Path Loss Simulation of an Infrared Optical Wireless System for Aircraft Svilen Dimitrov∗‡ , Raed Mesleh∗ , Harald Haas†∗ , Mario Cappitelli‡ , Michael Olbert‡ and Erhard Bassow§ ∗ Jacobs University Bremen, Campus Ring 1, 28759 Bremen, Germany, e-mail: s.dimitrov &
[email protected] ‡ EADS Deutschland GmbH, Nesspriel 1, 21129 Hamburg, Germany, e-mail: mario.cappitelli &
[email protected] † Institute for Digital Communications, The University of Edinburgh, Edinburgh EH9 3JL, UK, e-mail:
[email protected] § Airbus Deutschland GmbH, L¨ uneburger Schanze 30, 21614 Hamburg, Germany, e-mail:
[email protected]
Abstract—In this paper, the infrared optical wireless path loss inside an aircraft cabin is estimated. To this end, a Monte Carlo ray-tracing (MCRT) simulation is performed in a geometric computer-aided design (CAD) cabin model. Position, azimuth (AZ), elevation (EL) and field of view (FOV) properties of transmitters and receivers are defined. Key path loss parameters such as the path loss exponent and the standard deviation of the shadowing component are determined for particular line-ofsight (LOS) and non-line-of-sight (NLOS) cases. A LOS path loss exponent of 1.92 and a shadowing standard deviation of 0.81dB are obtained. In NLOS conditions, however, the path loss exponent varies depending on the nature of the particular NLOS case considered. The presented scenarios yield NLOS path loss exponent values of 2.26 and 1.28, and shadowing standard deviation values of 1.27dB and 0.7dB, respectively. Index Terms—Path loss, shadowing, LOS, NLOS, infrared wireless, optical communications, Monte Carlo ray-tracing.
I. I NTRODUCTION Power and bandwidth are limited resources in modern radio frequency (RF) wireless communication systems. Therefore, alternative or complementary wireless transmission techniques are currently being explored. With the advent of non-coherent high-power light emitting diodes (LEDs) and highly sensitive photodiodes (PDs), optical wireless communication has become a viable candidate for medium range data transmission [1, 2]. Compared to RF, optical wireless offers an almost unlimited bandwidth, license-free operation, low-cost front ends and expected high data rates. In addition, it is free of any health concerns as long as eye safety regulations are fulfilled [3]. Since there is no interference with RF-based technology, an optical wireless system can be deployed in an aircraft cabin. As a result, cabling and weight are reduced and a variety of applications are provided ranging from onboard inter-system communication to flight entertainment. Optical wireless transmission using on-off keying (OOK) or pulse position modulation (PPM) is shown to achieve high data rates in LOS communication scenarios [2, 4, 5]. However, due to the diffuse multipath nature of the channel and the expected blockage of LOS links because of movement inside an aircraft cabin, alternative modulation techniques are needed which exhibit an inherent robustness to multipath fading [6]. A well known such technique is orthogonal frequency division multiplexing (OFDM) [1, 7, 8]. The inherent high peak-toaverage-power ratio (PAPR) in OFDM which is a significant problem in OFDM-based RF systems is actually beneficial for
optical wireless communication when using direct detection (DD). It is constructively exploited in a way that the OFDM signal envelope variations are utilized to modulate the intensity of the LEDs [7]. Still, the development of an OFDM-based infrared wireless system requires a thorough characterization of the communication channel. On the one hand, the estimation of the channel impulse response allows for the optimization of the guard interval and the number of subcarriers/OFDM symbol duration. On the other hand, the path loss estimation in the setup facilitates the design of the radiation/detection properties of the transmitter and receiver units. There have been several studies of the infrared wireless channel, based on direct measurements [6, 9, 10] or ray-tracing simulations [11–14]. It is shown in [9] that non-directed channels can be well characterized by their path loss and delay spread. Indoor measurement and simulation results [6, 11–14] demonstrate a channel delay spread in the order of 30-50ns. A measurement conducted in [9] shows that the optical path loss ranges between 50dB and 80dB. Further measurements in [10] suggest that the path loss variation along the path is smooth and a curve-fitting algorithm can be used to interpolate the intermediate values. A radiosity simulation described in [14] determines the path loss for a particular transmitter and receiver placement in a room. However, to the best of the authors’ knowledge, there does not exist a model of the optical wireless path loss as a function of distance for typical indoor environments including shadowing. In this paper, a MCRT simulative approach is presented which allows for the establishment of a statistical path loss model which describes the light power distribution in an aircraft cabin along chosen paths with predefined light sources. Particular scenarios of LOS and NLOS cases are studied for a combination of AZ, EL and FOV properties of selected offthe-shelf components. The rest of the paper is organized as follows. Section II presents the simulation model and required inputs. Section III explains the methodology. Simulation results are discussed in Section IV. Finally, Section V concludes the paper. II. S IMULATION MODEL The infrared irradiation simulation is performed with the software tool Specter by Integra Inc [15]. This software utilizes a MCRT algorithm as a primary lighting simulation
978-1-4244-4148-8/09/$25.00 ©2009 This full text paper was peer reviewed at the direction of IEEE Communications Society subject matter experts for publication in the IEEE "GLOBECOM" 2009 proceedings.
