車身側滑角為描述車輛發生側滑現象之關鍵參數，其估計準確性將直接影響到車輛安全系統觸發之正確性，而本研究將以運動學模型為基礎，整合三種車輛上常見的感測器包含慣性感測器、GNSS (Global Navigation Satellite System)感測器以及相機，來獲得車身側滑角之估測值，此方法在計算上相對簡單，且感測器皆使用車輛之標準配備，能夠節省計算效能且降低成本，因此將針對多感測器整合來準確估算側滑角作為研究目的。在整合部分，首先須對感測器進行訊號前處理，根據雙天線GNSS系統搭配即時動態定位技術(Real Time Kinematic, RTK) 使其能獲得公分等級的定位精度，而視覺方面，使用了基於圖像金字塔的 Lucus-Kanade 光流估計法來估算車輛動態行為，並設計聯邦式卡爾曼濾波器(Federated Kalman filter, FKF)來整合三種感測器之測量值，其中由於感測器存在隨機大誤差，因此設計了一個故障檢測機制來防止隨機大誤差而導致的估計錯誤。最後將感測器裝置在簡易推車平台上，進行實際實驗之數據收集，並透過後處理計算來獲得側滑角估計值，其中設計兩種情境，包含直線與 U 型迴轉之運動。而結果顯示，相較於單一感測器所估算之側滑角，整合後之估計準確性有明顯之改善。

The vehicle body sideslip angle is a key parameter to describe the vehicle side slip phenomenon, and its estimation accuracy will directly affect the correctness of the
vehicle safety system triggering. Based on the kinematic model, this research will
integrate the three common sensors on the vehicle, includes inertial sensors, GNSS
(Global Navigation Satellite System) sensors and cameras to estimated vehicle body
sideslip angle. This method is relatively simple in calculation, and the sensors are all standard equipment of the vehicle. It can save computing performance and reduce cost. Therefore, the purpose of this study is to estimate the vehicle sideslip angle by multisensor intergration. In the integration part, the sensor must be pre-processed first. According to the dual-antenna GNSS system with Real Time Kinematic (RTK), it can obtain centimeterlevel positioning accuracy. In terms of vision, the Lucus-Kanade optical flow estimation method based on Gaussian Pyramid is used to estimate vehicle dynamic behavior. A Federated Kalman filter (FKF) is designed to integrate the measurement values of these sensors. Since the sensor has large random errors, a fault detection mechanism is designed to prevent estimation errors caused by large random errors. Finally, the sensor is installed on a simple cart platform to collect the data of the actual experiment, and obtain the estimated value of the sideslip angle through postprocessing calculation. Two scenarios are designed, including the movement of a straight line and a U-turn. The results show that, compared with the sideslip angle estimated by a single sensor, the estimation accuracy after integration is significantly improved.

[1] Matsumoto, R., "Vehicle traction control system for preventing vehicle turnover on curves and turns," ed: Google Patents, 1990.
[2] Erke, A. J. A. A. and Prevention. (2008). Effects of electronic stability control (ESC) on accidents: A review of empirical evidence. vol. 40, no. 1, pp. 167-173.
[3] Liu, W., Xiong, L., Xia, X. and Yu, Z. (2018). Vehicle sideslip angle estimation: a review.
[4] Chindamo, D., Lenzo, B. and Gadola, M. J. a. s. (2018). On the vehicle sideslip angle estimation: a literature review of methods, models, and innovations. vol. 8, no. 3, p. 355.
[5] Jin, X. and Yin, G. J. J. o. t. F. I. (2015). Estimation of lateral tire–road forces and sideslip
angle for electric vehicles using interacting multiple model filter approach. vol. 352, no. 2, pp. 686-707.
[6] Zhang, B., Du, H., Lam, J., Zhang, N. and Li, W. J. I. T. o. I. E. (2016). A novel observer design for simultaneous estimation of vehicle steering angle and sideslip angle. vol. 63, no. 7, pp. 4357-4366.
[7] Singh, K. B. J. e. (2019). Vehicle sideslip angle estimation based on tire model adaptation. vol. 8, no. 2, p. 199.
[8] Li, J. and Zhang, J. J. M. P. i. E. (2016). Vehicle sideslip angle estimation based on hybrid Kalman filter. vol. 2016.
