Autonomous Agent Self-Simulation for Adaptive Mission Planning

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Autonomous vehicles face limitations in adapting to changing environmental circumstances beyond the scope of anticipated world states, as their ability to emulate human behavior is restricted by pre-designed actions and environmental conditions.

Innovation Solution

An autonomous agent capable of online mission self-simulation, utilizing a strategy manager and self-simulator within a faster-than-real-time processing environment to select and switch between action sets, projecting future behaviors and optimizing mission objectives based on current conditions and mission status.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If autonomous vehicles operate based on pre-designed actions and predetermined routines, then the vehicle can execute missions reliably, but the vehicle cannot adapt to changing environmental circumstances beyond anticipated world states

Engineering Contradiction:
Improvemission execution reliabilityVSAvoidadaptability to changing environmental circumstances
Core Design Contradiction:
ReliabilityVSAdaptability or versatility

Solution Approach 1:

The system performs preliminary simulation of multiple possible future outcomes before executing actions. The digital twin simulates various action sets and environmental conditions in advance, allowing the autonomous vehicle to select the most appropriate action based on predicted outcomes, thus balancing reliability with adaptability

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system continuously monitors actual environmental conditions and compares them with simulated predictions. When discrepancies are detected between predicted and actual outcomes, the system adjusts its action selection in real-time, enabling adaptation to unforeseen environmental changes while maintaining reliable mission execution

Inventive Principle:
Principle #23Feedback

2Device complexity

If the vehicle behavior is strictly determined by pre-programmed actions, then the system complexity is reduced, but the scope of possible world states the vehicle can handle is limited

Engineering Contradiction:
Improvesystem complexityVSAvoidscope of handleable world states
Core Design Contradiction:
Device complexityVSAdaptability or versatility

Solution Approach 1:

The system creates a digital twin (virtual copy) of the autonomous vehicle and its environment. This copy allows simulation of numerous world states and scenarios without adding physical complexity to the actual vehicle. The digital twin handles the complexity of multiple world states while the physical vehicle remains relatively simple

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The system transitions from static pre-programmed actions to dynamic action selection based on real-time environmental conditions. The action set is no longer fixed but adapts dynamically based on simulated predictions and actual sensor data, expanding the scope of handleable world states without proportionally increasing system complexity

Inventive Principle:
Principle #15Dynamics

3Adaptability or versatility

If multiple action sets are maintained for different environmental conditions, then the vehicle can adapt to various conditions, but the difficulty of selecting the appropriate action set increases

Engineering Contradiction:
Improvebehavior adaptation capabilityVSAvoidaction set selection difficulty
Core Design Contradiction:
Adaptability or versatilityVSDifficulty of detecting and measuring

Solution Approach 1:

The system uses feedback from digital twin simulations to automatically evaluate and rank multiple action sets based on predicted outcomes. This feedback mechanism objectively determines which action set is most appropriate for current conditions, eliminating the difficulty of manual selection and enabling seamless adaptation

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The autonomous vehicle performs self-evaluation of multiple action sets through its own digital twin. The system independently simulates and compares outcomes without external intervention, automatically selecting the optimal action set based on simulated performance metrics, thus reducing selection difficulty

Inventive Principle:
Principle #25Self-service

4Reliability

If real-time simulation of multiple action sets is performed, then optimized action selection is achieved, but the computational processing time increases

Engineering Contradiction:
Improveaction selection optimalityVSAvoidcomputational processing time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system simulates only the most promising action sets identified through preliminary filtering rather than exhaustively simulating all possible actions. This partial simulation approach maintains action selection optimality while significantly reducing computational time by focusing resources on the most relevant options

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP4439420A1Autonomous agent with online mission self-simulation
Publication Date: 2024.10.02 ROCKWELL COLLINS INC
  • EP4439420A1 patent drawingFigure 1
  • EP4439420A1 patent drawingFigure 2
  • EP4439420A1 patent drawingFigure 3

AI summary

An autonomous agent (102) of a team (100) of autonomous agents (e.g., semi- or fully autonomous vehicles) includes a self-simulator incorporating a faster than real time (FTRT) processing environment for online simulation of each agent of the team. Based on the current mission status and one or more action sets determining the behaviors of the autonomous agents (102; 104; 106; 108; 110), the behaviors of each agent of the team are projected forward within the FTRT environment to determine mission status metrics relevant to the effectiveness of a particular action set towards optimal completion of mission objectives currently assigned to the team. Based on the mission status metrics, the self-simulator can select and provide an action set for optimized completion of mission objectives. For example, the self-simulator can recommend switching to a different preloaded action set or, in some cases, construct an optimized action set selected from multiple preloaded action sets tested in the FTRT environment.