||Development of an aviation meteorological information and unusual weather detection system prototype based on the ADS-B 1090MHz data
||Institute of Civil Aviation
Automatic Dependent Surveillance-Broadcast (ADS-B)
Global Forecast System (GFS)
unusual aviation weather detection algorithm
Support Vector Machine (SVM)
航空氣象一直以來與飛航安全有密不可分的關係，許多飛航事故都肇因於航空氣象之異常狀況，例如：低空風切(Low-level wind shear)及微爆氣流(Microburst)。為了達到預警的效果，現行的航空異常氣象偵測系統為低空風切警報系統(Low-Level Wind Shear Alert System, LLWAS)，然而此系統架設的氣象觀測站皆分佈於鄰近地表之位置，其偵測到的氣象異常狀況缺乏高程氣象資訊。因此，本論文將利用廣播式自動回報監視(Automatic Dependent Surveillance-Broadcast, ADS-B)系統所提供的高程氣象資訊來增強現行的氣象偵測系統。另一方面，本論文也利用航機的ADS-B系統驗證氣象異常狀況與航機資料之間的關聯性並建立異常氣象偵測系統之雛型。
為了取得高程氣象資訊以增強現有氣象偵測系統，本論文首先對航機廣播之ADS-B Mode-S ES氣象參數進行分析。由於ADS-B Mode-S ES氣象參數中磁向角及真空速資料需要經過修正才可以進一步計算出正確的高程風速向量資訊，本論文藉由全球預報系統(Global Forecast System, GFS)與航機的氣象資料設立修正資料庫，並依據不同的航機機型做參數之修正。接著，為了驗證航機氣象資料與氣象異常狀況之關聯性，本論文將利用三種氣象異常演算法來呈現在航機遭遇異常氣象之狀況，三種氣象異常演算法分別是氣象參數跳動演算法、風速向量差演算法及觀測航機異常位移之方法。最後，本論文將利用機器學習之支持向量機(Support Vector Machine, SVM)演算法來設立異常氣象狀況的資料庫，經由上述之氣象異常演算法及特徵選擇技術來挑選機器學習之特徵向量，藉此建立異常氣象偵測系統之雛型以判斷氣象異常之發生。
在本篇論文中，我們利用航機廣播的ADS-B Mode-S ES資料提供高程的氣象資訊，並且利用GFS建立航機參數的修正資料庫以獲得正確的高程風速向量資訊。並且本論文藉由三種不同氣象異常演算法及機器學習之SVM演算法建立出異常氣象偵測系統之雛型。
Aviation meteorology is closely associated with flight safety, as many accidents have occurred due to unusual weather conditions. However, the current aviation unusual weather detection systems, such as the low level wind shear alert system, can only detect unusual weather conditions near the ground surface, because all ground-based meteorological monitoring stations are at almost the same level. It is thus necessary to augment the current system, and by doing so all aircraft with ADS-B systems can serve as weather sensors in the air to collect vertical weather information. As a result, the objective of this research is deriving the vertical wind profile from ADS-B Mode-S ES data, evaluating the correlation between the decoded ADS-B Mode-S ES data and the unusual weather conditions, and building an unusual weather detection system prototype.
In order to acquire the vertical wind profile, this thesis analyzes the weather parameters decoded from the ADS-B Mode-S ES data. Because the weather parameters of magnetic heading and true airspeed need certain adjustments to get the accurate vertical wind profile, this thesis establishes an adjustment database for these two parameters with the assistance of the Global Forecast System (GFS). Next, this thesis discusses the feasibility of the vertical wind profile for use with three unusual weather algorithms, which are the meteorological parameter fluctuation processing algorithm, wind shear detection algorithm and abnormal movement detection algorithm. This thesis also presents the correlation between ADS-B Mode-S ES data and the unusual weather conditions. Furthermore, this thesis utilizes the above algorithms to find the features used in the machine learning algorithm, Support Vector Machine (SVM), and applies a feature selection technique to establish the learning model. This learning model can classify whether the feature sets derived from the ADS-B Mode-S ES data are in the normal condition or the unusual weather condition, and thus serves as an unusual weather detection system prototype.
The weather parameters derived from the ADS-B Mode-S ES data are adjusted in order to acquire the accurate vertical wind profiles. The three unusual weather algorithms evaluate the correlation between ADS-B Mode-S ES data and unusual weather conditions. Finally, this thesis uses the machine learning method to build an unusual weather detection system prototype to detect unusual weather conditions encountered by aircraft.
Table of Contents V
List of Tables VIII
List of Figures IX
CHAPTER 1 INTRODUCTION AND OVERVIEW 1
1.1 Motivation and objectives 2
1.2 Automatic Dependent Surveillance - Broadcast 5
1.3 Global Forecast System 8
1.4 Machine learning method 10
1.5 Previous works 12
1.6 Thesis organization 14
CHAPTER 2 VERTICAL WIND PROFILES ACQUISITION PROCESS 15
2.1 Method of acquiring the vertical wind profiles 16
2.2 Adjustment of magnetic heading 19
2.3 Adjustment of true airspeed 22
2.4 Results of overall adjustment 25
2.4.1 Magnetic heading adjustment database 26
2.4.2 True airspeed adjustment database 26
2.4.3 Verification of the adjusted results 27
2.5 Interim summary 30
CHAPTER 3 ALGORITHMS AND MACHINE LEARNING IN UNUSUAL WEATHER CONDITIONS 31
3.1 Unusual weather algorithms 32
3.1.1 Meteorological parameter fluctuation processing algorithm 32
3.1.2 Wind shear detection algorithm 36
3.1.3 Abnormal movement detection algorithm 41
3.2 Machine learning in unusual weather conditions 45
3.2.1 Process of machine learning method 45
3.2.2 Feature selection process 48
3.3 Interim summary 50
CHAPTER 4 EXPERIMENTAL RESULTS AND ANALYSES 51
4.1 Experiment setup 52
4.2 Acquisition of vertical wind profiles 55
4.3 Machine learning model establishment and classification results 57
4.3.1 SVM learning model establishment 57
4.3.2 SVM classification results 59
4.4 Interim summary 67
CHAPTER 5 CONCLUSIONS AND FUTURE WORKS 68
5.1 Conclusions 68
5.2 Future works 70
 J. C. Chen and H. F. Yuan, "Improving Low Level Wind Shear Alert System (LLWAS)," Journal of Aviation Safety and Management, vol. 1, pp. 85-102, March 2014.
