Secured ADAS Perception Resource Manager for ML Compute Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
ADAS systems in vehicles are vulnerable to attacks that manipulate sensor data, and managing limited computing resources between perception and security tasks is challenging.
Innovation Solution
A resource manager allocates and manages computing resources between perception and security ensembles of machine learning models, disabling less relevant models when resource thresholds are exceeded to ensure adequate functionality and security.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple ML models run concurrently for both perception and security tasks, then system functionality and security are improved, but computing resource consumption increases
Solution Approach 1:
The system dynamically adjusts the operational state of ML models based on real-time resource availability. The resource manager continuously monitors computing resources and transitions models between enabled and disabled states, making the system adaptable to changing resource conditions while maintaining security when resources permit
Solution Approach 2:
The system changes the operational parameter of ML models (enabled/disabled state) based on resource thresholds. When resource consumption exceeds predefined thresholds, the system modifies the operational parameters by disabling less critical models, thereby controlling resource usage while preserving essential functionality
2Productivity
If resource thresholds are set low to conserve computing resources, then resource efficiency is improved, but system security and perception capability deteriorate
Solution Approach 1:
The system applies different resource allocation strategies to different ML models based on their criticality. Essential perception models maintain higher resource allocation while less critical security models are more aggressively managed. This local differentiation allows efficient resource usage without uniformly compromising security capabilities
3Reliability
If all ML models are kept enabled to maintain security, then security monitoring is improved, but device complexity and resource management difficulty increase
Solution Approach 1:
The resource manager implements automated self-service mechanisms by continuously monitoring resource consumption and autonomously making decisions about model enabling/disabling. This self-service approach eliminates the need for manual resource management intervention, reducing operational complexity while maintaining security monitoring through automated model selection
Data Source
AI summary
Techniques and systems are provided for managing computing resources, comprising: obtaining resource usage information based on computational resources used by one or more perception ensembles of machine learning (ML) models and computational resources used by one or more security ensembles of ML models, wherein one or more ML models of a functional ensemble of ML models are configured to perform one or more perception tasks, and wherein one or more ML models of a security ensemble of ML models are configured to perform one or more security tasks; and disabling the ML model based on a comparison between the resource usage information to a first threshold.