Design and Simulation of Time-energy Optimal Anti-swing Trajectory Planner for Autonomous Tower Cranes

TL;DR

This paper presents a novel time-energy optimal anti-swing trajectory planner for autonomous tower cranes, utilizing multi-objective evolutionary algorithms to improve safety and efficiency in construction site lifting operations.

Contribution

It introduces a trajectory planning module based on differential flatness and multi-objective optimization, enhancing the CALP system for autonomous crane lift planning.

Findings

The proposed planner effectively reduces swing and energy consumption.

NSGA-II and GDE3 algorithms are compared to select the best optimization engine.

Simulations demonstrate the planner's reliability and efficiency in real-world scenarios.

Abstract

For autonomous crane lifting, optimal trajectories of the crane are required as reference inputs to the crane controller to facilitate feedforward control. Reducing the unactuated payload motion is a crucial issue for under-actuated tower cranes with spherical pendulum dynamics. The planned trajectory should be optimal in terms of both operating time and energy consumption, to facilitate optimum output spending optimum effort. This article proposes an anti-swing tower crane trajectory planner that can provide time-energy optimal solutions for the Computer-Aided Lift Planning (CALP) system developed at Nanyang Technological University, which facilitates collision-free lifting path planning of robotized tower cranes in autonomous construction sites. The current work introduces a trajectory planning module to the system that utilizes the geometric outputs from the path planning module and optimally scales them with time information. Firstly, analyzing the non-linear dynamics of the crane operations, the tower crane is established as differentially flat. Subsequently, the multi-objective trajectory optimization problems for all the crane operations are formulated in the flat output space through consideration of the mechanical and safety constraints. Two multi-objective evolutionary algorithms, namely Non-dominated Sorting Genetic Algorithm (NSGA-II) and Generalized Differential Evolution 3 (GDE3), are extensively compared via statistical measures based on the closeness of solutions to the Pareto front, distribution of solutions in the solution space and the runtime, to select the optimization engine of the planner. Finally, the crane operation trajectories are obtained via the corresponding planned flat output trajectories. Studies simulating real-world lifting scenarios are conducted to verify the effectiveness and reliability of the proposed module of the lift planning system.