Sensor Performance Detection via Trajectory Deviation Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional vehicles equipped with automated driving systems face challenges in detecting deteriorating sensor performance, which can lead to critical safety issues as sensors may continue to operate with reduced functionality without generating error messages, potentially affecting autonomous driving functions.
Innovation Solution
A method utilizing artificial intelligence to calculate and simulate setpoint trajectories, comparing actual and simulated trajectories to detect sensor performance deviations, and alternately deactivating sensors to identify and address decreased performance, leveraging existing AI for automated driving and car-to-x communication for data exchange.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If sensors operate with partial functionality, then the sensors can continue to function without complete failure, but the performance and detection precision deteriorate without generating error messages
Solution Approach 1:
The system implements feedback by continuously monitoring the relationship between setpoint trajectories (calculated by AI) and actual trajectories (derived from sensor data). When deviations occur, the system activates alternative sensors to verify whether the deviation is due to sensor deterioration, creating a closed-loop feedback mechanism that detects performance degradation without requiring complete sensor failure
Solution Approach 2:
The system performs preliminary action by proactively comparing AI-calculated setpoint trajectories with actual sensor-based trajectories before complete sensor failure occurs. This early detection mechanism allows the system to identify deteriorating sensors and switch to alternative sensors or safe operating modes before the deterioration becomes critical
2Reliability
If conventional error message monitoring is used, then complete sensor failures are detected, but gradual performance deterioration remains undetected
Solution Approach 1:
The system performs preliminary monitoring by continuously comparing setpoint and actual trajectories using AI calculations and alternative sensor verification, detecting performance deterioration trends before they lead to complete failure. This proactive approach enables early intervention and sensor replacement scheduling, reducing the time loss associated with undetected gradual degradation
Solution Approach 2:
The system establishes continuous feedback loops that monitor trajectory deviations and activate alternative sensors to verify sensor health. This feedback mechanism provides real-time detection of performance deterioration, enabling timely response before complete failure occurs, thereby reducing the detection time gap that exists in conventional error-message-only systems
3Reliability
If multiple sensors are used for automated driving, then system redundancy and safety are improved, but the complexity of detecting which sensor is deteriorating increases
Solution Approach 1:
The system applies segmentation by isolating and testing individual sensors through sequential activation and deactivation. When a trajectory deviation is detected, the system systematically deactivates each sensor in turn and re-evaluates the trajectory calculation, allowing precise identification of the deteriorating sensor without requiring complex simultaneous analysis of all sensors, thus managing monitoring complexity while maintaining system safety
Solution Approach 2:
The system uses AI-based trajectory calculation as an intermediary mechanism to indirectly assess sensor health. Instead of directly monitoring each sensor's output quality, the system compares AI-calculated setpoint trajectories with actual sensor-based trajectories, using this intermediary comparison to detect sensor deterioration in a simplified manner that does not require complex direct sensor-to-sensor comparison
Data Source
AI summary
A method for detecting a decreasing performance of at least one sensor is described, in particular in a vehicle, wherein a setpoint trajectory of the vehicle is calculated, the setpoint trajectory is driven by the vehicle or is simulated by an artificial intelligence, the actual trajectory driven by the vehicle or the simulated trajectory is compared with the setpoint trajectory, a performance of the at least one sensor is tested by the control unit if a deviation of the actual trajectory or the simulated trajectory from the setpoint trajectory is determined, in which test each sensor of the vehicle is alternately deactivated and with the aid of at least one alternative sensor the setpoint trajectory is driven by the vehicle or is simulated by an artificial intelligence. Furthermore, a control unit, a computer program and a machine-readable storage medium are described.