Micro Air Vehicles for Optical Surveillancevehicle possibilities include conventional main...

• DAVIS, KOSICKI, BOROSON, AND KOSTISHACKMicro Air Vehicles for Optical Surveillance

VOLUME 9, NUMBER 2, 1996 THE LINCOLN LABORATORY JOURNAL 197

Micro Air Vehicles forOptical SurveillanceWilliam R. Davis, Jr., Bernard B. Kosicki, Don M. Boroson, and Daniel F. Kostishack

■ We present a study of micro air vehicles (MAVs) with wingspans of 7.4 to 15cm. Potential applications for MAVs, both military and civilian, are numerous.For most military applications, MAVs would be controlled by local users,operating covertly, to supply real-time data. This article focuses on a militarysurveillance application that uses either visible or mid-wavelength infraredimaging sensors. We present concepts for these sensors as well as for a miniatureKa-band communications link. MAV flight control would require miniaturemotion sensors and control surface actuators based on technology underdevelopment by the micro electromechanical systems community. As designed,the MAV would fly in a low Reynolds-number regime at airspeeds of 10 to 15m/sec. Propulsion would be provided by a combination of an electric motorwith either an advanced lithium battery or fuel cell, or by a miniature internal-combustion engine, which is a more efficient option. Because of the closecoupling between vehicle elements, system integration would be a significantchallenge, requiring tight packaging and multifunction components to meetmass limitations. We conclude that MAVs are feasible, given about two to threeyears of technology development in key areas including sensors, propulsion,aerodynamics, and packaging. They would be affordable if manufactured inquantity by using microfabrication techniques.

I , researchers have developedincreasingly sophisticated unmanned air vehicles(UAV) for military applications. In the Persian

Gulf War, for example, UAVs served in surveillancemissions and as decoys to distract enemy air defenses.Increased demands for intelligence are spawning thedevelopment of a smaller next-generation UAV calledthe micro air vehicle, or MAV. Small enough to fit inthe palm of your hand, an MAV would have an oper-ating range of several kilometers and transmit de-tailed pictures back to a portable base station, asshown in Figure 1. Several MAVs and their base sta-tion could be carried by a single person—an impos-sible scenario with the much larger UAVs, which havewingspans of 2 to 35 m.

On the basis of recent advancements in key tech-nologies of propulsion, flight control, communica-

tions, and sensors, we believe that MAVs with wing-spans of 7.4 to 15 cm, or 3 to 6 in, could be devel-oped in two to three years. These MAVs would be tentimes smaller than the smallest UAV currently flyingfor defense applications, the Self-Navigating DroneExpandable/Recoverable, or SENDER, from the Na-val Research Laboratory. Figure 2 compares the size oftwo proposed MAVs with four existing UAVs, includ-ing SENDER.

Potential capabilities for MAVs range from a fixed-wing surveillance MAV that uses a data link and line-of-sight control to an advanced MAV that hovers andnavigates independently and carries multiple sensors.Because of their small size and low power, such MAVswould be quite covert. In addition, exploiting micro-fabrication technology would make possible the pro-duction in large quantities of MAVs at low unit cost.


198 THE LINCOLN LABORATORY JOURNAL VOLUME 9, NUMBER 2, 1996

In December 1992, RAND [1] reported on astudy conducted for the Defense Advanced ResearchProjects Agency (DARPA) that considered the use ofa wide range of microdevices for defense applications.They projected that flying vehicles with a 1-cm wing-span and with payloads less than 1 g were feasible inten years. We were motivated by the RAND study tolook at MAVs in more detail.

Creating MAVs offers us the challenge of integrat-ing several technologies under development at Lin-coln Laboratory into a single vehicle. We began in1994 by considering all the key MAV subsystems, in-cluding sensors, and their integration into a practicalvehicle. Figure 3 shows payload mass versus wingspanfor several UAVs, with predicted MAV payloads fall-ing within the trends extrapolated from UAVs. Ourinitial efforts focused on determining the smallest ve-hicles possible within two to three years that would bebuilt with extensions of existing technologies. Thesetechnologies included microelectronics fabrication offocal-plane arrays and radio frequency (RF) compo-nents; the relatively new field of micro electrome-chanical systems (MEMS); and high-performancepropulsion systems.

Our initial design concept resulted in a model of a7.4-cm wingspan MAV, shown in Figure 4. Most ofthe model’s 10.5-g mass comes from a propulsion sys-tem comprising an advanced lithium battery and anelectric drive motor. The vehicle would be equippedwith a 21-GHz data link and a high-definition visiblecamera (1-g mass) that uses a silicon charge-coupleddevice (CCD) array of 1000 × 1000 pixels. We arecontinuing the process of evaluating subsystem tech-nologies and are now beginning to design a 15-cm,fixed-wing MAV.

Making the Most of MAVs

The MAV has a variety of potential uses in militaryoperations, including local reconnaissance, fire con-trol, and detection of intruders. Law enforcement or-ganizations could use MAVs for hostage rescue, bor-der patrol, traffic surveillance, and riot control. Formost of these applications, a swarm of MAVs couldprovide wide-area coverage.

Much of the appeal of the MAV for covert opera-tions comes from its small size. To determine how“invisible” the MAV would be on the battlefield, weexamined the various means of detection available to

FIGURE 1. Micro air vehicle (MAV) used for reconnaissance. A soldier carrying the MAV and a portable base stationcould remotely monitor the MAV under hostile conditions. Sensor data could be transferred in real time or stored onboard the MAV.

Data

Commands

Base station

MAV



FIGURE 2. Size comparison of existing unmanned air vehicles (UAVs) and proposed MAVs. The profile of a soldier, forscale, represents six feet. The smallest known UAV for defense applications currently flying is the Naval Research Labo-ratory Self-Navigating Drone Expandable/Recoverable (SENDER), which has a 1.2-m wingspan. The proposed MAVshave wingspans of 7.4 cm and 15 cm, or 3 and 6 in.

potential adversaries. To the human eye, an MAV inflight would resemble a small bird. MAV radar signa-tures would be similar to those of small birds and arethus likely to be lost in clutter. Furthermore, the pro-jected MAV airspeed of 10 to 15 m/sec is below theminimum detectable velocity for most radars. Infra-red search-and-track units would be able to detect anMAV only at short ranges because of its low power.For an electrically powered MAV, the acoustic signa-ture would be dominated by the aerodynamic noiseof the propeller, and would be audible only at closerange. An MAV powered by an internal-combustionengine with a muffler could achieve similar acousticperformance.

