【摘要】 设计了一种可部署在便携式设备Jetson TX2上的机器人实验教学平台，该平台主要采用了机器人操作系统（ROS）和PyTorch框架进行开发。它采用主从机的通信方式，能够方便实现基于深度学习的目标检测和机器人控制的交互。重点设计了人体跟踪与手势识别两种智能识别模式，使得机器人不仅能够通过识别手势来做出相应动作，还能够通过识别特定人物实现跟随移动。实验表明，这款机器人在Jetson TX2平台上能够进行实时且准确的目标检测，具有良好的应用前景。更多还原

【Abstract】 This paper develops a robotic experimental teaching platform deployable on the portable Jetson TX2 device.This platform utilizes the robot operating system(ROS) and the PyTorch framework for its development. It employs a master-slave communication architecture, facilitating seamless interaction for deep learning-based object detection and robotic control. This paper designs two intelligent recognition modes: human tracking and gesture recognition, enabling the robot to perform actions based on recognized gestures and to follow specific individuals. Experimental results demonstrate that the robot is capable of real-time and accurate object detection on the Jetson TX2 platform.更多还原