Predictive ADAS Evasion Planning via Post-Crash Trajectory Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current driver assistance systems lack the ability to predict the movement of traffic objects involved in collisions, which limits their effectiveness in avoiding or mitigating collisions, especially in complex traffic scenarios where objects abruptly change velocity and direction.
Innovation Solution
A method and system that acquire data to determine the likelihood of a crash and predict the post-crash movement behavior of involved objects using crash situation models, including physical collision models, to plan evasion trajectories for the host vehicle, incorporating sensors like radar and lidar, and communication systems for vehicle-to-vehicle information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional driver assistance systems are used without post-crash movement prediction, then the system complexity remains low, but the collision avoidance capability deteriorates in scenarios where objects change velocity and direction after collision
Solution Approach 1:
The system performs preliminary prediction of post-crash movement trajectories before the host vehicle needs to react. By anticipating how other traffic objects will move after a collision (using physics-based models to predict velocity and direction changes), the system prepares evasion trajectories in advance, enabling earlier and more effective collision avoidance actions.
Solution Approach 2:
The system changes the prediction model parameters dynamically based on the traffic situation. When a crash is detected or anticipated, the system switches from standard movement prediction to crash-specific physics models that account for velocity changes, direction changes, and post-impact dynamics, thereby improving prediction accuracy without always maintaining high computational complexity.
2Reliability
If early reaction is initiated to avoid collision, then the probability of effectively avoiding collision increases, but the requirement for accurate prediction of other traffic objects' movement increases
Solution Approach 1:
The system performs preliminary prediction of post-crash movement trajectories before the host vehicle needs to react. By anticipating how other traffic objects will move after a collision (using physics-based models to predict velocity and direction changes), the system prepares evasion trajectories in advance, enabling earlier and more effective collision avoidance actions.
Solution Approach 2:
The system introduces an intermediary prediction layer that estimates post-crash movement based on detected crashes or near-crashes. This intermediary model acts as a mediator between raw sensor data and evasion trajectory planning, providing smoothed and physically-consistent predictions that improve reliability even when direct observation of post-crash movement is limited by noise or occlusion.
3Measurement precision
If crash situation models including physical collision models are applied to predict post-crash movement, then the accuracy of movement prediction improves, but the computational complexity increases
Solution Approach 1:
The system applies high-complexity physics-based crash models only locally to the specific traffic objects involved in a detected crash or near-crash situation, rather than applying complex models to all objects continuously. This localized application of sophisticated prediction models improves accuracy where needed while maintaining lower overall computational complexity for the rest of the traffic scene.
4Reliability
If the host vehicle plans evasion trajectory based on predicted post-crash movement, then the effectiveness of evasion maneuver improves, but the response time required for trajectory calculation increases
Solution Approach 1:
The system performs preliminary prediction of post-crash movement trajectories before the host vehicle needs to react. By anticipating how other traffic objects will move after a collision (using physics-based models to predict velocity and direction changes), the system prepares evasion trajectories in advance, enabling earlier and more effective collision avoidance actions.
Solution Approach 2:
The system dynamically adapts the trajectory planning process based on the predicted post-crash movement. As new sensor data arrives and predictions are updated, the evasion trajectory is continuously refined to account for changing positions and velocities of other traffic objects, ensuring the maneuver remains effective while adapting to real-time conditions.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
The invention regards the field of Advanced Driver Assistant Systems (ADAS) and the field of predictive emergency ADAS configured to predict trajectories of other traffic objects in order to avoid collisions with those traffic objects. In a first step, the method estimates the probability for at least two traffic objects to collide and, in a second step, predicts a potential movement of the traffic objects presumably involved in the collision after the collision. This information about a potential movement of the traffic objects can be used to plan a suitable trajectory path of an evasion manoeuvre for a host-vehicle employing the predictive emergency ADAS according to the invention. The invention proposes a system that detects the collision situation and applies crash situation models to predict the future dynamic properties of other traffic objects that collide or are about to collide with each other. The invention proposes to apply specific crash models like collision models, models derived from crash simulations and tests or models of human reactions in emergency situations.