Autonomous Fleet Route Planning With Batched Conflict Resolution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing multi-agent pathfinding techniques in automated systems, such as logistics facilities and warehouses, face increased runtime complexity due to the number of conflicts between autonomous transport vehicles, especially when traffic becomes severe, leading to inefficiencies in collision-free path planning.
Innovation Solution
Implementing a priority-based divide and search autonomous transport vehicle path planning process that divides route legs into batches, prioritizes them based on deferral penalties, and plans paths in a sequence to minimize conflicts, allowing for rapid path planning within large fleets of autonomous vehicles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If priority-based search techniques are used to resolve conflicts between autonomous transport vehicles, then collision-free path planning is achieved, but runtime increases cubically as the number of conflicts increases
Solution Approach 1:
The patent divides the fleet of autonomous transport vehicles into multiple groups based on their destinations and route legs. Each group is processed separately through the priority-based search algorithm, rather than processing all vehicles simultaneously. This segmentation reduces the number of conflicts that need to be resolved in each processing iteration, thereby reducing the cubic runtime complexity while still ensuring collision-free paths for all vehicles.
2Productivity
If the number of autonomous transport vehicles increases, then fleet capacity and throughput are improved, but the number of conflicts to resolve increases rapidly
Solution Approach 1:
The patent segments the large fleet into multiple smaller groups based on destination and route leg characteristics. This allows the system to handle larger overall fleet sizes by processing vehicles in manageable batches, reducing the complexity of conflict resolution while maintaining high fleet throughput. The segmented approach prevents the cubic complexity from scaling with the total fleet size.
Solution Approach 2:
The patent implements dynamic priority assignment where priorities are not fixed but are determined on-demand based on current traffic conditions, vehicle positions, and destination requirements. This dynamic approach allows the system to adapt to changing fleet compositions and traffic patterns, efficiently handling variable fleet sizes without proportionally increasing conflict resolution complexity.
3Stability of the object's composition
If centralized multi-agent processes are used for path planning, then coordinated routing is achieved, but runtime is dominated by the number of conflicts between agents
Solution Approach 1:
The patent segments the centralized planning process into multiple passes, where each pass handles a specific group of vehicles with similar destinations or route characteristics. This maintains the coordinated routing benefit of centralized control while reducing the runtime burden by processing fewer conflicts in each pass, avoiding the cubic complexity that would result from processing all vehicles simultaneously in a single centralized process.
Data Source
AI summary
An automated storage and retrieval system includes a storage array, a plurality of autonomous guided bots, and a controller connected to each autonomous guided bot to assign a series of tasks to each autonomous guided bot, which series of tasks includes a task to an autonomous guided bot moving the autonomous guided bot from initial location to a different final location via bot routes. The controller is configured with a bot route planner that has a resolver that seeks conflicts between autonomous guided bots on the bot routes describing bot paths, and resolves each bot route to determine the bot route, at least one bot route being determined based on bot priority, and the resolver sequences the bot routes into a sequence of bot route leg batches, where route legs, forming each bot route in entirety, are divided into corresponding batch of the sequence of route leg batches.