Attitude estimation of camera with six degrees of freedom based on deep learning of cyclic convolution neural network
Download as PDF
Wu Fan, Zong Yantao, Tang Xiaqing
Firstly, the mathematical simulation model of Four-rotor system is established, then the control problems of trajectory tracking and attitude stabilization in flight are analyzed, and a six-degree-of-freedom attitude estimation method for camera based on deep learning of cyclic convolution neural network is designed. The outer loop is a spatial position loop, the input is the difference between the actual position and the desired trajectory path, and the output is the attitude control angle. After converting to angular velocity, the outer loop is used as the input of the inner loop attitude loop for attitude stabilization control of four rotors. Over the years, it has been widely used in motion simulators, parallel machine tools, precision positioning platforms and various entertainment occasions. Under this development trend, a new and efficient system is designed by applying the 6-DOF parallel platform to simulated target tracking. Thus the hydraulic system is controlled to realize the control of the platform and the target tracking task.
Six degrees of freedom, attitude estimation, PAC controller, neural network, in-depth learning