Research On The Techniques Of Virtual Gyroscope Based On Redundant Sensors And Its Applications

Gyroscope is a sensor to detect angular velocity signal,widely used in aerospace,military,industrial and consumer fields.With the development of gyroscope applications,the performance requirements of the gyroscope continue to increase.However,in practice,low-cost gyroscopes are difficult to provide high-precision angular velocity detection.Virtual gyro technology can make use of multiple low-cost sensors to improve the detection accuracy of angular velocity through data fusion,and has broad application prospects.Aiming at the problems of robustness and configuration optimization in virtual gyroscope,this paper proposes a robust virtual gyroscope and a new virtual gyro configuration based on inertial sensors and the fusion algorithm.A redundant sensor inertial system based on virtual gyroscope is also proposed.Firstly,aiming at problem that the virtual gyro technology is not enough robust to the filter noise parameters and signal outliers,a virtual gyro based on H? filter is proposed to improve the robustness of the system and the adaptability of the environment signal.A novel virtual gyro based on outliers robust filtering algorithm is proposed to eliminate the influence of single outliers and outliers on the system output.Secondly,a novel virtual gyroscope based on inertial sensor array is proposed.Atypical configuration of multiple gyroscopes and accelerometers mounted on designated positions is proposed.The angular acceleration obtained by merging the accelerometers and the outputs of multiple gyroscopes as the observations of the Kalman filter are used to establish a new virtual gyro fusion algorithm.The results of simulation and prototype show that the virtual gyroscope greatly improves the dynamic performance,the random walk and the bias instability are reduced by 1.57 times and 65.5%,respectively,in the static test.The accuracy the dynamic test is improved by 1.525 times.Finally,a new inertial system with redundant sensors is proposed.The inertial system is composed of four inertial measurement units(IMU)with three-axis gyroscope and three-axis accelerometer.With the fusion algorithm of angular velocity and acceleration signal based on H? filter and attitude estimation algorithm,the experimental prototype of the inertial system is developed.Experiments show that the inertial system with redundant sensors greatly improves the attitude angle detection accuracy.The roll/pitch angle accuracy is 0.06°,and the yaw angle accuracy is 0.5°,better than the inertial system using the same level sensors in industry.