Autonomous Vehicle Trajectory Scoring With Dynamic Actor Margins
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous driving systems face computational inefficiencies when determining optimal trajectories for lane changes and in-lane maneuvers due to the need to process vast amounts of information and calculate numerous constraints, often resulting in suboptimal or unfeasible solutions, especially in dense traffic conditions.
Innovation Solution
A computer-implemented method for autonomous vehicles that generates continuous scores for candidate trajectories based on dynamic margins between the vehicle and predicted trajectories of other actors, using a distance table regeneration and margin scoring system to evaluate risks and passenger comfort, allowing for strategic maneuvering and improved trajectory planning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing autonomous driving systems evaluate a finite number of candidate trajectories using discretization or random sampling, then the computational cost is reduced, but the system fails to find feasible or optimal trajectories when they exist
Solution Approach 1:
The system changes the parameter of trajectory evaluation from discrete sampling to continuous scoring. By implementing a continuous margin-based scoring system that evaluates trajectories based on dynamic margins to actors, the system achieves both computational efficiency and reliability. The continuous score allows for nuanced evaluation without requiring exhaustive sampling of all possible trajectories.
Solution Approach 2:
The system performs preliminary action by pre-calculating distance tables and margin scores for candidate trajectories before final selection. This allows the system to efficiently evaluate and rank trajectories in advance, ensuring that feasible and optimal trajectories are identified without requiring exhaustive real-time computation during critical decision-making moments.
2Measurement precision
If the system calculates cost functions for numerous constraints in dense traffic conditions, then more accurate trajectory evaluation is achieved, but the computational cost becomes infinitely expensive
Solution Approach 1:
The system extracts only the essential margin information needed for trajectory evaluation rather than computing all possible constraint cost functions. By focusing on dynamic margins to actors and using pre-computed distance tables, the system extracts the critical safety and comfort metrics without the computational burden of evaluating every possible constraint combination in dense traffic scenarios.
3Adaptability or versatility
If the system explores more variations of lane change maneuvers including transition start/end locations, times, and steering profiles, then the ability to plan successful maneuvers in dense traffic improves, but the computational cost increases significantly
Solution Approach 1:
The system applies partial action by evaluating a focused set of margin-based scores for candidate trajectories rather than exhaustively computing all possible maneuver variations. The continuous margin scoring provides sufficient discrimination to identify successful maneuvers without requiring complete exploration of all transition parameters, achieving effective adaptability with reduced computational effort.
Data Source
AI summary
Provided are autonomous vehicles (AV), computer program products, and methods for maneuvering an AV in a roadway, including receiving forecast information associated with predicted trajectories of one or more actors in a roadway, determining a relevant trajectory of an actor based on correlating a forecast for predicted trajectories of the actor with the trajectory of the AV, regenerate a distance table for the relevant trajectory previously generated for processing constraints, generate a plurality of margins for the AV to evaluate, the margins based on a plurality of margin types for providing information about risks and effects on passenger comfort associated with a future proximity of the AV to the actor, classifying an interaction between the AV and the actor based on a plurality of margins, and generating continuous scores for each candidate trajectory that is also within the margin of the actor generated for the relevant trajectory.