1. EDGE™ MAV Control System Project # P09122 Erik Bellandi – Project Manager Ben Wager – Lead Engineer Garrett Argenna – Mechanical Engineering Michael Pepen – Electrical Engineering Tahar Allag – Electrical Engineering Ramon Campusano – Computer Engineering Stephen Nichols – Computer Engineering

2. EDGE™ Contents Background Project Planning Concept Development –Control System –Logic Controller –Sensors –Test Stand Future Work Risk Assessment

3. EDGE™ Background Past –Focused on small scale surveillance. Future –MAV rules have changed so now focus is on autonomy with small size being secondary. –Fly autonomously indoors and outdoors –Goal is to compete in the EMAV 2010 competition MIT Autonomous UAV Aerobatics Project MAV 2006 Model

4. EDGE™ Project Planning

5. EDGE™ Project Overview & Deliverables Product Description / Project Overview To design and build a flight control system for the Micro Aerial Vehicle, that will most quickly lead to a fully autonomous system. Key Business Goals / Project Deliverables Primary Goals: – Make the MAV as autonomous as possible. Stabilize Flight Adaptable Fully Tested and Integrate with Platform Secondary Business Goal: – Able to compete in the 2010 EMAV Competition.

6. EDGE™ Identify Customer Needs Needs Hierarchy 1.Control Capability 1.Be as autonomous as possible. 2.Create a stable flight. 1.Command the control surfaces appropriately. 3.Have a video relay system. 4.Process data from all inputs.. 2.Adaptability 1.Calibrated for the platform characteristics. 2.Compensate for environmental conditions. 3.Compensate for various payloads. 4.Have interchangeable sensors. 3.Receive Inputs 1.Work simultaneously with remote input. 2.Measure the current conditions. 3.Have GPS capability. 4.Weight and Size 1.Be light weight 2.Fit within MAV platform 5.Independence 1.Be independent of the platform.

7. EDGE™ Identify Customer Needs Relative Importance of Needs (1=Highest) #NeedNeeds ToImportance 1.1Autonomous as PossibleAs many autonomous functions as resources permit 7 1.2Create Stable FlightCompensate for instability1 1.2.1Command SurfacesMove Surface appropriate direction and amount1 1.3Video RelayCapture and relay surroundings to fly remotely10 2.1Process DataReceive Data and determine action4 2.2Calibrated for platformControl based on Aerodynamics of platform8 2.3Compensate for Environment Correct and Recover from environmental disturbance 3 2.4Compensate for PayloadAdjust aerodynamics based on payloads11 2.5Interchangeable SensorsUpgradeable and Replaceable12 3.1Simultaneous with RemoteWork concurrently and assist during remote input 5 3.2Measure ConditionsCollect data of all in-flight conditions2 3.3GPS CapabilityMeasure and program position13 4.1Light WeightMinimize weight <0.5kg9 4.2Fit with MAV PlatformAll onboard components must fit within platform9 5.1Independent from platformNot reliant on other projects, configurable and testable 6

8. EDGE™ Establish Target Specifications List of Metrics NumberMetricImportanceUnits 1Recover from 5mph gust4Mph, m/s 2Fly straight and level within a meter over a distance of 50 m 5m, ft 3Have at least 6 changeable parameters8# 4Weight less then 0.5 kg.7kg, lb 5Fit within MAV platform 2.25”x2.25”x8”6in, cm 6All testing matrices completed1# 7Receive and process remote signal2Y/N 8Transmit data to ground unit9List 9Process and use data from all sensors3Y/N – List 10Determine it’s position within 1 meter10m, ft 11Fly a designated pattern within 2 meters11m, ft

9. EDGE™ Concept Development

10. EDGE™ Overall System Architecture

11. EDGE™ Control System Concept Requirements: –Receive All Inputs (Pilot Input & Sensor Input) –Create Stable Flight –Command Surfaces (Elevons, Elevator, Rudder & Thrust) –Compensate for Environment (Disturbance) –Adaptable for Different Platforms Concept: PID Feedback Control for Each Input

12. EDGE™ Overall Control System Concept PID Feedback Control for Each Input Logic Controller Functions

13. EDGE™ Preliminary System Model Logic Controller Functions

14. EDGE™ Flight Dynamics Model

15. EDGE™ Logic Controller Selection Criteria – Control Capability – Adaptability – Inputs Receivable – Weight & Size – Cost – Complexity – Time to get working

16. EDGE™ Logic Controller Concepts –Last Year’s O-Navi Controller –Purchase different commercial fully developed board –Design and build from parts O-Navi Microcontroller

17. EDGE™ Logic Controller Design Logic Controller Design Concepts –MCU only –MCU and FPGA –FPGA only

18. EDGE™ Detailed System Diagram

19. EDGE™ FPGA Selection Selection Criteria –Familiarity –Price –Manual solderability –Power efficient –I/O pins Selection: Altera Cyclone III- EP3C16E144C8N –Package:EQFP –Logic elements:15408 –I/O pins:84 –Cost:$26.70 –CMOS process:65nm

20. EDGE™ FPGA System Diagram

21. EDGE™ Sensor Concepts Required Measurements –3-Axis Translation –3-Axis Rotation –Airspeed –Altitude –Angle of Attack

22. EDGE™ IMU Selection Selection Criteria –Cost –Dimension –Gyro range (deg/sec) –Acceleration range (g) –band width –Power Usage –Output Selection: Analog Devices- ADIS16350 –Resolution:14bit –Measurement :300 (deg/sec) –Interface:I2C/SPI –Voltage:5VCurrent:33mA –Price$528.00 –DOF6 axis

