Autonomous Vehicle Supervisor Monitoring for Re-Engagement Safety

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing Supervisor Performance Monitoring Systems (SPMS) are inadequate for autonomous vehicles (AVs) as they do not account for the additional time required for human supervisors to re-engage with vehicle operations and are based on assumptions about human reaction times that are unsuitable for AVs, leading to potential safety failures.

Innovation Solution

The Autonomous Vehicle Performance Monitoring System (AVPMS) monitors both autonomy and human supervision, incorporating models of human re-engagement time and adjusting alert thresholds to enhance safety by providing multiple levels of alerts and responses tailored to AVs, including vehicle maneuvers and warnings.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If autonomous systems operate without human intervention, then productivity and efficiency are improved, but safety and reliability deteriorate due to inability to handle unexpected situations

Engineering Contradiction:
Improveoperational efficiencyVSAvoidsafety
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The system implements continuous feedback loops where performance data from autonomous operations is collected, analyzed, and used to update monitoring models. Human supervisors receive real-time feedback about system performance and anomalies, enabling them to intervene when necessary while maintaining high automation levels.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

A human-in-the-loop monitoring system serves as an intermediary between autonomous operations and safety-critical decisions. The monitoring model acts as a mediator that continuously evaluates system state and alerts human supervisors only when anomalies are detected, allowing autonomous operation to proceed uninterrupted during normal conditions.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Reliability

If comprehensive monitoring of autonomous systems is implemented, then safety and reliability are improved, but device complexity and computational resources increase

Engineering Contradiction:
Improvesafety monitoringVSAvoidmonitoring system complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent extracts and separates the monitoring function from the autonomous system operations themselves. The monitoring model is trained independently on historical performance data and operates as a separate evaluation layer that assesses system state without interfering with core autonomous functions, thereby reducing complexity coupling.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The system creates a virtual copy or digital twin of the autonomous system's expected behavior through the monitoring model. This model replicates normal operational patterns and compares actual system state against the copied expected behavior, enabling comprehensive monitoring through pattern recognition rather than complex rule-based systems.

Inventive Principle:
Principle #26Copying

3Measurement precision

If extensive performance data is collected and analyzed, then measurement precision and anomaly detection are improved, but loss of time and computational energy increase

Engineering Contradiction:
Improveanomaly detection accuracyVSAvoiddata processing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The monitoring model is pre-trained offline on extensive historical performance data to learn normal operational patterns and anomaly signatures. This preliminary action transfers computational burden from real-time operation to offline training, enabling fast real-time inference with high precision without delaying autonomous system operations.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system applies monitoring and analysis selectively rather than continuously at full capacity. The monitoring model focuses computational resources on evaluating specific critical parameters and only triggers detailed analysis when anomaly indicators are detected, reducing average computational load while maintaining high detection precision when needed.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP3894973B1Human-supervised autonomous systems performance and safety monitoring methods and apparatus
Publication Date: 2026.05.06 EDGE CASE RESEARCH INC
  • EP3894973B1 patent drawingFigure 1
  • EP3894973B1 patent drawingFigure 2
  • EP3894973B1 patent drawingFigure 3

AI summary

Human supervisor failures in autonomous system operations are detected. Perception sensor data input is obtained from a sensor. A violation of a safety envelope by at least one external object is detected within the perception sensor data input. A response to the violation of the safety envelope is triggered. The safety envelope has predetermined dimensions that are based upon an expected time to mitigate the violation of the safety envelope.