Optical Wireless on Swarm UAVs for High Bit Rate Applications

Optical Wireless on Swarm UAVs for High Bit Rate Applications Ch. Chlestil*, E. Leitgeb*, S. Sheikh Muhammad*, A. Friedl*, K. Zettl*, N. P. Schmitt**, W. Rehm**, N. Perlot*** *Institute of Broadband Communications, Graz University of Technology, Graz, Austria **EADS Corporate Research Centre, Munich, Germany ***Institute for Communications and Navigation, DLR Munich, Germany

Abstract-Unmanned Aerial Vehicles (UAV) flying in swarm formations carrying a variety of sensors for monitoring and surveillance is in fact a future application for both civil and military use. Different requirements of guidance and control of UAV were already identified; however, the communication facilities between mobile platforms and air-to-ground links are restricted to low bit rate and radio based technology. In the near term data rates in the area of 100 Mbit/s to 1 Gbit/s will be needed to handle multiple sensor information in real-time and parallel. Thus, UAV in combination with Free Space Optics (FSO) is a new challenge and will be needed everywhere, where big amounts of data have to be delivered in real-time. With up to 2.5 Gbit/s the optical solutions beat all the other ones like for example microwaves. I. INTRODUCTION One of the crucial factors to ensure the interaction among swarm UAVs is the reliable, high performance wireless communication link among UAVs. The air-to-air UAV communication system enables the sharing of sensor and map information among UAVs, while an air-toground communication system provides mission information to the ground station for mission control and display [3]. The key point is that actions of each UAV can reduce the risk in the environment for all other UAVs. Simulations [2] showed that ignoring the cooperation in the assignment results in mission plans that have much lower expected performance/value. Furthermore, techniques that model this probability but ignore its coupling to each UAVs mission can result in very poor performance of the entire fleet. Reference [3] is investigating Radio Frequency (RF) links using Orthogonal Frequency Division Multiplex (OFDM) based multi-carrier transmission as a future communication means for swarm UAVs. A key aspect in maintaining high performance in the communication channel is the rapid relative movement between UAVs and their high speed over ground. Such high mobility leads to a much shorter coherence time and a larger Doppler spread in the multi-path fading channel. As a direct result, the orthogonality among subcarriers is lost and Inter-Carrier-Interference (ICI) degrades the performance [3]. Although simulation results in [3] shows that the performance degradation of OFDM system due to large Doppler spread and ICI is well tolerated,

communication scenarios by using FSO inbetween swarm UAVs have not been deeply investigated so far. Laser links have a number of advantages over RF, not least in the area of security. The performance of laser systems is not affected by Electromagnetic Interference (EMI) and is not subject to interference with traditional wireless devices such as microwave or radio system, nor are laser based systems affected by the Fresnel zones [1]. Since current low cost FSO systems with data rates up to 2.5 Gbit/s are transmitting base band signals (intensity modulation/direct detection), the light detection is not affected by Doppler spread. Data rates in excess of 1 Gbit/s, using other than RF links, will be needed to exploit sensor capabilities, as well as to reduce RF spectrum saturation, in the near term. Current data capacities of 274 Mbit/s are stressed when carrying multiple sensors simultaneously. Classes of sensors that particularly tax links are radar imagers when full phase history is sent to a ground station for post processing and multispectral sensors with high resolution and wide fields of view [4]. The following section II outlines possible UAV network architectures, covering different aspects of reliability and availability. Section III is dedicated to the technical issues of mobile FSO, including results of power budgets for various scenarios and section IV handles important advantages of using FSO over other wireless technologies. II. UAV NETWORKING ARCHITECTURES For the application of FSO scenarios in relation to reliability and availability, different kind of networks have to be considered. A. Ring Architecture In Fig 1 a UAV FSO-network in a ring architecture is shown. All UAVs described in the network architectures do have bidirectional FSO links. In the event of a broken link (failure) between two UAV FSO-links, an indirect link can be used. Instead, the information is sent in the other direction of the ring network. Thereby, a partial security against failure can be achieved. The installation of additional links increases the availability and the security against failure (redundant links). An Optical Repeater has then to be used, if there is no line of sight between transmitter and receiver. Fig 1 also shows the link to the optical ground station, which is indeed a very high sophisticated technical

solution (see chapter III and Reference [10] for more information on this topic).

