Autonomous Vehicle Risk Monitoring for Sensor Failure Response
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for pre-deployment verification and validation of autonomous vehicle components are insufficient in ensuring safety, reliability, and optimal performance due to the dynamic nature of environments, necessitating a more holistic approach to run-time operational monitoring.
Innovation Solution
A method and system for run-time operational monitoring of autonomous vehicles using multiple monitoring parameters to evaluate the connection statuses and consistency of imaging sensors, adjusting severity levels based on external factors, and implementing remedial actions when thresholds are exceeded, including sensor fusion adjustments and learning-enabled model switching.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If pre-deployment verification and validation of components and sensors are performed, then manufacturing precision is improved, but reliability under dynamic environments deteriorates
Solution Approach 1:
The system performs preliminary calibration of sensors and components during manufacturing, establishing baseline performance metrics before deployment. This preliminary action ensures that components meet specification thresholds while enabling runtime comparison to detect deviations caused by dynamic environmental factors.
Solution Approach 2:
The patent implements continuous feedback loops during runtime operation that monitor sensor outputs and component performance against expected ranges. When deviations are detected, the system provides feedback to adjust operational parameters or trigger remedial actions, thereby maintaining reliability despite dynamic environmental changes.
2Reliability
If multiple monitoring parameters are used to evaluate sensor connection statuses and consistency, then reliability is improved, but device complexity increases
Solution Approach 1:
The monitoring system is segmented into modular components, each responsible for specific aspects such as sensor connection status monitoring, data consistency verification, and threshold evaluation. This segmentation allows the complex monitoring function to be implemented through independent, manageable modules that can be configured based on specific operational needs.
Solution Approach 2:
The patent implements a universal monitoring framework that can evaluate multiple sensor types and component categories using a common set of principles and algorithms. The system applies consistent monitoring logic across diverse sensors (imaging sensors, LIDAR, radar) and components (processing units, communication modules), reducing overall system complexity through standardization.
3Reliability
If dynamic risk assessment and remedial actions are implemented, then reliability is improved, but use of energy increases
Solution Approach 1:
The system dynamically adjusts the intensity and frequency of monitoring activities based on operational context and detected risk levels. During normal operation, monitoring operates at baseline levels to conserve energy. When anomalies or high-risk conditions are detected, the system dynamically increases monitoring intensity and triggers remedial actions only when necessary, optimizing the balance between safety and energy consumption.
Solution Approach 2:
The patent implements parameter-based thresholding where monitoring intensity and remedial action triggers are adjusted based on changing operational parameters such as vehicle speed, environmental conditions, and sensor confidence levels. This allows the system to maintain high reliability when needed while reducing computational overhead during low-risk scenarios.
Data Source
AI summary
A method and system for controlling a vehicle are provided. The method comprises: in parallel to a first processing pipeline for controlling the vehicle, executing a risk management processing pipeline comprising: acquiring first monitoring data from a first monitoring source used in first processing pipeline; acquiring second monitoring data from a second monitoring source used in the first processing pipeline; generating a cumulative risk assessment value based on a combination of the first monitoring data and the second monitoring data; in response to the cumulative risk assessment value being above a pre-determined main threshold, generating a warning notification; and in response to a frequency of occurrence of the generating the warning notification being above a pre-determined frequency threshold, triggering, independently from the first processing pipeline, a remedial action to be performed by the vehicle. The present technology may allow increasing safety of operation of the vehicle.