Autonomous Driving Control-State Detection via Reference Model Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current autonomous driving systems face challenges in detecting unexpected control states during real-world road testing, particularly in reproducing extreme traffic conditions and subtle corner cases, leading to inefficiencies and potential safety issues due to limitations in virtual testing methods.
Innovation Solution
A computer-implemented method and apparatus for detecting unexpected control states in autonomous driving systems by generating environmental data, determining a first control state based on this data, and comparing it with a reference model-defined second control state that adheres to traffic rules, using a state machine to identify inconsistencies and violations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If large-scale road testing is conducted in real-world traffic, then safety verification is improved, but time consumption and cost increase
Solution Approach 1:
The patent creates virtual copies of real-world traffic scenarios through a reference model that simulates expected control states. Instead of testing exclusively in physical road environments, the system generates virtual test cases that replicate traffic conditions, allowing safety verification to be performed in silico rather than requiring extensive physical road testing.
Solution Approach 2:
The reference model pre-defines expected control states and traffic rule compliance before actual testing occurs. By establishing what correct behavior should be in advance through the state machine model, the system can efficiently compare actual autonomous vehicle performance against predetermined safety criteria, reducing the need for iterative real-world testing.
2Loss of time
If virtual testing methods are used, then time and cost are reduced, but ability to detect unexpected control states deteriorates
Solution Approach 1:
The system implements a feedback mechanism where the actual control states produced by the autonomous driving system are continuously compared against the expected control states from the reference model. This comparison generates feedback information that identifies deviations and unexpected control states, maintaining high detection accuracy while operating in the efficient virtual environment.
Solution Approach 2:
The patent introduces an intermediary comparison mechanism that bridges the virtual reference model and actual system output. This intermediary layer analyzes the differences between expected and actual control states, enabling the virtual testing system to detect unexpected behaviors with high precision by examining deviations from the reference model's predetermined correct behavior.
3Productivity
If reference model comparison is implemented, then detection efficiency is improved, but system complexity increases
Solution Approach 1:
The patent segments the safety verification process into distinct modular components: the reference model that defines expected states, the actual system that produces control states, and the comparison mechanism that identifies deviations. This segmentation allows each component to be developed and maintained independently, managing overall system complexity while enabling efficient parallel operation and high detection throughput.
Data Source
AI summary
The disclosure describes various embodiments for detecting an unexpected control state of an autonomous driving system. According to an embodiment, an exemplary method of detecting an unexpected control state of an autonomous driving system include the operations of generating environmental data of a vehicle; determining, by the autonomous driving system, a first control state based on the environmental data of the vehicle; determining, by a reference model, a second control state based on the environmental data, wherein the reference model defines at least one scenario each corresponding to a plurality of expected control states and a state switching condition, and in each of the expected control states corresponding to the scenario, an action of the vehicle in the scenario obeys a traffic rule; and determining the unexpected control state of the autonomous driving system by comparing the first control state with the second control state.