Open Access
Add to BibSonomyAbstract
In a multiuser bidirectional visible light communication (VLC), a large number of LEDs or an LED array needs to be allocated in an efficient manner to ensure sustainable data rate and link quality. Moreover, in order to support an increasing or decreasing number of users in the network, the LED allocation is required to be performed dynamically. In this paper, a novel smart LED allocation scheme for efficient multiuser VLC networks is presented. The proposed scheme allocates RGB LEDs to multiple users in a dynamic and efficient fashion, while satisfying illumination requirements in an indoor environment. The smart LED array comprised of RGB LEDs is divided into sectors according to the location of the users. The allocated sectors then provide optical power concentration toward the users for efficient and reliable data transmission. An algorithm for the dynamic allocation of the LEDs is also presented. To verify its effective resource allocation feature of the proposed scheme, simulations were performed. It is found that the proposed smart LED allocation scheme provides the effect of optical beamforming toward individual users, thereby increasing the collective power concentration of the optical signals on the desirable users and resulting in significantly increased data rate, while ensuring sufficient illumination in a multiuser VLC environment.
© 2015 Optical Society of America
1. Introduction
Light emitting diodes (LEDs) are becoming popular as illumination devices. The LEDs are now considered as a strong candidate for a last-mile wireless communication technology. LED based visible light communications (VLC) have many advantages such as energy efficiency, high security, the absence of electromagnetic interference and higher data rate as compared with other wireless communication technologies. Researches have been focused on addressing important issues such as high data speed, bit-error rate (BER) performance and multiuser (MU) bidirectional transmission [1–5]. These advancements in the field of VLC have now led to a short-range high-speed wireless communication technology [1, 3].
In the LED based VLC systems, the received optical power determines integrity of the transmitted data. Although high received optical power is desired, the optical power is predominantly dependent upon the line-of-sight (LOS) path transmission and the distance between the LED light source and the receiver. Unfortunately, it is inappropriate in indoor VLC environments to increase the launching optical power to compensate the power drop as in optical fiber communications [6], because the increase in the optical power affects the illumination. Hence, optical power concentration schemes in the form of resource allocation are desired instead. As a matter of fact, a successful deployment of VLC in multiuser environments is hinged upon an efficient resource allocation that makes the most of available resources to offer high speed data transmission with minimal errors.
Recognizing the importance of the resource allocation, a few interesting methods were reported in the literature. A discrete multitone (DMT) based multiple access resource allocation technique was proposed for unidirectional communication only [4]. In [5], the researchers proposed a novel user allocation scheme in a MU bidirectional VLC environment, but failed to present any resource allocation scheme. Another group of researchers proposed a spatial light modulator (SLM) based optical beam-forming scheme to focus LED light onto desired receivers [7]. This scheme entails an expensive SLM involving extra mechanical control and complex circuitry. This mechanically operating device attached to the LEDs is not desirable for indoor applications. In an effort of VLC standardization, the IEEE 802.15.7 standard focuses on the point-to-point communication and it does not contemplate multiuser transmissions yet nor specify resource allocation schemes in the MU bidirectional VLC network [8].
In this paper, we propose the resource allocation scheme in that the LED array is divided dynamically into sectors for efficient data transmission in a changing number of users of the VLC network. The sectors are then able to concentrate the optical power on each user and therefore give an effect of optical beamforming. Each sector is defined with a color value for the ease of identification and is allocated an equal number of LEDs from the LED array. Each user is allocated with a particular sector on the basis of its location within the room. It is important to note that the proposed scheme operates in a bidirectional transmission environment where an RGB LED array, infrared (IR) receivers and a control unit are employed. The location of the user can be identified by the control unit with the knowledge of the transmitting sector used for downlink transmission and the specific IR photodetector (PD) used as the uplink receiver. Simulations are performed to verify the effectiveness of the proposed scheme. It is found that the proposed scheme with sufficient illumination provides efficient resource allocation in an indoor VLC environment.
Section 2 provides details about the proposed system. Theoretical analysis is presented in Section 3 and results are discussed and analyzed in Section 4. Conclusions have been drawn in Section 5.
