In this paper a new hybrid wireless access system incorporating high bandwidth line-of-sight free space optical wireless and radio frequency localization is proposed and demonstrated. This system is capable of supporting several gigabits/second up-stream and down-stream data transmission and ideally suited for high bandwidth indoor applications such as personal area networks. A radio frequency signal is used to achieve localization of subscribers, offering limited mobility to subscribers within a practical office scenario. Even with the modest transmitted power of 5dBm, we demonstrate satisfactory performance of bit error rates better than 10−9 over the entire room in the presence of strong background light. Using simulations, the effectiveness of the proposed system architecture is investigated and the key performance trade-offs identified. Proof-of-concept experiments have also been carried out to validate simulation model, and initial experimental results successfully demonstrate the feasibility of the system capable of supporting 2.5Gbps over a 1-2m optical wireless link (limited by the length of the sliding rail used in the experiment) with a 45 degrees diffused beam in an indoor environment for the first time.
©2010 Optical Society of America
Optical indoor wireless communications have attracted enormous attention because of the unregulated large bandwidth available, immunity to electromagnetic interference, and the possibility of frequency reuse and security at physical layer where optical beam does not penetrate walls or opaque objects [1–3]. However, there are also several major drawbacks associated with it, such as additive noise due to sunlight and artificial background light in built environments and transmission power restricted by the eye and skin safety regulations .
There are two major types of indoor optical wireless communication systems, namely direct line-of-sight (LOS) system and diffused system. The diffused beam method allows the system to operate even when barriers are placed between transmitters and receivers, and is capable of providing mobility to subscribers. However, it suffers from ineffective use of transmitted power and multipath dispersion, which causes pulse spreading and severe inter-symbol interference (ISI). These disadvantages can be overcome by direct LOS approach. It is more effective in power consumption and multipath dispersion minimization, leading to higher data rate. Its drawbacks are tight alignment requirement, immobility of the transceivers, and interruptions in transmission caused by shadowing.
Until now, most research has focused on diffused beam systems due to their capacity to easily support mobility for applications in personal area networks. Several improvements have been studied for better system performance, such as transmitter beam diversity technique , line strip multibeam transmitter systems (LSMS) [6,7], receiver diversity technique [8–10], and imaging receiver instead of non-imaging receiver method [11,12]. However, due to the inherent diffusive nature, multipath spread is unavoidable, so the potential bit rate of the diffused system is still highly limited. On the contrary, although direct LOS system has the potential to transmit data at much higher speed, little research has been done to date. Due to strict requirement of precise alignment required to maintain LOS communication, even communication systems with very limited mobility largely remained impossible.
In this paper we propose a system that combines radio frequency (RF) and optical wireless (OW) communications together [13,14]. We also propose the use of comparatively wider beam than that often used in the conventional direct LOS system. The RF signal is used for subscriber localization and optical signal is then directed to that known position of the subscriber for high-speed data transmission. Therefore LOS link is always available and subscribers can enjoy both high-speed data transmission and limited mobility within a personal area network. Furthermore, the room employing our proposed system is connected to a central office (CO) by a fiber distribution network and optical fiber ends located in the ceiling serve as optical sources. Therefore this system can be simply connected to the existing backbone networks and even higher bandwidth could easily be achieved by using wavelength division multiplexing (WDM) technology. In addition to the down-stream communication system, OW up-stream communication system is also considered. To the best of our knowledge, this is the first time an OW system for up-stream communication is proposed.
