Driving State Estimation Using Absolute Value Binning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional driving state estimation devices face high computing loads due to the large number of bins in traveling state distributions, making it difficult to implement them in devices with low computing power, such as smartphones or cheap in-vehicle controllers.
Innovation Solution
The solution involves converting distribution data into absolute values and classifying them into fewer bins across different time ranges to calculate frequency distributions for estimating the driving state, reducing the computational burden.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If bins of traveling state distribution are set separately for positive and negative value ranges, then measurement precision of driving state estimation is improved, but device complexity and computing load increase
Solution Approach 1:
The patent merges the positive and negative value ranges into a single bin structure by using absolute values of distribution data. Instead of maintaining separate bins for positive and negative steering angle deviations, the invention converts all deviation values to absolute values and classifies them into unified bins, thereby reducing the total number of bins from approximately double to the original count while preserving estimation accuracy
Solution Approach 2:
The patent applies parameter transformation by converting distribution data values from signed (positive/negative) to absolute values. This parameter change allows the system to use a single set of bins for both positive and negative deviations, effectively halving the computational burden of frequency distribution calculations while maintaining the ability to detect driving state changes
2Measurement precision
If more bins are used for frequency distribution, then measurement precision of traveling state distribution is improved, but productivity and processing speed decrease
Solution Approach 1:
The patent combines the frequency distribution calculations for positive and negative value ranges into a single unified distribution. By merging the binning operations and frequency counting processes for both ranges, the system reduces the total number of processing operations while maintaining sufficient resolution for detecting driving state changes
3Reliability
If separate bin structures are maintained for positive and negative ranges, then reliability of driving state detection is improved, but ease of operation and implementation in low-power devices deteriorates
Solution Approach 1:
The patent transforms the data representation from signed to absolute values, enabling a simplified bin structure that is easier to implement in low-power devices. This parameter change reduces memory requirements, simplifies the binning logic, and decreases computational complexity while maintaining detection reliability through the use of frequency distribution analysis on absolute deviations
Data Source
AI summary
The driving assistance unit acquires distribution data (a steering angle prediction error) for the traveling state distributions (a first traveling state distribution, a second traveling state distribution). Next the driving assistance unit converts value of the distribution data (a steering angle prediction error) into their absolute values. Next, the driving assistance unit classify the absolute value of the distribution data (a steering angle prediction error) in different time ranges into bins as plural segmented data ranges, based on the distribution data (a steering angle prediction error) whose values are converted into their absolute values and calculates a frequency distributions of the distribution data (a steering angle prediction error) as plural traveling state distributions (the first traveling state distribution, the second traveling state distribution). Next, the driving assistance unit estimates (determines driving instability degree) a driving state of a driver based on the plural calculated traveling state distributions.