慣性感測元件於水下載具導航之特性評估

PERFORMANCE EVALUATION OF A MEMS-BASED INERTIAL SENSORS FOR UNDERWATER VEHICLE

隨著微機電（MEMS）製程技術的日益發展，微型慣性感測元件擁有體積小、製造成本低以及低耗電性等優點。然而就快速且可靠的定位系統（positioning system）的需求而言，目前MEMS慣性感測元件仍存有訊號飄移的問題，以至於單獨應用於導航定位方面的準確度仍嫌不足。因此本文主要著重於MEMS慣性感測元件的前置校正，探討造成感測元件訊號飄移的成因，並藉由不同的校正方法如線性度校正、小波訊號處理等進一步降低誤差量隨著積分而逐漸放大的現象，以利於在無法有效接收外界參考訊號的情形下透過MEMS慣性感測元件也能達到定位或是導航等目標，最後由軌跡實測的結果顯示慣性感測模組經由校正後能達成較佳的定位效果，其整體平均誤差量約可縮小至5到10cm左右。 The development of a micro-electromechanical systems (MEMS) inertial navigation system for the navigation ofunderwater vehicles, both autonomous underwater vehicles (AUVs) and remotely-operated vehicles (ROVs), is extremelychallenging. The navigation estimation of underwater vehicles is easily influenced by error noise arising in theMEMS-based inertial sensors. Therefore, this paper presents a integrated calibration method to overcome the challenges ofMEMS-based inertial sensors for underwater navigation. The MEMS-based inertial sensors module is composed of amicro-accelerometer, a micro-gyroscope, and signal processing circuits. In general, the sources of error noise can becategorized into two groups, deterministic and stochastic: the former primarily include bias errors, misalignment andnonlinearity; the latter include temperature effects and signal drifting. Linearity calibration is used to modify deterministicerrors and wavelet analysis can suppress stochastic noise. The integrated calibration method therefore includes biascompensation, linearity calibration and wavelet signal processing to enhance the accuracy and performance of theMEMS-based inertial navigation system. The experimental results demonstrate that the output signal can be correctedsuitably by means of the proposed method. Overall, the results confirm the strong potential of MEMS-based inertialsensors for use in underwater vehicle navigation applications.