Although emissions from the MAV omnidirec-tional communications downlink could be detectedby an adversary, a broadband-radar warning receiverwill have limited detection capability because of thelow power emissions of the MAV. An electronic sup-port-measures receiver, on the other hand, can inter-cept the communications downlink if the receiveremploys a narrowbeam search antenna and the down-link frequency is known. There are several ways tolimit detection, however. One strategy is to usespread-spectrum techniques; another approach is tooperate the MAV autonomously, storing data onboard until a later time when conditions are favorablefor transmission.

SENDER

Exdrone

Micro air vehicles

Pioneer

Pointer



Design capabilities for MAVs are tightly coupledto their missions, most of which could be carried outby using fixed-wing aircraft that can circle areas of in-terest. The vehicle must fly 10 to 15 m/sec—fastenough to overcome head winds—and have an en-durance of 20 to 60 min to provide adequate rangeand mission time. For information-gathering mis-sions, simple acoustic, seismic, or magnetic sensorscan detect the presence of personnel, vehicles, andstructures. Additional sensors can permit the MAV todetect chemical, biological, and nuclear contami-nants in the atmosphere. Nonimaging sensors can de-tect light sources or measure local temperature. Vis-ible and infrared imaging systems can provide usefuldata for surveillance applications.

The simplest design is an MAV that can remainwithin the line of sight of a small base station thattracks the vehicle, maintains the communicationslink, and performs navigation calculations. A vehiclethat flies behind buildings or hills—beyond the lineof sight—must depend on some other approach to

communications and needs an independent means ofnavigation. One configuration that meets these re-quirements stores data on board with later readoutwhen the vehicle returns to line of sight. Anotherconfiguration includes an overhead communicationsrelay. Without a line of sight for navigation, alterna-tive navigation approaches such as dead reckoning,inertial navigation, and the Global Positioning Sys-tem (GPS) might be tapped, with the latter two de-pending on the availability of small components.

Intelligence gathering around or within buildingsrequires a hovering vehicle with a sophisticated navi-gation system. Alternatively, the MAV might be ableto perch, or fasten itself to a fixed object, or turn intoa crawler for local sensing. Combined hovering-flyingvehicle possibilities include conventional main rotor-tail rotor helicopters, coaxial rotors, propulsion-driven rotors, ducted fans, and tail-sitter airplanes.For some applications, the vehicle would need to befully autonomous and able to respond to the data re-ceived by onboard sensors.

FIGURE 3. Payload versus wingspan for existing UAVs and proposed MAVs. Predicted capabilities forMAVs fall within the trends extrapolated from larger vehicles. Note that the MAVs proposed in this articleare an order of magnitude smaller than existing UAVs.

10–4

10–3

10–2

10–1

100

101

102

103

10–1 100 101

Pay

load

mas

s (k

g)

Wingspan (m)

Exdrone

GlobalHawk

SENDER

MAVs

Gnat 750HunterPioneer

Pointer

PredatorDarkstar



Baseline Surveillance Application

The remainder of this article presents a baseline 15-cm fixed-wing MAV concept and discusses vehiclesubsystems and the status of the technology needed toimplement them. Baseline variations that provide ad-ditional performance, such as increased range, are alsopresented. To establish performance requirements, wechose a surveillance mission. A high-resolution, vis-ible charge-coupled device (CCD) camera is thebaseline sensor. We also discuss a larger and heaviermid-wavelength (3 to 5 µm) infrared (MWIR) cam-era that would increase the airframe size.

We assumed that a single MAV operates withinclear line of sight of a controlling base station atranges up to five kilometers. The base station tracksthe position of the MAV, performs navigation calcu-lations, and receives data from the MAV sensor. Wealso considered variations on this approach, includingthe possibility of GPS navigation. Table 1 summarizesadditional requirements chosen for the baseline MAV.

To detect people, the MAV requires a low operatingaltitude of about one hundred meters. Note that evenwith a narrowbeam receive antenna at the ground sta-tion, multipath effects could degrade link perfor-mance significantly and the low operating altitudemight not be maintained beyond a communicationsrange of about one kilometer. Longer ranges wouldrequire higher altitudes to keep the MAV within lineof sight and minimize multipath effects.

Aerodynamics and Vehicle Configuration

The aerodynamic design of MAVs offers unique chal-lenges because of the relatively low Reynolds-numberflight regime in which they fly. The Reynolds numberis a nondimensional similarity parameter that relatesinertial forces to viscous forces, and is proportional tothe flight speed times a characteristic length, such asthe wing chord. MAVs would operate at lowReynolds numbers of 20,000 to 50,000. Little re-search has been conducted on aircraft in the low-Reynolds-number regime, and both analytical and ex-

FIGURE 4. Model of Lincoln Laboratory concept of the smallest possible MAV (7.4-cmwingspan) with a visible imager for reconnaissance missions. This bottom view of themodel shows the downlooking camera port in the nose.



perimental research is required to develop the MAV.We do know that in this regime viscous forces aremore significant than those experienced by conven-tional aircraft in flight, and the MAV would experi-ence increased drag, reduced lift-to-drag ratios, andreduced propeller efficiency. Wing boundary-layerairflow would be laminar rather than turbulent, asfound in conventional aircraft, and boundary-layerseparation effects must be taken into account.

To compensate for these factors, we are examiningdrag minimization, which reduces the propulsion re-quirements, and specialized airfoil design. Despite theaerodynamic penalties of small size, an advantage ac-crues because as the size of the vehicle decreases, thevolume, and therefore the mass, decreases more rap-idly than the wing area required to generate lift.

Figure 5 illustrates some of these aerodynamic ef-fects. Curves of constant wing aspect ratio—wing-span squared divided by wing area—are also includedin the figure. The lift coefficient CL for minimumdrag is plotted as a function of the parasitic drag coef-ficient CD0

. Lift and drag are calculated by multiply-ing their respective coefficients by the dynamic pres-sure and wing area. The quantity CD0

accounts for allthe drag on the vehicle except the induced drag,which is the drag associated with generating lift. Fig-ure 5 shows that the values of CD0

for MAVs are con-siderably higher than for other aircraft because of theReynolds-number effect. We can expect to get CL val-ues of about 0.6 to 0.8.

The lift-to-drag ratio is an important measure ofthe propulsive power required to fly, and equals theratio of CL to CDtotal

, where CDtotal equals the parasitic

drag plus the induced drag. The lift-to-drag ratio forMAVs is only 5 to 8, while SENDER and conven-tional jet transports have a value of about 15, and sail-planes have larger values of 30 to 50. Because theMAV wingspan is constrained, a low aspect ratio ofabout 3 would be needed to provide enough wingarea to lift the vehicle. Similar aerodynamic consider-ations affect the performance of the propeller, andMAV propeller efficiencies would be about 50 to60%, compared with 80% or greater for conventionalaircraft.

