Evolutionary Algorithm-Based Energy-Aware Path Planning with a Quadrotor for Warehouse Inventory Management

https://doi.org/10.28991/HIJ-2023-04-04-012
Evolutionary Algorithms Inventory Management Path Planning Quadrotor Warehouse.

Authors

  • C. J. P. De Guzman
    carlos_james_deguzman@dlsu.edu.ph
    Department of Mechanical Engineering, De La Salle University, 2401 Taft Ave. Malate, Manila,, Philippines
  • A. Y. Chua Department of Mechanical Engineering, De La Salle University, 2401 Taft Ave. Malate, Manila,, Philippines
  • T. S. Chu Department of Mechanical Engineering, De La Salle University, 2401 Taft Ave. Malate, Manila,, Philippines https://orcid.org/0000-0001-5775-1532
  • E. L. Secco Robotics Laboratory, School of Mathematics, Computer Science & Engineering, Liverpool Hope University,, United Kingdom
Vol. 4 No. 4 (2023): December
Research Articles

Downloads

Quadrotors have been vital for automating warehouse processes. However, a significant gap in recent studies is that they use a single quadrotor with limited battery life, considering that their objective involves navigation in a large-scale environment such as a warehouse. Using an energy consumption model to enable more efficient navigation can be explored. Conventional data-driven energy models and path planning algorithms are insufficient for describing the various motions that a quadrotor can perform in warehouse operations, such as changes in yaw. This study aims to design a novel exhaustive data-driven energy consumption model and evolutionary algorithm-based path planning algorithm to consider various quadrotor movements involved in warehouse operations. The quadrotor is tasked with performing a set of movements to each be represented as a power equation in terms of their velocity. The obtained equations were subsequently used as the primary optimization objective for the path planning algorithm, which included yaw angle objectives and constraints. A set of experiments was performed with Crazyflie quadrotors to verify the model and the algorithm. The results showcased the accuracy of the energy consumption model, which was kept at a maximum difference of 0.6%. The designed path planning algorithm obtained greater energy efficiency in the generated paths compared to other state-of-the-art evolutionary algorithms with similar objectives and constraints.

 

Doi: 10.28991/HIJ-2023-04-04-012

Full Text: PDF