Sequencing automated multi-agent wall construction: UAV case scenario

Date
2023
Journal Title
Journal ISSN
Volume Title
Publisher
University of Delaware
Abstract
Unmanned autonomous vehicles (UAVs), also known as drones, are pervasive in modern society for both recreational and business purposes. Construction sites routinely employ UAVs to capture imagery used for photo documentation, progress reporting, surveying, and inspection. Drones can also carry infrared cameras and LIDAR, increasing opportunities for their use on construction sites. They are not limited, however, to remote sensing and photogrammetric applications. Research is pushing the boundaries in automation of onsite construction tasks to support and enable the construction industry to match the rising demand. Construction automation and robotics promise the solution with the use of technologies, such as UAVs, for physical construction, which not only eliminates risks associated with workers’ safety but can provide improved quality and productivity. Accordingly, this dissertation investigates how can construction of masonry walls be better automated using robotics. We introduce a novel algorithm, termed CMCP (Cooperative Masonry Construction Planner), a greedy randomized adaptive search procedure metaheuristic approach that can determine a near-optimal wall construction plan for a swarm of UAVs. The problem is formulated as mixed-integer linear programming with added wall-building, precedence and concurrence rule constraints, that ensure bricks are built in the correct order and help prevent collision between cooperating agents during construction. The proposed approach introduces efficiency benefits and enhances productivity when dealing with multiple agents building a wall, where the plan is indeterministic and ensures that all available agents are utilized efficiently. The approach is validated on simulation and compared with several approaches, including MILP and naive traditional human wall construction.
Description
Keywords
Construction automation, GRASP, Multi-agent orienteering problem, Wall-building planner, Human wall construction
Citation