2. Room configuration
In this paper, we consider a practical office environment, which is a 10m × 8m × 3m room as shown in Fig. 1 . The receiver end of the down-stream system, either fixed or mobile, are placed on a plane that is 1m above the ground. This plane is called communication floor (CF). This room consists of eight rectangular cubicles with surfaces parallel to the room walls. The size of each cubicle is 2.5m × 3m × 1.5m. All the partitions are opaque so the signals incident on them are either absorbed or blocked. Furthermore, the cubicles are equipped with tables and chairs. It is obvious that in such scenario, shadowing due to physical obstacles result in the worst signal reception. In addition, the room is assumed to be well-illuminated by eight 100W tungsten floodlights creating an environment of worst case background light. These lamps are positioned at coordinates of (x, y, z) = (2, 2, 3), (4, 2, 3), (6, 2, 3), (8, 2, 3), (2, 5, 3), (4, 5, 3), (6, 5, 3), and (8, 5, 3). Measurements of these lamps in  show that an excellent model for them is a generalized Lambertian emitter, with a radiant intensity (W / Sr) given by
3. Down-stream system architecture
The structure of the down-stream system is explained in this section and the up-stream system configuration will be discussed later in section 5.
3.1 Overall architecture of the proposed system
In our system, we do not use a separate optical source such as a LD or a LED in each room as supposed in previous research [14–17]. Instead a fiber distribution network connecting several rooms to a CO is utilized, as shown in Fig. 2 . The optical sources, control systems and complex operations are all dealt with in the CO, so a separate complicated transmitter with control system is no longer needed in each room, leading to reduced cost and complexity. Furthermore, we combine RF and optical wireless techniques together. The high bit rate communication data is transmitted in the optical domain, and we only use RF (the low cost and widely deployed WiFi technique) signals for subscriber localization.
WiFi is a low cost and widely deployed RF technique. In addition to applications in wireless local area networks (WLANs), it is also capable of providing localization based services  and such function is currently being included in the IEEE 802.11TgV standard development . In our system, the room is divided into 20 cells, each with a size of 2m × 2m and a cell number from 1 to 20 as shown in Fig. 3 . We do not need the precise localization information of the subscriber, instead the rough position of the subscriber is sufficient. From [20–23], it has been shown that in an office scenario, a localization resolution of better than half a meter can be easily realized. Therefore it is relatively simple to achieve precision localization functionality using WiFi system in our application. As a result, in this paper we will only focus on the OW transmission part of our system.
3.2 Down-stream optical wireless system structure
In the room configuration described in section 2, shadowing effects exist in substantial positions. To overcome this shadowing problem and provide mobility to subscribers over the entire room, we use a comparatively wider beam to cover an entire cell at any one time with several optical fiber ends installed in the ceiling (four in this system), as illustrated in Fig. 3. Each of the four fiber ends is installed above the intersection of two adjacent cubicles and based on the coordinates defined in Fig. 1, these fiber ends are located at (2.5, 3, 3), (7.5, 3, 3), (2.5, 5, 3), and (7.5, 5, 3), respectively.
There are a range of distances and angles between the fiber transmitters and receivers for different subscriber locations. To accommodate for this, we propose the use of adaptive lens and steering mirrors as illustrated in Fig. 4 . With the operations described below, all the cells within the service area can be covered and beam width is always constant.
At the output plane of a single-mode fiber, the optical field distribution closely approximates a Gaussian profile . So we define three important parameters ω(z), R(z) and f representing the beam waist, the radius of curvature of the beam for any distance z, and the equivalent beam focal length, which are mathematically defined as
When a Gaussian beam transmits through a lens (adaptive lens here) with a focal length of F, the beam remains Gaussian. If the distance between the beam waist and lens is l, as shown in Fig. 5 , then after the lens, we have
Beyond the lens, the beam remains as a Gaussian beam throughout our free space link. If we set the position of the new Gaussian beam waist z = 0, then after transmitting for a distance z in free space, using Eq. (2), the waist of the beam spot is given by
From Eq. (5)–(8) and with the localization information provided by the WiFi system, the corresponding focal length F of the adaptive lens can be decided to keep the beam waist constant at all cells. Furthermore, with the WiFi localization system, the angle between the cell center and fiber end can also be calculated. As shown in Fig. 6 , the angle αx and αy, representing the angle associated with a cell with respect to the normal of the output plane of the fiber (the same as the ceiling), can be calculated as
