Redundant Vehicle Trajectory Validation Across Dual SoCs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional techniques fail to meet Automotive Safety Integrity Level (ASIL) D requirements for L2+ driving scenarios, particularly in urban and rural settings, due to the inability to limit steering errors and ensure rapid driver intervention in autonomous vehicles, which can lead to hazardous situations.
Innovation Solution
Implementing redundant algorithms and hardware components for vehicle trajectory validation, using different sets of vehicle sensors to independently validate computed trajectories, ensuring redundancy without full duplication, and leveraging trained machine learning models to project and verify vehicle paths.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional single-algorithm trajectory computation is used in L2+ driving assistance, then device complexity is reduced, but reliability of trajectory validation fails to meet ASIL-D requirements
Solution Approach 1:
The validation system is segmented into multiple independent components: a primary validation algorithm and a secondary validation algorithm, each executing on separate hardware threads or processors. This segmentation allows independent verification of the same trajectory computation, thereby improving reliability to meet ASIL-D requirements while maintaining manageable system complexity through modular architecture
Solution Approach 2:
Different validation algorithms are assigned to different computational resources with specialized functions. The primary validation algorithm handles standard trajectory verification, while the secondary validation algorithm provides redundant verification for critical safety checks. This local differentiation of validation quality across computational components enables ASIL-D compliance without uniformly increasing complexity throughout the entire system
2Reliability
If driver intervention is required for trajectory correction, then system autonomy is reduced, but safety in case of wrong steering is improved
Solution Approach 1:
The system performs preliminary validation of computed trajectories using multiple independent validation algorithms before the trajectory is executed. By pre-validating trajectories against road boundaries, obstacles, and safety constraints, the system prevents wrong steering before it occurs, maintaining high automation levels while ensuring safety through advance verification rather than requiring driver intervention
Solution Approach 2:
The redundant validation system continuously monitors trajectory computations and provides feedback on validation status. When validation fails or uncertainty is detected, the system generates alerts to the driver, creating a feedback loop that maintains automation while enabling safety-critical driver awareness and intervention only when necessary
3Reliability
If control limiters are applied to mitigate steering hazards, then safety is improved, but adaptability to different driving scenarios is reduced
Solution Approach 1:
The validation system dynamically adjusts its verification strictness and alerting behavior based on the current driving scenario. In high-risk scenarios such as rural roads with barriers or urban environments with oncoming traffic, the system intensifies validation and lowers alert thresholds. In lower-risk scenarios, validation remains active but operates with reduced intensity, thereby maintaining safety without unnecessarily limiting system adaptability across diverse driving conditions
Data Source
AI summary
Techniques are disclosed for validating a vehicle trajectory using redundancy in hardware and software components. A computed vehicle trajectory may be validated independently via two separate SoCs by projecting the 3D computed vehicle trajectory onto a 2D image acquired by a vehicle camera. Each SoC may perform an independent trajectory validation with the use of a trained machine learning model such as a deep neural network (DNN). The DNNs implemented by each SoC may perform trajectory validation using a separate set of camera inputs for the mapping and validation process. The vehicle implements the vehicle trajectory for control functions only when the trajectory is validated by both SoC trajectory validators, thus providing a robust trajectory validation process that complies with regulatory requirements such as Automotive Safety Integrity Level (ASIL) level D.