Autonomous Vehicle Policy Curation for Low-Latency Decision Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous vehicle systems face challenges in efficiently simulating and selecting optimal policies due to high computing resource requirements, leading to increased latency and inability to consider all candidate policies.
Innovation Solution
A system and method for dynamically curating policies by leveraging infrastructure devices, teleoperator input, and onboard computing to reduce computing resources and prioritize policies based on environmental conditions, allowing for real-time adjustments and optimal policy selection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a large number of candidate policies are simulated to ensure optimal policy selection, then the quality of policy decision improves, but the computing resource requirements and latency increase significantly
Solution Approach 1:
The system performs preliminary actions by pre-simulating policies offline and storing their outcomes in a database. During real-time operation, the curated policy selector retrieves pre-simulated policies from the database rather than simulating all policies from scratch, significantly reducing latency while maintaining decision quality.
Solution Approach 2:
The system extracts and separates the computationally intensive simulation process from the real-time decision-making process. By extracting simulation results and storing them separately in a database, the system can quickly retrieve appropriate policies without performing full simulations during time-critical moments.
2Measurement precision
If all candidate policies are simulated to ensure comprehensive evaluation, then the accuracy of policy selection improves, but the computing resource requirements increase
Solution Approach 1:
The system applies local quality by curating a subset of policies that are locally relevant to the current situation rather than evaluating all possible policies uniformly. The curated policy selector uses situation descriptors to identify and retrieve only the policies applicable to the specific context, reducing computational resources while maintaining evaluation accuracy for relevant policies.
Solution Approach 2:
The system performs partial action by simulating and evaluating only a curated subset of policies that are most relevant to the current situation, rather than performing exhaustive simulation of all possible policies. This partial evaluation approach consumes fewer computing resources while still achieving accurate policy selection for the given context.
3Adaptability or versatility
If the policy set is expanded to cover all possible scenarios, then the adaptability of the autonomous agent improves, but the complexity of policy management and simulation increases
Solution Approach 1:
The system segments the large policy set into smaller, situation-specific subsets using situation descriptors and tags. By organizing policies according to their applicable contexts and characteristics, the system manages complexity through structured segmentation while maintaining comprehensive scenario coverage through the segmented policy database.
Solution Approach 2:
The system implements dynamic policy curation by adapting the policy subset based on the current situation. The curated policy selector dynamically determines which policies to retrieve from the database based on real-time situation descriptors, making the policy management process adaptive and flexible rather than static and rigid.
Data Source
AI summary
A system for dynamic policy curation includes a computing system and interfaces with an autonomous agent. A method for dynamic policy curation includes collecting a set of inputs; processing the set of inputs; and determining a set of available policies based on processing the set of inputs. Additionally or alternatively, the method can include any or all of: selecting a policy; implementing a policy; and/or any other suitable processes.