基于卡尔曼滤波的两轮平衡姿态检测系统设计

Design of two-wheel balance attitude detection system based on Kalman filter

摘要:

针对传统两轮平衡检测系统易受噪声等环境因素影响，导致测量产生误差的现象，提出基于卡尔曼滤波的两轮平衡系统设计。选择SCA610-CA1H1G型号姿态传感器，以MEMS技术为基础设计姿态传感器结构，具备可选择的多种数据输出模式。选取36V电源将3个12V12AH的铅酸电池利用串联在一起，为系统提供充足电源。设计主控芯片Pentium II/III/4系列CPU，将它的状态信息描述和对外输出。设计基于卡尔曼滤波的去噪流程，构建误差修正模型，通过四元数转旋转矩阵转移到当前设备坐标系上处理，由此获取姿态矩阵从本体坐标系旋转到另一个坐标系的过程，对测量数值进行归一化处理，可得到无量纲标量。使用定时采样三阶逼近法，更新四元数，完成姿态检测。由实验结果可知，无论是在0-15dB噪声环境，还是15-30dB噪声环境下，系统监测精准度最高为92%，为汽车、航天领域提供设备支持。

Abstract:

Aiming at the phenomenon that the traditional two wheel balance detection system is easy to be affected by environmental factors such as noise, resulting in measurement error, a two wheel balance system design based on Kalman filter is proposed. Sca610-ca1h1g attitude sensor is selected and the structure of attitude sensor is designed based on MEMS technology. It has a variety of optional data output modes. Three 12v12ah lead-acid batteries were connected in series with 36V power supply to provide sufficient power for the system. The main control chip Pentium II / III / 4 Series CPU is designed, and its state information is described and output. The denoising process based on Kalman filter is designed, and the error correction model is constructed. The quaternion rotation matrix is transferred to the current equipment coordinate system for processing. The process of the attitude matrix rotation from the body coordinate system to another coordinate system is obtained. The measured values are normalized and the dimensionless scalar is obtained. The quaternion is updated and the attitude detection is completed by using the timing sampling third-order approximation method. According to the experimental results, the highest monitoring accuracy of the system is 92% in 0-15db noise environment or 15-30db noise environment, which provides equipment support for automobile and aerospace fields.