Driver Attention Guidance Using Latent Recognition Regions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing driving assistance systems provide fixed assistance information, making it difficult to adapt to changing road situations and effectively present appropriate information based on the driver's unconscious recognition of the surroundings.
Innovation Solution
A driving assistance device that detects latent recognition regions in a captured image using a deep learning prediction model to identify areas where the driver unconsciously focuses attention, and presents driving assistance information in these regions to guide conscious attention.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If fixed timing assistance information is provided, then the system structure is simple, but the adaptability to changing road situations deteriorates
Solution Approach 1:
The patent implements dynamic adjustment of assistance information timing based on detected latent recognition regions. The control unit varies the timing and content of assistance information provision according to the driver's unconscious recognition patterns, transforming the fixed timing system into a dynamic one that adapts to changing road situations and driver attention states.
Solution Approach 2:
The system changes the timing parameter of assistance information provision based on detected latent recognition regions. By adjusting when assistance information is provided according to the driver's unconscious attention patterns and road situation, the system achieves adaptability without requiring complete system redesign.
2Reliability
If assistance information is provided at predetermined timing, then the control logic is simple, but the effectiveness in calling driver attention deteriorates
Solution Approach 1:
The system uses detection results of latent recognition regions as feedback to adjust the timing and content of assistance information provision. The control unit continuously monitors the driver's unconscious recognition patterns and adjusts the assistance information timing accordingly, creating a closed-loop feedback system that improves attention guidance effectiveness.
Solution Approach 2:
The system provides assistance information before the driver's conscious recognition timing by detecting latent recognition regions in advance. This preliminary action allows the assistance information to be presented at the optimal moment when the driver is most receptive, improving the effectiveness of attention guidance.
3Adaptability or versatility
If fixed assistance information is provided, then the information presentation is consistent, but the situation-specific appropriateness deteriorates
Solution Approach 1:
The patent applies local quality by providing assistance information specifically targeted at the detected latent recognition regions rather than using uniform assistance for all situations. The control unit adjusts the content and timing of assistance information based on the specific road situation and driver attention patterns detected in each local context.
Data Source
AI summary
A driving assistance device includes a processor connected to a memory. The processor detects a latent recognition region that is unconsciously recognized by a human in an entire imaging region of a captured image. The captured image is captured by an external camera that is provided in a mobile body. The captured image is an image of at least a traveling direction in which the mobile body travels. The latent recognition region is detected based on a predetermined feature value included in the captured image. The processor presents predetermined driving assistance information in a region on a display device. The region is estimated in accordance with the detected latent recognition region.