(a) Front view.
(b) Perspective view, includes relative directions.
Fig. 1: Generalized model of the aircraft cabin. NLOS path loss is estimated along paths 1 and 2. LOS path loss is estimated along paths 3 and 4. Three transmitter positions and a simplified part of the cuboid array of receivers are depicted. The dimensions of the setup and the relative directions ”front”, ”back”, ”up”, ”down”, ”left” and ”right” are used for orientation in the cabin throughout the paper.
method. The following inputs are required: geometric model of the simulation setup, materials reflection characteristic, and definition of key properties of light sources and observers. A. Setup geometry and materials The geometric 3D model of the aircraft cabin is constructed with the CAD tool Rhinoceros 3D [16]. A generalized model of the setup is presented on Fig. 1. The reflection characteristics of the materials on the different objects in the cabin are defined through diffuse and specular reflection coefficients. These are obtained by material measurements over a sweep of wavelengths in the infrared spectrum until 1100nm and exhibit a high diffuse reflection with a small specular portion. Since the chosen commercially available infrared emitter has a very narrow light spectrum at 870nm, the focus of the simulation is on monochromatic light at this wavelength. B. Definition of transmitters and receivers The simulation focuses not only on LOS but also on NLOS path loss estimation. Therefore, an omnidirectional radiation pattern of the transmitter unit is designed to provide a high-power multipath signal component. Each transmitter unit consists of 16 point light sources as shown on Fig. 2a. Each light source is modeled according to the specifications of the Vishay infrared emitter TSFF 5210 [17] with FOV of ±10o and an optical power of 50mW. As a result, the transmitter unit has a total optical power of 800mW. According to the BS EN 62471:2008 standard for photobiological safety of lamps and lamp systems [3], non-coherent diffuse continuous-wavemodulated LEDs belong to the exempt group classification and pose no photobiological hazard for the hyman eye, if the irradiance does not exceed 100W/m2 at a distance of 0.2m
(a) Radiation pattern of the transmitter unit. 16 LEDs with FOV of ±10o are placed around a circle with diameter of 2cm to form an omnidirectional radiation pattern.
(b) Detection pattern of the receiver unit. 6 PDs with FOV of ±45o are arranged in a cubic formation to cover all receive directions.
Fig. 2: Transmitter and receiver units.
from the optical source in the direction of maximal directivity within 1000s. Since the respective irradiance of the designed transmitter unit at the specified distance results in 27W/m2 , the eye safety regulations are fulfilled. Three transmitter units are positioned in the setup as shown on Fig. 1. To estimate the path loss along predefined paths, the irradiance in the setup volume is sampled by means of a planar irradiance observer which is a collection of receiver points. Each receiver point in the setup is modeled according to the specifications of the Vishay PD TESP 5700 [18] with a FOV of ±45o and a radiant sensitive area of 5.7mm2 . Six PDs are placed together to form a single receiver unit as shown on Fig. 2b. It covers all six receive directions and allows for estimation of the path loss in a particular direction. With the help of equidistantly separated irradiance observer planes a three
978-1-4244-4148-8/09/$25.00 ©2009 This full text paper was peer reviewed at the direction of IEEE Communications Society subject matter experts for publication in the IEEE "GLOBECOM" 2009 proceedings.
