Task and Path Planning for Multi-Agent Pickup and Delivery Minghua - PowerPoint PPT Presentation

. Background . . . . . . . . . . Task Assignment . Prioritized Path Planning Hybrid Path Planning Experimental Results Task and Path Planning for Multi-Agent Pickup and Delivery Minghua Liu Hang Ma Jiaoyang Li Sven Koenig AAMAS May 16, 2019 . . . . . . . . . . . . . . . . . . . . . . . . . . . . Task and Path Planning for Multi-Agent Pickup and Delivery

. . . . . . . . . . . . Background . Task Assignment Prioritized Path Planning Hybrid Path Planning Experimental Results Real-World Applications Aircraft-Towing Vehicles Amazon Warehouse Robots Agents have to operate in a common environment, continuously attend to pickup and delivery tasks one by one, and plan collision-free paths to execute the tasks. . . . . . . . . . . . . . . . . . . . . . . . . . . . Task and Path Planning for Multi-Agent Pickup and Delivery

. . . . . . . . . . . . Background . Task Assignment Prioritized Path Planning Hybrid Path Planning Experimental Results Multi-Agent Pickup and Delivery (MAPD) Problem A team of agents ( a i ) have to execute a batch of tasks ( t i ) in a known environment. Each task: a pickup location ( s i ), a delivery location ( g i ), and a release time . Each agent: a unique parking location . . . . . . . . . . . . . . . . . . . . . . . Task and Path Planning for Multi-Agent Pickup and Delivery . . . . . s 1 g 3 s 2 a 2 a 1 g 2 s 3 g 1

. . . . . . . . . . . . . . Background Task Assignment Prioritized Path Planning Hybrid Path Planning Experimental Results Multi-Agent Pickup and Delivery (MAPD) Problem Free Agents Task Agents pickup location delivery location . . . . . . . . . . . . . . . . . . . . . Task and Path Planning for Multi-Agent Pickup and Delivery . . . . . s 1 a 2 a 1 g 1 current location → pickup location

. . . . . . . . . . . . . . . Background Task Assignment Prioritized Path Planning Hybrid Path Planning Experimental Results Multi-Agent Pickup and Delivery (MAPD) Problem Free Agents Task Agents . . . . . . . . . . . . . . . . . . . . Task and Path Planning for Multi-Agent Pickup and Delivery . . . . . s 1 s 1 a 1 a 2 a 2 a 1 g 1 g 1 current location → pickup location pickup location → delivery location

. . . . . . . . . . . . Background . Task Assignment Prioritized Path Planning Hybrid Path Planning Experimental Results Multi-Agent Pickup and Delivery (MAPD) Problem Assign tasks to agents and plan collision-free paths for them to execute their tasks. Unlike the traditional multi-agent path fjnding (MAPF) problem, multiple tasks can be assigned to each agent and the agent has to visit multiple locations to execute a task. . . . . . . . . . . . . . . . . . . . . . . . . . . . Task and Path Planning for Multi-Agent Pickup and Delivery

. . . . . . . . . . . . . . . Background Task Assignment Prioritized Path Planning Hybrid Path Planning Experimental Results Objective Finish executing all tasks as quickly as possible. Makespan : the earliest time step when all tasks have been executed. . . . . . . . . . . . . . . . . . . . . . . . . . Task and Path Planning for Multi-Agent Pickup and Delivery

. . . . . . . . . . . . Background . Task Assignment Prioritized Path Planning Hybrid Path Planning Experimental Results Online Version & Offmine Version Online version 1 : each task becomes known only after its release time. Offmine version : the tasks and their release times are known a priori. 1 H. Ma et al. “Lifelong Multi-Agent Path Finding for Online Pickup and Delivery Tasks”. In: AAMAS . 2017, pp. 837–845. . . . . . . . . . . . . . . . . . . . . . . . . . . . Task and Path Planning for Multi-Agent Pickup and Delivery

. Experimental Results . . . . . . . . Background Task Assignment Prioritized Path Planning Hybrid Path Planning MAPD Algorithms . G-TAPF 3 Using answer set programming. Scales only to 20 agents or tasks for a simplifjed warehouse variant. TA-Prioritized & TA-Hybrid (offmine) Scales to hundreds of agents and thousands of tasks . Smaller makespans than CENTRAL. Complete for well-formed MAPD instances. 2 Ma et al., “Lifelong Multi-Agent Path Finding for Online Pickup and Delivery Tasks”. 3 V. Nguyen et al. “Generalized Target Assignment and Path Finding using Answer Set Programming”. In: IJCAI . 2017, pp. 1216–1223. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Task and Path Planning for Multi-Agent Pickup and Delivery CENTRAL etc. 2 (online)

. . . . . . . . . . . . . . . . Background Task Assignment Prioritized Path Planning Hybrid Path Planning Experimental Results TA-Prioritized & TA-Hybrid . . . . . . . . . . . . . . . . . . . . . . . . Task and Path Planning for Multi-Agent Pickup and Delivery 1. Assign the tasks to agents by solving a special TSP . TA-Prioritized: Improved TA-Hybrid: MAPF and Prioritized Path Planning anonymous MAPF 2. Plan collision-free paths for agents to finish their assigned tasks.

