Radar Signatures Of Small Consumer Drones

Radar Signatures of Small ConsumerDronesChenchen J. Li and Hao LingThe University of Texas at Austin1

Motivation and Objective The proliferation of small consumer drones has raisedmuch recent interest in their regulation and monitoring. One potential way to detect and identify these drones isusing a ground-based radar. Objective: To investigate the radar signatures of thesesmall drones.DJI Phantom3DR SoloDJI Inspire2

Approach Carry out laboratory measurement of small consumer drones. Examine their radar signatures in the form of inverse syntheticaperture radar (ISAR) images. ISAR imaging provides not only radar cross section (RCS), butalso maps the dominant scattering in 2-D.Scientific Questions:1. Will the small size and low reflectivity of the drone bodyresult in a very low RCS?2. Will the spinning propeller blades result in significantDoppler artifacts [1]?[1] P. Pouliguen et al., IEEE Trans. Antennas Propagat., 2002.3

Laboratory Measurement SetupVector network analyzer (VNA) S11 measurement.Drone mounted and rotated on a turntable.Start with baseline scenario and then deviate.Calibrated results in terms of absolute RCS using acalibration sphere. Horn mismatch and room clutter are reduced bysubtracting the moving average. 3 GHz of bandwidth. – 5 cm of down-range resolution.4

FrequencySwathISAR Image Formation 2-D image generated using 2-D inverseFourier transform of frequency/angle data.FrequencyresponseDown-rangePolar- Rect. Rotation Angle(or Time)2-D IFFTISARimageCross-range𝐼𝑚𝑎𝑔𝑒 𝑟, 𝑐𝑟 𝑤ℎ𝑒𝑟𝑒1.𝐸 0 𝑓, ϕ 𝑒 4567 𝑒 45897 𝑑𝑘, 𝑑𝑘𝐾, 𝐾-4𝜋𝑓𝑘, cos ϕ𝑐4𝜋𝑓𝑘- sin ϕ𝑐 Angular swath chosen for equal down-range/ cross-range resolution of 5 cm. Slide swath along angle to generate an ISARmovie.5

Baseline Scenario DJI Phantom 2 (35 cm diagonal). Azimuth scan at zero-elevation angle: results in a top-view ofthe drone. 12-15 GHz, blades stationary, VV-pol, no camera.θ %θ"#ISAR movie available at: http://users.ece.utexas.edu/ ling/DroneISARMovie.gif6

ISAR Snapshots ISAR image snapshots at different look angles with an outlineof drone overlaid. Maximum RCS in each snapshot is listed.EL 0 , AZ 8 EL 0 , AZ 45 Max -13.8 dBsmEL 0 , AZ 90 Max -19.4 dBsmEL 0 , AZ 135 Max -18.8 dBsmMax -9.3 dBsmEL 0 , AZ 172 Max -16.3 dBsm7

Deviations From BaselineBlades SpinningFrequency ChangePolarization ChangeBaseline ScenarioMounted CameraElevation ScanLarger Drones8

Deviations From BaselineRotating BladesBlades Stationary Max -14.6 dBsmEL 0 , AZ 102 Max -15.3 dBsmNo significant differences between blades rotating and stationary.Frequency Change EL 0 , AZ 102 Blades Spinning12-15 GHzEL 0 , AZ 194 Max -17.2 dBsm3-6 GHzEL 0 , AZ 194 Max -27.5 dBsmOn average, RCS at 3-6 GHz about 12 dB lower than 12-15 GHz.9

Polarization ChangeDeviations From Baseline HH-PolEL 0 , AZ 6 Max -13.5 dBsmEL 0 , AZ 6 Max -15.3 dBsmHH-Pol: weaker battery return, stronger motor return.Camera Mounted VV-PolWithout CameraWith CameraEL 0 , AZ 60 EL 0 , AZ 60 CameraMax -12.0 dBsmMax -12.0 dBsmCamera can only be seen at specific look angles.10

Deviations From Baseline Elevation scan instead of azimuth scan.Captures the shape of the drone in another imaging plane.Instead of top-view of the drone, captured the side-view.In practice, collected by flipping drone on its side and rotating.Azimuth ScanElevation ScanEL 0 , AZ 90 EL -90 , AZ 0 Max -9.3 dBsmMax -9.0 dBsm11

