The invention discloses an anti-swing control method of a flexible cable type parallel hoisting device, and relates to the technical field of automatic control, a binocular vision system and a reinforcement learning control model are included, the binocular vision system collects image information of a hoisted object in real time, kinematics parameters such as speed are calculated by combining depth information and a background model, and the reinforcement learning control model is used for reinforcement learning control. And inputting kinematics parameters to the reinforcement learning control model, and finally outputting an optimal evaluation function by the trained reinforcement learning model to realize rapid swing elimination. Compared with other control methods, the method has the core advantages that autonomous learning can be achieved, reinforcement learning has smaller dependence on a mathematical model, and long-term value evaluation of the strategy is exploration based on a large amount of data; the complexity of the movement process and environment modeling of the crane is high, and the optimal solution can be obtained more easily by adopting data-driven reinforcement learning.
本发明公开了一种柔索式并联起吊装置的防摇摆控制方法,涉及自动化控制技术领域,包括双目视觉系统和强化学习控制模型,双目视觉系统实时采集被吊运物体图像信息,再结合深度信息和背景模型,计算出速度等运动学参数,输入运动学参数至强化学习控制模型,最终经过训练后的强化学习模型输出最优评价函数,实现快速消除摆动;对比其他控制方法,核心优势在于可以自主学习,强化学习对数学模型的依赖性更小,对于策略的长期价值评估是基于大量数据的探索;起重机的运动过程和环境建模复杂度较高,采用数据驱动的强化学习能够更轻松地获得最优解。
Anti-swing control method of flexible cable type parallel lifting device
一种柔索式并联起吊装置的防摇摆控制方法
28.03.2023
Patent
Elektronische Ressource
Chinesisch
Gantry crane lifting appliance electronic anti-swing system and anti-swing control method
Europäisches Patentamt | 2023
|