23. EDGE™ GPS Selection Selection Criteria –Accuracy –Voltage Supply –Power Consumption –Battery Backup –Built in Antenna –Baud rate Selection: Tyco Electronics (Vincotech) V23993-A1082-A –Accuracy:<2.5m –Voltage: 1.75-1.85VCurrent 35mA –Antenna :Included –Baud rate:4800-34400bpsUpdates:<0.1s –Dimensions: 0.55 x 0.45 x 0.095'' Weight: <0.05oz –Channels:12 –Price:$55.60 –Package connection –Dimensions –Weight –Price –Acquisition rate –Channel Tracking A1082-A

24. EDGE™ Airspeed Sensor Selection Selection Criteria –Differential Pressure Sensor: Cost Sensitivity Active Range Linearity Dimensions Selection: Freescale MPXV7002 –Range: 0 - 0.3 PSI-D –Sensitivity: 1 V/kPa –Cost:$15.78

25. EDGE™ Altimeter Sensor Selection Selection Criteria –Absolute Pressure Sensor: Cost Sensitivity Active Range Linearity Dimensions Selection: Freescale MPXH6130A –Range: 2.2 – 18.9 PSIA –Sensitivity: 39.2 mV/kPa –Cost:$15.09

26. EDGE™ Airspeed Sensor Selection Selection Criteria Calculation: –Bernoulli: Need sensor with range close to 0 to 0.015 psi Smallest Range Available: 0 to 0.3 psi Sensitivity = 1 V/psi For 1mV electronics accuracy: Low Speed: ΔP = 1 Pa, v = 3 mph, ΔP = 2 Pa, v = 1.81 m/s = 4 mph Resolution: 1 mph At Cruise: v = 30 mph, ΔP = 109 Pa, ΔP = 108 Pa, v = 29.88 mph Resolution :0.12 mph at cruise Cruise: v = 30 mph = 13.4112 m/s Standard Density:ρ = 1.21 kg/m 3

27. EDGE™ Altimeter Sensor Selection Selection Criteria Calculation –Hydrostatic Pressure: Need Absolute Pressure Sensor Standard P atm = 101.3 kPa = 14.69 psi Say altitude of 1000 ft = 300 m assuming ρ = constant = 1.21 kg/m 3 Considering normal variations or pressure and temperature want margin: Want range around: 75 kPa < P < 125 kPa = 10.8 psi < P < 18.13 psi Closest Range: 2.2 psi to 18.9 psi Sensitivity: 39.2 mv/kPa For 100 ft change: ΔP = -0.36 kPa For 10 ft change: ΔP = -0.036 kPa For 1 mV change: Δh = 7 ft

28. EDGE™ Test Stand Architecture

29. EDGE™ Test Stand Concept Multi-Axis Controlled Test Stand

30. EDGE™ Test Stand Motor Selection Selection Criteria –Cost –Holding Torque (oz-in) –Step Angle (deg) –Power (w) –Resistance (ohms) –Weight (g) Selection: Danaher Motion: 26M048B1B-V19 –Bipolar Holding Torque (oz-in):3.00 –Weight (g):57.2 –Step Angle (deg):1.00 –Cost:$26. 26M048B1B-V19

31. EDGE™ Test Stand Motor Selection Selection Criteria Calculation –Torque Inner Motor: I = 17 lb m in 2 (from CAD Model) Outer Motor: I = 150 lb m in 2 Need to calculate max α we want for the test stand Say max roll rate = 10 rpm (F-18 = 120 rpm) If reaches roll rate by 45 deg (1/8 rev) with constant α Using Rotational Kinematics: Solving for α: Calculating Torque:

32. EDGE™ MSD I Future Work

33. EDGE™ Control System Review & Finalize Non-Linear Plant Model Finish Feedback Conversions Model Sensors Linearize Plant and Sensor Models Develop Continuous Control Gains Discretize System Model Develop Discrete Control Gains Generate Control Law Code

34. EDGE™ Logic Controller Review Component Documentation Familiarize with NIOS II –Instantiate NIOS II core on FPGA –Store program code in Flash Implement Serial Protocols Investigate SD Card Data Storage Potential Begin Prototyping All Component Communication

35. EDGE™ Sensors Temperature Sensor Selection Determine Pitot-Tube Hardware and Location Review Sensor Documentation Develop Sensor Power Strategy Research Sensor Modeling Theory Develop PCB layout software knowledge Research Method for Digitizing Analog Sensors

36. EDGE™ Test Stand Research & Select Motor Drivers Research & Select Encoders Select Power Supply Select Transceiver Module Develop Communication Module Develop User Interface Refine Test Stand Design –Structure –Wire Routing

37. EDGE™ Other Develop Power Budget –Primary and Secondary Systems –Test Stand IMU Stage Motors

38. EDGE™ Current Schedule & Progress

39. EDGE™ MSD I Projected Progress Complete Tasks Listed in Future Work Finish Detail Design Early –Start Ordering Sensors and Components Early EE’s Can Start Modeling Sensors CE’s Can Start Checking Communication

40. EDGE™ Risk Assessment RiskProbabilitySeverityOverall RiskMitigation Component Interfacing LowHighMedThoroughly research all components and datasheets Damage when interfacing electronics LowHighMedAgain thoroughly research components and datasheets and Difficulty Discretizing Control System LowHighMedResearch digital controls and consult with faculty Having hardware soon enough to prototype and test Med Complete component selection as soon as possible and order Test Stand SafetyLowHighMedTest stepper motor driver with motor unattached, test procedures, protective cover for test stand, and emergency stop Other team’s delays prevent integration Low Test system with test fixture and flight testing with either OTS kit plane or previous year’s MAV platform.