Autonomous Vehicle Path Planning Evaluation by Candidate Path Convergence
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
There is a need for more efficient solutions to develop, verify, and validate path planning functions for Automated Driving Systems (ADSs) to ensure safer and more performant systems, while allowing for more frequent releases without compromising safety or increasing costs, size, or power consumption.
Innovation Solution
A computer-implemented method for closed-loop evaluation of path planning modules in ADS-equipped vehicles, which involves obtaining candidate paths from multiple modules, determining convergence criteria, selecting and executing a candidate path, and evaluating the modules based on the executed path, while also considering exposure needs and operational design domains.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional field tests and verification methods are used for path planning modules, then safety and reliability are ensured, but development time and costs increase significantly
Solution Approach 1:
The system performs preliminary evaluation of path planning modules through simulation before deployment to real vehicles. Multiple modules are evaluated in advance using virtual test environments, allowing safety verification to occur before actual field testing, thus reducing development time while maintaining reliability standards
Solution Approach 2:
The system creates virtual copies of path planning modules and test scenarios in a simulation environment. Instead of testing directly on real vehicles, the system uses digital twins and virtual replicas to perform comprehensive safety verification, significantly reducing the time and cost associated with physical field tests
2Reliability
If multiple path planning modules are tested extensively in the field, then validation coverage improves, but the cost and complexity of the testing system increases
Solution Approach 1:
The system implements a universal evaluation platform that can assess multiple different path planning modules using the same simulation infrastructure. The virtual test environment serves multiple functions including safety verification, performance evaluation, and validation across different modules, reducing the need for separate dedicated test systems for each module
Solution Approach 2:
The evaluation system automatically performs comprehensive validation of path planning modules through autonomous simulation execution. The system self-manages the testing process, including automatic scenario generation, module evaluation, and result analysis, reducing the need for complex manual testing procedures and specialized test equipment
3Reliability
If conventional validation approaches are used, then thorough testing is achieved, but the frequency of releases decreases
Solution Approach 1:
Comprehensive validation testing is performed in advance through simulation before modules are ready for release. The system conducts thorough safety and performance evaluation during the development phase, allowing modules to be released more frequently to real vehicles since extensive field testing is no longer required
Solution Approach 2:
The system replaces physical field testing with virtual simulation-based validation. By substituting mechanical field tests with digital simulation environments, the system achieves thorough testing without the time delays and logistical constraints of physical deployment, enabling more frequent releases
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A computer-implemented method and related aspects for closed-loop evaluation of path planning modules for a vehicle equipped with an Automated Driving System (ADS) are disclosed. The method comprises obtaining a candidate path from each of a plurality of path planning modules of the ADS, wherein each path planning module is configured to generate candidate paths to be executed by the ADS of the vehicle. The method further comprises determining a fulfilment of one or more convergence criteria by the obtained candidate paths by comparing the obtained candidate paths with each other and determining a level of convergence between the candidate paths. Moreover, in an instance where the obtained candidate paths fulfil the one or more convergence criteria, the method further comprises selecting one of the obtained candidate paths and controlling the vehicle so to execute the selected candidate path, and evaluating each of the plurality of path planning modules based on the executed candidate path. However, in an instance where the obtained candidate paths do not fulfil the one or more convergence criteria. The method further comprises determining an exposure need of each path planning module in view of a predicted scene or scenario in the surrounding environment of the vehicle that the vehicle is expected to be exposed to while executing any one of the obtained candidate paths. Furthermore, the method comprises selecting one candidate path based on the determined exposure need and the determined level of convergence and controlling the vehicle so to execute the selected candidate path, and evaluating the path planning module that generated the selected candidate path based on the executed candidate path.