3.3 Operation algorithm of the down-stream system
The overall operational principle of the down-stream system is as follows:
- a. Use the WiFi system to roughly locate the subscriber and feedback the cell number information;
- b. According to the position of the subscriber (cell number), decide which fiber should be used for high speed down-stream data transmission. As illustrated in Fig. 3, for cell 1, 2, 6, and 7, transmission fiber 1 is selected, for cell 4, 5, 9, and 10, transmission fiber 2 is selected, for cell 11, 12, 16,and 17, transmission fiber 3 is selected, and for cell 14, 15, 19, and 20, transmission fiber 4 is selected. For cell 3, 8, 13, 18, the situation is more complicated, as shadowing caused by cubicle partitions exist in each of the cells. However, in these cells we only need to monitor the received optical power and the direct LOS link can be always available as well. For example, when the subscriber is in cell 3, we can first assign fiber 1 for data transmission, and at the same time monitor the received optical signal power. If it is almost 0, then assign fiber 2 for data transmission;
- c. With the information of cell number and fiber chosen, calculate the distance and angle between the fiber and cell center;
- d. Change the position of the steering mirror to point at the cell center;
- e. Adjust the focal length of the lens to enable the beam to cover the entire cell;
- f. Start high speed downlink transmission;
- g. Continuously monitor the subscriber’s position through the WiFi localization system. If the subscriber is moving from one cell to another, then first decide if we need to change the fiber used for data transmission. If not, then first increase the width of the beam to cover a larger area to provide enough time for adaptive adjustment, followed by calculation of the new distance and angle, and then use the steering mirror to change the orientation of the beam to point at the center of the new cell. After finishing these, decrease the divergent angle of the beam to cover only the new cell, just as before. If we need to change the fiber utilized, then first calculate the distance and angle between the new fiber end and new cell center, adjust the steering mirror and adaptive lens in front of the new fiber end accordingly, turn on the new fiber to provide transmission link, and finally turn off the previous one.
4. Simulation results of down-stream system and discussions
4.1 Noise analysis
In our proposed system, since we use direct LOS link instead of diffused beam, multipath spread is no longer a problem. So to reduce system complexity and cost, we choose the simplest non-imaging single compound parabolic concentrator (CPC)  in front of the receiver. The CPC with a wide FOV is chosen here. Such a CPC will collect more background light and consequently degrade the system performance. However, due to the size of the room considered, at room corners the signal light will enter the CPC with a large incident angle. Therefore such a CPC is necessary to enable the mobility of subscriber over the entire room. In addition, it is also possible to further improve the system performance by using more advanced receiver and the idea of adaptive receiver is extremely interesting. The adaptive receiver just employs a CPC with a small FOV to reject strong background light. To achieve ubiquitous coverage inside the room, the orientation of the receiver can be changed according to localization information from the WiFi system or even simply with the received intensity footprint. Such a receiver will be studied further in the future.
The dominant noise in our system is shot noise induced by the ambient light and the receiver preamplifier induced noise [26–29]. In our system, on-off-keying (OOK) modulation scheme is employed since it is the simplest and most mature technology for OW systems. Therefore the noise variance σ02 and σ12 associated with the transmitted signal ‘0’ and ‘1’ respectively are given by
The preamplifier used in our system is a p-i-n FET transimpedance receiver proposed in . The principle noise sources in this preamplifier are Johnson noise associated with the FET channel conductance, Johnson noise from the load or feedback resistor, shot noise arising from gate leakage current and 1/f noise. The preamplifier shot noise variance is given by
Furthermore, the background light induced shot noise can be calculated by
To minimize the received background light power, an optical bandpass filter is employed in front of the photodetector. The received background light power is illustrated in Fig. 7 . It is obvious that the maximum received background light power occurs at positions (2, 2, 1), (4, 2, 1), (6, 2, 1), (8, 2, 1), (2, 5, 1), (4, 5, 1), (6, 5, 1), and (8, 5, 1) because these positions are directly under a lamp. In other positions, the received background light power is much smaller (< 1.5µW (−28.2dBm)).
