Abstract

Collaborative Robotics (cobots) are transforming enterprises by operating alongside humans in dynamic settings; yet, enhancing their efficacy presents challenges. Energy efficiency, work accuracy and collision avoidance can occasionally be at odds, necessitating intricate optimisation. This work presents the Dragonfly Algorithm (DA), a metaheuristic derived from dragonfly swarming, aimed at enhancing cobot performance through Multi-Objective Optimisation (MOO). The DA models five swarming interactions: separation (collision avoidance), alignment (synchronized movement), cohesion (group stability), attraction (goal targeting) and distraction. These tendencies equilibrate exploration and exploitation, rendering Differential Algorithms advantageous for intricate optimisation challenges. The proposed approach optimizes energy efficiency, enhances work productivity via trajectory planning and mitigates collision risks to bolster safety. Pareto front optimization alters the decision analysis for multi-objective scenarios to uncover non-dominated solutions, providing decision-makers with flexible trade-offs. Dragonflies symbolize collaborative robot configurations, and swarm interactions enhance solutions. Fitness functions assess energy, efficiency, and safety metrics to promote convergence towards optimal solutions. Simulations in industrial and service environments demonstrate the superiority of the Differential Algorithm (DA) compared to Genetic Algorithms (GA) and Particle Swarm Optimization (PSO). The statistics indicate a 10-15% decrease in energy consumption, a 20-30% enhancement in work efficiency, and a 5-8% reduction in accidents. DA exhibits rapid convergence and adaptability, rendering it suitable for real-time applications. This study emphasizes the DA’s capacity to enhance the efficiency, safety, and flexibility of cobots in intricate, multi-objective scenarios. Future research will incorporate machine learning for adaptive parameter optimization and broaden applications to heterogeneous multi-cobot systems.

Keywords:Multi-objective optimization; Collaborative robots; Task Allocation; Dragonfly algorithm; Industrial automation; Path planning