Steering Torque Signal Filtering for Driver Intervention Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting steering wheel intervention states in vehicles are unreliable due to noise in sensor measurements and can be fooled by external influences, leading to incorrect determinations of driver control.
Innovation Solution
A computer-implemented method using a Kalman filter to process steering torque signals and shifts, generating a clean estimate of steering torque that is compared to limit values to determine driver intervention, with optional use of band-pass filters and adaptive Kalman filters to improve accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a torque sensor is used to measure steering force, then the ability to detect driver control is improved, but the system can be outsmarted by external influences such as hanging a weight on the steering wheel
Solution Approach 1:
The system continuously monitors steering torque and compares it against dynamically adjusted threshold values. When the torque exceeds the threshold, the system generates a feedback signal to deactivate the driver assistance function, creating a closed-loop control system that responds to driver input while filtering out spurious external influences through the threshold comparison mechanism
Solution Approach 2:
The patent introduces an intermediary evaluation mechanism that sits between the torque sensor and the final control decision. This intermediary layer processes the raw torque signal through threshold comparison and temporal filtering, acting as a mediator that translates physical torque into reliable control state determination, thereby protecting against direct manipulation of the sensor output
2Measurement precision
If capacitive sensors are used on the steering wheel rim, then hands-on detection is improved, but the system fails when drivers wear gloves and cannot distinguish between holding the wheel and controlling the vehicle
Solution Approach 1:
The patent replaces the capacitive sensing mechanism (which detects electrical properties of the hand) with a mechanical torque sensing approach. By measuring the physical torque applied to the steering wheel column, the system can detect driver control intent regardless of whether the driver is wearing gloves or not, as the torque measurement directly reflects mechanical interaction rather than electrical contact
Solution Approach 2:
The system changes the detection parameter from electrical capacitance (which is blocked by gloves) to mechanical torque (which penetrates through gloves). This parameter transformation allows the sensing system to operate reliably under different driving conditions, including when drivers wear gloves, by measuring a physical quantity that is not affected by the presence of protective gear
3Speed
If raw steering torque signals are used directly for intervention detection, then the response time is minimized, but noise in sensor measurements leads to false positives and false negatives
Solution Approach 1:
The system performs preliminary filtering of the steering torque signal before it is used for intervention detection. By pre-processing the raw sensor signal to remove noise and spurious components, the system prepares a cleaned signal that can be reliably compared against threshold values, thereby reducing false positives and false negatives while maintaining responsive detection capability
Solution Approach 2:
The patent converts the harmful noise in the sensor signal into a beneficial filtering opportunity. By analyzing the noise characteristics and using threshold comparison, the system learns to distinguish between legitimate driver input and spurious signals. The noise that would normally degrade performance is instead used to calibrate the detection thresholds, making the system more robust to signal degradation
Data Source
Figure 1
Figure 2
AI summary
The invention relates to a computer-implemented method for detecting a steering wheel intervention status of a driver of a motor vehicle, wherein the motor vehicle has at least one automated driving function intervening in a steering arrangement of the motor vehicle, wherein the steering torque in the steering assembly is detected and a steering torque signal is generated from same, wherein, in a first branch, a steering torque shift is determined from the steering torque signal, wherein an adaptive Kalman filter is used, the adaptive Kalman filter using the steering torque signal and the steering torque shift as measured values and filtering out the steering torque shift from the steering torque signal in order to determine a steering torque estimated value, wherein the steering torque estimated value is compared with at least one threshold value, and the comparison is used to determine whether the driver intervenes in the steering assembly.