Driving Support Data Processor Sensor Selection Range
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The decrease in detection accuracy of sensors used in vehicle driving behavior models leads to erroneous detection and undetection, affecting the reliability of driving behavior estimation and automatic driving control.
Innovation Solution
A data processor and driving support system that selects detection results within a predetermined selection range narrower than the sensor's detectable range, using a sensor profile to improve detection accuracy and reduce the influence of decreased sensor performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If the sensor's detectable range is used for driving behavior estimation, then the quantity of detection data is increased, but the detection accuracy decreases due to sensor performance limitations
Solution Approach 1:
The patent extracts only the high-accuracy detection results from within the sensor's detectable range by introducing a selection unit that filters detection results based on a predetermined selection range. This extraction principle resolves the contradiction by selecting only those detection results that meet accuracy requirements, discarding the less accurate portions while retaining sufficient data quantity for reliable driving behavior estimation.
2Measurement precision
If a narrower selection range is used to improve detection accuracy, then the measurement precision is improved, but the quantity of detection data is reduced
Solution Approach 1:
The patent applies parameter changes by dynamically adjusting the selection range parameters based on sensor characteristics and driving conditions. The selection range is optimized to maintain detection accuracy while preserving sufficient data quantity, resolving the contradiction through parameter optimization rather than simple range reduction.
3Productivity
If all detection results within the sensor range are used for model training, then the productivity of model training is improved, but the reliability of the model decreases due to inclusion of low-accuracy detection results
Solution Approach 1:
The patent applies preliminary action by pre-filtering detection results through the selection unit before they are used for model training. This preliminary filtering ensures that only high-accuracy detection results are input to the driving behavior model, improving model reliability without significantly impacting training productivity, as the filtering process is efficiently integrated into the data processing pipeline.
Data Source
AI summary
A driving support device as an example of a data processor executes processing for estimating a driving behavior of a vehicle by using a driving behavior model trained based on detection results by a sensor. A detected-information input unit acquires detected information including the detection results. From the detection results included in the detected information input to the detected-information input unit, a selection unit selects a detection result that falls within predetermined selection range narrower than a range detectable by the sensor. A processing unit executes the processing, based on the detection result selected by the selection unit.