This paper presents an improved zero vibration method for the swing control of bridge-type grab ship unloader. With the method, the concepts of equivalent frequency and the equivalent damping ratio are proposed to cope with the changeable length of rope, and the optimal path planning is considered to avoid collision and improve efficiency. Numerical simulation results of a case study indicate that the maximum residual swing angle of the grab can be limited to a small range to ensure safety using the improved zero vibration method, whereas the traditional zero vibration method with average frequency and zero damping ratio gets poor results of swing control. After that, the sensitivities of the max residual swing angle to the changes of some main design parameters (damping coefficient, deviation of the center of gravity of the grab in rope direction, and time delay of the system) and operating parameters (position deviation of the trolley, initial length deviation of the rope, and initial swing angle) are analyzed. The results obtained display that the residual swing angle is sensitive to the deviation of grab’s center of gravity, the deviation of trolley’s position, and the initial swing angle under the same control parameters, but insensitive to the damping coefficient, the time delay of the system, and the initial length deviation of the rope. This can help to select the appropriate parameter values ​​or adaptive range in an actual unloader.@en

To address the problem of training on small datasets for action recognition tasks, most prior works are either based on a large number of training samples or require pre-trained models transferred from other large datasets to tackle overfitting problems. However, it limits the research within organizations that have strong computational abilities. In this work, we try to propose a data-efficient framework that can train the model from scratch on small datasets while achieving promising results. Specifically, by introducing a 3D central difference convolution operation, we proposed a novel C3D neural network-based two-stream (Rank Pooling RGB and Optical Flow) framework for the task. The method is validated on the action recognition track of the ECCV 2020 VIPriors challenges and got the 2nd place (88.31%). It is proved that our method can achieve a promising result even without a pre-trained model on large scale datasets. The code will be released soon. @en