Tracker Trajectory Validation for Autonomous Vehicle Collision Checking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face inaccuracies in collision checking due to discrepancies between generated and actual vehicle locations, leading to potential unsafe operations.
Innovation Solution
Implement a secondary computing device to generate a tracker trajectory based on an offset from a reference trajectory, verifying its accuracy by comparing predicted and actual vehicle positions, and issuing alerts for model updates to improve collision checking reliability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a vehicle computing system generates trajectories based on vehicle locations, then the vehicle can be controlled to follow desired paths, but differences between generated and actual vehicle locations result in inaccuracies in collision checking
Solution Approach 1:
The system continuously monitors the difference between generated vehicle locations and actual vehicle locations, using this feedback to validate collision checking accuracy. When discrepancies are detected, the system can trigger alerts and initiate model updates to improve the accuracy of location tracking and collision detection.
Solution Approach 2:
The system performs validation of the tracker trajectory model before full deployment by comparing predicted positions with actual positions. This preliminary validation identifies potential inaccuracies in collision checking, allowing the system to update models proactively before unsafe conditions arise.
2Reliability
If a secondary computing device generates tracker trajectories to validate reference trajectories, then collision checking reliability is improved, but device complexity increases
Solution Approach 1:
A secondary computing device acts as an intermediary between the primary trajectory generation system and the validation process. This intermediary generates tracker trajectories based on offset relationships, enabling independent validation of collision checking without requiring complete system redesign.
Solution Approach 2:
The system creates a copy of the trajectory generation process through the secondary computing device, which generates tracker trajectories that mirror the reference trajectory generation. This copying approach allows validation without adding significant complexity, as the secondary device uses the same algorithms and data structures.
3Reliability
If the system continuously validates tracker trajectory accuracy by comparing predicted and actual positions, then model performance decreases are identified, but loss of time occurs in processing
Solution Approach 1:
The system performs partial validation by checking key position points along the trajectory rather than continuously validating every position. This approach identifies model performance decreases while minimizing processing time, focusing computational resources on critical validation points.
Solution Approach 2:
The validation process skips detailed analysis when the tracker trajectory closely matches the reference trajectory, only performing thorough checks when discrepancies exceed thresholds. This allows the system to quickly process valid trajectories while dedicating more time to investigating potential issues.
Data Source
AI summary
Collision avoidance and error determination for a component of an autonomous vehicle comprising receiving a first trajectory, such as to return a vehicle to an intended trajectory, that a vehicle is predicted to follow, based on an offset between the vehicle and a second trajectory associated with the vehicle, such as a reference trajectory. The first trajectory predicts a first movement characteristic (e.g., a position) of the vehicle at a point in time. A second movement characteristic is received, representing an actual movement characteristic of the vehicle at that point in time. A first error between the first and second movement characteristics is determined. Based at least in part on the first error, performance of a model for generating trajectories that a vehicle is predicted to follow is validated.