To determine the configurations that best satisfythe demands of aerodynamic efficiency, flight con-

trol, and vehicle-subsystem packaging, we are consid-ering a variety of vehicle types and platforms. In addi-tion to the canard configuration shown in Figure 4,the possibilities for fixed-wing MAVs include conven-tional wing tail and flying wings, both illustrated inFigure 6, as well as multiple-wing combinations. Theconventional wing-tail configuration has effectivecontrol surfaces and predictable stability. It shouldhave a relatively good lift-to-drag ratio, although theinterference drag at the wing-fuselage juncture could

FIGURE 5. Comparison of lift and drag parameters forconventional aircraft, SENDER, sailplanes, and MAVs.Because of its small size and airspeed, the MAV experi-ences higher drag coefficients than other aircraft.

Table 1. Baseline MAV Performance Goals

Airspeed 10 to 15 m/sec

Endurance 20 to 60 min

Downlink rate 2 Mb/sec

Communications 5 km range

Navigation method Ground-station tracking(line of sight)

Visible sensor 1000 × 1000-pixel CCD;40° × 40° field of view

0

0.5

1.0

1.5

0 0.05 0.10

Parasitic drag coefficient CD0

MAV

Lif

t co

effi

cien

t CL fo

r m

inim

um

dra

g

Sailplane

SENDER

Jet transport

1

3

5

Wing aspect ratio = 10



be high. The flying wing would probably have a lowerlift-to-drag ratio, but its angle-of-attack range wouldbe larger, which is advantageous in turbulent condi-tions that cause large angle-of-attack variations. A dis-advantage is that the flying wing could be more diffi-cult to stabilize in flight.

Propulsion

Having explored the predicted lift-to-drag ratio andpropeller efficiency for the MAV, we now consider thepower required to fly one. Figure 7 shows the flightpower (airspeed times thrust), shaft power, and elec-tric power needed for a range of MAV sizes, with therequirements for our baseline 15-cm wingspan high-lighted. For conservative choices of lift-to-drag ratioand CL, the baseline flight power is 1.25 W. For a pro-peller efficiency of 50%, the baseline shaft power is2.5 W. If we use an electric motor with 60% effi-ciency, the baseline electrical power is 4.2 W. Thesevalues, however, provide only enough power for levelflight, and they must be doubled so that the MAV canturn, climb, and fly in gusty air.

To produce this power, we considered a variety of

efficient and lightweight propulsion systems, includ-ing electric motors powered by batteries or fuel cells,internal-combustion engines, turbines, compressedgas, and power plants using flywheels or capacitorsfor energy storage. The majority of these systemsproved inadequate. Compressed gas is not likely toprovide enough endurance, and flywheels and capaci-tors require significant development to be practical.Microsize turbines under development at MIT couldoffer robust performance and be used to generatethrust or electrical power; however, they require morethan three years of development [2]. Fuel cells, par-ticularly those combining atmospheric oxygen withhydrogen generated by using chemical hydride ormethanol oxidation [3, 4], have promise, but nonehave been built in the MAV-size range. Consequently,we focused on the most promising near-term candi-dates for power—battery-driven electric propulsionand internal-combustion engines.

Battery-driven electric propulsion has three advan-tages: it avoids the need for consumable fuel, is morereliable than internal-combustion engines, and isquiet. Small electric motors with adequate power

FIGURE 6. Promising MAV aerodynamic configurations. The conventional wing-tail configuration has effectivecontrol surfaces, predictable stability, and relatively good lift-to-drag ratio. However, interference drag at thewing-fuselage juncture could be high. The flying wing would have a larger angle-of-attack range than the conven-tional design; this range is advantageous in turbulent conditions. The flying wing could, however, be more diffi-cult to stabilize in flight because of its lower lift-to-drag ratio.

Flying wingConventional wing tail



densities are now available. New magnetic materialsunder development will also improve performance.

Electric propulsion, however, poses several ob-stacles. Currently available batteries in the requiredsmall sizes are not designed for the high dischargerates needed for MAV propulsion. Rather, they are in-tended for powering electronics at low discharge ratesover long periods. Lithium chemistries offer adequateenergy, but the case that contains the reaction con-tributes to battery mass. It should be possible to re-duce the case mass significantly while still safely con-taining the battery chemicals. With one or two years’development, a smaller battery should be able to pro-duce power densities of about 350 mW/g and energydensities of about 800 J/g that would provide a one-hour endurance for our 15-cm fixed-wing baselinevehicle. A hovering MAV would require significantlyhigher power densities, which batteries will be un-likely to achieve for some time.

Internal-combustion engines offer the possibilityof greater power and energy densities that would beadequate for hoverers and improve the performanceof fixed-wing MAVs. However, internal-combustionengines must be muffled for covert operations. Figure

8 compares total propulsion-system mass (includingfuel for a one-hour mission) for internal-combustionpropulsion, and for electric propulsion based on theprojections in the previous paragraph. We assumedthat our baseline MAV maneuvers 20% of the time;thus the average shaft power requirement is 3 W. Aninternal-combustion engine of the required size forthe baseline MAV is projected to offer a significantpower advantage. The smallest available model-air-plane engines, shown for comparison, use less ener-getic fuels than the hydrocarbon fuels that could beburned by an MAV engine. We predict the develop-ment of miniature engines using energetic fuels inabout one to two years.

As with any aircraft development, we would un-dertake an iterative design process to determine thevehicle gross mass and the distribution of massamong subsystems. For our 15-cm baseline MAV,aerodynamic performance and propulsion require-ments are the biggest factors in this process, severelyconstraining the mass available to other subsystems.Table 2 shows a preliminary mass distribution for thebaseline MAV. The mass allotments for the payloadand the flight-control subsystems and their electrical

FIGURE 7. Required propulsion power for a range of MAV sizes and a particularchoice of aerodynamic parameters. The lines indicate the power required to drivean electric motor, to drive a propeller shaft, and to fly.

10–2

10–1

100

101

4 5 6 7 8 9 10 15 20

Po

wer

(W

)

Wingspan (cm)

Electric power

Shaft power

Flight power

4.2 W

2.5 W

1.25 W

Sea level cruise at 13 m/secLift/drag = 5Lift coefficient = 0.6Propeller efficiency = 50%Electric motor efficiency = 60%Wing loading = 0.66 g/cm2

Baseline MAV



power sources are extremely small. This distributionresults in significant design challenges that are dis-cussed in the following sections on flight control,communications and navigation, and optical sensors.