Figure 8 illustrates the noise variance in our proposed system over the entire room. Similarly, the peak of the noise variance also occurs in the positions directly under a lamp. It can also be seen from Fig. 8 that the noise variance varies severely in different locations. However, as we will see later, the peak of the noise variance occurs approximately at the same position as the peak of received optical signal power, leading to a steadier system performance over the entire room.
4.2 Multipath dispersion analysis
In wireless communication systems multipath dispersion induces severe ISI, leading to degraded system performance. However, as will be shown later, multipath dispersion is not a limiting factor in our system.
In our proposed system, only the reflected light with an incident angle smaller than the FOV of CPC results in multipath dispersion. This can be characterized by the impulse response h(t) of the channel 1]. Only the first order and second order reflections are considered because the third order reflections and higher have minimal impact on the system. The reflectivities of the wall, the floor and the ceiling are 0.8, 0.3 and 0.8, respectively  and the simulated impulse response (1ns impulse) when the subscriber is in one of the room corner (0, 0, 1) is illustrated in Fig. 9 . It is clear that the strength of the signal reflected back and captured by the CPC is more than 60dB lower than that of the signal from the direct LOS channel. Therefore multipath dispersion in our system is so weak that can be safely discarded.
4.3 Signal to noise ratio (SNR) and bit-error-rate (BER) performances
SNR is a significant parameter for system performance indication. It is defined as 
Utilizing Eq. (13)–(20), when the transmitted power at the fiber output plane is 5mW (7dBm) and the bit rate is 1Gbps, the SNR and BER performances of the system over the entire room are shown in Fig. 10 . Here we fix the beam waist at the receiver end at 2m with the help of an adaptive lens in the fiber transmitter. From Fig. 10(a), at (2, 2, 1), (4, 2, 1), (6, 2, 1), (8, 2, 1), (2, 5, 1), (4, 5, 1), (6, 5, 1), and (8, 5, 1), the SNR is lower than the surrounding area. This is because of lamps located just above each of these positions, inducing stronger background noise. However, the SNR difference over the entire room is not large (within 6dB) since the fiber transmitters are also located near these positions as well. In addition, the system performance is asymmetrical, resulting from the fact that the locations of the lamps are not symmetrical (along y = 2m and y = 5m line). Consequently in the vicinity of y = 8m, the background light induced shot noise is much lower resulting in a relatively higher received SNR. From Fig. 10(b), it is clear that a BER<1.2 × 10−8 can be achieved over the entire room and a BER<10−9 performance is achieved in most parts of the room except for cases directly under a lamp. Therefore this indicates that satisfying performance and mobility can be provided over the entire room.
Figure 11 shows the performance of our proposed system when the bit rate increases to 2.5Gbps. The transmitted power at the output plane of the fiber remains at 5mW (7dBm). It is obvious that a SNR>10dB and a BER<1.5 × 10−4 can be achieved in most parts of the room except for areas directly under the lamps. Furthermore, similar to Fig. 10, the SNR and BER are relatively constant over the entire room. As a result, error-free data transmission with mobility can be achieved over the entire room.
Figure 12 shows the performance of our proposed system plotted as a function of bit rates and the minimum transmitted power needed to obtain a BER<10−9. It can be seen that for 1Gbps system, the power needed is 6.6mW (8.2dBm), and even when the bit rate increases to 5Gbps, a transmitted power of 14.8mW (11.7dBm) is sufficient for achieving satisfying performances over the entire room.
5. Up-stream system structure and simulation results
The basic structure of the up-stream system is the same as that of the down-stream system described above. Adaptive lens and steering mirror are also employed in front of the up-stream transmitter, as shown in Fig. 13 . According to the localization information of the subscriber provided by the WiFi system, focal length of the adaptive lens and orientation of the steering mirror are adjusted to roughly point at the fiber ends at the ceiling. The optical signal is then coupled into the fiber directly and transmitted back to the CO for processing. In Fig. 14 the transceiver structure at the ceiling is illustrated. The concentrator and lens are used to collect up-stream signal light and couple it into the fiber.
