本工程为越疆（Dobot）双臂AI训练平台(Xtrainer)的配套程序，可通过直接操纵小型主手来直观、灵巧和精确地控制从手机械臂运动，实现可靠、高效、高质量的演示数据采集。同时提供了一个基础的模仿学习数据采集、算法训练与推理的示范代码程序、以及整套平台的API接口（包含主手关节角度数据读取/锁定控制、从手机械臂关节角度和笛卡尔坐标读取/控制、夹爪开合度控制、相机RGB/深度图像获取等接口） 通过本硬件平台和二次开发接口，可实现灵活、丰富的机器人应用研究。

This project is a supporting program for the Dobot dual-arm AI training platform. It can intuitively, flexibly and accurately control the movement of the slave manipulator by directly manipulating the small master hand, and realize reliable, efficient and high-quality demonstration data collection. At the same time, it provides a basic imitation learning data collection, algorithm training and reasoning demonstration code program, as well as the API interface of the whole platform (including master hand joint angle data reading/locking control, slave manipulator joint angle and Cartesian coordinate reading/control, gripper opening and closing control, camera RGB/depth image acquisition, etc.). Through this hardware platform and secondary development interface, flexible and rich robot application research can be realized.

[Third-Party Components]

This project includes components derived from other open-source projects. For detailed information about these components, including license information, please refer to the [THIRD-PARTY-LICENSES] file.