ONLINE and multi-objective trajectory planner for robotic systems

Abstract

Advancements in automation technology have led to the increased utilization of industrial robots in manufacturing processes. Trajectory planning, which is crucial in robotics, involves designing smooth trajectories that adhere to constraints. Trajectory planning methods can be classified as either kinematic or dynamic, with dynamic models providing improved capacity utilization but requiring greater complexity. Given the need for efficient real-time implementation with low computational demands, the kinematic method is indispensable. The challenge lies in finding a balance between swift movements and minimal vibration, as smoother trajectories often necessitate higher-order polynomials, resulting in longer execution times and more intricate computations. Addressing the trade-off between speed and smoothness is crucial, as trajectory planning effectiveness depends on balancing energy efficiency, smooth motion, and computational complexity. A novel trajectory planner for point-to-point movements has been developed to enable rapid and smooth motion by integrating the advantages of minimum acceleration and minimum jerk trajectories. This innovative approach, tailored for a diverse range of robotic systems, generates a multi-objective and optimized trajectory as a single segment for seamless online implementation. Simulation and experimental tests were conducted to evaluate the proposed trajectory planner, comparing its performance against commonly used methods in terms of velocity, energy consumption, and smoothness.