Linear Array Trajectory Evaluation for Low-Latency Vehicle Planning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Evaluating proposed trajectories for autonomous vehicles requires fast and responsive computation to ensure safe navigation, but existing methods face challenges due to increased latency and resource demands, limiting the number of objects that can be evaluated and potentially leading to unsafe operations.
Innovation Solution
A method involving a first processing unit determining prediction data for objects within a region of interest (ROI) and converting it into a linear array, which is then processed by a second processing unit, such as a GPU, to efficiently evaluate proposed vehicle trajectories, reducing computational resources and latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional trajectory evaluation methods are used to evaluate proposed trajectories for autonomous vehicles, then comprehensive object evaluation can be performed, but latency increases and computational resources are excessive
Solution Approach 1:
The patent divides the trajectory evaluation process into two distinct stages: a first stage that performs comprehensive object evaluation and trajectory scoring, and a second stage that performs validation and refinement. This segmentation allows the system to maintain accurate evaluation while reducing overall latency by distributing computational tasks across different processing units and time periods.
Solution Approach 2:
The first processing unit performs preliminary trajectory evaluation and object assessment before the second processing unit conducts validation. By pre-computing trajectory scores and object relevances in advance, the system reduces the computational burden during the final decision-making phase, thereby reducing latency while maintaining evaluation comprehensiveness.
2Reliability
If comprehensive object evaluation is performed for trajectory planning, then safety is improved, but computational resource demands increase
Solution Approach 1:
The patent segments computational tasks between a first processing unit that handles comprehensive object evaluation and a second processing unit that performs validation. This division allows comprehensive safety evaluation to be performed while distributing computational resource consumption across multiple units, preventing any single unit from being overwhelmed.
Solution Approach 2:
The system performs evaluation on a subset of relevant objects identified by the first processing unit, rather than evaluating all possible objects. The second processing unit validates only the most critical trajectory candidates, performing partial evaluation that maintains safety while reducing overall computational resource demands.
3Reliability
If multiple trajectories are evaluated to ensure safe navigation, then collision avoidance improves, but processing speed decreases
Solution Approach 1:
The first processing unit pre-evaluates multiple trajectories and identifies relevant objects in advance, creating a shortened list of candidates for the second processing unit. This preliminary action allows comprehensive multi-trajectory evaluation to occur without sacrificing processing speed, as the second unit only needs to validate the most promising candidates.
Solution Approach 2:
The patent extracts and separates the validation function from the full evaluation process. The second processing unit takes out only the critical validation tasks from the complete trajectory evaluation pipeline, allowing rapid verification of collision avoidance for the most relevant trajectories without repeating the full evaluation for each candidate.
Data Source
AI summary
Techniques for storing predicted state using a linear array and for querying the linear array to evaluate a proposed trajectory are described herein. In some cases, a first processing unit (e.g., a central processing unit (CPU)) determines prediction data associated with an object in the vehicle environment and sends the prediction data to a second processing unit (e.g., a graphics processing unit (GPU)). In some cases, after the second processing unit receives the prediction data, the second processing unit stores the prediction data using a linear array. In some cases, the second processing unit can use the linear array to evaluate a proposed trajectory for the vehicle and determine whether to validate the proposed trajectory for the vehicle. In some cases, based on the validation of the proposed trajectory by the second processing unit, the vehicle is controlled based on the proposed trajectory.