The noise in the up-stream system is slightly different from that in the down-stream system. The background light in the up-stream system does not come from the lamps directly, but comes from the reflections of walls and ground. As a result, in the simulation we divide the walls and ground into considerable reflection elements each with a size of 5cm × 5cm, calculate their contribution to the noise independently (each element can be seen as a Lambertain source), and finally sum them up to obtain the background light induced noise.
The transmitter of the up-stream system is closer to the user, so the possible transmitted power is highly limited by the laser eye and skin safety regulations. In our up-stream system, a wavelength (different from the down-stream signal wavelength) in the 1550nm band is used, so the highest possible transmitted power is 7mW (8.45dBm) .
Suppose the subscriber is in cell 1, which is the worst case scenario since it is located at one of the room corners (the farthest from the fiber ends). In our system, wherever the subscriber is in this cell, the localization information available is the same - the angle and distance between fiber end and cell center. So when the subscriber is in one of the cell corners the largest possible distance from the up-stream beam center to the fiber end is 1.4m. Consider the situation where the subscriber is initially at one of the cell corners and walk along the diagonal of the cell to another corner. When the up-stream beam waist is 1m, and the bit rate is 400Mbps, for different transmitted power within the safety limitation, the received SNR at the fiber end is shown in Fig. 15 . Here we assume the coupling efficiency of the light into the fiber is only 30%, which is easy to realize. It is obvious that the SNR performance improves considerably when the transmitted power increases. Therefore we choose 5mW (7dBm) transmitted power in the up-stream system to leave some margin. If we fix the up-stream beam waist at 1m and transmitted power at 5mW, the SNR performance with respect to different bit rates is shown in Fig. 16 . When the bit rate decreases, the system performance improves. Since the up-stream system does not require relatively high data rate, we choose 200Mbps based on the results in Fig. 16 to provide mobility over the entire room.
The beam waist of the signal also influences the performance of the system. In Fig. 17 we investigate the SNR performance of the system for different beam waists when the bit rate is maintained at 200Mbps and the transmitted power is 5mW. It is obvious that when the beam waist increases, the peak value of the received SNR decreases while the minimum SNR increases. When the beam waist is 1.9m, a SNR>15.6dB or equivalently a BER<10−9 is achieved. Since cell 1 is the worst position for up-stream data transmission over the entire room, we would expect better performance in other cells which warrants the mobility requirement can be achieved over the entire room.
Applications would require portable and compact transponder devices integrating - a CPC and a PD for the down-stream communication and a light source, a lens and a steering mirror are required for the up-stream operation. The size, cost and weight of the CPC and PD are reasonable, and the lens and steering mirror can be realized with compact micro-electro-mechanical systems (MEMS) which are already commercially available. Output power and bandwidth requirements make laser diodes as the potential candidates for the light source and cost-effective integration of laser diodes with other devices will be critical to success of such implementation.
6. Experiment setup and results
6.1 Experiment setup
To demonstrate the feasibility of our proposed hybrid system, a proof-of-concept experiment has been carried out. The setup of the experimental system is shown in Fig. 18 . The optical signal is generated in a CO and transmitted to the room via a fiber distribution network. This is emulated by a 5.6km single-mode fiber in our experiments. Then the light from a pigtailed fiber goes through a lens to increase its divergence before propagating through free space until it reaches the CPC at the receiver end. Here we use a coupling system consists of several lenses and a fiber collimator to couple the light onto a PD with a small sensitive area. Although a large sensitive area PD will simplify the system, its bandwidth is highly limited. An optical power meter with a large free space optical detection head is used for signal power detection and calibration. The sensitive area of the detector head is larger than the exit area of CPC therefore almost all the power collected by the CPC can be measured.
The beam waist in the experiments is chosen to be 2m at the receiver end, consistent with our proposed scheme. Here we only use a simple lens configuration with fixed focal length instead of the more complex adaptive lens proposed in the simulations since the beam waist has already been fixed. In a real scenario when the subscriber is moving in the room, light incident on the CPC originates from different directions further compounded by the different distances between the CPC and the beam center, therefore we quantitatively characterize these parameters as a function of system performance by manually and systematically tilting the transmitter to emulate these.
