姿態演算在感測器的應用發展中佔有非常重要且不可或缺之地位，隨著當前各種感測器趨於微小化的時代，各項產品與功能相應而生，如今在我們的手機、穿戴裝置、導航系統、無人飛行器等，都必定會使用到姿態感測，因此如何計算出精確的姿態是非常重要的。
Kalman filter姿態演算法中，角速度計的資料主要是用來預測姿態，其不受外力影響，因此能夠得到較為平滑的姿態角，但會因為離散積分的算法使得姿態角產生累積誤差，而加速度計的資料主要是用來更新姿態，其值非常容易受到外力影響，導致更新的姿態角產生大量誤差，但在微小外力的情況下，能夠得到較為正確姿態角。
在角速度與加速度所估計出的姿態之間有權重值的分配，若權重值較為偏向角速度資料，則會導致累計誤差的現象增加，反之則是受到外力干擾的情況明顯，使用單一且固定的權重值將無法在兩者之間取得最佳解。
為了解決此問題，我們提出利用物體運動的慣性來預估附體加速度，並且將讀取的加速度數據減去此附體加速度後可以得到外力加速度，之後藉由估計出的外力加速度計算出權重值，再將讀取的加速度與估計的附體加速度分別乘上計算出的權重值，最後利用狀態切換的方式來估計出最終的附體加速度後傳入Kalman filter進行更新姿態。
本篇論文提出的方法能夠有效的估計出外力，並藉由自適應的權重值來提升計算出的姿態角精確度。

Attitude estimate algorithm takes a critical and indispensable part in the sensor fusion. With the sensors getting smaller, various product and application are developed. Now we have smart phone, wearable device, Navigation system, drone, etc. All this product needs an attitude estimation. Therefore, how to get a good attitude is very important.
In the Kalman filter algorithm, gyro data is used to predict the attitude. It won’t be affected by external force so we can get a smooth attitude from gyro data. But we will get accumulative error because of the discrete integral of gyro data. The acceleration data is used to update the attitude. It is sensitive for external force so we will get a lot of noise after update the attitude. But in little external force situation, we can get a good estimate of attitude.
There is a weight value between gyro and acceleration. If the weight is close to the gyro data, we will get accumulative error from discrete integral. On the contrary, we will get a lot of noise from external force. Only use a weight value won’t get the best attitude from gyro and acceleration data.
We propose a method to solve this problem by using inertia of object motion to estimate body gravity. After that we use acceleration data minus estimate body gravity to get external force. Then use external force to calculate adaptive weight value. Finally, we use state switch method to estimate final body gravity then input to Kalman filter to update attitude.
In our propose, we can estimate external force and using adaptive weight value to calculate attitude more accurately.

[1] S. O. H. Madgwick, A. J. L. Harrison, and R. Vaidyanathan, "Estimation of IMU and MARG orientation using a gradient descent algorithm," in 2011 IEEE International Conference on Rehabilitation Robotics, 29 June-1 July 2011 2011, pp. 1-7, doi: 10.1109/ICORR.2011.5975346.
[2] R. Mahony, T. Hamel, and J. Pflimlin, "Nonlinear Complementary Filters on the Special Orthogonal Group," IEEE Transactions on Automatic Control, vol. 53, no. 5, pp. 1203-1218, 2008, doi: 10.1109/TAC.2008.923738.
[3] Z. Tan, Y. Wu, and J. Zhang, "Fused attitude estimation algorithm based on explicit complementary filter and Kalman filter for an indoor quadrotor UAV," in 2018 Chinese Control And Decision Conference (CCDC), 9-11 June 2018 2018, pp. 5813-5818, doi: 10.1109/CCDC.2018.8408147.
[4] Z. Wu, M. Yao, H. Ma, and W. Jia, "Improving Accuracy of the Vehicle Attitude Estimation for Low-Cost INS/GPS Integration Aided by the GPS-Measured Course Angle," IEEE Transactions on Intelligent Transportation Systems, vol. 14, no. 2, pp. 553-564, 2013, doi: 10.1109/TITS.2012.2224343.
[5] Y. Pititeeraphab, T. Jusing, P. Chotikunnan, N. Thongpance, W. Lekdee, and A. Teerasoradech, "The effect of average filter for complementary filter and Kalman filter based on measurement angle," in 2016 9th Biomedical Engineering International Conference (BMEiCON), 7-9 Dec. 2016 2016, pp. 1-4, doi: 10.1109/BMEiCON.2016.7859621.
[6] K. Feng et al., "A New Quaternion-Based Kalman Filter for Real-Time Attitude Estimation Using the Two-Step Geometrically-Intuitive Correction Algorithm," (in eng), Sensors (Basel), vol. 17, no. 9, p. 2146, 2017, doi: 10.3390/s17092146.
[7] V. V. Avrutov, P. M. Aksonenko, P. Henaff, and L. Ciarletta, "3D-calibration of the IMU," in 2017 IEEE 37th International Conference on Electronics and Nanotechnology (ELNANO), 18-20 April 2017 2017, pp. 374-379, doi: 10.1109/ELNANO.2017.7939782.