Flight Control

Aerodynamic and propulsion factors limit the MAVto narrow ranges of airspeed and angle of attack, ren-dering the vehicle more vulnerable to gust upset.Flight control allows the MAV to fly at low airspeedin the presence of wind gusts and turbulence, stabi-lizes the vehicle with the aid of appropriate sensors,and provides aerodynamic controls. Figure 9 illus-trates the elements in the MAV flight-control system.Because the MAV airframe dynamic modes, such asDutch roll and the short-period longitudinal mode,will occur at higher frequencies compared with largervehicles, the MAV will need some means of augment-ing the natural stability of the airframe. In addition,the MAV should have the capability to fly itself topreprogrammed waypoints selected by the operator.

Microsize pressure gauges and accelerometers arecurrently available [5–7] and miniature magneticcompasses may also be feasible soon. Most useful for

this application, however, are rate sensors. Microchipangular-rate sensors are now being produced [8, 9],and will be useful for MAVs as soon as they are matedto miniaturized readout electronics. Drift rates fromthese sensors will be adequate for vehicle stabilizationapplications.

The ability to generate aerodynamic forces andmoments is also required to stabilize and maneuverthe MAV. These controls could be achieved with con-

FIGURE 8. Mass projections for propulsion system for a one-hour mission. Inter-nal-combustion engines that use energetic fuels would offer significant advan-tages over the best projections for battery-motor combinations. The smallestmodel-airplane engines are too large for the MAV.

Table 2. Baseline MAV Mass Distribution

Component Mass (g)

Airframe 6

Propulsion 36

Flight control 2

Payload Communications 3 Visible sensor 2

Total mass 49

100

101

102

103

100 101 102

Average shaft power (W)

Pro

ject

ed p

rop

uls

ion

-sys

tem

mas

s (g

)

Diesel internal-combustionengine and hydrocarbon fuel

Available internal-combustion model-airplane engines

Baseline MAV requirement

Electric motor and battery



ventional discrete hinged surfaces such as ailerons andelevators; distributed micro-actuated control surfaces;or wings that change shape or warp. All methods re-quire micromechanical actuators. Because of recentadvances in MEMS, a number of different actuatorcandidates should be available in the next one to twoyears [10–19]. Examples include integrated force ar-rays, which generate electrostatic attraction force, andseveral approaches using piezoelectric crystals. Theseactuators can generate linear forces or be used in theconstruction of rotary machines that produce torque.They have the advantage of employing fabrication ap-proaches that lend themselves to high productionrates. Tiny conventional electromagnetic actuators,such as those used in watches, may also be tapped forsome first-generation MAVs [20].

The flight-control sensors and actuators must beintegrated into the flight-control system by using adigital processor with the necessary signal interfaces.A custom microcontroller chip that also serves as thecentral processor for the communications and opti-cal-sensor subsystem will accomplish this function.

Communications and Navigation

For nonautonomous operation the communicationssystem must provide flight- and payload-controlcommands to the MAV and receive data transmittedfrom onboard sensors. For our baseline MAV, thecommunications system also tracks the position ofthe MAV from the ground.

Using the Ka-band for communications provides agood compromise of antenna size, antenna beam-width, and propagation losses. The 21-GHz band waschosen because of its availability and the existence ofcircuit technologies for satellite communications inthat band. A half-dipole antenna at this frequency isonly 0.7 cm long, readily fitting within the verticalstabilizer of the vehicle and providing omnidirec-tional coverage. With current gallium arsenide(GaAs) monolithic microwave integrated circuits(MMIC) technology, we can build an onboard trans-ceiver with 25 mW of transmit power. This trans-ceiver requires 200 mW from the vehicle, and a massof about 2 g. Development of this transceiver would

FIGURE 9. MAV flight-control system. Flight control requires sensors that measure motion (roll, pitch, and yaw) of theMAV, and aerodynamic control inputs that stabilize and maneuver the MAV in wind gusts and turbulence.

Position accuracy

Line-of-sight jitter

Wind gusts

Flaps (discrete, distributed) Warped wings

MotorsMEMS technology

Control surfaces

Actuators

Flight-control sensors

Micro gyroscopesAccelerometersPressure gaugesMagnetic compasses



require a custom stripped-down architecture withinMMIC capabilities. The onboard receiver portion ofthe transceiver would require most of these design re-sources, even though its data rate is low. Simple oscil-lators, power amplifiers, and phase or frequencymodulators would be straightforward for transmitterdesign in the range of several megabytes per second.

For a minimum system with an operating range of1 km, a ground station equipped with a 13-cm dishantenna could accommodate a video downlink at 2Mb/sec and a command uplink at 1 kb/sec. The dishantenna at the ground site is mounted on a drive thatallows it to track the azimuth and elevation of the ve-hicle. Range is derived with the two-way link. Theazimuth, elevation, and range information is used todetermine the vehicle location to within about 7 m inthree dimensions. The navigation calculations andvideo display are performed on a laptop computer.

Range capability could be increased by using alarger dish antenna or increasing the onboard powerconsumption, which is only a small portion of the to-tal power for the baseline system. The data rate couldbe improved by using proportionally more power.Another power-usage adjustment involves adding lowprobability of detection or anti-jam capabilities. Forcommunication ranges out to about 10 km, the netresult of these adjustments is a system consisting of aground station (with a dish antenna proportionallylarger for the longer distance) and several MAVs thatcould be carried in a knapsack.

This simple line-of-sight communications systemlimits operation to a minimum elevation angle ofabout 6° above the horizon—an MAV altitude ofabout 100 m at 1 km—and can be blocked by terrain,trees, or buildings. One alternative would be to use anoverhead communications relay that would allow theMAV to fly close to the ground or at least below thedirect line of sight. Such a relay function could be ac-complished with a second flying vehicle such as aUAV. An MAV is not a good candidate for the relayvehicle unless the carrier frequency is much lower,and the relay craft is close to the mission MAV.