Larger Drones: 3DR Solo3DR Solo (46 cm diagonal)12-15 GHzEL 0 , AZ 90 Max -14.1 dBsm Drone shape and sizecaptured.Maximum RCS smaller thanPhantom 2 due to bodyshape.Similar trends as Phantom2.3-6 GHzEL 0 , AZ 90 Max -24.2 dBsm12

Larger Drones: DJI Inspire 1DJI Inspire 1 (56 cm diagonal)12-15 GHzEL 0 , AZ 270 Max -3.0 dBsm Drone shape and sizecaptured.Additional feature fromthe horizontal frame.Highest maximum RCS ofthe three drones.Similar trends asPhantom 2 and 3DR Solo.3-6 GHzEL 0 , AZ 270 Max -13.7 dBsm13

Recap Overall RCS level is low, but the drone size and shape can becaptured in the ISAR imagery. Non-plastic portions dominate their radar signatures (such asmotors, battery pack, and carbon-fiber frame). Drone propellers did not contribute a significant returnrelative to the drone body (static or spinning). Data collection was under idealized conditions, but it shouldbe feasible to collect such data from an actual drone in flight. Next: Carry out in-flight measurement of the small drone. Scientific Question: Can focused ISAR images be generatedfrom these small drones in flight?14

In-Situ Measurement Using a UWB Radar PulsON 440 (P440) ultra-wideband (UWB) radar byTime Domain Corporation. Emits short pulses at a pulse repetition frequency of10 MHz. Equivalent frequency bandwidth from 3.1 to 5.3 GHzcentered at 4.3 GHz.15

In-Flight Measurement Setup Measurement setup on the ground includes P440 radar,circulator, and single horn antenna. Phantom 3 Adv. is used since it has extractable GPS flight data. Collect range profiles (at 100 Hz) as the drone flies by in astraight line.16

Motion CompensationRaw Range ProfilesAfter Motion Comp.0111.5-5-109.5-159-208.58-257.50.5 11.52Time (s)2.53-30Down-range (m)Down-range Time (s)2.5-303Motion compensation is necessary to remove translational motion andretain only the rotational motion.Images generated through blind motion compensation (alignment of theRCS centroid).Baseline images also generated with aid from “ground truth,” GPS flightdata.Angle estimate based on ϕ cosGH (𝑅/𝑅LMN). Images formed using samek-space imaging as before.17

Phantom 3 Adv. ResultsBlind Motion CompensationAZ 270 GPS-Assisted Motion CompensationAZ 270 Images obtained from blind motion compensation are on-par withimages obtained from GPS-assisted motion compensation. Images are focused but narrower bandwidth of radar and limitednumber of scatterers on the drone make it challenging to discernthe shape.18

Larger Drone: DJI Inspire 1ISAR movie available at: http://users.ece.utexas.edu/ ling/DroneISARMovie Inspire.gif19

Compared to Laboratory MeasurementIn-Situ Meas. (3.1-5.3 GHz)AZ 270 Lab Meas. (3-6 GHz)EL 0 , AZ 270 Max -13.7 dBsmImages after blind motion compensation are comparable tothose obtained in laboratory measurement.20

Compared to Phantom 3In-situ Meas. Inspire 1In-situ Meas. Phantom 3AZ 270 AZ 270 Size difference, from Phantom 3, is observed.21

Conclusion ISAR images can capture the drone shape and sizedespite its small size and low reflectivity. Spinning propellers do not contribute significantDoppler clutter. Focused images can be generated from in-flightmeasurement. Radar is a potential candidate for tracking andclassification of small consumer drones.22

AcknowledgmentColton Bostick of UAV Direct and Lenny Tropiano ofFlyTheSkyDrone Productions for providing drones formeasurement.Alan Petroff of Time Domain Corporation for technicalsupport on the ultra-wideband radar.This work was supported in part by the National ScienceFoundation under Grant ECCS-1232152.23

Examine their radar signatures in the form of inverse synthetic aperture radar (ISAR) images. ISAR imaging provides not only radar cross section (RCS), but also maps the dominant scattering in 2-D. Scientific Questions: 1. Will the small size and low reflectivity of the drone body result in a very low RCS? 2.