Decentralized Multi-Agent Trajectory Planning for Collision-Free AMRs

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current decentralized motion and path planning methods for autonomous agents, such as Autonomous Mobile Robots (AMRs), face challenges in cluttered environments due to unreliability, lack of completeness and optimality guarantees, and high computational complexity, especially in semi-structured and dynamic settings.

Innovation Solution

A decentralized algorithm that generates collision-free trajectories using a multi-stage approach, including efficient stochastic planning, adaptation for neighboring AMRs' sub-goals, and dynamic constraint consideration, ensuring collision-free paths in 2D and 3D spaces within a shared environment.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If decentralized methods are used for path planning, then computational complexity is reduced and scalability is improved, but reliability and completeness guarantees are lost

Engineering Contradiction:

Improvecomputational complexityVSAvoidcompleteness guarantee

Core Design Contradiction:

Device complexityVSReliability

Solution Approach 1:

The path planning problem is segmented into multiple independent planning horizons (current horizon and future horizons). Each AMR independently plans for its current horizon based on known trajectories of other AMRs, while maintaining awareness of future horizons. This segmentation enables decentralized computation while preserving reliability through systematic collision checking across all horizons.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The algorithm performs preliminary collision checking between the current AMR's planned trajectory and future trajectories of neighboring AMRs before finalizing the path. By anticipating potential conflicts in advance and adjusting the current plan accordingly, the system guarantees collision-free operation without requiring centralized coordination.

Inventive Principle:

Principle #10Preliminary action

2Ease of operation

If decentralized reactive collision avoidance is used, then individual robot path generation is simplified, but reliability in cluttered environments deteriorates

Engineering Contradiction:

Improvepath generation simplicityVSAvoidcollision avoidance reliability

Core Design Contradiction:

Ease of operationVSReliability

Solution Approach 1:

Instead of reacting to collisions after they occur, the algorithm performs preliminary collision checking by comparing the current AMR's planned trajectory with future trajectories of neighboring AMRs. This advance planning ensures collision avoidance reliability while maintaining decentralized operation and computational efficiency.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system incorporates feedback mechanisms where each AMR continuously monitors the trajectories of neighboring AMRs and adjusts its own path planning accordingly. This feedback loop ensures reliable collision avoidance in cluttered environments while preserving the simplicity of decentralized individual path generation.

Inventive Principle:

Principle #23Feedback

3Manufacturing precision

If centralized control is used for multi-AMR coordination, then global trajectory optimization is achieved, but scalability and computational efficiency deteriorate

Engineering Contradiction:

Improvetrajectory optimizationVSAvoidcentral controller complexity

Core Design Contradiction:

Manufacturing precisionVSDevice complexity

Solution Approach 1:

The centralized coordination problem is segmented into decentralized individual planning tasks. Each AMR independently optimizes its trajectory by considering future horizons of other AMRs, eliminating the need for a complex central controller while achieving globally optimized collision-free paths through distributed computation.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Each AMR performs self-service trajectory optimization by independently planning its path while considering the future trajectories of neighboring AMRs. This self-service approach achieves global optimization without requiring complex centralized control, improving scalability and computational efficiency.

Inventive Principle:

Principle #25Self-service

4Productivity

If decentralized planning is used, then scalability is improved, but global solution completeness is lost

Engineering Contradiction:

ImprovescalabilityVSAvoidglobal solution completeness

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The planning problem is segmented into current and future horizons, allowing each AMR to independently plan for its current horizon while considering future horizons of others. This segmentation enables scalable decentralized operation while maintaining global solution completeness through systematic collision checking across all horizons.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Each AMR incorporates feedback from future trajectories of neighboring AMRs into its current planning decisions. This feedback mechanism ensures that decentralized planning achieves global solution completeness by preventing conflicts before they occur, while maintaining scalability through independent local computation.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS20220236736A1Decentralized trajectory planning for multi-agent coordination

Publication Date: 2022.07.28 INTEL CORP

AI summary

Techniques are disclosed for a decentralized path and motion planning of autonomous agents within an environment. The planning may include determining if an active neighboring autonomous agent is present and selectively controlling the autonomous agent to operation in in an independent path planning operation mode and in a coordinating path planning operation mode, based on the detection of the neighboring agent(s).