Expanding the wrench feasible workspace of quadrotor-based mobile cable-driven parallel manipulators using multi-objective optimization and machine learning

Abstract

Quadrotor-based Reconfigurable Cable-Driven Parallel Manipulators (QRCDPMs) have previously been introduced as having significant potential in extending the feasible workspace of Cable-Driven Parallel Manipulators (CDPMs). However, such configurations have limitations in executing positive Z-direction trajectories resulting in non-optimal workspace and performance. This study proposes a novel Quadrotor-based Enhanced Reconfiguration Cable-Driven Parallel Manipulator (QERCDPM) that significantly enhances the Wrench Feasible Workspace (WFW). This allows for extended safe platform operation along all XYZ directions by adjusting the bottom cable exit point. Simulation results indicate that the proposed setup achieves a 96.39 % ratio of safe workspace to available workspace in various cable exit points configurations. Tension Distribution analysis for different trajectories demonstrates the proposed setup's superior cable tension management, increased flexibility and reachability. A Random Forest based Particle Swarm Optimization - Modified Frequency Bat Algorithm (RFPSOMFB) optimization algorithm is proposed, which outperforms existing methods in terms of run time and solution quality. The findings hold significance for applications requiring safe and efficient trajectory execution in dynamic and constrained environments, contributing to the evolution of multi-robot based intelligent and distributed systems.