1. Owen Macmann, Aerospace Engineering, Pre-Junior, University of Cincinnati Devon Riddle, Aerospace Engineering, Junior, University of Cincinnati Mahogany M. Williams, Computer Engineering, Senior, Wilberforce University ASSISTED BY: Wei Wei, Graduate Research Assistant Dr. Kelly Cohen, Faculty Mentor Rotorcraft Handling Qualities and System Identification 1 1

2. Motivation & Operational Goals of UC’s UAV Rotorcraft Program We want to use available quadrotor technology to help with structural firefighting In order to deploy quadrotors in live situations, we need to develop a controller/autopilot for them. We will need a mathematical model to do this. 2 2

3. How can we get this model? We can use a process called system identification. – Relates inputs to outputs. – Results in a mathematical model. We will need experimental data gathered from flight testing – Inputs from the physical RC controller – Outputs from the inertial measurement unit 3 3

4. Objectives Objective 1: Objective 1: Study the flight characteristics of the rotorcraft and learn how to pilot AR parrot drone. Objective 2: Objective 2: Utilize CIFER software and “System Identification” to create a dynamic model of the rotorcraft. Objective 3: Objective 3: Prepare a detailed flight test and modeling report. 4 4

5. Initial Testing 5 5

6. What kinds of inputs? 4 inputs total – Yaw – Pitch – Roll – Thrust 6 6

7. What kinds of outputs? Inertial measurement unit measures 9 outputs simultaneously: – 3 attitudes Pitch, yaw, and roll angles – 3 angular rates – 3 accelerations 7 7

8. System Identification 8 8

9. What is System Identification? System Identification is the process of obtaining a mathematical model via extraction from test data. Using such models, we can predict the dynamic behavior of the motion of the quadrotor. The main goal of this project is to apply state-of-the-art System Identification techniques to develop the dynamic model of the radio-controlled AR Parrot Quadrotor Drone system 9 9

10. What is System Identification? 10

11. What is System Identification? Inputs System Outputs 4 signal inputs9-12 motions 11

12. AR Drone Power Data Time History Circuit Board Quadrotor System Identification Process 12

13. Data Consistency 1313

14. Why Check for Consistency? 14

15. How Can We Check for Consistency? We can obtain a spectrum of frequencies comprising a time- dependent signal. This process is called Fourier Analysis. It is realized by the Fourier Transform. 15

16. How Can We Check for Consistency? The Fourier Transform: Inverse Fourier Transform: We can attenuate or remove parts of the frequency spectrum and then use the Inverse Fourier Transform to obtain a “smoother” signal. 16

17. Unsmoothed “Roll” Input 17

18. Unsmoothed Rolling Angle Output 18

19. Unsmoothed Input vs. Output 19

20. Smoothed “Roll” Input 20

21. Smoothed Rolling Angle Output 21

22. Smoothed Input vs. Output 22

23. “Roll” Input Comparison 23

24. Roll Angle Output Comparison 24

25. Frequency-Response Method Frequency Sweep Inputs Aircraft Data Consistency & Reconstitution Multivariable Spectral Analysis Frequency Response & Partial Coherence Transfer Function & State Space Modeling 25

26. Timeline Week 1Week 2Week 3Week 4Week 5Week 6Week 7Week 8 Learn to pilot Flight Testing Physical Model State Space Vali- dation Final Report Journal Paper Due Final Day 26

27. Questions? 27