ADS Perception Development Using Federated Worldview Evaluation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current Automated Driving Systems (ADS) face challenges in efficiently developing and verifying perception modules, leading to high costs and complexity, especially in managing vast amounts of data and ensuring safety and performance improvements.
Innovation Solution
A method and system for performance evaluation of a perception-development module in ADS-equipped vehicles, involving the storage of perception data, formation of a baseline worldview, matching and evaluation of perception data, and updating of perception model parameters using an optimization algorithm to minimize a cost function.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional field tests and data management methods are used for ADS development, then system safety and performance can be verified, but development costs and time consumption increase significantly
Solution Approach 1:
The patent creates virtual copies of the real-world environment through simulated worlds that replicate physical test scenarios. These virtual copies allow extensive testing and verification without the time and resource constraints of physical field tests, enabling parallel testing across multiple scenarios simultaneously.
Solution Approach 2:
The system performs preliminary actions by pre-generating and storing multiple simulated worlds representing various driving scenarios before actual testing begins. This allows the ADS to be tested against pre-prepared scenarios, eliminating the need to conduct time-consuming field tests for each scenario and enabling faster iterative development.
2Reliability
If extensive field tests are conducted to verify ADS safety, then system reliability improves, but development costs increase
Solution Approach 1:
The patent replaces expensive physical field tests with virtual copies of test environments. These simulated worlds can be generated and reused multiple times without additional cost, allowing comprehensive safety verification at a fraction of the cost of traditional field testing programs.
Solution Approach 2:
The simulated worlds serve multiple functions: they can be used for initial system verification, continuous performance evaluation, and testing of multiple ADS configurations simultaneously. This multi-functionality eliminates the need for separate testing programs for different scenarios, reducing overall development costs while maintaining comprehensive safety verification.
3Measurement precision
If more sensors and computational hardware are added to improve perception accuracy, then measurement precision improves, but device complexity increases
Solution Approach 1:
The patent uses virtual copies of sensor data and environmental conditions in simulated worlds to train and evaluate perception algorithms. This allows the system to achieve high measurement precision through software-based simulation rather than adding more physical sensors, thereby improving perception accuracy without increasing hardware complexity.
4Manufacturing precision
If vast amounts of perception data are collected for training and verification, then model accuracy improves, but data management complexity and storage requirements increase
Solution Approach 1:
The patent generates synthetic perception data through simulated worlds, creating virtual copies of real-world sensing scenarios. This synthetic data can be generated on-demand in unlimited quantities without the logistical challenges of collecting, storing, and managing vast amounts of real-world data, thereby improving model accuracy while simplifying data management infrastructure.
Data Source
AI summary
The present invention relates to methods and systems that utilize the production vehicles to develop new perception features related to new sensor hardware as well as new algorithms for existing sensors by using federated learning. To achieve this, the production vehicle's own worldview is post-processed and used as a reference, towards which the output of the software (SW) or hardware (HW) under development is compared. Through this comparison, a cost function can be calculated and an update of the SW parameters can be locally updated according to this cost function.