Crane Gesture Detector

Project Objectives In the field of construction there are some operators called crane helpers which indicate to the crane operator through some hand signals what he has to do and of course there are some other operators that work on other activities But they have to be aware about the crane activities so they constantly must see the signals performed by the crane helpers. And all these activities represent an imminent risk and not only for the crane operator but also for the people close to him The main point of our project is create an application based on activity recognition through smartphone sensors to detect the signal performed by a human operator or a crane helper through a machine learning algorithm. Sensors used: Two cell phones are used, of which the accelerometer and the gyroscope are used. Location of the sensors: The sensors are located in the hands and one at the height of the bicep Collection mechanism: For the collection of the data, five series were made, in each of the series 6 signals were made defined for the experiment, for the exercise 5 people were used.项目目标 在建筑施工领域，存在一类被称为起重机辅助员（crane helpers）的操作人员，他们通过手势信号向起重机驾驶员传达作业指令；此外还有负责其他作业的操作人员，同样需要知悉起重机的作业动态，因此需持续关注起重机辅助员发出的手势信号。此类作业场景不仅对起重机驾驶员，也对周边人员均存在即时安全风险。本项目的核心目标是开发一款基于智能手机传感器的活动识别应用，借助机器学习算法识别人工操作人员或起重机辅助员发出的手势信号。 所用传感器 本次实验采用两部智能手机，采集其内置的加速度计（accelerometer）与陀螺仪（gyroscope）数据。 传感器部署位置 两部手机的传感器分别部署于手部，其中一处传感器位于肱二头肌高度位置。 数据采集机制 数据采集阶段共开展五组实验，每组实验中受试者需完成实验预先设定的6种手势信号；本次实验共招募5名受试者参与。