dimensional array of 200×200×100 receiver units is defined throughout the entire aircraft cabin. The receiver units are separated by approximately 3cm along the respective ”frontback”, ”left-right” and ”up-down” dimensions. A simplified part of this cuboid array is presented on Fig. 1. The limitations of the simulation are related to the MCRT simulation stopping criterion of stochastic confidence [15] and the definition of the simulation input parameters [19], such as tessellation accuracy of the CAD cabin model, measurement tolerance of the materials reflection coefficients, and modeling accuracy of the radiation/detection patterns of the LEDs and PDs. The summation of all the above mentioned error sources actually defines the total error of the MCRT simulation. III. M ETHODOLOGY To estimate the infrared wireless channel path loss under consideration of the transmitter and receiver characteristics, the generic model for RF path loss [20] is assumed. Thus, the path loss exponent and the standard deviation of the shadowing component are fitted to the following equation: P T L(d)[dB] = 10 log10 PR (d) (1) = L(d0 ) + 10n log10 d + Xσ , d0 where L(d0 )[dB] is the path loss at reference distance d0 [m] which accounts for system losses, n is the path loss exponent, d[m] is the distance between transmitter and receiver, and Xσ [dB] is a log-normally distributed random variable with zero mean and standard deviation σ that accounts for shadowing effects. The optical power radiated from the transmitter and the optical power impinging on the receiver are denoted by PT [W] and PR [W], respectively. Given a particular communication system in an environment with high diffuse reflection, it is expected to encounter LOS paths and NLOS paths with short range single reflection or NLOS paths with long range multiple reflections due to obstruction. Thus, these distinctive scenarios of the infrared wireless signal propagation are the main target of the MCRT irradiation simulation. The lines along which the path loss is estimated are shown on Fig. 1. Paths 3 and 4 are chosen for LOS path loss estimation. Path 3 represents a LOS condition with FOV alignment between the operational transmitter 2 and the receivers along the path, enhanced by a single reflection component coming from the hatrack in proximity. In the case of transmitter 3, the LOS path 4 collects, in addition, a second single reflection component coming from the sidewall. Therefore, the path loss exponent along the two paths is estimated when only the signal arriving at the receiver units in the ”back” receive direction is counted as shown on Fig. 3a. The same receive direction is considered for LOS shadowing component estimation. Along paths 3 and 4, receiver units within the FOV of the respective transmitter are selected from circles which are orthogonal to and centered on the respective path. A simplified illustration is shown on Fig. 3b. The selected receiver units on each circle
(a) LOS path loss exponent estimation along paths 3 and 4. Only the LOS signal falling within the FOV of the ”back” receivers is counted.
(b) LOS shadowing estimation along paths 3 and 4. Receivers within the FOV of the transmitter and at distances from 1.2m to 5.6m away from it are selected from circles around the path. Some of the circles are illustrated.
Fig. 3: LOS path loss estimation scenarios.
are approximately equidistant to the transmitter and have the same FOV orientation towards it. The circles on path 4 have a radius of 0.6m in order to include the receivers which are closest to the wall and experience the highest shadowing. For the sake of consistency, the same radius is chosen for the circles around path 3. To minimize the effect of the FOV on the LOS path loss, receivers located at distances greater than 1.2m are considered in the estimation of the LOS shadowing. The NLOS path loss is estimated along paths 1 and 2. Path 1 represents a NLOS condition due to azimuthal misalignment of the operational transmitter 1 and the receivers along the path which is enhanced by the short range single reflection component coming from the sidewall. Consequently, to estimate the path loss exponent only the signal arriving at the receiver units along the path in the ”right” receive direction
978-1-4244-4148-8/09/$25.00 ©2009 This full text paper was peer reviewed at the direction of IEEE Communications Society subject matter experts for publication in the IEEE "GLOBECOM" 2009 proceedings.
