Autonomous Vehicle Trajectory Limiting at Traffic-Control Stops
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Trajectory planning for autonomous vehicles is computationally expensive due to the need to ensure safe avoidance of objects and obstacles, dynamic constraints, and collision and proximity analysis, which can be resource-intensive, especially in dense urban environments.
Innovation Solution
A method and system for generating and following planned trajectories for autonomous vehicles that determine a baseline trajectory, identify a stopping point based on traffic controls, and generate constraints only up to that point, allowing for reduced computational complexity by ignoring constraints beyond the stopping point, thereby optimizing trajectory length and computation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If trajectory planning ensures safe avoidance of all objects and obstacles with full collision and proximity analysis, then safety is improved, but computational cost and processing time increase significantly
Solution Approach 1:
The patent segments the trajectory planning process into two distinct phases: a global planning phase that performs comprehensive collision and proximity analysis for the entire trajectory to ensure safety, and a local tracking phase that follows the pre-computed trajectory with reduced computational requirements. This segmentation allows the system to maintain high safety standards while improving real-time computational efficiency.
Solution Approach 2:
The patent performs comprehensive collision and proximity analysis in advance during the global planning phase before the vehicle executes the trajectory. By pre-computing the safe trajectory with full safety constraints satisfied, the system eliminates the need for continuous real-time collision checking during execution, thereby reducing computational load while maintaining safety guarantees.
2Measurement precision
If trajectory planning considers dynamic constraints and predicted responses of other objects, then trajectory accuracy and safety are improved, but computational complexity increases
Solution Approach 1:
The patent divides trajectory planning into global planning that handles complex dynamic constraints and local tracking that follows the pre-computed path. The global phase computes the accurate trajectory considering all dynamic constraints and predicted object responses, while the local phase simply tracks this trajectory with minimal computational overhead.
Solution Approach 2:
The system pre-computes the accurate trajectory satisfying all dynamic constraints and predicted object responses before execution. This preliminary computation ensures high trajectory accuracy while allowing the execution phase to operate with reduced computational complexity by following the pre-determined path.
3Reliability
If trajectory length is extended to ensure safe avoidance and stopping distance, then safety is improved, but computational resources and processing time increase
Solution Approach 1:
The patent performs comprehensive safety analysis including collision and proximity checks for the entire extended trajectory in the global planning phase before execution. By pre-computing the safe trajectory with adequate stopping distances and avoidance margins, the system ensures safety is maintained while the execution phase experiences reduced processing time since no real-time safety checks are needed.
Data Source
AI summary
Aspects of the disclosure provide a method of generating and following planned trajectories for an autonomous vehicle. For instance, a baseline for a planned trajectory that the autonomous vehicle can use to follow a route to a destination may be determined. A stopping point corresponding to a traffic control that will cause the autonomous vehicle to come to a stop using the baseline may be determined. Sensor data identifying objects and their locations may be received. A plurality of constraints may be generated based on the sensor data. A planned trajectory may be generated using the baseline, the stopping point, and the plurality of constraints, wherein constraints beyond the stopping point are ignored.