6.2 Experiment results
In the experiment 1Gbps OOK signal (231-1 PRBS) is transmitted. The transmitted power is 8.4dBm, the distance between free space transmitter and receiver is 104cm (limited by the length of sliding rail used), and the CPC has a FOV = 45°. Although the distance between transmitter and receiver is different from that assumed in the simulations, the beam waist parameter at the receiver end is the same. The received power and BER with respect to different distances from the beam center are shown in Fig. 19 . When the distance from beam center is smaller than 65cm, BER is smaller than 10−12 so it is not measured. However, SNR of such signals are measured and the corresponding BER is calculated. It can be seen that BER increases with the increasing distance from the beam center. This is simply due to the lower received signal power while the noise remains almost constant. Furthermore, error-free operation (BER<10−9) can be achieved even at the beam boundaries. From these results it is obvious that when the subscriber is moving inside the area covered by signal beam, error-free performance can always be achieved with mobility incorporated. With the WiFi localization system, it is possible to establish error-free operation over the entire room with a transmission power of 8.4dBm. The minimum power needed to achieve error-free operation in the experiments is slightly higher than that obtained in the simulations (8.2dBm). This is due to the fact that in our experiments the testbed is directly under and nearer to the background lamps. Therefore there is stronger background light induced noise in the experiments.
The beam waist in the experiments is 2m and with such a large beam width we are able to minimize service interruption when subscribers are moving across the room as well as relax the speed and precision localization requirements, but on the expense of higher transmission power. Therefore the tradeoff between the beam waist and minimum power needed to achieve error-free operation even at beam boundaries is also examined, and the results are shown in Fig. 20 . It can be seen from Fig. 20(a) that when the transmitted power is 3.2dBm, 5.9dBm and 8.4dBm for 1m, 1.5m and 2m beam waists respectively, error-free operation (BER<10−9) can be achieved even at the beam boundaries. The minimum power needed to provide error-free performance with respect to different beam widths is shown in Fig. 20(b). From the result we can deduce that the maximum possible beam width for our system is 2.05m which corresponds to ~8.45dBm transmitted power. This limitation is imposed by the eye and skin safety regulations which limit maximum free space transmission power of 7mW (~8.45dBm). In a practical scenario, different beam widths can be chosen according to the system requirements while meeting the safety regulations.
When the beam waist is fixed at 1m and transmitted power at 5.5dBm, the BER with respect to distance from beam center is shown in Fig. 21 and the eye diagrams of the received signals when the distances from beam center is 25cm and 50cm are also shown. It can be seen that the eyes are clearly open that error-free operation can be achieved at beam boundaries (50cm from beam center). Although the beam waist here is only 1m, not 2m as proposed in simulations, it is still reasonable and WiFi system can easily achieve such localization precision. Therefore, the feasibility of such a 2.5Gbps OW communication system with mobility feature is successfully demonstrated.
In this paper, a high speed hybrid OW/RF communication system with mobility feature for indoor applications is proposed. With the WiFi RF technique used for localization of the subscriber, both down-stream and up-stream systems are studied. Simulation results of the down-stream system show that with modest transmission power several gigabits/second bit rate can be achieved. OW up-stream system is studied for the first time and it is shown that bit rate of several hundred megabits/second can be easily achieved.
Furthermore, proof-of-concept experiments are carried out as well to demonstrate the feasibility of our proposed system and study various tradeoffs. The experimental results fit the simulations well and bit rate as high as 2.5Gbps has been successfully transmitted for the first time in indoor OW communication systems with mobility feature.
Finally, it is important to note here that despite the state of the art RF gigabit wireless system can achieve 5Gbps at 0dBm, it requires complex signal processing and advanced modulation format. Furthermore, for RF gigabit wireless system operated at 60GHz band, only ~7GHz license-free band is available. On the contrary, our proposed OW system does not require any complex signal processing functionality and advanced modulation format while achieving a much higher bandwidth available for future usage. As a result, with further improvements this system has the potential to achieve higher bit rate and better performance than its RF counterpart.
