Autonomous Vehicle Software Message Testing for Non-Determinism
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicle software exhibits non-deterministic behavior under the same circumstances, leading to unpredictable driving decisions, which is critical to address for reliable operation.
Innovation Solution
Run multiple simulations of the software and compare output messages from decision-making modules to identify and flag differences indicative of non-determinism, filtering out irrelevant messages and setting tolerances to focus on significant deviations, aggregating data to detect anomalies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple simulations are run and compared to detect non-determinism, then software reliability is improved, but processing time and computational resources increase
Solution Approach 1:
The patent segments the simulation output into discrete messages with specific fields that can be independently compared. By dividing the output comparison into field-level operations (timestamp fields, decision fields, control fields), the system efficiently identifies non-determinism without requiring complete re-analysis of entire simulation datasets, thus reducing processing time while maintaining reliability detection capability
Solution Approach 2:
The patent performs preliminary actions by pre-defining tolerance thresholds for acceptable differences between simulation runs and pre-identifying critical fields that indicate non-determinism. This preliminary configuration allows the comparison process to focus only on relevant differences within acceptable bounds, significantly reducing the computational burden of multiple simulation comparisons while maintaining reliable detection of actual non-deterministic behavior
2Measurement precision
If detailed comparison of all output fields is performed, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent applies local quality by differentiating the comparison depth applied to different fields based on their importance. Critical fields (such as decision-making outputs and control commands) undergo precise field-level comparison, while less critical fields use coarser comparison methods or are excluded entirely. This selective approach maintains high measurement precision for safety-critical parameters while reducing overall system complexity
Solution Approach 2:
The patent changes parameters by introducing tolerance thresholds that dynamically adjust the stringency of field comparisons. For numerical fields, tolerance ranges define acceptable variation without indicating non-determinism. This parameter-based approach enables precise detection of meaningful differences while filtering out negligible variations, thereby maintaining measurement precision without requiring overly complex comparison logic
Data Source
AI summary
The disclosure relate to detecting differences indicative of non-determinism in software used to control autonomous vehicles. For instance, a simulation may be run using the software at least two times. The output of the simulations may be tracked. The output of the simulations includes messages may be generated by a decision making module of the software that affect vehicle behavior. A pair of messages generated from each of the at least two times the simulation is run may be compared. Based on the comparison, a difference indicative of non-determinism in the software may be detected.