Jammer Acquisition with GPS Exploration and Reconnaissance JÄGER

Jammer Acquisition with GPS Exploration and Reconnaissance JÄGER SCPNT PRESENTATION Adrien Perkins James Spicer, Louis Dressel, Mark James, and Yu-Hsuan Chen

!Motivation NextGen Airspace Increasing use of GPS! Need reliable signal during landing 2

Motivation Newark Int l Airport Example personal privacy device jamming Need to locate and disable jammer quickly 3

Motivation Requirements and Simplifications Requirements! Quickly determine location of jamming device! Navigate in a GPS denied environment! Use only a single vehicle! Pinpoint location Initial Simplifications! Stationary jammer! Single jammer! Constant jamming 4

!Methodology Jammer Detection Independent of GPS! Bearing to jammer!! Signal strength + heading!! Directional antenna + magnetometer! Multicopter vs. Plane!! Take measurements at fixed points strainsecuritysystems.squarespace.com 5

!Methodology Path Planning Represent problems as Partially Observable Markov Decision Process (POMDP)!! Grid-based, offline solution to a probabilistic model of sensor and vehicle motion!! Based on real-time signal measurements and current belief of sensor location, take action that localizes the jammer in the shortest amount of time! Onboard computer actively communicating with flight computer Flight Computer Computer 6

!Methodology Positioning Signals of Opportunity Hardened GPS 7 Aviation Signals! Alternate Position Navigation and Time (APNT)! Automatic Dependent Surveillance Broadcast (ADS-B)! Universal Access Transceiver (UAT)! 1090! Distance Measuring Equipment (DME) wifi Digital TV 3G / 4G Cell Network

!Methodology Aviation Signals Possible Signals!! Automatic Dependent Surveillance Broadcast (ADS-B)! Universal Access Transceiver (UAT) and 1090!! Distance Measuring Equipment (DME)! Initial testing with UAT signals! Leverage known altitude from barometer 8

!System Design System Diagram Weight: 2 kg Power: 12 W Weight: 0.5 kg Power: 8 W! All communication through autopilot! Each system is completely independent and modular!! Each has own power source 9

!System Design Vehicle Selection! Target Performance!! 20 minute flight!! 4 kg payload! DJI S1000 Spreading Wings Octocopter!! Frame Weight: 4.2 kg!! Max Takeoff Weight: 11.0 kg! Power!! Max of 4000 W!! 1.5 kg battery payload 10

!System Integration Path Planning Two way communication with Autopilot! Hunt Mode for autopilot!! Waits for commands from outside source!! Reports completion of maneuvers! Ground station for testing Flight Computer Computer 11

!System Integration Path Planning vehicle autopilot Path planning Ground station Safety pilot Not Human Position Command execution Command Path planning to human 1 Path planning to human 12

!System Integration Path Planning vehicle autopilot Path planning Ground station Safety pilot Not Human Path planning 1 to human Command execution Command Human generated commands 2 Human generated commands 13

!System Integration Path Planning vehicle autopilot Path planning Ground station Safety pilot Not Human Path planning 1 to human Command execution Position Command 2 Full integration and automation 3 Full integration and automation 14

!System Integration APNT Electromagnetic Interference (EMI) 2 sources of interference onboard!! Telemetry antenna at 915 MHz!! Computer at 1.575 GHz GPS signal before shielding (16 db peak) 15

System Integration APNT Electromagnetic Interference (EMI) 16

!System Integration APNT Electromagnetic Interference (EMI) Before shielding (16 db peak) 17 After shielding (-2 db peak)

!Testing Flight Testing Split series of flights by hardware being flown Series 2:! Pixhawk + APNT Series 1:! Pixhawk Series 4:! Pixhawk + APNT + Path Planning Series 3:! Pixhawk + Path Planning 18

Testing Autopilot Only Test vehicle performance! Stability Weight simulation Tethered Testing 19

!Testing UAT Ground Test Static and dynamic tests 12 km 40 km 20

Results APNT Ranging +235 Pseudorange from San Jose Pseudorange(m) +177 +118 +59 min = 39656 0 100 200 300 400 500 600 700 +292 Pseudorange from Woodside UAT GPS Pseudorange(m) +219 +146 +73 min = 11262 0 100 200 300 400 500 600 700 21

Results Ranging Histo of Pseudorange from San Jose, mean=76.2575m, std=11.6779m Pseudorange Error Gaussian Fit 10-2 10-3 20 40 60 80 100 120 Histo of Pseudorange from Woodside, mean=138.5428m, std=12.6833m Pseudorange Error Gaussian Fit 10-2 22 10-3 80 100 120 140 160 180 200

Results Static 37.4326 37.4324 GPS 37.4322 UAT 37.432 lattitude(degree) 37.4318 37.4316 37.4314 37.4312 37.431 23 37.4308-122.183-122.1825-122.182-122.1815-122.181 longtitude(degree)

Results Accuracy 24

Results Dynamic 37.4325 GPS 37.432 UAT lattitude(degree) 37.4315 37.431 25-122.183-122.1825-122.182-122.1815-122.181 longtitude(degree)

!Testing Flight Testing! APNT!! Constant altitude!! Loiter at each vertex (static)!! Dynamic flight! Jammer Acquisition!! Wifi router to mimic jammer 26 Wifi

Future Future Work Flight testing of APNT at Camp Roberts Continued path planning Integration Jammer acquisition testing Hardened GPS vs various signals of opportunity! Explore the options 27

Questions?