long-term reliable operation
3Measurement precision
If anomaly detection and model update mechanisms are added, then measurement precision is maintained, but device complexity increases
Solution Approach 1:
The calculation unit is designed with multi-functionality, where the model processor performs multiple tasks: evaluating sensor signals, detecting anomalies, determining whether model updates are needed, and executing model updates. This consolidation of functions into a single processing unit maintains precision while minimizing the increase in device complexity
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A calculation unit for a sensor device comprises an interface configured for obtaining sensor signals of the sensor device and a processing processor configured for processing at least one sensor signal by use of a senor model. The calculation unit comprises a model processor configured for evaluating sensor signals for a presence of an anomaly in the sensor signals that indicates an ageing of the sensor device. The model processor is configured for updating the sensor model based on the presence of the anomaly to at least partially compensate for the anomaly to obtain an updated sensor model. The processing processor is configured for using the updated sensor model for future sensor signals.