Fig. 3. Meshed Architecture

Fig. 1. Ring Architecture

III. TECHNICAL ASPECTS FOR MOBILE FSO COMMUNICATION

B. Star Architecture In Fig 2 the UAV formation and the FSO links form a star architecture network are shown. The UAV in the middle of the formation acts as an Optical Multipoint Unit (OMU). Five users are permanently connected by their optical transceiver units to the Optical Multipoint station. The advantage of this configuration are shorter distances between any two FSO-units, because the OMU is used as a repeater. In general, the OMU is in the centre of the area, but the architecture has the disadvantage of a single point of failure. If the OMU fails, a system breakdown of the whole installation is caused. To improve the reliability of this architecture, a redundant Multipoint Unit has to be installed.

Fig. 2. Star Architecture

C. Meshed Architecture For high reliability, the optimum network architecture is a meshed network, because it combines the advantages of a star and a ring architecture. Different connections are possible. In Fig 3 an expanded version of the ring architecture is shown. An information flow from one UAV to another UAV can be realized in different forms. The information can be sent in the other direction (anticlockwise) of the ring-network. Hence, more security against failure can be achieved and a network with high reliability can be realised. However, an increase of the complexity of the optical network architecture also means a dramatically increase in the technical realization of the routing scheme.

The greatest challenges when transmitting data over collimated laser beams between airborne terminals are loss of sight, atmospheric influences and geometrical losses. A. Weather Influences In clear air weather conditions one has to consider turbulent motion of the atmosphere in the presence of moisture and temperature gradients. These effects give rise to disturbances in the atmosphere’s refractive-index in the form of cells called optical turbules [5]. The index of refraction, one of the most significant parameters of the atmosphere for optical wave propagation, is very sensitive to small-scale temperature fluctuations. In particular, temperature fluctuations combined with turbulent mixing induce a random behaviour in the field of atmospheric index of refraction. The refractive-index structure constant Cn2 is a measure of the strength of the fluctuations in the refractive index. Values of Cn2 typically range from 10-17 m-2/3 in the upper atmosphere up to 10-13 m-2/3 near the ground. Over short time intervals at a fixed propagation distance and constance height above the ground it may be reasonable to assume that Cn2 is essentially constant [6]. For a plane wave, a low turbulence and a specific receiver, the scintillation variance σχ2 [dB2] can be expressed by the following equation σ χ2 = 23.17 ⋅ k 7 / 6 ⋅ Cn2 ⋅ l 11/ 6 k=

2π

(1)

λ

2

where σχ is the scintillation-variance, k is the wave number and l defines the link length. Scintillation peak with peak amplitude is equal to 4σχ and attenuation related to scintillation is equal to 2σχ [7]. For weak fluctuations and planar waves, factors of 1.23 instead of 23.17 for (1) and (2) can be found in Reference [5]. With (1) we can estimate the attenuation of the signal in dependence on scintillation with ⎛ 2π ⎞ ascin = 2 23.17 ⋅ ⎜ 109 ⎟ ⎝ λ ⎠

7/6

⋅ Cn2 ⋅ l 11/ 6

(2)

This expression suggests that larger wavelengths would experience a smaller variance, all other factors being equal [7]. As Table 1 depicts, the attenuation per kilometer decreases within the wavelength. Especially the moderate attenuation values between 1330 nm and 1550 nm

wavelength gives the advantage of choosing optical components in this optical window for free space communication. It should be noted that Table I considers a worst case scenario of the scintillation variance; additionally only one transmitter (mono beam) systems and narrow focused beams in the magnitude of below 7 mrad were assumed. TABLE I

