Autonomous Vehicle Scenario Reasoning With Episodic Likelihoods
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current symbolic or logic-based reasoning systems in autonomous vehicles are inflexible and slow when processing incomplete or ambiguous information, particularly in dynamic environments with unexpected events, leading to inefficient trajectory planning.
Innovation Solution
A method and system that infer multiple scenarios based on the current environment state, using historical likelihoods and logic-based likelihoods, with tokenization to determine similarity to past events, and applying abduction, deduction, or a combination of both to select the most appropriate scenario for the autonomous vehicle's operation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If symbolic or logic-based reasoning systems are used to generate binary hypotheses, then the system maintains logical consistency and reliability, but the system becomes inflexible and slow when processing incomplete or ambiguous information
Solution Approach 1:
The patent introduces an intermediary module that translates natural language traffic rules into logical constraints. This intermediary layer allows the system to handle incomplete or ambiguous information by interpreting it through natural language understanding before applying strict logical reasoning, thus maintaining both reliability and adaptability
Solution Approach 2:
The reasoning system is segmented into multiple independent components: a hypothesis generation module that creates multiple possible scenarios, an evaluation module that assesses each scenario against logical constraints, and a selection module that chooses the most plausible scenario. This segmentation allows parallel processing of multiple hypotheses, improving both speed and flexibility
2Reliability
If logic-based reasoning generates binary hypotheses, then the system ensures deterministic outcomes, but the system stalls when encountering unexpected events or conflicting data from independent agents
Solution Approach 1:
The system dynamically adjusts the number of hypotheses generated and the depth of logical evaluation based on the complexity and uncertainty of the traffic situation. For routine situations, the system quickly evaluates a limited set of hypotheses, while for unexpected events, it generates and evaluates more hypotheses in parallel, maintaining deterministic outcomes while improving decision-making speed
Solution Approach 2:
The system pre-loads and caches logical constraints representing common traffic rules and scenarios. When encountering a new situation, it quickly retrieves relevant pre-processed constraints rather than deriving them from scratch, significantly reducing computation time while maintaining deterministic evaluation
3Adaptability or versatility
If the system processes multiple scenarios with historical likelihoods, then the system improves adaptability to real-time traffic conditions, but the system complexity increases
Solution Approach 1:
The system employs a universal hypothesis evaluation framework that can handle multiple types of scenarios (normal traffic flow, unexpected events, ambiguous situations) using the same logical constraint checking mechanism. This multi-functional approach improves adaptability without proportionally increasing system complexity, as the same core engine serves multiple purposes
Data Source
AI summary
A vehicle and a system and a method of operating the vehicle. The system includes a reasoning engine, an episodic memory, a resolver and a controller. The reasoning engine infers a plurality of possible scenarios based on a current state of an environment of the vehicle. The episodic memory determines a historical likelihood for each of the plurality of possible scenarios. The resolver selects a scenario from the plurality of possible scenarios using the historical likelihoods. The controller operates the vehicle based on the selected scenario.