. . . . . . . . . . . . Background . Task Assignment Prioritized Path Planning Hybrid Path Planning Experimental Results Goal Ignores collisions and computes one task sequence for each agent. The task sequence specifjes which tasks are assigned to the agent and in which order the agent should execute them. First construct a directed complete graph for a MAPD instance and then solve a special TSP on it to compute good task sequences. . . . . . . . . . . . . . . . . . . . . . . . . . . . Task and Path Planning for Multi-Agent Pickup and Delivery a 1 : [ t 5 t 1 t 4 ] a 2 : [ t 6 t 2 t 9 ] a 3 : [ t 3 t 8 t 7 ]

. . . . . . . . . . . . . . . Background Task Assignment Prioritized Path Planning Hybrid Path Planning Experimental Results Constructing the Graph Two types of vertices: . . . . . . . . . . . . . . . . . . . . Task and Path Planning for Multi-Agent Pickup and Delivery . . . . . τ 9 α 3 τ 2 τ 3 τ 6 τ 8 α 2 τ 7 τ 4 α 1 τ 1 τ 5 A directed complete graph G ′ = ( A ∪ T , E ′ ) . α i ∈ A represents agent a i , τ i ∈ T represents task t i .

. . . . . . . . . . . . . . . Background Task Assignment Prioritized Path Planning Hybrid Path Planning Experimental Results Constructing the Graph A Hamiltonian cycle can be converted to task sequences, one for each agent. . . . . . . . . . . . . . . . . . . . . Task and Path Planning for Multi-Agent Pickup and Delivery . . . . . τ 9 α 3 τ 2 a 1 : [ t 5 t 1 t 4 ] τ 3 τ 6 a 2 : [ t 6 t 2 t 9 ] τ 8 α 2 τ 7 τ 4 a 3 : [ t 3 t 8 t 7 ] α 1 τ 1 τ 5

. . . . . . . . . . . . . . Background Task Assignment Prioritized Path Planning Hybrid Path Planning Experimental Results Constructing the Graph The sum of the edge weights of each part is a lower bound on the execution time of the corresponding task sequence. parking location to the pickup location of its fjrst task t j . . . . . . . . . . . . . . . Task and Path Planning for Multi-Agent Pickup and Delivery . . . . . . . . . . . . τ 9 α 3 τ 2 a 1 : [ t 5 t 1 t 4 ] τ 3 τ 6 a 2 : [ t 6 t 2 t 9 ] τ 8 α 2 τ 7 τ 4 a 3 : [ t 3 t 8 t 7 ] α 1 τ 1 τ 5 Each directed edge has an integer weight w ( u , v ) . E.g., w ( α i , τ j ) is calculated as the travel time of agent a i from its

. . . . . . . . . . . . . . . Background Task Assignment Prioritized Path Planning Hybrid Path Planning Experimental Results Solving a Special TSP Objective: makespan (the largest execution time of all task sequences). . . . . . . . . . . . . . . . . . . . . Task and Path Planning for Multi-Agent Pickup and Delivery . . . . . τ 9 α 3 τ 2 a 1 : [ t 5 t 1 t 4 ] τ 3 τ 6 a 2 : [ t 6 t 2 t 9 ] τ 8 α 2 τ 7 τ 4 a 3 : [ t 3 t 8 t 7 ] α 1 τ 1 τ 5

. . . . . . . . . . . . Background . Task Assignment Prioritized Path Planning Hybrid Path Planning Experimental Results Prioritized Planning collision-free paths for the agents to execute all of their tasks according to their task sequences. 4 J. P. Van den Berg and M. H. Overmars. “Prioritized Motion Planning for Multiple Robots”. In: IROS . 2005, pp. 430–435. . . . . . . . . . . . . . . . . . . . . . . . . . . . Task and Path Planning for Multi-Agent Pickup and Delivery TA-Prioritized uses an improved version of prioritized planning 4 to plan

. . . . . . . . . . . . . . . Background Task Assignment Prioritized Path Planning Hybrid Path Planning Experimental Results Original Prioritized Planning Plan paths for the agents, one by one, in decreasing order of the estimated execution times of their task sequences. . . . . . . . . . . . . . . . . . . . . . . . . . Task and Path Planning for Multi-Agent Pickup and Delivery 1 2 3 estimated execution time