receiver in the specified directions are selected from concentric spheres around the position of the operational transmitter 1. Furthermore, only those receiver units on the spheres are counted which are located in the small cuboid volume in the neighborhood of path 2. IV. S IMULATION RESULTS AND DISCUSSION The infrared signal distribution in the setup is simulated up to a MCRT stochastic confidence of 3%. The output for irradiance in mW/m2 is multiplied by the photosensitive area of the PD to obtain the received power, PR , at every point in the mesh grid. Using eqn. (1) with transmit power, PT , of 800mW, the path loss values in dB are obtained. A. LOS path loss estimation (a) Path loss exponent estimation along paths 1 and 2. Only the signal reflected by the sidewall is counted along path 1. The receivers along the path 2 sum the multiple reflections of the signal in the ”front”, ”back”, ”left” and ”right” receive directions.
(b) NLOS shadowing estimation along path 1. Receivers are selected from concentric circles around the projection of transmitter 1 onto path 1 in a vertical plane. They are located at distances from 0.4m to 5.6m away from the transmitter. Some of the circles are illustrated.
Fig. 4: NLOS path loss estimation scenarios.
is considered as shown on Fig. 4a. This receive direction is chosen when estimating the NLOS shadowing along the path. Thus, receiver units on concentric circles around the projected position of the transmitter onto the vertical plane along the path are selected as shown on Fig. 4b. The NLOS path loss scenario along path 2 studies the impact of obstruction, as well as transmitter and receiver misalignments in EL as opposed to AZ (as in all previous cases). Thus, in order to estimate the path loss exponent, the receiver units along the path collect the signal that impinges in the ”back”, ”front”, ”left” and ”right” receive directions as illustrated on Fig. 4a. To estimate the shadowing component,
Since transmitters 2 and 3 are positioned on paths 3 and 4, respectively, the LOS path loss is estimated for distances from 0.04m to 5.6m away from the respective operational transmitter. The two resulting path loss graphs are presented on Fig. 5. As expected, the path loss exhibits a linear behavior in log-domain. On Figs. 5a and 5b the exponent from eqn. (1) is the slope of the path loss graph, obtained after linear regression. Therefore, on paths 3 and 4 the path loss exponent amounts to 1.94 and 1.92, respectively. Hence, the single reflection component coming from the wall along path 4 does not contribute significantly, and the signal power is dominated by the strong LOS component. According to the LOS shadowing estimation methodology, described in section III, the two resulting path loss scatter plots are presented on Figs. 5a and 5b. Following the generally accepted assumption that shadowing effects can be modelled with a log-normal distribution, the standard deviation of the shadowing component around paths 3 and 4 is evaluated to 0.57dB and 0.81dB, respectively. On the two plots, the path loss points are more scattered around the mean curve at shorter distances. This is because with the increase of distance the single reflection from the near objects becomes much less intensive than the LOS component and a better FOV alignment is achieved between the transmitter and the receivers along the path. Since path 4 exhibits a higher shadowing because of the sidewall in proximity as compared to path 3, it is used to develop the model for LOS path loss in the aircraft cabin. Assuming that the shadowing is present along the whole path, the LOS path loss equation as a function of separation distance d is given as follows: d + Xσ , (2) L(d)[dB] = 26.99 + 19.2 log10 0.04 where σ ≈ 0.81dB and d ≥ 0.04m. B. NLOS path loss estimation Because of the horizontal displacement between the operational transmitter 1 and path 1, the NLOS path loss is estimated for distances from 0.4m to 5.61m away from the transmitter. The resulting path loss graph is presented on Fig. 6a and it exhibits a path loss exponent of 2.26. The scattering of the NLOS shadowing component along path 1 is displayed on
978-1-4244-4148-8/09/$25.00 ©2009 This full text paper was peer reviewed at the direction of IEEE Communications Society subject matter experts for publication in the IEEE "GLOBECOM" 2009 proceedings.
(a) Path loss exponent and shadowing estimation along path 3.
(a) Path loss exponent and shadowing estimation along path 1.
(b) Path loss exponent and shadowing estimation along path 4.
(b) Path loss exponent and shadowing estimation along path 2.
Fig. 5: LOS path loss estimation.
Fig. 6: NLOS path loss estimation.
978-1-4244-4148-8/09/$25.00 ©2009 This full text paper was peer reviewed at the direction of IEEE Communications Society subject matter experts for publication in the IEEE "GLOBECOM" 2009 proceedings.