Autonomous operation is desirable when the lineof sight to the base station cannot be preserved. Inthis mode, an air vehicle climbs periodically to trans-mit data to the user. Autonomous operation requires

a means of navigation independent of the base stationwhen the MAV is out of sight. GPS is an obviouschoice, but further development is required to reducethe size, mass, and power requirements of a GPS re-ceiver. GPS works by receiving a simultaneous num-ber of satellite transmissions and, through some fairlysophisticated signal processing, by deducing thereceiver’s position in three dimensions. A GPS an-tenna operates at L-band, where the characteristic an-tenna dimension is larger than the MAV unless rela-tively heavy dielectric materials are used in its design.Also, current GPS receiver power consumption is toolarge for MAVs. Efficient receivers plus the requiredsignal processing can take hundreds to thousands ofmilliwatts with present designs, although improve-ments in these areas are expected in a few years.

We considered two additional navigation schemes:dead reckoning and inertial navigation. Dead reckon-ing is not a good candidate because it requires fairlyaccurate knowledge of the winds aloft in order to cal-culate absolute position. Inertial navigation, whichuses microrate sensors and accelerometers plus carefulfiltering algorithms to deduce absolute position, haspotential. Tiny inertial navigation sensors are underdevelopment, but achieving drift rates suitable for po-sition determination is probably still several yearsaway. Another possibility, if compatible with the mis-sion, is to navigate with prepositioned beacons withinthe line of sight of the vehicle.

Optical Sensors

Without optical sensors, an MAV would be just apocket-sized, high-tech model airplane, unsuitablefor surveillance operations. Like all MAV compo-nents, these sensors must meet small mass and powerrequirements: sensor mass must be under 2 g andpower consumption under 100 mW. These param-eters are one to several orders of magnitude smallerthan for any commercial cameras available today. Inaddition, surveillance missions require high-resolu-tion sensors with the ability to see in the completerange of outdoor light levels, from noonday sunlightto overcast starlight. These optical sensors must havehigh resolution (approximately 1000 × 1000 pixels)for recognition of human figures at the mission alti-tude of 100 m. Other operational requirements are



driven by two important environmental factors:movement of the aircraft, which could cause imageblur, and relatively high operating temperature. Tomeet these requirements, Lincoln Laboratory hasconsidered visible and infrared sensors.

Visible sensors use an object’s reflected radiation toproduce an image. The visible imager is sensitive tothe visible spectrum (400 nm to 700 nm) and thenear-infrared spectrum (700 nm to 1000 nm). Thelatter range is typically utilized in night-visiongoggles. Although imaging capability at night is desir-able for the MAV, current night-vision technologythat uses high-voltage image intensifiers is too heavyto implement, has a limited dynamic range, and doesnot work well in daylight conditions. Current re-search efforts at Lincoln Laboratory focus on a super-sensitive silicon imager that will be capable of re-sponding to the full range of desired light levels.

Infrared sensors use an object’s emitted radiationand, to a lesser degree, its reflected light to produce animage. Because the emitted radiation depends on anobject’s temperature and emissivity, and not solar illu-mination, infrared sensors are sensitive during nightconditions. One disadvantage to this technology isthat sensitive infrared imagers operate at cryogenictemperatures and require a cooling unit that increasesthe MAV size and mass. Another disadvantage is thatan infrared image requires more interpretation than avisible-band image. Warmer objects are prominent,but some terrains have low temperature contrast,which makes placing an object into context with itssurroundings difficult. Researchers are addressing thisproblem by combining infrared and visible informa-tion to produce more easily interpreted images.

Visible Sensor

A visible silicon imager built with current technologyfor the baseline MAV can address noonday sunlightto partial moon illumination, which is most of thedesired light-level range. Operation of the cameradown to overcast-starlight night conditions is not fea-sible with current visible CCD technology because ofthe large f-number and small lens-size optics for theMAV. Consequently, infrared imaging or a visible im-ager with larger optics would have to be used for theseextremely low light-level conditions.

Several important environmental factors influencethe design of the optical sensor. The first of these isaircraft movement, which has two components: for-ward movement of approximately 15 m/sec, andmovement caused by turbulence. The degree ofmovement determines the maximum exposure timebefore image resolution is degraded. The high opera-tional temperature of the device (ambient air tem-perature reaches up to 115°F) contributes to the gen-eration of dark current in the visible imager. Darkcurrent can increase noise in the image in addition tothe read noise. Although cooling the visible sensorwould enhance performance, the cooling unit wouldalso exceed the MAV mass and power requirements.For a given temperature, there is a trade-off betweenlimiting the time to read out the device, thereforelimiting the dark current, and conversely maximizingthe time to read the device, and therefore limiting thebandwidth necessary for the output amplifier, thusalso limiting the thermal read noise. The impact oftemperature on the imaging device also affects our de-cision to attempt to integrate all control and readoutelectronics on one imaging chip.

Candidates for the visible imager include threetypes of CCD devices: a full-frame CCD with framestore (CCD-1), the same device but with the addi-tional feature of a built-in electronic shutter (CCD-2)[21], and an interline transfer CCD with diode-arraylight-sensitive elements (CCD-3). CMOS imager de-vices (active pixel arrays) are also candidates [22–23].Table 3 summarizes the important properties of thefour candidate architectures. The properties are listedapproximately in decreasing order of importance forthe MAV application.

All the CCD devices can be made with the re-quired resolution. The large pixel size of the CMOSdevice, however, would require a large lens and unac-ceptably increase the size and weight of the entirecamera. The CMOS imager does have the advantagesof standard integrated-circuit fabrication techniquesand low operating power requirements. The twoback-illuminated CCD candidates have the bestquantum efficiency, share the lowest read noise, andtherefore are the most sensitive detectors for night ap-plication. CCD-1 has the largest packet size amongthe CCDs; therefore, it is best equipped for a large



dynamic range and can handle large amounts of darkcharge. Because of its relatively simple fabricationprocess, it is also low in dark current.

The shutter property was originally thought to beimportant because of the short image exposure times(1 to 8 msec) dictated by MAV motion. The shuttermechanism for the CCD-1 device is the rapid move-ment of charge into a frame-store array. However, thisaction takes 1 msec, which equals the shortest shuttertime expected and indicates the potential for imagesmear. (Shutter time is less than 1 µsec for the otherimager candidates.) A simulation was carried out toassess whether the frame-shift shutter method causedunacceptable image degradation. The conclusion isthat this method caused only minor degradation toan average aerial image, and therefore is not an im-portant limitation.

For the MAV, a remote frame store is needed tocompensate for the light leakage associated with elec-tronic shutters. In normal applications the exposuretime is comparable to the time needed to read out the

image. For the MAV imager, however, the readouttime is approximately 1 sec, a factor of about 1000larger than the exposure time. This longer readouttime puts severe requirements on the shutter leakageand is the reason that all three CCD candidates areequipped with a frame-store region that is remotefrom the imaging region. However, the CMOS de-vice is not readily able to be equipped with a remoteframe-store region, and therefore image corruptionby shutter leakage is a risk in this device.

