Ij

An Analysis of Unmanned Airborne Vehicle Relay Coverage

in Urban Environments

Carmen Cerasoli

The MITRE Corporation

Eatontown, NJ

[1, 2, 3]. In this study, the connectivity provided by a radio

ABSTRACT

relay UAV over a relatively small 500 m2 urban area was

Unmanned Airborne Vehicles, UAVs, can be used as radio

assessed. A ray tracing model [4] was used to compute

relay platforms in environments characterized by poor RF

path losses between the UAV and ground based receivers.

connectivity. Those environments can be urban, forested

Excellent connectivity was found for low flying UAVs in

the limited urban area considered here if the UAV

or mountainous regions where no Line of Sight (LOS)

exists between ground transmitters and receivers. The

transmitting powers are comparable to ground platforms.

effectiveness of a UAV air relay was assessed in an urban

Under those conditions, nearly 100% of the outdoor

area using the ray tracing method. UAVs were placed at

receivers have the required signal strength to maintain

various positions and heights over an approximately 500

desired data rate connectivity. This optimistic result is not

square meter urban area, and excess path loss relative to

applicable to the UAV relay for two reasons. SWAP

unblocked LOS propagation was computed for outdoor

restrictions put a limit on available transmit power for a

receivers. UAV altitudes were 500, 1000 and 2000 meters

communications relay, and there is the possibility of

and frequencies were 400 and 1800 MHZ.

friendly interference. Communication systems operating in

the vicinity can be victim to the UAV transmissions if

Excess path loss is the appropriate metric when UAV

unblocked propagation paths exist between UAV and

transmitter power is limited due to Size, Weight, and

potential victim receivers.

Power (SWAP) constraints and the need to minimize

friendly interference. We found that a UAV at 2000 m

A well designed UAV relay will minimize transmit power

provided coverage for over 90% of the ground receivers

to address the two issues raised above. Power will be set

within 10 dB of LOS path loss. Most of the connectivity

near the level allowing connectivity to ground receivers

was obtained via unblocked LOS paths due to the small

with LOS paths. Additional transmit power can provide

urban area considered. Diffraction around buildings

connectivity to receivers blocked by buildings. Therefore,

played a larger role in providing connectivity than

a practical metric is the received signal power relative to

reflections at 400 MHz and the roles were approximately

that for an unblocked LOS path. Relative Received Power

equal at 1800 MHz. The percentage coverage results were

(RRP) was used throughout this paper along with

found to be stable to within +/-5% when poorly known

Cumulative Distribution Functions (CDF) for the number

building and ground electrical properties or ray tracing

of ground receivers within a given dB range of RRP.

computational parameters were varied over reasonable

values.

In Section II, the method used to compute RRP is

described while Section III provides RRP results for

Channel characteristics for time of arrival, time delay

varying UAV altitudes and positions above the city.

spread and angle of arrival were also computed. The

Channel statistics for time of arrival, delay spread and

multi-path nature of the NLOS paths suggested the

angle of arrival are given in Section IV, and discussion and

possible use of Multiple Input, Multiple Output (MIMO)

summary are presented in Section V. Appendix I describes

techniques.

the effects of frequency on coverage along with the roles

of reflections and diffractions. Appendices II and III

present the coverage sensitivity to building and ground

electrical properties and ray tracing computation

I. INTRODUCTION

parameters

Present and planned combat missions are taking place in

urban settings where building blockages can reduce point

to point radio connectivity. One way to obtain improved

connectivity is to use airborne relays carried aboard UAVs

II. METHOD

The urban area and UAV placements are shown

schematically in Figure 1, where Rosslyn VA buildings

were used. The readily available building geometry file

encompassed only an area of approximately 200 m2. A

larger computational area was created by tiling four sets of

Rosslyn buildings together and additional buildings were

120

added outside the edge of a 500 by 500 m receiver array.

Figure 2. Histograph of Rosslyn building heights.

Receiver spacing was 2.5 m and the buildings

encompassed approximately 25% of the area, resulting in

RRPs and their associated CDFs were computed using the

over 29,000 outdoor receiver positions. The UAV was

stated UAV heights and positions. The CDFs provided the

positioned at the city center and eight compass points.

percentage of ground receivers at or above a given RRP

Three UAV altitudes were used, 500, 1000 and 2000 m,

level. This metric was computed as a function of UAV

and the 2000 m altitude was consistent with the Future

position, height and transmit frequency. The sensitivity of

Combat System Class IV UAVs. Frequencies were 400

coverage percentage to building and ground properties and

and 1800 MHz consistent with military radio spectrum use

ray tracing computational parameters was also assessed.

and the UAV mounted antenna transmitted isotropically.

Building heights were color coded in Figure 1, and heights

III. RELATIVE RECEIVED POWER &

were typically in the 20 m range with a few tall (>80 m)

PERCENT COVERAGE

buildings as shown in the histograph of Figure 2.

