Autonomous Action Planning With Candidate Objectives for Abnormal Events

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional autonomous systems face challenges in reacting to unexpected events, experiencing long computation times for action strategy establishment, and employing static strategies that are unsuitable for dynamic environments, leading to potential interruptions and inefficiencies.

Innovation Solution

The development of a method for generating and executing adaptive action strategies by an autonomous system, which involves accessing databases of event descriptions and candidate objectives, selecting appropriate objectives based on activation conditions, and executing a series of actions through progressive task unit structures that include failure recovery modules, allowing for real-time adjustments and stochastic decision processes.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If conventional autonomous systems use static action strategies, then system simplicity is maintained, but the system cannot adapt to unexpected events or dynamic environments

Engineering Contradiction:
Improveadaptability to unexpected eventsVSAvoidsystem complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The action strategy is segmented into multiple candidate objectives, each with its own progressive task unit structure. This allows the system to evaluate multiple pre-defined strategies and select the most appropriate one for the current situation, enabling adaptability without requiring a completely complex dynamic generation system.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Multiple candidate objectives and their corresponding progressive task unit structures are prepared in advance. When an unexpected event occurs, the system can quickly select from these pre-prepared strategies rather than generating them in real-time, reducing computation time while maintaining adaptability.

Inventive Principle:
Principle #10Preliminary action

2Reliability

If conventional systems perform comprehensive computation for action strategy establishment, then optimal strategies are achieved, but computation time becomes excessively long causing interruptions

Engineering Contradiction:
Improveaction strategy optimalityVSAvoidcomputation time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

Candidate objectives and progressive task unit structures are computed and prepared in advance offline. During online operation, the system only needs to evaluate which pre-computed strategy is most appropriate for the current event, dramatically reducing computation time while maintaining strategy quality.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The comprehensive action strategy is divided into multiple candidate objectives with progressive task unit structures. This segmentation allows the system to evaluate smaller, manageable strategy components rather than computing one large strategy from scratch, reducing online computation time.

Inventive Principle:
Principle #1Segmentation

3Reliability

If autonomous systems use detailed progressive task unit structures with failure recovery modules, then task accomplishment reliability is improved, but device complexity increases

Engineering Contradiction:
Improvetask accomplishment reliabilityVSAvoidstrategy structure complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The task unit structure is segmented into hierarchical levels with clear progression. Each level has specific sub-tasks and failure recovery modules, allowing the system to handle complexity in a structured, manageable way while maintaining high reliability through progressive validation at each level.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Failure recovery modules are built into the progressive task unit structure in advance. When failures occur, the system can immediately execute pre-planned recovery actions without needing to compute new strategies, ensuring task accomplishment reliability while keeping the overall system manageable.

Inventive Principle:
Principle #11Beforehand cushioning (Prior cushioning)

4Speed

If autonomous systems implement real-time event detection and response mechanisms, then responsiveness to abnormal events is improved, but computation overhead increases

Engineering Contradiction:
Improveresponse speed to abnormal eventsVSAvoidcomputation energy
Core Design Contradiction:
SpeedVSUse of energy by moving object

Solution Approach 1:

Event detection rules and candidate objective activation conditions are established in advance. When events occur, the system only needs to check whether pre-defined activation conditions are met, rather than performing comprehensive analysis, significantly reducing real-time computation energy while maintaining fast response.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP4113062B1System and method for generation of action strategies by an autonomous system
Publication Date: 2024.12.11 ECOLE NAT SUPERIEURE DINGS DE CAEN
  • EP4113062B1 patent drawingFigure 1
  • EP4113062B1 patent drawingFigure 2
  • EP4113062B1 patent drawingFigure 3

AI summary

Systems and methods for generating an action strategy to be executed by an autonomous system are disclosed. The action strategy comprises a series of actions to be performed by the autonomous system to accomplish a corresponding active objective in response to detecting an abnormal event, the abnormal events occurring or having occurred in an environment where the autonomous system is configured to operate. The method comprises accessing a first database populated with event descriptions corresponding to abnormal events and accessing a second database populated with candidate objectives. Each candidate objective defines a task accomplishable by the autonomous system and comprises an activation condition and a progressive task unit structure describing a hierarchy of actions to be performed in order to accomplish the corresponding candidate objective. An execution of a candidate objective generating an action strategy from the progressive task unit structure of the active objective and executing the action strategy.