 "WEATHER-RELATED AVIATION ACCIDENT STUDY," February 2 2010.
 "Statistical Summary of Commercial Jet Airplane Accidents," 2015.
 (May 29, 2015). Surface weather observations. Available: http://www.cwb.gov.tw/V7/eservice/docs/overview/observation/metro/sfc_obs.htm
 NCEP. (May 29, 2015). Global Forecast System (GFS). Available: http://www.emc.ncep.noaa.gov/index.php?branch=GFS
 C. C. Chang and C. J. Lin, "LIBSVM: a library for support vector machines," ACM Transactions on Intelligent Systems and Technology (TIST), vol. 2, p. 27, 2011.
 J. Scardina, "OVERVIEW OF THE FAA ADS-B LINK DECISION," FAA2002.
 G. J. Liou, "Assessment of Using UATADS-B Onboard Data to Simulate 3D Aviation Weather Flow Information," Department of Aeronautics and Astronautics, National Cheng Kung University, 2014.
 RTCA, "ADS-B MOPS (Minimum Operational Performance Standards for 1090 MHz Extended Squitter Automatic Dependent Surveillance - Broadcast and Traffic Information Services - Broadcast)," ed, 2003.
 S. d. Haan, "An improved correction method for high quality wind and temperature observations derived from Mode-S EHS," M. o. I. a. t. Enviornment, Ed., ed: Royal Netherlands Meteorological Institute, 2013.
 J. Brownlee. (2013, May 24). A Tour of Machine Learning Algorithms. Available: http://machinelearningmastery.com/a-tour-of-machine-learning-algorithms/
 Y. S. Abu-Mostafa, M. Magdon-Ismail, and H. T. Lin, Learning from data vol. 4: AMLBook Singapore, 2012.
 NCAR. (May 30, 2016). Low Level Wind Shear Alert System (LLWAS). Available: https://www.ral.ucar.edu/projects/low-level-wind-shear-alert-system-llwas
 CAA, "Introduction to LLWAS of RCTP & RCSS in Taiwan."
 NCAR. (May 30, 2016). Windshear and Turbulence Warning System - Hong Kong. Available: https://www.ral.ucar.edu/projects/windshear-and-turbulence-warning-system-hong-kong
 J. Piccola, "Flight Test Guide For Certification Of Transport Category Airplanes," U. S. D. o. Transportation, Ed., ed: Federal Aviation Administration, 2012.
 A. D. Hartkamp, K. De Beurs, A. Stein, and J. W. White, "Interpolation techniques for climate variables," 1999.
 O. Babak and C. V. Deutsch, "Statistical approach to inverse distance interpolation," Stochastic Environmental Research and Risk Assessment, vol. 23, pp. 543-553, 2009.
 S. d. Haan, "High‐resolution wind and temperature observations from aircraft tracked by Mode‐S air traffic control radar," Journal of Geophysical Research: Atmospheres, vol. 116, 2011.
 N. Doris, A. Izidor, and P. Boris, "Influence of Airspeed Measurement Error–Implications for Dead Reckoning Navigation," Transactions of FAMENA, vol. 39, pp. 13-22, 2015.
 J. Neter, M. H. Kutner, C. J. Nachtsheim, and W. Wasserman, Applied linear statistical models vol. 4: Irwin Chicago, 1996.
 C. P. Pu, M. L. Hsu, F. Yu, and L. Liu, "A Case Study of the Closed Low, Cold Front, Thunderstorm and Low-Level Wind Shear at Sung-Shan Airport," Journal of Aviation Safety and Management, vol. 1, pp. 227-243, July 2014.
 A. K. Blackadar, "The vertical distribution of wind and turbulent exchange in a neutral atmosphere," Journal of Geophysical Research, vol. 67, pp. 3095-3102, 1962.
 FAA, "PILOT WINDSHEAR GUIDE," 1988.
 "Flight Operations Briefing Notes - Windshear Awareness," October 2007.
 P. Juszczak, D. Tax, and R. P. Duin, "Feature scaling in support vector data description," in Proc. ASCI, 2002, pp. 95-102.
 X. G. Zhang, X. Lu, Q. Shi, X. Q. Xu, E. L. H.C., L. N. Harris, et al., "Recursive SVM feature selection and sample classification for mass-spectrometry and microarray data," BMC bioinformatics, vol. 7, p. 1, 2006.
 Y. Wang, J. Wong, and M. Andrew, "Anomaly intrusion detection using one class SVM," in Information Assurance Workshop, 2004. Proceedings from the Fifth Annual IEEE SMC, 2004, pp. 358-364.