Figure 3 shows a contour plot of RRP and the UAV

position. The UAV altitude was 500 m and situated to the

UAV Positions Around Edge

East of the urban area, while transmitter frequency was

of Coverage Area at Eight

400 MHz. A large number of receivers possess a LOS to

Compass Points

the UAV with RRPs ~ 0, while the shadowing behind

buildings can also be seen. Although contour plots are

UAV at Center

useful, a better metric was the CDF for RRP shown in

Figure 4. Again, UAV altitude was 500 m; the blue and

black lines are for the UAV at all cardinal and

Zuav

intercardinal points, while the red line was for the UAV

centered over for urban area.

Using the CDF shown in Figure 4, one finds

approximately 74% of the ground receivers had signal

strengths greater than -10 dB of the free space LOS value

for all eight compass points. When the UAV was

North

positioned directly over the city, the percent coverage with

RRP greater than -10 dB was approximately 87%. The

East

RRP = -10 dB value will be used as the criterion for most

Area Containing Receiver

of this analysis.

Array with 2.5 meter

Spacing

Buildings Outside of Receiver

Area Provide Reflection

Surfaces

Figure 1. Aerial view of the urban area and UAV

placement. Building heights are color coded and the

red rectangular surface area depicts receiver array.

2 of 7

with increasing UAV altitude as building shadowing

Zuav = 500m

diminished and coverage approached the LOS values.

East

+10

NLOS

Blocked Path

Isotropic

Zuav = 500m

Zuav = 1000m

Zuav = 2000m

Radiator

LOS

100

Unblocked Path

-40

NLOS

LOS

ESW

Compass Point

Figure 5. Percent coverage based on -10 dB RRP

as a function of compass position. Also shown are

Buildings

coverage values for LOS paths, dashed red, and

North

city centered UAV, solid blue.

East

UAV to South

Figure 3. Contour map of Relative Received

Power over ground receiver array. UAV position

at 500 m altitude East is shown.

100

UAV to East

UAV to West

Center

N, E, S & W

NE, NW, SE & SW

UAV at Center

+20

-20

-40

-60

10 < RRP < -5

RRP (dB)

-5 < RRP < -15

Figure 4. Relative Received Power Cumulative

-15 < RRP < -25

UAV to North

Density Function for Zuav = 500 m at all eight

-25 < RRP < -35

compass points plus city centered UAV.

Figure 6. Contour map of RRP for the UAV

The effects of UAV altitude and compass point position

positioned at the four cardinal compass points plus

are summarized in Figure 5. The solid red lines show

UAV at city center.

percent coverage at the RRP > -10 dB level for the three

Building shadowing can be seen in Figure 6 where the

UAV altitudes at the eight compass points. The deviation

shadowed regions moved as the UAV was positioned at

from the mean value was small and decreased as Zuav

the cardinal compass points and the urban center. Zuav was

increased. The dashed red lines depict the -10 dB coverage

500 m to emphasize shadowing. A very conservative

for LOS paths averaged over the compass points. The

metric can be defined by computing the minimum RRP for

NLOS paths (reflected and diffracted) increased coverage

all eight compass points. This metric guarantees a level of

from 12% to 4% as Zuav increased from 500 to 2000 m.

RRP independent of the UAV compass point position.

The solid blue line depicts percent coverage from a

Contour plots of the minimum RRP values over the eight

centered UAV. By Zuav = 2000 m, the coverage from the

compass positions for the three UAV altitudes are shown

UAVs at the city's edge was approaching the value for a

in Figure 7. The percent coverage at -10 dB RRP is given

city centered UAV. As expected, these results were

as a bar chart below each respective contour plot, while

strongly geometry dependent. Coverage steadily improved

coverage values for a city centered UAV are depicted by

3 of 7

the red lines. The benefit of increasing UAV altitude is

-20

clearly seen. For Zuav = 2000 m coverage approached the

RRP = 0 dB

city centered value.

Even using this conservative

minimum RRP metric, good coverage of ~75% receivers

-140

within -10 dB of LOS signal strength was attained for

-20

planned Class IV UAV operation independent of the

RRP = -10 dB

precise UAV position. Of course, this optimistic result was

a result of the relatively small urban area considered.

-140

Zuav =2000

Zuav = 500 m

Zuav =1000 m

-20

RRP = -20 dB

-140

τ (µsec)

7.0

8.0

Figure 8. Impulse response at three receiver

positions where RRP was 0, -10 and -20 dB.

100

UAV at Center

The probability density function for τM and τRMS at ground

receivers is shown in Figure 9, where the LOS arrival time

is denoted in the τM plot. Time delay spreads were modest

with more than 99% of the receivers having τRMS below

0.05 microseconds. The relatively