在線性加速度下使用之姿態估測器

Abstract

以往使用量測慣性的感測器所組合成的姿態估測器，大部分都只能在沒有運動加速度的狀態下使用，可是實際上我們想量測姿態的對象，是很難不具有運動加速度的。在本論文中，我們提出新的物體姿態估測方法，利用平面式純加速規慣性量測單元、磁場感應器、固定點距離偵測感應器搭配卡爾曼估測器，設計出可以估測到具有加速度運動的系統姿態、角速度與加速度。 在具有運動加速度的狀態下估測姿態，若加速度快速變化時，我們搭配記憶褪去式卡爾曼估測器的設計，提高系統追跡變化訊號之能力，大大增進狀態收斂的速度。 本論文目前的成果，利用Matlab做系統的模擬，共計使用了七個單軸加速規、一個三軸磁場感應器和一個距離感應器，可以成功估測出具有運動加速度的物體姿態，即使物體具有變加速度運動，也可以追跡成功。在估測精度方面，穩態時的姿態四元數精度可達到7.80×10-5，加速度估測精度2.87×10-3 m/s2，角速度估測精度0.78 rad/s。

In the past, the estimation of an object in motion is done by using inertial sensor. The conventional approach can not apply to the object under a linear acceleration. Although, in most cases, the object to be measured work under linear acceleration. In this thesis, we proposed a novel design of orientation determination, using coplanar gyro-free inertial measurement unit, magnetic field inductor and a distance sensor. These sensor accompany with Kalman filter technique, can determine attitude, angular velocity and acceleration of the object in motion. When we determine orientation under linear acceleration, if the acceleration changes fast, we design observer that accompany with “Fading Memory” technique, that will be raised tracking ability of variable signal, and improve speed of state convergence. According to our simulation by Matlab, the proposed design determines attitude in motion successfully, which use seven single-axis accelerometers, a 3-axis magnetometer and a distant sensor. Even though the acceleration is changed fast, this observer tracks successful. Tracking error standard deviation on simulative data are shown, 7.80×10-5 in quaternion of attitude, 2.87×10-3 m/s2 in acceleration and 0.78 rad/s in angular velocity.