In collaboration with Beijing Uline Intelligent Technology, our team develops an unmanned cleaning ship for debris sensing and water-quality monitoring. While Uline provides guidance and our hardware group deliver the physical testbed, my algorithm group focuses on localization, perception, and autonomous navigation. Unlike the Auto Patrol Robot project, where we were heavily involved in structural and hardware testing and trouble shooting, this project allows us to concentrate on the algorithm layer, making the development process really comfortable and rewarding.
Prototype Auto Navigation
        This stage focused on validating core vessel stability and control. Two paddlewheels were mounted on the sides of the boat, driven by a base control program running on the domain controller. The program continuously monitors chassis status including battery status and wheel conditions, while managing control inputs. It supported flexible control methods switching between handheld telecontroller, remote drive-wheel controller, and cmd_vel commands from ROS topics aiming for autonomous navigation. As the algorithm developer, our group participated in the final PID adjustment and threshold refining stage, then immediately advanced to the next phase of self-driving development.