NUMERICAL VALUES OF ATTENUATION IN DEPENDENCE ON THE CN2

Cn2 Wavelength [nm] 650 690 780 830 850 950 980 1330 1540 1550 10591

weak 10-16

Turbulence intermediate 10-14

strong 10-13

Attenuation

Attenuation

Attenuation

[db/km] 0.64 0.62 0.58 0.56 0.55 0.52 0.51 0.42 0.39 0.39 0.13

[db/km] 6.43 6.21 5.78 5.58 5.50 5.15 5.06 4.23 3.89 3.89 1.26

[db/km] 20.33 19.64 18.28 17.63 17.39 16.30 16.00 13.39 12.29 12.29 3.99

Since the requirements on Pointing, Acquisition and Tracking (PAT) increases with the link length [1], we reduce our investigations of airborne free space optical communications to short links not exceeding 3 km. Swarm UAVs flying in these short distances mean also an advantage and increase of reliability and availability for the link when precipitation occurs. We know that penetration of light through a dense fog is much more difficult than through a heavy shower [8]. TABLE II TRANSMITTER SPECIFICATIONS AND FOG ATTENUATION

Receiver Lens Diameter (mm)

100

Transmitter Lens Diameter (mm)

10

Receiver Power Sensitiv. (dBm)

-36

Optical Losses (dB)

0.5

Turbulent Attenuation (dB)

0.39

Wavelength (nm)

1550

Fog Attenuation

Kruse

V (%)

0.05

V(km)

1

λ

1550

λ0

550

q

0.585

aspec

-7.09668

The proposed fog model for short distance link is the Kruse model, where V% is the reference visibility of the transmissiometer, V(km) is the visibility, λ denotes the wavelength and λ0 the reference wavelength and q is the attenuation coefficient of the Kruse model. In visible and near IR up to about 2.5 µm, this formula relates

attenuation to visibility V (in km) for a given wavelength λ (in nm). γ (λ ) ≈ β a (λ ) =

ln (τ TH ) ⎛ λ ⎞ ⎜ ⎟ V ⎜⎝ 550 nm ⎟⎠

−q

=

3.912 ⎛ λ ⎞ ⎜ ⎟ V ⎜⎝ 550 nm ⎟⎠

−q

(3)

The attenuation adB over the link path distance dLINK in km can then be calculated from the measured transmission τ or the extinction coefficient γ(λ) (in km-1) according to 10 ⎛1⎞ a dB = 10 log ⎜ ⎟ = γ (λ ) d LINK ⎝ τ ⎠ ln(10)

(4)

Additional results can be found in [8]. Our simulation results [1] for 1 km distance between two UAVs and a divergence angle of 50 mrad give a system power of 11 mW for clear sky conditions. For moderate fog conditions 113 mW of transmitter power is needed. For 2 km distance the required power increases to 44 mW for clear sky and 4.6 W for foggy weather conditions. Scintillation losses are highly dependent on the Cn2 parameter, however, the attenuation due to scintillation is a very small factor at weak Cn2 index compared to a fog event or geometrical losses. Table I shows a very high attenuation in the region of 12 – 20 dB for turbulent air with a strong Cn2 index. However, these calculations were only done for a single transmitter unit. Hence, for a multibeam arrangement as it is used in terrestrial systems, the turbulent cells have even a smaller impact on the propagation path. By far more critical is the impact of the weather situation. Temporary attenuations of more than 300 db/km in heavy fog have already been reported. The situation can be explained by a theory describing the interaction of electromagnetic waves and particles, which can be simplified depending on the ratio of particle size and wavelength, to geometrical optics for interaction with comparatively large particles causing wavelengthindependent absorption (for optical wavelengths), and to Rayleigh-scattering for comparatively small particles. If the particle size is of the same order as the wavelength, then we have to consider the scattering theory of Mie. The most critical condition appears when both are approximately in the same order, such as optical wavelengths in the order of 1 µm and haze or fog that consists of water droplets at diameters from about 1 to 15 µm [9]. B. Geometrical Issues on Beam Spreading For a mobile high bandwidth use of FSO technology on mobile platforms, optical wireless has to fight with the drawback of line of sight between the stations. To overcome the standard solution of using very high priced tracking turrets, we investigated the use of non adjustable laser mounts on UAV. Two approaches will be presented including the use of spherical emitters and highly directional emitters with large beam divergence. Link budgets were made for both approaches for different spot diameters. The required optical power is calculated based on assumptions of receiver sensitivity, BER and datarate. Omnidirectional beam spreading could be achieved by using spherical FSO nodes and multiple transmitters. Therefore transceivers have to be tessellated around a sphere in a specific angle to cover a wide beam area. The

advantage of such a spherical arrangement is its ease of use. No high speed steering turrets are necessary and the light beam is spread over all directions.