The pixel-readout binning function is planned foruse in low-light-level conditions, to improve the sig-nal-to-noise ratio and therefore the resolution at lowlight levels. The CMOS device is not equipped to binphotocharge in a noiseless way, because charge is con-verted to voltage (and therefore read noise is added) atevery pixel site.

The last two entries in Table 3 deal with requiredoperating voltage levels and signal output. Any MAVof reasonable complexity and sophistication requiresan internal regulated power supply. Therefore, the ex-

Table 3. Candidate Device Architectures for MAVs

CCD-1: Full Frame, CCD-2: Full Frame CCD-3: Interline CMOS

No Shutter with Shutter Transfer with Frame Store Active Pixel

Resolution 1000 × 1000 pixels 1000 × 1000 pixels 1000 × 1000 pixels 1000 × 1000 pixels

Pixel size 5 × 5 µm 5 × 5 µm 5 × 5 µm <20 × <20 µm

Quantum efficiency >85% >85% 20% + lenslet array 20–25% *

Read noise <10 e– at 1 MHz <10 e– at 1 MHz <10 e– at 1 MHz 14 e– at 0.1 MHz

Packet size 40,000 e– 30,000 e– 15,000 e– 64,000 e– *

Dark current 100 pA/cm2 300 pA/cm2 50 pA/cm2 500 pA/cm2

Shutter Move to frame store Electronic Electronic Electronic

Frame store Yes Yes Yes No

Noiseless binning Yes Yes Yes No

Voltage levels 11 V 21 V 5 V 5 V

Signal output Analog-to-digital or Analog-to-digital or Analog-to-digital or Analog-to-digitalcharge-to-digital charge-to-digital charge-to-digital converter

converter converter converter

* Extrapolated or estimated from Reference 22.



istence of larger operational voltages for CCD-1 andCCD-2 is not an important penalty. The output of allfour candidates is planned to be produced with con-ventional CMOS amplifiers and analog-to-digitalconverters. A new device, a direct charge-to-voltageconverter [24], is currently being investigated as analternative. This device could perform the signal con-version function at a lower power than conventionaltechniques. However, it would operate only on chargesignals, which precludes the CMOS device. The con-verter does require integration directly on the CCDchip, but CCD and CMOS integrated fabricationprocesses with sufficient capability have already beendemonstrated.

As mentioned above, our strategy for designingMAV imaging sensors is to reduce the pixel size of thefocal-plane array, thus minimizing the size of the op-tics, and to incorporate additional functions, such ascharge-to-digital conversion and clocking, on thesame chip as the focal-plane array. Figure 10 shows aconcept that incorporates this approach. The visiblesensor is based on a silicon CCD focal plane with a1000 × 1000 array of 5-µm pixels. The optics wouldbe built with microfabrication techniques, resultingin overall camera dimensions of about one cubic cen-timeter, or the size of a dime. The mass of the com-plete camera is under 1 g, and power requirements areunder 25 mW.

The 1000 × 1000 pixel array provides image reso-lutions equivalent to high-definition television. Thesample image shown in the figure is derived from aphotograph taken at an altitude, aspect angle, andwidth of field of view representative of conditionsseen by an MAV CCD sensor. The photograph wasdigitized to form an image representative of the num-ber of pixels (in the horizontal dimension) and 4-bitgray scale envisioned for the CCD sensor. The result-ing image provides sufficient detail to recognize thepresence of vehicles and personnel on the ground.

The image contains 4 Mb of data that must bestored or transmitted to the MAV operator. An up-date rate of 0.5 frame/sec should be adequate forflight speeds of 10 to 15 m/sec, which would require acommunications link capability of 2 Mb/sec (assum-ing no image compression). Frame rates could be in-creased with a more capable communications system.

Infrared Sensor

A candidate infrared-camera design has been devel-oped with off-the-shelf technologies. Figure 11 showsa 3-to-5-µm-band infrared camera based on a plati-num silicide (PtSi) CCD focal-plane array with 512 ×485 pixels. Other infrared imager technologies withhigher quantum efficiencies such as indium anti-monide (InSb) and mercury cadmium telluride(HgCdTe) were considered, but the PtSi arrays havethe smallest possible pixel sizes and lowest noise. Thelonger wavelength range necessitates larger and morecomplex optics than the visible camera, and 3 g of liq-uid nitrogen is required to cool the CCD for aboutone hour. (Liquid nitrogen could be generated in thefield with a portable mechanical refrigerator.) Thecomplete camera mass is under 16 g with power un-der 150 mW from using small pixels and combiningfunctions on the CCD chip. While this camera massgreatly exceeds the payload mass limit of Table 2, sec-ond-generation MAVs with advanced propulsion sys-tems could be capable of carrying such a sensor for ex-tremely dark night-vision missions.

The infrared image in Figure 11, while only 256pixels wide, indicates the image quality from this in-frared sensor. The view spans a parking lot and roads;automobiles and a pedestrian can be easily identified.A 2-Mb/sec communications link accommodates arate of 2 frames/sec for the full 512 × 485-pixel array.

The two cameras described here are within the cur-rent state of the art for imaging sensors. The smallpixel size has been demonstrated, as has the combin-ing of processing functions on CCD arrays. The re-maining step is the investment of resources to designand fabricate the custom CCD and CMOS camerachips needed for this application.

Systems Integration

Systems integration for the MAV to meet low massand volume requirements and to permit low-costmass production requires close interaction amongmultiple disciplines. Because conventional aircraft in-tegration technology does not apply at the MAVscale, new approaches must be developed.

Systems integration affects the selection of compo-nents and the design of the vehicle. For example, bat-



FIGURE 10. Visible sensor for the MAV. (a) Advanced silicon CCD technology permits the packaging ofa 1000 × 1000-pixel imager and associated output electronics in a single chip, resulting in a camera thesize of a cubic centimeter and weighing less than 1 g. (b) With resolution comparable to that of high-definition TV, the simulated image shows an example of the detail that could be obtained from the vis-ible-light camera mounted in an MAV.