This work was supported in part by NICTA. NICTA is funded by the Australian Government as represented by the Department of Broadband, Communications and the Digital Economy and the Australian Research Council through the ICT centre of Excellence program.
References and links
1. F. R. Gfeller and U. Bapst, “Wireless in-house data communication via diffuse infrared radiation,” Proc. IEEE 67(11), 1474–1486 (1979). [CrossRef]
2. J. R. Barry, J. M. Kahn, W. J. Krause, E. A. Lee, and D. G. Messerschmitt, “Simulation of multipath impulse response for indoor wireless optical channels,” IEEE J. Sel. Areas Comm. 11(3), 367–379 (1993). [CrossRef]
3. S. T. Jovkova and M. Kavehard, “Multispot diffusing configuration for wireless infrared access,” IEEE Trans. Commun. 48(6), 970–978 (2000). [CrossRef]
4. J. M. Kahn, J. R. Barry, M. D. Audeh, J. B. Carruthers, W. J. Krause, and G. W. Marsh, “Non-directed infrared links for high-capacity wireless LANs,” IEEE Personal Commun. 1(2), 12–25 (1994). [CrossRef]
5. A. P. Tang, J.M. Kahn, and K.-P. Ho, “Wireless infrared communication links using multi-beam transmitters and imaging receivers,” in Proceedings of IEEE International Conference on Communications (Dallas, Texas, 1996), pp.180–186.
6. A. G. Al-Ghamdi and J. M. H. Elmirghani, “Spot diffusing technique and angle diversity performance for high speed indoor diffuse infra-red wireless transmission,” IEE Proc., Optoelectron. 151(1), 46–52 (2004). [CrossRef]
7. A. G. Al-Ghamdi and J. M. H. Elmirghani, “Line strip spot-diffusing transmitter configuration for optical wireless systems influenced by background noise and multipath dispersion,” IEEE Trans. Commun. 52(1), 37–45 (2004). [CrossRef]
8. J. M. Kahn and J. R. Barry, “Wireless infrared communications,” Proc. IEEE 85(2), 265–298 (1997). [CrossRef]
9. J. B. Carruther and J. M. Kahn, “Angle diversity for nondirected wireless infrared communication,” IEEE Trans. Commun. 48(6), 960–969 (2000). [CrossRef]
10. K. L. Sterckx, J. M. H. Elmirghani, and R. A. Cryan, “Pyramidal fly-eye detection antenna for optical wireless systems,”, in Proceedings of IEE Colloquium on Optical Wireless Communications (London, 1999), pp.1–5.
11. G. Yun, and M. Kavehrad, “Spot-diffusing and fly-eye receivers for indoor infrared wireless communications,” in Proceedings of IEEE International Conference on Selected Topics in Wireless Communications (London, 1992), pp.262–265.
12. P. Djahani and J. M. Kahn, “Analysis of infrared wireless links employing multibeam transmitters and imaging diversity receivers,” IEEE Trans. Commun. 48(12), 2077–2088 (2000). [CrossRef]
13. K. Wang, A. Nirmalathas, C. Lim, and E. Skafidas, “Indoor gigabit optical wireless communication system for personal area networks”, accepted by the 23rd Annual Meeting of IEEE Photonics Society (Denver, 2010), TuO3.
14. K. Wang, A. Nirmalathas, C. Lim, and E. Skafidas, “Gigabit optical wireless communication system for indoor applications,” accepted by Asia Communications and Photonics Conference and Exhibition (Shanghai, China, 2010), FP7.
15. J. B. Carruthers, “Multipath Channels in Wireless Infrared Communications: Modeling, angle diversity and estimation,” Ph.D. dissertation, Univ. of California, Berkeley (1997).