Fig. 6a, as well. It has a standard deviation of 1.27dB around the linear regression mean curve. Therefore, the NLOS path loss equation along path 1 yields: d + Xσ , (3) L(d)[dB] = 53.61 + 22.6 log10 0.4 where σ ≈ 1.27dB and d ≥ 0.4m. Since the operational transmitter 1 and path 2 are displaced horizontally and vertically, the NLOS path loss is estimated for distances from 1.45m to 5.78m away from the transmitter. The path loss exponent on Fig. 6b amounts to 1.28. The scattering of the NLOS shadowing component along path 2 is also given on Fig. 6b. It shows a standard deviation of 0.7dB. Thus, the NLOS path loss as a function of transmitter-receiver separation distance along path 2 is determined as follows: d + Xσ , (4) L(d)[dB] = 64.22 + 12.8 log10 1.45 where σ ≈ 0.7dB and d ≥ 1.45m. The two presented NLOS communication scenarios show a difference in the resulting path loss exponents. Consequently, the short range single reflection scenario provides lower path loss at shorter distances. However, because of the high diffuse reflectivity of the materials in the aircraft cabin, the situation is the opposite at longer distances, where the long range multiple reflections scenario demonstrates a lower path loss. V. C ONCLUSION In this paper, a comprehensive MCRT irradiation simulation has been performed to determine the path loss of infrared wireless communication inside an aircraft cabin. Measured reflections coefficients of the materials in the cabin are used for simulation modeling. This approach allows for the characterization of propagation paths of interest, without having to resort to expensive and time consuming measurements. Specifically, the path loss exponent and the standard deviation of shadowing have been determined for particular LOS and NLOS scenarios. The obtained path loss parameters can be used to establish a realistic link budget, i.e., to determine the number of LEDs in a transmitter unit and PDs in a receiver unit, and to design a cellular optical system. The results show that NLOS in an aircraft cabin degrades the performance by approximately 10dB as compared to LOS. Nonetheless, the level of degradation will still enable communication links in NLOS scenarios, if proper link budget margins are considered. Also, the path loss fluctuations due to shadowing are not significant and robust communication links can be expected. ACKNOWLEDGMENT We gratefully acknowledge the support for this work from EADS Germany and Airbus Germany. In addition, we acknowledge the support from the German Federal Ministry of Economics and Technology (BMWi) under grant 20K0806G as part of the SINTEG project (”Gef¨ordert vom Bundesministerium fur Wirtschaft und Technologie aufgrund eines Beschlusses des Deutschen Bundestages”).
R EFERENCES [1] Y. Tanaka, T. Komine, S. Haruyama, and M. Nakagawa, “Indoor Visible Communication Utilizing Plural White LEDs as Lighting,” in Proc. of the 12th IEEE International Symposium on Personal, Indoor and Mobile Radio Communications, vol. 2, San Diego, CA, USA, Sep. 30 – Oct. 3, 2001, pp. 81–85. [2] T. Komine and M. Nakagawa, “Fundamental Analysis for Visible– Light Communication System using LED Lights,” IEEE Transactions on Consumer Electronics, vol. 50, no. 1, pp. 100–107, Feb. 2004. [3] BS EN 62471:2008, Photobiological Safety of Lamps and Lamp Systems, BSI British Standards Std., Sep. 2008. [4] M. D. Audeh, J. M. Kahn, and J. R. Barry, “Performance of Pulseposition Modulation on Measured Non-directed Indoor Infrared Channels,” IEEE Transactions on Communications, vol. 44, no. 6, pp. 654– 659, June 