[9] Stephant, J., Charara, A. and Meizel, D. J. I. T. o. i. e. (2004). Virtual sensor: Application to vehicle sideslip angle and transversal forces. vol. 51, no. 2, pp. 278-289.
[10] Pi, D., Chen, N., Wang, J. and Zhang, B. J. I. J. o. A. T. (2011). Design and evaluation of sideslip angle observer for vehicle stability control. vol. 12, no. 3, pp. 391-399.
[11] Antonov, S., Fehn, A. and Kugi, A. J. V. S. D. (2011). Unscented Kalman filter for vehicle state estimation. vol. 49, no. 9, pp. 1497-1520.
[12] Stéphant, J., Charara, A. and Meizel, D. J. C. E. P. (2007). Evaluation of a sliding mode observer for vehicle sideslip angle. vol. 15, no. 7, pp. 803-812.
[13] Farrelly, J. and Wellstead, P. (1996). Estimation of vehicle lateral velocity. Proceeding of the 1996 IEEE International Conference on Control Applications IEEE International Conference on Control Applications held together with IEEE International Symposium on Intelligent Control, pp. 552-557. IEEE
[14] Selmanaj, D., Corno, M., Panzani, G. and Savaresi, S. M. J. C. E. P. (2017). Vehicle sideslip estimation: A kinematic based approach. vol. 67, pp. 1-12.
[15] Bevly, D. M., Gerdes, J. C. and Wilson, C. J. V. S. D. (2002). The use of GPS based velocity measurements for measurement of sideslip and wheel slip. vol. 38, no. 2, pp. 127-147.
[16] Bevly, D. M. J. J. D. S., Meas., Control. (2004). Global positioning system (GPS): A low-cost velocity sensor for correcting inertial sensor errors on ground vehicles. vol. 126, no. 2, pp.255-264.
[17] Bevly, D. M., Ryu, J. and Gerdes, J. C. J. I. T. o. I. T. S. (2006). Integrating INS sensors with GPS measurements for continuous estimation of vehicle sideslip, roll, and tire cornering stiffness. vol. 7, no. 4, pp. 483-493.
[18] Yoon, J.-H. and Peng, H. J. I. T. o. I. T. S. (2013). Robust vehicle sideslip angle estimation through a disturbance rejection filter that integrates a magnetometer with GPS. vol. 15, no. 1, pp. 191-204.
[19] Yoon, J.-H., Eben Li, S. and Ahn, C. J. I. j. o. a. t. (2016). Estimation of vehicle sideslip angle and tire-road friction coefficient based on magnetometer with GPS. vol. 17, no. 3, pp. 427-435.
[20] Botha, T. R. and Els, P. S. J. J. o. T. (2015). Digital image correlation techniques for measuring tyre-road interface parameters: Part 1–Side-slip angle measurement on rough terrain. vol. 61, pp. 87-100.
[21] Johnson, D. K., Botha, T. R. and Els, P. S. J. J. o. T. (2019). Real-time side-slip angle measurements using digital image correlation. vol. 81, pp. 35-42.
[22] Melzi, S., Sabbioni, E. J. M. S. and Processing, S. (2011). On the vehicle sideslip angle estimation through neural networks: Numerical and experimental results. vol. 25, no. 6, pp.2005-2019.
[23] Sasaki, H. and Nishimaki, T. J. S. t. (2000). A side-slip angle estimation using neural network for a wheeled vehicle. pp. 1026-1031.
[24] Du, X., Sun, H., Qian, K., Li, Y. and Lu, L. (2010). A prediction model for vehicle sideslip angle based on neural network. 2010 2nd IEEE International Conference on Information and Financial Engineering, pp. 451-455. IEEE
[25] Chindamo, D. and Gadola, M. (2018). Estimation of vehicle side-slip angle using an artificial neural network. MATEC web of conferences, vol. 166, p. 02001. EDP Sciences
[26] Bonfitto, A., Feraco, S., Tonoli, A. and Amati, N. J. V. s. d. (2020). Combined regression and classification artificial neural networks for sideslip angle estimation and road condition identification. vol. 58, no. 11, pp. 1766-1787.
[27] Boada, B. L., Boada, M. J. L., Gauchía, A., Olmeda, E., Díaz, V. J. J. o. M. S. and Technology. (2015). Sideslip angle estimator based on ANFIS for vehicle handling and stability. vol. 29, no. 4, pp. 1473-1481.