16. A. G. Al-Ghamdi and J. M. H. Elmirghani, “Performance comparison of LSMS and conventional diffuse and hybrid optical wireless techniques in a real indoor environment,” IEE Proc., Optoelectron. 152(4), 230–238 (2005). [CrossRef]
17. H.-H. Chan, K. L. Sterckx, J. M. H. Elmirghani, and R. A. Cryan, “Performance of optical wireless OOK and PPM systems under the constraints of ambient noise and multipath dispersion,” IEEE Commun. Mag. 36(12), 83–87 (1998). [CrossRef]
18. J. Bellon, M. J. N. Sibley, D. R. Wisely, and S. D. Greaves, “Hub architecture for infra-red wireless networks in office environments,” IEE Proc., Optoelectron. 146(2), 78–82 (1999). [CrossRef]
19. IEEE, P802.11 Task Group V, “Status of Project IEEE 802.11v”, http://grouper.ieee.org/groups/802/11/Reports/tgv_update.htm.
20. T. Ozugur, J. A. Copeland, M. Naghshineh, and P. Kermani, “Next-generation indoor infrared LANs: issues and approaches,” IEEE Personal Commun. 6(6), 6–19 (1999). [CrossRef]
21. C.-C. Lim, Y. Wan, B.-P. Ng, and C.-M. S. See, “A real-time indoor WiFi localization system using smart antennas,” IEEE Trans. Consum. Electron. 53(2), 618–622 (2007). [CrossRef]
22. B. Crow, T. Widjaja, J. Kim, and P. Sakai, “IEEE 802.11 wireless local area networks,” IEEE Commun. Mag. 35(9), 116–126 (1997). [CrossRef]
23. I. T. Haque, I. Nikolaidis, and P. Gburzynski, “A scheme for indoor localization through RF profiling”, in Proceedings of 2009 IEEE International Conference on Communications (Dresden, Germany, 2009), pp. 1–5.
24. A. Paolo, B. Luigi, and R. Rocco, “Adaptive localization techniques in WiFi environments,” in Proceedings of 5th International Symposium on Wireless Pervasive Computing (Fisciano, Italy, 2010), pp.289–294.
25. A. Papapostolou, and H. Chaouchi, “Exploiting multi-modality and diversity for localization enhancement: WiFi & RFID usecase,” in Proceedings of IEEE 20th International Symposium on Personal, Indoor and Mobile Radio Communications (Paris, France, 2009), pp.1903–1907.
26. P. LoPresti, H. Refai, and J. Sluss, “Adaptive power and divergence to improve airborne networking and communications,” in Proceedings of the 24th Digital Avionics Systems Conference (Washington D.C., 2005), pp.1.B.1–1.1.6.
27. E. Hossain, R. Muhida, and A. Ali, “Efficiency improvement of solar cell using compound parabolic concentrator and sun tracking system,” in Proceedings of IEEE Canada Electric Power Conference (Vancouver, Canada, 2008), pp.1–8.
28. F. Alsaadi and J. M. H. Elmirghani, “Performance evaluation of 2.5 Gbit/s and 5 Gbit/s optical wireless systems employing a two dimensional adaptive beam clustering method and imaging diversity detection,” IEEE J. Sel. Areas Comm. 27(8), 1507–1519 (2009). [CrossRef]
29. F. Khozeimeh and S. Hranilovic, “Dynamic spot diffusing configuration for indoor optical wireless access,” IEEE Trans. Commun. 57(6), 1765–1775 (2009). [CrossRef]
30. B. Leskovar, “Optical receivers for wide band data transmission systems,” IEEE Trans. Nucl. Sci. 36(1), 787–793 (1989). [CrossRef]
31. F. Alsaadi and J. M. H. Elmirghani, “Adaptive mobile line strip multibeam MC-CDMA optical wireless system employing imaging detection in a real indoor environment,” IEEE J. Sel. Areas Comm. 27(9), 1663–1675 (2009). [CrossRef]
32. AS/NZS 2211.1:2004, Safety of laser products (Standards Australia International Ltd and Standards New Zealand, 2004).