Fig. 4. Sperical FSO Mount

This is basically not possible with commercial tracking systems because they are always designed for a single transmitting channel. Additional problems appear especially when UAVs should communicate among each other on a distinct channel without interference. Then it is necessary to separate each communication link by an optical CDMA scheme for example. A more sophisticated system has to be developed to separate the communication channels to allocate bandwidth among multiple users. Otherwise, adjacent beams are not allowed to overlap since it will cause interference. Several link power budgets were calculated in regard to spherical emitters, each of the transmitters having a lens diameters of 100 mm and a receiver diameter of 10 mm. The divergence angle assumed to be 50 mrad, which gives spot diameter of 50 m in 1000 m distance and approximately 100 m in 2000 m link distance. Spherical Emitter Power vs. Link Distance 1600,00

1400,00

Transmitter Power[mW]

1200,00

1000,00

Clear Sky Fog

800,00

600,00

400,00

200,00

0,00 500

550

600

650

700

750

800

850

900

950

1000

1050

1100

1150

Link Distance [m]

Fig. 5. Required transmitter power for various link distances

Fig 5 shows a significant increase of required transmitter power for link distances over 900 m for foggy weather conditions. This can be explained by generating more laser power to overcome the fog attenuation on each transmitting channel. At clear sky conditions, the atmospheric attenuation on a short distance link is small and the required laser power in Fig 5 shows a smooth slope. Thus, additional transmitter sections introduce a quite large multiplication factor in foggy weather conditions. So far, only communication links between UAVs have been considered. These solutions have been regarded as low cost FSO system, since no tracking and acquisition

will be used to transmit data. Low cost systems are categorized as Type-1 and Type-2 systems and have high reliability and availability in the last mile area [11]. In order to analyze data in real time, at least one UAV needs an additional backup or down-link to a ground station. Such systems are very sophisticated solutions (Type-3), primarily used in satellite communication systems, including strategies to extend the distance for reliable operation as much as possible. The beam divergence has values less than 1 mrad leading to low geometrical losses, concentrating all incoming light on a small area around the receiver. Automatic tracking of the beam is implemented and adaptive optics may help to compensate atmospheric fluctuations or allow higher optical output over a large area. The lack of continuous line of sight to a terrestrial remote platform for retrieving data requires the need for alternative solutions. Enhanced data harvesting technology, where UAVs visit fixed waypoints to rapidly and reliably deploy data to the base station, becomes an increasingly important aspect. Results for laser-based downlinks and long range terrestrial links can be found in [10]. IV ADVANTAGES OF FSO ON SWARM UAV FSO technology has a variety of tactical advantages over RF based communication systems. One of the benefits is that the enemy cannot detect microwave activities, if the communication is based on FSO. It is necessary to directly interrupt the beam in order to access information, and this is both exceedingly difficult to achieve and easily detectable. Additionally it causes no interference with nearby RF sources. The other solution, only recording data and not sending it, can result in losses of information, if the plane was destroyed by the enemy’s air defense. With FSO information wouldn’t get lost till the losing of the vehicle. However, it is a precondition that a line of sight always exists. UAVs can be used for both military and civil applications. The main scenarios for the use in military are to monitor and sense the battlefield, to find out about the military goals behind the enemy’s lines, to keep peace in critical regions, to support forces, carrying weapons or other useable things for supply. UAVs can also be used for civil purposes. Applications like observing important buildings like airports, pipelines (oil companies in Russia already use UAVs to protect important pipelines), bridges, dams or observing the traffic would be conceivable. Especially in times of high pollution because of cars and factories, air-observing-systems are highly recommended. With special airplanes for delivering of meteorological data, prognostics can become better and surer.

monitored by the enemy. Communication devices are available for first intercepting RF signals and then identifying, manipulating or destroying the enemy’s facilities. REFERENCES [1] E. Leitgeb, Ch. Chlestil, A. Friedl, K. Zettl, S.S. Muhammad, „Feasibility Study: UAVs”, Study 2005, TU-Graz/EADS.

