US 12,282,314 B2
Hypergraph search for real-time multi-robot task allocation in a smart factory
Zixiang Nie, Tampa, FL (US); and Kwang-Cheng Chen, Tampa, FL (US)
Assigned to University of South Florida, Tampa, FL (US)
Filed by University of South Florida, Tampa, FL (US)
Filed on Jan. 28, 2022, as Appl. No. 17/587,532.
Claims priority of provisional application 63/143,978, filed on Feb. 1, 2021.
Prior Publication US 2022/0253048 A1, Aug. 11, 2022
Int. Cl. G05B 19/418 (2006.01); G06N 5/01 (2023.01)
CPC G05B 19/41865 (2013.01) [G05B 19/41895 (2013.01); G06N 5/01 (2023.01)] 15 Claims
OG exemplary drawing
 
1. A system comprising:
a multi-robot system (MRS) comprising a plurality of robots, wherein the MRS is configured to perform a manufacturing task, wherein the plurality of robots comprises a plurality of production robots and a plurality of autonomous mobile robots (AMRs);
a computing device in electronic communication with the MRS configured to:
perform a multi-robot task allocation (MRTA) for the MRS, wherein the MRTA is configured to generate task assignments for each of the plurality of production robots and transportation paths for each of the plurality of AMRs,
adjust the task assignments and transportation paths to assign flows in real-time through iterative execution of a hypergraph search algorithm that is optimized for energy efficiency within the MRS and based at least in part on a hypergraph production robot model corresponding with the MRS, wherein each hyper-vertex of the hypergraph production robot model represents a respective AMR, wherein each directed edge set of the hypergraph production robot model represents a transportation path set with corresponding quantized energy consumption as weights, and wherein the hypergraph search algorithm is configured to: (a) perform biased sampling in favor of higher energy efficiency solutions from an entire solution space, and (b) stop the iterative execution based on marginal distribution bounds estimations determined based on the weights.