1996. [5] M. Akbulut, C. Chen, M. Hargis, A. Weiner, M. Melloch, and J. Woodall, “Digital Communications Above 1 Gb/s Using 890-nm Surface-Emitting Light-Emitting Diodes,” IEEE Photonics Technology Letters, vol. 13, no. 1, pp. 85–87, Jan. 2001. [6] J. Kahn, J. Barry, W. Krause, M. Audeh, J. Carruthers, G. Marsh, E. Lee, and D. Messerschmitt, “High–Speed Non–Directional Infrared Communication for Wireless Local-Area Networks,” in Proc. of the 26th Asilomar Conference on Signals, Systems and Computers, vol. 1, California, USA, Oct. 26–28, 1992, pp. 83–87. [7] M. Z. Afgani, H. Haas, H. Elgala, and D. Knipp, “Visible Light Communication Using OFDM,” in Proc. of the 2nd International Conference on Testbeds and Research Infrastructures for the Development of Networks and Communities (TRIDENTCOM), Barcelona, Spain, Mar. 1–3, 2006, pp. 129–134. [8] H. Elgala, R. Mesleh, H. Haas, and B. Pricope, “OFDM Visible Light Wireless Communication Based on White LEDs,” in Proc. of the 64th IEEE Vehicular Technology Conference (VTC), Dublin, Ireland, Apr. 22–25, 2007. [9] J. B. Carruthers and J. M. Kahn, “Modeling of Nondirected Wireless Infrared Channels,” in Proc. of the IEEE Conference on Communications.: Converging Technologies for Tomorrow’s Applications, vol. 2, Dallas, TX, USA, Jun. 23–27, 1996, pp. 1227–1231. [10] M. R. Pakravan, M. Kavehrad, and H. Hashemi, “Indoor Wireless Infrared Channel Characterization by Measurements,” IEEE Transactions on Vehicular Technology, vol. 50, no. 4, pp. 1053–1073, Jul. 2001. [11] J. Barry, J. Kahn, W. Krause, E. Lee, and D. Messerschmitt, “Simulation of Multipath Impulse Response for Indoor Wireless Optical Channels,” IEEE Journal on Selected Areas in Communications, vol. 11, no. 3, pp. 367–379, Apr. 1993. [12] V. Pohl, V. Jungnickel, and C. von Helmolt, “A Channel Model for Wireless Infrared Communication,” in Proc. 11th IEEE International Symposium on Personal, Indoor and Mobile Radio Communications PIMRC 2000, vol. 1, London, UK, Sep. 18–21, 2000, pp. 297–303. [13] S. Rodriguez, R. Perez-Jimenez, F. Lopez-Hernandez, O. Gonzalez, and A. Ayala, “Reflection Model for Calculation of the Impulse Response on IR-wireless Indoor Channels Using Ray-tracing Algorithm,” Microwave Optical Technology Letter, vol. 32, no. 4, pp. 296–300, Jan. 2002. [14] M. Bertrand, O. Bouchet, and P. Besnard, “Personal Optical Wireless Communications: LOS/WLOS/DIF Propagation Model and QOFI,” in Proc. 6th International Symposium on Communication Systems, Networks and Digital Signal Processing CSNDSP 2008, Graz, Austria, 25– 25 Jul. 2008, pp. 179–182. [15] Integra Incorporated, “Specter,” Retrieved Dec. 21, 2008 from http://www.integra.jp/en/specter/index.html. [16] Robert McNeel and Associates, “Rhinoceros 3D,” Retrieved Jan. 15, 2009 from http://www.rhino3d.com/. [17] Vishay Semiconductors, “Datasheet: TSFF5210 High Speed Infrared Emitting Diode, RoHS Compliant, 870 nm, GaAlAs Double Hetero,” Retrieved Dec. 22, 2008 from http://www.vishay.com/docs/81090/81090.pdf, Aug. 2008. [18] ——, “Datasheet: TESP5700 Silicon PIN Photodiode, RoHS Compliant,” Retrieved Dec. 22, 2008 from http://www.vishay.com/docs/81573/tesp5700.pdf, Sep. 2008. [19] M. Cappitelli, M. Olbert, C. Mussmann, and R. Greule, “GUM in Simulation und Messung,” in Proc. Licht2008 Ilmenau, Ilmenau, Germany, 10–13 Sep. 2008, pp. 487–491. [20] T. S. Rappaport, Wireless Communications: Principles and Practice, 2nd ed. Prentice Hall PTR, 2002.
978-1-4244-4148-8/09/$25.00 ©2009 This full text paper was peer reviewed at the direction of IEEE Communications Society subject matter experts for publication in the IEEE "GLOBECOM" 2009 proceedings.