Fig. 6. UAV swarm exchanging data with surveillance plane

Fig 6 shows an interaction scenario between an UAV swarm and an airborne surveillance station. The UAVs are operating in a low altitude level and the high altitude airplane is receiving information via optical links. Other major tasks could be defined e.g. in monitoring traffic situations, in identifying and tracking individual vehicles, in identifying episodic behavior of vehicles, in providing assistance to emergency services, in serving as a mobile sensory platform with real-time information gathering and processing capabilities and more. Another disadvantage when considering RF Technology is that the enemy may use devices that will monitor or jam enemy tactical radio frequency and radar transmissions. Such communication devices are available to first intercept, identify, and locate tactical communications and then manipulate or destroy the enemy’s facilities. But not only in the presence of jamming signals by the enemy, RF technology has some drawbacks concerning interferences and distortions in the signal path. A source antenna radiates RF energy in more than one definite direction. Multipath propagation occurs when RF signals take different paths from a source to a destination. As a result, part of the signal encounters delay and travels a longer path to the destination; distortions of the desired waveform arise and affect the decoding capability and data rate of the receiver. CONCLUSIONS This paper addressed the investigation of FSO communication links on swarm UAVs. We considered only short links between UAVs since costs for the technical realization of high sophisticated optical systems would otherwise raise dramatically. The impacts of turbulent motion of air in the atmosphere on the propagation path of laser beam is negligible compared to foggy events. For short links especially the broadening of the beam and a omnidirectional beam arrangement is an interesting alternative solution to expensive and heavy tracking systems. Additionally, omnidirectional multibeam systems installed on UAVs have the capability to establish a variety of network scenarios for enhanced reliability and availability. FSO installed on UAVs have a number of advantages; working in a “silent mode” where no RF is emitted and providing low detectability of the light beam by the enemy are by far the most interesting aspects. RF technology has the disadvantage of being jammed or

[2] Mehdi Alighanbari; How, J.P.; „Cooperative task assignment of unmanned aerial vehicles in adversarial environments” American Control Conference, 2005. Proceedings of the 2005, 8-10 June 2005 Page(s):4661 - 4666 vol. 7 [3] Zhiqiang Wu; Kumar, H.; Davari, A., „Performance evaluation of OFDM transmission in UAV wireless communication” System Theory, 2005. SSST '05. Proceedings of the Thirty-Seventh Southeastern Symposium on, 20-22 March 2005 Page(s):6 – 10 [4] Department of Defence, „Unmanned Aircraft Systems Roadmap 2005 -2030”, DOD 2005 [5] L. C. Andrews, R. L. Phillips, „Laser Beam Propagation through Random Media“, SPIE, 1998 [6] H. Willebrand, B.S. Ghuman, „Free-Space Optics: Enabling Optical Connectivity in Today's Networks” December 21, 2001, Sams [7] S. Sheikh Muhammad, E. Leitgeb, P. Kohldorfer, „Channel modeling for free space optical links” Proc. of ICTON 05, Barcelona, Spain, 2005 [8] M. Gebhart, E. Leitgeb, S. Sheikh Muhammad, B. Flecker et al., „Measurement of light attenuation in dense fog conditions for FSO applications” Proc. of SPIE symposium on Optics and Photonics, San Diego, California, USA, 2005 [9] E. Leitgeb, M. Gebhart et al., „Impact of atmospheric effects in freespace optics transmission sytems,“ Proc. of SPIE, Vol. 4976-28, Jan. 2003, San Jose, USA [10] F. David, D. Giggenbach et al., „Preliminary results of a 61 km Ground-to-Ground Optical IM/DD Data Transmission Experiment“ Proc. of SPIE, Vol. 4635 (2002) [11] E. Leitgeb et al., „Free Space Optics – Extension to Fibre-Networks for the Last Mile” Proc. of the 15th Annual IEEE / LEOS-Meeting Nov. 2002, Glasgow