Vehicle Collision Risk Prediction for Pre-Crash Occupant Protection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional collision safety technologies are inadequate in predicting vehicle collision risks and occupant injuries due to limited perception capabilities, operation uncertainty, and the inability to provide timely and accurate injury information, leading to severe injuries and inefficiencies in rescue and maintenance processes.
Innovation Solution
A vehicle collision risk prediction device comprising a perception component, basic component, prediction component, and decision component that acquires comprehensive data to calculate collision, vehicle injury, and human injury risks, using machine learning algorithms and models to provide real-time risk assessment and control instructions for pre-collision warnings and system activation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If conventional collision safety systems use limited sensors to detect collisions, then the system complexity is reduced, but the measurement precision and reliability of collision detection deteriorates
Solution Approach 1:
The patent combines multiple perception components (cameras, radar, LIDAR, ultrasonic sensors) into an integrated perception system that fuses data from multiple sources. This merging of sensors improves collision detection precision and occupant state assessment without requiring a single complex sensor system, thereby resolving the contradiction between system complexity and measurement precision.
Solution Approach 2:
The perception components are designed to serve multiple functions: detecting collisions, assessing occupant states, predicting injury risks, and providing data for both active safety interventions and post-accident analysis. This multi-functionality reduces the need for separate specialized systems, maintaining manageable complexity while achieving high measurement precision across multiple parameters.
2Loss of time
If conventional systems respond only after collision occurs, then the response time is shortened, but the effectiveness of occupant protection deteriorates due to high energy transmission
Solution Approach 1:
The system performs preliminary assessment of collision risk and occupant state before the actual collision occurs. By predicting potential injury risks and identifying high-risk scenarios in advance, the system can prepare and activate protective measures (such as pre-tensioning seat belts or preparing airbag deployment) before impact, reducing the energy transmitted to occupants and improving protection effectiveness.
Solution Approach 2:
The system continuously monitors occupant states (seating position, restraint system status, physiological signals) and uses this feedback to dynamically adjust protection strategies. Real-time feedback allows the system to optimize response timing and intensity based on actual conditions, ensuring protective actions are taken at the optimal moment to minimize injury while avoiding premature activation.
3Ease of manufacture
If conventional systems use standardized dummy models for safety testing, then the testing process is simplified, but the adaptability to real human variation deteriorates
Solution Approach 1:
The system transitions from static, standardized safety assessment to dynamic, real-time occupant characterization. By continuously monitoring and adapting to each occupant's actual physical characteristics, seating posture, and movement patterns during the collision event, the system captures the diversity of human responses to impact forces, improving adaptability while maintaining testing feasibility through automated sensor-based measurement.
4Speed
If conventional systems activate restraint systems based on basic collision detection, then the activation speed is increased, but the manufacturing precision of injury prediction deteriorates
Solution Approach 1:
The system performs preliminary calculations of injury risk metrics (such as HIC, OLC, and other biomechanical parameters) based on initial collision detection data and predicted occupant kinematics. These preliminary injury predictions are generated in parallel with the decision to activate restraint systems, allowing both the activation decision and the injury assessment to be prepared simultaneously before final execution, thus maintaining speed while improving prediction precision.
Data Source
Figure 1
Figure 2~3
Figure 4~5
AI summary
Disclosed is a vehicle collision risk prediction device and method. The device includes a perception component, a basic component, a prediction component, and a decision component. The perception component is configured to collect human, environment and vehicle related information. The basic component is configured to predict position parameters of an occupant from a current moment to a future moment according to scene parameters of a scene that the occupant is currently in. The prediction component is configured to predict reliability criteria of expectation parameters according to the scene parameters and expectation parameters when a hypothetical collision occurs. The decision component is configured to acquire the expectation parameters and the corresponding reliability criteria, assess a comprehensive collision risk while taking into account criteria of extreme values and change trends of predicted signals in the time domain, and output a control instruction. In this way, a command with explicit guidance may be sent for application scenarios such as a vehicle pre-warning during a pre-collision stage, pre-collision system activation, active intervention of an auxiliary/automatic driving system, ignition of an airbag controller during a collision stage, and driving risk evaluation and judgement.