Traffic Safety System Group-Based Collision Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing traffic safety systems face challenges in predicting and mitigating collisions among multiple traffic participants within a target area due to high processing loads, which can inhibit real-time support and lead to increased complexity for drivers.
Innovation Solution
A traffic safety support system that includes a recognizer to identify traffic participants, a predictor to forecast potential collisions by grouping and analyzing movement data, and a notifier to provide targeted notifications based on prediction results, reducing processing load and enhancing collision avoidance measures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If information on an enormous number of traffic participants existing in the target traffic area is aggregated to the server, then comprehensive grasp of traffic participant flow is improved, but processing load at the server increases
Solution Approach 1:
The system segments the traffic monitoring task by dividing traffic participants into multiple moving groups based on spatial proximity and movement characteristics. Instead of processing all individual traffic participants separately, the server groups them into clusters (moving groups) and processes group-level information, significantly reducing the data volume and processing load while maintaining comprehensive situational awareness.
Solution Approach 2:
The system applies partial action by focusing computational resources on predicting collisions only for moving groups that are in close proximity or have intersecting trajectories. Rather than performing collision prediction for all possible pairs of traffic participants, the system selectively processes only those groups with potential collision risk, reducing overall processing requirements while maintaining safety.
2Measurement precision
If collision prediction is performed for all traffic participants, then collision detection accuracy is improved, but processing time increases
Solution Approach 1:
The system segments the collision prediction task by first predicting collisions at the moving group level, then only for groups identified as having potential collision risk does it proceed to predict collisions among individual traffic participants within those groups. This two-stage segmentation approach maintains high detection accuracy for critical situations while avoiding unnecessary processing of non-critical traffic participants.
Solution Approach 2:
The system performs preliminary collision prediction at the moving group level before conducting detailed individual participant collision analysis. By first identifying which moving groups have potential collision risk, the system prepares a filtered subset of candidates for more detailed analysis, ensuring that accurate collision detection is applied only where necessary and reducing overall processing time.
3Measurement precision
If detailed prediction for each traffic participant is performed, then notification accuracy is improved, but device complexity increases
Solution Approach 1:
The prediction system is segmented into two hierarchical levels: a first predictor that operates at the moving group level to identify potential collision risks, and a second predictor that operates at the individual traffic participant level within identified risk groups. This segmentation allows the system to maintain high notification accuracy for at-risk participants while avoiding the complexity of detailed individual prediction for all traffic participants in the target area.
Solution Approach 2:
The system applies detailed prediction partially, focusing computational resources only on traffic participants within moving groups that are predicted to collide. Rather than performing complex detailed prediction for all traffic participants, the system selectively applies the more complex second-stage prediction only where necessary, maintaining notification accuracy for critical cases while reducing overall system complexity.
Data Source
AI summary
A coordination support device includes a target traffic area recognizer configured to recognize traffic participants, a predictor configured to predict futures of a plurality of the traffic participants, and a coordination support information notifier configured to notify notification targets. The predictor includes a group classifier configured to divide the plurality of traffic participants into a plurality of moving groups G1 to G9 and acquire moving group information, a first predictor configured to predict whether or not a collision will occur among the moving groups G1 to G9 on the basis of the moving group information and specifies collision sites in the moving groups that are predicted to collide, and a second predictor configured to determine two or more specific traffic participants among the plurality of traffic participants on the basis of the collision sites and predict whether or not a collision will occur among the specific traffic participants.