Visible-light camera

0.85 cm

Imager

(a)

1.2

cm 40° × 40°field of view

Timing electronics

Output electronics

100-m altitude, 45° aspect

(b)

Simulated visible-light camera image

ApertureAngular resolutionPixel countFrame rateMassPower

0.26 cm0.7 mrad (7 cm at 100-m range)1000 × 10000.5 frame/sec<1 g<25 mW



FIGURE 11. Mid-wavelength infrared (3 to 5 µm) camerafor day and night conditions. (a) The camera, based onplatinum silicide (PtSi) CCD technology, is larger thanthe visible camera of Figure 10 because of requirementsfor larger optics and liquid nitrogen cooling of the focalplane. (b) A sample image of a parking lot taken at arange of 200 m shows a pedestrian and automobiles.

teries need to serve multiple functions, such as con-tributing to the vehicle structure. Electronic func-tions have to be combined, which can entail using asingle custom application-specific integrated circuitfor the entire vehicle. Mass can be reduced by thin-ning electronic circuitry and using interconnectionsprinted onto the vehicle shell in place of intercon-necting wiring. The close proximity of vehicle sub-systems also provides challenges such as control ofheat dissipation, vibration from internal-combustionengines, and electromagnetic interference from elec-tric motors.

Conclusions

An MAV could provide significant new capabilities toa wide range of users. Several MAVs and a base stationcould be transported and operated by a single indi-vidual, providing real-time data directly to the localuser. The MAV promises to be particularly useful forcovert operations. A variety of vehicle configurationsand sensors could be used for many possible missions.We conclude that about two to three years of aggres-sive development in the appropriate technologies willproduce a working MAV with an imaging sensor.Propulsion is the most significant challenge. Otherkey technologies include aerodynamics, flight con-trol, communications, sensor development, and sub-system integration.

Acknowledgments

This work was sponsored by the Department of theAir Force and DARPA. Many Lincoln Laboratorypersonnel contributed to the concepts in this article.In particular, Walter Morrow and Milan Vlajinac sawthe merit of pursuing micro air vehicles and suppliedvision and guidance to the overall effort. In addition,the authors would like to express particular apprecia-tion to Marvin Pope, who led the original effort andis responsible for many of the ideas presented here.

The following Lincoln Laboratory personnel con-tributed to the article’s technical content: Brian Aull,Gregory Berthiaume, Jamie Burnside, MichaelCantella, Brian Edwards, Mike Fedor, DavidGonzales, David Harrison, Robert Hull, MichaelJudd, David Johnson, Robert Reich, TimothyStephens, Gary Swanson, and Douglas White.

Sample day/night camera image

ApertureAngular resolutionPixel countFrame rateMassPower

0.77 cm1.3 mrad (13 cm at 100-m range)512 × 4852 frames/sec<15.5 g<150 mW

Day/night camera

(b)

LN2reservoir

LensInsulation Cold shield

3.5 cm

Focalplane

Filter

Window

40° × 40°field of view

Camera dewar

2 cm

(a)



R E F E R E N C E S1. R.O. Hundley and E.C. Gritton, “Future Technology-Driven

Revolutions in Military Operations—Results of a Workshop,”RAND National Defense Research Institute, Dec. 1992.

2. A.H. Epstein and S.D. Senturia, “Macro Power from MicroMachinery,” Science 276, 23 May 1997, p. 1211.

3. S. Surampudi, S.R. Narayanan, E. Vamos, H. Frank, G.Halpert, A. LaConti, J. Kosek, G.K. Surya Prakash, and G.A.Olah, “Advances in Direct Oxidation of Methanol Fuel Cells,”J. Power Sources 47 (3), 1994, pp. 377–385.

4. W.G. Tashcek and C.E. Bailey, “High Energy Metal HydrideFuel Cell Power Source,” Defense Technical Information Cen-ter Report ADA056491, Defense Logistics Agency, CameronStation, Alexandria, Va., 1978.

5. C. Ajluni, “Low-Pressure Sensor Opens Wide ApplicationsFrontier,” Electron. Des., 16 Dec. 1996, pp. 59–64.

6. B.E. Boser and R.T. Howe, “Surface Micromachined Acceler-ometers,” IEEE J. Solid-State Circuits. 31 (3), 1996, pp. 366–375.

7. R. Allan, “Silicon MEMS Technology Is Coming of Age Com-mercially,” Electron. Des., 20 Jan. 1997, pp. 75–88.

8. J. Bernstein, S. Cho, A.T. King, A. Kourepenis, P. Maciel,and M. Weinberg, “A Micromachined Comb-Drive TuningFork Rate Gyroscope,” Proc. IEEE Conf. on Micro Electro Me-chanical Systems, Fort Lauderdale, Fla., 7–10 Feb. 1993, pp.143–148.

9. M. Weinberg, J. Bernstein, J. Borenstein, J. Campbell, J.Cousens, B. Cunningham, R. Fields, P. Greiff, B. Hugh, L.Niles, and J. Sohn, “Micromachining Inertial Instruments,”SPIE 2879, 1996, pp. 26–36.

10. P. Dario, R. Valleggi, M.C. Carrozza, M.C. Montesi, andM. Cocco, “Microactuators for Microrobots: A Critical Sur-vey,” J. Micromech. Microeng. 2 (3), 1992, pp. 141–157.

11. C. Liu, T. Tsao, Y.-C. Tai, T.-S. Leu, C.-M. Ho, W.-L. Tang,and D. Miu, “Out-of-Plane Permalloy Magnetic Actuators forDelta-Wing Control,” Proc. IEEE Conf. on Micro Electro Me-chanical Systems, Amsterdam, 29 Jan.–2. Feb. 1995, pp. 7–12.

12. C.H. Ahn and M.G. Allen, “A Fully Integrated Micromag-netic Actuator with a Multilevel Meander Magnetic Core,”IEEE Solid-State Sensor and Actuator Workshop, Hilton Head,S.C., 22–25 June 1992, pp. 14–18.

13. K. Matsuzaki, T. Matsuo, and Y. Mikuriya, “Comparison ofElectrostatic and Electromagnetic Motors Based on Fabrica-tion and Performance Criteria,” 1994 5th Int. Symp. on MicroMachines and Human Science Proc., Nagoya, Japan, 2–4 Oct.1994, pp. 77–81.

14. H. Guckel, T.R. Christenson, K.J. Skrobis, T.S. Jung, J. Klein,K.V. Hartojo, and I. Widjaja, “A First Functional CurrentExcited Planar Rotational Magnetic Micromotor,” Proc. IEEEConf. on Micro Electro Mechanical Systems, Fort Lauderdale,Fla., 7–10 Feb. 1993, pp. 7–11.

15. W.P. Robbins, D.L. Polla, T. Tamagawa, D.E. Glumac, andW. Tjhen, “Design of Linear-Motion Microactuators UsingPiezoelectric Thin Films,” J. Micromech. Microeng. 1 (4), 1991,pp. 247–252.