[28] Wei, W., Shaoyi, B., Lanchun, Z., Kai, Z., Yongzhi, W. and Weixing, H. J. M. P. i. E. (2016). Vehicle sideslip angle estimation based on general regression neural network. vol. 2016.
[29] Liu, J., Wang, Z., Zhang, L., Walker, P. J. I. C. T. and Applications. (2020). Sideslip angle estimation of ground vehicles: a comparative study.
[30] Li, X., Song, X. and Chan, C. J. M. (2014). Reliable vehicle sideslip angle fusion estimation using low-cost sensors. vol. 51, pp. 241-258.
[31] Li, X., Xu, N., Li, Q., Guo, K. and Zhou, J. J. P. o. t. I. o. M. E., Part D: Journal of Automobile Engineering. (2020). A fusion methodology for sideslip angle estimation on the basis of kinematics-based and model-based approaches. vol. 234, no. 7, pp. 1930-1943.
[32] Li, X., Chan, C.-Y. and Wang, Y. J. I. T. o. V. T. (2015). A reliable fusion methodology for simultaneous estimation of vehicle sideslip and yaw angles. vol. 65, no. 6, pp. 4440-4458.
[33] Boada, B. L., Boada, M. J. L., Diaz, V. J. M. S. and Processing, S. (2016). Vehicle sideslip angle measurement based on sensor data fusion using an integrated ANFIS and an Unscented Kalman Filter algorithm. vol. 72, pp. 832-845.
[34] Acosta, M., Kanarachos, S. J. N. C. and Applications. (2018). Tire lateral force estimation and grip potential identification using Neural Networks, Extended Kalman Filter, and Recursive Least Squares. vol. 30, no. 11, pp. 3445-3465.
[35] 维基百科编者, "惯性导航系统," in 维基百科，自由的百科全書, ed.
[36] 張弘毅（2010）。整合GPS與MEMS感測器於自行車導航系統之應用。國立臺灣科技大
學機械工程系碩士，台北市。
[37] Wikipedia, c., "Geographic coordinate conversion," in Wikipedia, The Free Encyclopedia., ed.
[38] Mallick, S. (2020). Geometry of Image Formation. Available:
https://learnopencv.com/geometry-of-image-formation/
[39] Camera coordinate. Available: https://reurl.cc/ZAmG76
[40] Radial Distortion Correction. Available: http://www.baspsoftware.org/radcor_files/hs100.htm
[41] Zhang, Z. (2000). A flexible new technique for camera calibration. vol. 22, no. 11, pp. 1330-1334.
[42] Jianbo, S. and Tomasi. (1994). Good features to track. 1994 Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, pp. 593-600. DOI: 10.1109/CVPR.1994.323794
[43] Bouguet, J.-Y. J. I. c. (2001). Pyramidal implementation of the affine lucas kanade feature tracker description of the algorithm. vol. 5, no. 1-10, p. 4.
[44] Lucas, B. D. and Kanade, T. (1981). An iterative image registration technique with an application to stereo vision. Vancouver, British Columbia
[45] Fischler, M. A. and Bolles, R. C. J. C. o. t. A. (1981). Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography. vol. 24, no.6, pp. 381-395.
[46] Carlson, N. A. and Berarducci, M. P. J. N., Journal of the institute of navigation. (1994).
Federated Kalman filter simulation results. vol. 41, no. 3, pp. 297-322.
[47] 维基百科编者, "卡尔曼滤波," in 维基百科，自由的百科全書, ed.
[48] Xunzhong, W., Jun, Z. and Kai, Q. J. J. o. A. (2006). Study of a robust federated filtering algorithm based on fault factor function. vol. 27, no. 1, pp. 57-60.
[49] Yue, Z., Lian, B., Tang, C., Tong, K. J. M. S. and Technology. (2020). A novel adaptive federated filter for GNSS/INS/VO integrated navigation system. vol. 31, no. 8, p. 085102.
[50] Jones, J. Table: Chi-Square Probabilities. Available:
https://people.richland.edu/james/lecture/m170/tbl-chi.html
[51] SiK Telemetry Radio V3. Available: http://www.holybro.com/product/transceiver-telemetryradio-v3/
[52] BFS-U3-16S2M-CS. Available: https://www.apostar.com.tw/products/9c0f04d2.php
[53] Diddi, M. Spinnaker_sdk_camera_driver. Available:
https://github.com/neufieldrobotics/spinnaker_sdk_camera_driver