Open Autonomy Kernel for Autonomous Fault Diagnosis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current automation efforts in complex connection-based systems, such as engineering plants on naval vessels, rely heavily on manual interpretation and response to sensor data, requiring significant human diagnosis and input, especially in damage control scenarios, which limits manpower efficiency.
Innovation Solution
The Open Autonomy Kernel (OAK) architecture introduces a goal-directed commanding approach, using intelligent subsystem management control agents and declarative model-based reasoning to enable real-time detection, identification, and diagnosis of fault conditions, allowing for dynamic and adaptable coordination across a distributed network.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If expert diagnostic knowledge is used in the form of coded rules or procedures interpreted by the system, then automation capability is improved, but significant amounts of human diagnosis and input are still required
Solution Approach 1:
The control agents are equipped with model-based reasoning capabilities that enable them to autonomously perform detection, identification, and diagnosis of fault conditions without requiring human intervention. The system serves itself by having control agents continuously monitor subsystem status, compare it against declarative models, and autonomously generate diagnostic conclusions and response actions.
Solution Approach 2:
The system divides the complex diagnostic task into smaller sub-tasks distributed across multiple control agents, each responsible for specific subsystems. Each control agent independently performs monitoring, analysis, and diagnosis for its assigned subsystem, allowing parallel processing and reducing the need for centralized human oversight.
2Extent of automation
If traditional rule-based automation systems are used, then some diagnostic functions are automated, but the system lacks adaptability to unanticipated fault conditions
Solution Approach 1:
The system uses declarative models that can be dynamically updated and modified to reflect changing system conditions and new fault patterns. The model-based reasoning approach allows the control agents to adapt their diagnostic reasoning by updating model parameters and relationships based on observed system behavior, enabling response to previously unanticipated fault conditions.
Solution Approach 2:
The system transitions from static rule-based automation to dynamic model-based reasoning where control agents continuously update their understanding of system state and fault conditions. The declarative models allow the system to dynamically adapt its diagnostic and response strategies based on real-time observations, enabling flexibility in handling unexpected situations.
3Reliability
If manual interpretation and response to sensor data is used, then system monitoring is achieved, but significant manpower levels are required
Solution Approach 1:
The control agents autonomously perform the complete monitoring and diagnostic cycle without human intervention. They continuously monitor sensor data, interpret it using model-based reasoning, identify fault conditions, and generate response actions, completely replacing the need for human operators in routine monitoring tasks.
Solution Approach 2:
The control agents act as intelligent intermediaries between the physical subsystems and human operators. They translate raw sensor data into meaningful diagnostic information and automated response actions, filtering and processing information so that human operators only need to review high-level summaries when necessary, significantly reducing manpower requirements.
Data Source
AI summary
The Open Autonomy Kernel (OAK) addresses critical infrastructure requirements for next generation autonomous and semi-autonomous systems (24), including performance tracking, anomaly detection, diagnosis, fault recovery, and plant “safing”. OAK combines technologies in automated planning and scheduling, control agent-based systems (22), and model based reasoning to form a portable software architecture (26), knowledge-base, and open Application Programming Interface (API) to enable integrated auxiliary subsystem autonomy.