Projects
Group Projects
Automatically tracking system using monocular vision algorithm PnP
1. Demonstrate how to build an automatic tracking system based on vision methods. The system has successfully been used on robots made by YSU-Eagle.
2. This system is mainly compiled by C++ programming language, including vision algorithm codes and serial communication codes.
[report(in Chinese)] [Youtube]
Mathematical modelling
Computer Vision
SLAM
Particle filter SLAM using odometry, 2D laser scans and RGBD images [report]
Visual-Inertial SLAM with EKF [report]
Others
A wind panel control device
This device performs PID feedback through the angle of the wind panel to control the motors on both sides. It allows the panel to be in any position while maintain or rotate it at a constant speed. [Youtube]
A plate-ball control system
This device recognizes the black ball with the camera above the plate and transmits the position information to the control system below. The control system then adjusts the pitch angle of the plate by two motors to move the ball across the plate according to the programmed movement trajectory.
A single inverted pendulum system
Based on the mechanical model of the inverted pendulum, the DC motor and speed reducer on the frame are PID controlled according to the pendulum angle. When the end of the bar is kept vertical, the rotating rod is made to rotate in uniform or variable velocity.
|