16. D.L. Polla, P.J. Schiller, and L.F. Francis, “Microelectro-mechanical Systems Using Piezoelectric Thin Films,” SPIE2291, 1994, pp. 108–124.

17. T. Niino, S. Egawa, H. Kimura, and T. Higuchi, “ElectrostaticArtificial Muscle: Compact, High-Power Linear Actuatorswith Multiple-Layer Structures,” Proc. IEEE Conf. on Micro

Electro Mechanical Systems, Oiso, Japan, 25–28 Jan. 1994, pp.130–135.

18. S.M. Bobbio, M.D. Kellam, B.W. Dudley, S. Goodwin-Johansson, S.K. Jones, J.D. Jacobson, F.M. Tranjan, andT.D. DuBois, “Integrated Force Arrays,” Proc. IEEE Conf. onMicro Electro Mechanical Systems, Fort Lauderdale, Fla., 7–10Feb. 1993, pp. 149–154.

19. J.D. Jacobson, S.H. Goodwin-Johansson, S.M. Bobbio, C.A.Bartlett, and L.N. Yadon, “Integrated Force Arrays: Theoryand Modeling of Static Operation,” J. Microelectromech. Syst.4 (3), pp. 139–150.

20. M.A. Gottschalk, “Miniature Motors Deliver Big Perfor-mance,” Design News 52 (9), 5 May 1997, pp. 67–68.

21. R.K. Reich, R.W. Mountain, W.H. McGonagle, J.C.-M.Huang, J.C. Twichell, B.B. Kosicki, and E.D. Savoye, “Inte-grated Electronic Shutter for Back-Illuminated Charge-Coupled Devices,” IEEE Trans. Electron Dev. 40 (7), 1993, pp.1231–1237.

22. R.H. Nixon, S.E. Kemeny, C.O. Staller, and E.R. Fossum,256x256 CMOS Active Pixel Sensor Camera-on-a-Chip,”1996 IEEE Int. Solid-State Circuits Conf., San Francisco, 8–10Feb. 1996, pp. 178–179.

23. B. Ackland and A. Dickson, “Camera on a Chip,” 1996 IEEEInt. Solid-State Circuits Conf., San Francisco, 8–10 Feb. 1996,pp. 22–25.

24. S.A. Paul and H.-S. Lee, “A 9-b Charge-to-Digital Converterfor Integrated Image Sensors,” IEEE J. Solid-State Circuits 31(12), 1996, pp. 1931–1944.



. , .leads the Optical SystemsEngineering group, whichdevelops optical and aerome-chanical systems for aircraftand spacecraft. He also leadsthe Lincoln Laboratory MAVprogram.

Bob joined Lincoln Labora-tory in 1977 after spendingfour years in the U.S. AirForce developing airbornehigh-energy laser systems,working principally in struc-tural dynamics and beampointing. At Lincoln Labora-tory, Bob continued his workin high-energy laser systemsand managed the developmentof expendable radio frequency(RF) and infrared countermea-sures that protect aircraftagainst missiles. He also par-ticipated in the developmentof reentry decoys for ballisticmissiles. Throughout his careerhe has been involved in flighttesting, conducting measure-ments and qualifying payloadsfor a wide variety of aircraft.He has also managed severalstudies concerning spacecraftand UAVs.

Bob received a B.S. degree inengineering mechanics fromLehigh University, and earnedS.M. and Sc.D. degrees inaeronautics and astronauticsfrom MIT.

. joined Lincoln Laboratory in1983 with the responsibilityfor charge-coupled device(CCD) technology develop-ment and yield improvement.He now serves as associateleader of the Microelectronicsgroup, where he is responsiblefor silicon process and devicetechnology.

Before joining LincolnLaboratory, Bernard spent tenyears in managerial positionsat Sperry Research Center,General Instruments Micro-electronics, and FairchildSemiconductor, involved firstin metal nitride oxide silicon(MNOS) device developmentand pilot production, and thenin process and product engi-neering for microprocessorproduction, and finally inadvanced silicon technologydevelopment. Prior to that, hespent six years on the technicalstaff at Bell Telephone Labora-tories at Murray Hill, where heconducted research on growth,structure, and dielectric andelectroluminescent propertiesof various thin-film materialsand structures. He becameinvolved in CCD and processtechnology shortly after thesedevices were invented.

Bernard is currently a seniormember of the IEEE. He hasbeen an author or coauthor ofmore than twenty technicalpublications and patents. Hereceived a B.A. degree inphysics from Wesleyan Univer-sity and M.A. and Ph.D.degrees in solid state physicsfrom Harvard University.

. is the associate leader of theOptical CommunicationsTechnology group, where hehas led several projects devel-oping technologies for space-borne laser communicationssystems. He has also worked intraditional satellite communi-cations areas, such as adaptivearrays, onboard signal process-ing, integration and testing ofcomplex systems, signal andreceiver design, and datacommunications.

Don joined Lincoln Labora-tory in 1977, after receivingB.S.E, M.A., and Ph.D. de-grees in electrical engineeringfrom Princeton University.Outside of his professionalinterests, Don is a pianist andtheatrical music director.

. is associate group leader of theAdvanced Systems and Sensorsgroup, responsible for thedevelopment of advancedinfrared sensor technology andseeker experimental systemsfor ballistic missile defenseinterceptor programs of theArmy, Navy, and BallisticMissile Defense Organization.

Dan joined Lincoln Labora-tory in 1967 and worked onelectronic countermeasuresand penetration aid technolo-gies for the Reentry SystemsProgram. He served threeyears, beginning in 1974, astest director and assistantleader at the ALCOR radar ofthe Kwajalein Missile Range.In 1977, he joined theGround-Based Electro-OpticalDeep Space Surveillanceprogram (GEODSS) to de-velop electrooptic sensors. In1982, he joined the SpaceSurveillance group as groupleader to work on the applica-tion of advanced space surveil-lance sensor technologies andsystems to strategic defenseballistic missile systems. Forthe past eight years as associategroup leader, Dan has pursuedthe development and transferof advanced electrooptictechnologies in ballistic missiledefense interceptor seekerprograms for missile defenseapplications.

He holds B.S., M.S. andPh.D. degrees in electricalengineering, all from CarnegieMellon University.

Date post:	31-Aug-2020
Category:	Documents
Upload:	others
View:	3 times
Download:	0 times

Micro Air Vehicles for Optical Surveillancevehicle possibilities include conventional main...

Documents