Driving Assistance Apparatus Estimation Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing driving assistance systems face reduced estimation accuracy for future driver actions due to reliance on past vehicle states, especially when encountering special or rare situations.
Innovation Solution
A driving assistance apparatus that stores past driving situations with associated action characteristic values, occurrence frequencies, factor influence values, and change frequencies, allowing for accurate estimation of future driver actions by analyzing these factors and performing assistance accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the future driving action of the driver is estimated on the basis of the past vehicle state, then the driving assistance can be performed, but the estimation accuracy is reduced because the past vehicle state includes special vehicle state
Solution Approach 1:
The patent extracts and removes special vehicle states from the past vehicle state data used for estimation. By identifying and excluding these outlier states that would skew the estimation results, the system maintains reliable driving assistance performance while improving estimation accuracy by relying only on normal, representative driving patterns.
Solution Approach 2:
The patent changes the parameters used for estimation by introducing occurrence frequency and factor influence values as weighting factors. Instead of treating all past vehicle states equally, the system adjusts the estimation parameters to give higher weight to frequently occurring states and lower weight to rare special states, thereby resolving the contradiction between maintaining assistance performance and improving accuracy.
2Quantity of substance
If all past vehicle states are used for estimation, then more data is available for driving assistance, but the estimation accuracy is reduced due to inclusion of special vehicle states
Solution Approach 1:
The patent transforms the quality of data by introducing occurrence frequency as a parameter. Instead of simply filtering data, it changes how data is weighted in estimation calculations, allowing all past states to be used while adjusting their influence based on frequency, thus maintaining data quantity while improving accuracy.
Solution Approach 2:
The patent applies partial weighting to different data points based on their occurrence frequency. Rather than using all data equally or excluding special states entirely, it applies a partial weighting scheme where frequent states have higher influence and rare states have lower influence, optimizing both data utilization and estimation accuracy.
Data Source
AI summary
A driving assistance apparatus is provided with: a storing device for storing each of a plurality of past driving situations of a self-vehicle as a group of a plurality of driving situation factors, for storing, correspondingly to each of the plurality of past driving situations, an action characteristic value and a driving situation occurrence frequency, and for storing, correspondingly to each of the plurality of driving situation factors, a factor influence value and a factor change frequency; an estimating device for estimating a future driving action of the driver on the basis of the plurality of past driving situations, the action characteristic value, the driving situation occurrence frequency, the factor influence value, and the factor change frequency; and a performing device for performing driving assistance in accordance with the estimated driving action.