2. Smart LED allocation scheme
2.1. Proposed scheme
The proposed scheme operates in a bidirectional environment where the RGB LED array, IR PDs and a control unit are present. The key idea of the proposed scheme is to provide efficient VLC links and subsequently robust performance in a multiuser VLC environment by concentrating optical power toward each user. For uplink transmission, IR is used to avoid any interference from the RGB LED array based downlink [9].
The number of LEDs in the array is chosen in such a way that the LEDs should fulfill the necessity of sufficient illumination and can be divided into an integral number for an increasing or decreasing number of users in an indoor VLC environment. PDs are installed to provide the location of a particular user, thus enabling a particular number of LEDs to be allocated to that user.
For the ease of demonstration, suppose that the scheme supports a maximum of 10 users. The number of LEDs in the LED array is equal to the least common multiple (LCM) of all the integers from 1 to 10, i.e. 2520. Also, 8 IR PDs are installed that they are used for identifying the user’s location. A typical indoor environment with the LED array and PDs is shown in Fig. 1(a). Figure 1(b) shows its bottom view.
Fig. 1 (a) Typical indoor environment with LED array, photodetectors and control unit. (b) Bottom view.
When the number of users is larger than 10 or the number of LEDs is not equal to the LCM for the users, the scheme allocates the LEDs for each user in the following way. The number of LEDs per user, Nuser, is found as , where Ntotal is the total number of LEDs in the LED array, n is the number of users and floor(x) denotes round operation for each element of x to the nearest integer less than or equal to x. The remaining LEDs, Nrem, is, if any, given by Nrem=Ntotal − (Nuser × n) and these can be utilized either for illumination purpose or for distribution over the users, i.e. Nrem users will have (Nuser+1) LEDs and n − Nrem users will have Nuser LEDs allocated.
2.2. Frame structure and LED allocation
In the proposed scheme, the user initially requests for resource allocation by transmitting a request frame using IR transmitter. Figure 2(a) shows its structure. The request frame is an extended version of the frame structure defined by IEEE standards [8]. The frame structure consists of a preamble for synchronization, a PHY header (frame length, modulation and coding), data payload and user device ID.
This initial request for connection is received by the nearest IR receiver. If multiple detectors receive this signal, the control unit will select the one with the highest intensity. On the basis of the selected PD’s position, the control unit divides the LED array into sectors. In this way, a dynamic allocation of LEDs is achieved as per the location of the user. The notification of the allocation is then sent back to the user using a frame called allocation frame. The allocation frame comprises both the device ID and the transmitting sectors ID. Figure 2(b) depicts the allocation frame. After the reception of the allocation frame, the user sends acknowledgment using the similar frame structure. Therefore, the user now knows its sector’s information and also the control unit identifies the user’s location. With an increasing number of users, it is apparent that the number of LEDs allocated to each sector decreases; however, the optical power concentration toward the users increases.
It is worth noting that the allocation frame, the acknowledgment frame and the data frame can be differentiated by using different types of pre-defined sequence in the place of Ack/Allocation/Data bits as shown in Fig. 2(b). For example, a sequence of 10101010.... can be placed in the allocation frame. Likewise, a sequence of 110011001100... can be used in the acknowledgment frame. For the data frames, actual data bits can be placed.
A comprehensive allocation strategy for up to 10 users with colored sectors is now shown in Table 1. It is noteworthy to see the effect of an increasing number of users on the number of LEDs allocated to each user.
Table 1. LED allocation for up to 10 users
Each sector is identified with a particular color associated to that sector, i.e. with a color defined by weights allocated to each primary color (red, blue and green). For example, for 4 users, each user is allocated 630 LEDs and the colors for 4 users are red, green, blue and half-weight red plus half-weight green, respectively. The color for the fourth user is determined by the International Commission on Illumination (CIE) 1931 xy color coordinates [8]. This color definition of each sector will be conducive to implementing color based multiple user access schemes such as wavelength division multiple access (WDMA) [10] or color-clustered multiple access (CMA) schemes [2]. The proposed technique can be implemented using other mature multiple access schemes such as orthogonal frequency division multiple access (OFDMA) [5, 11], optical code division multiple access (OCDMA) [11] or time division multiplexing (TDM) [10] for multiple users in bidirectional VLC systems.
The sequence diagram associated with the proposed scheme is shown in Fig. 3.
As mentioned previously, the uniformly distributed PDs are used to detect the transmitted signals from the users. Following the location of the user, the LED array should be allocated uniformly according to Table 1. It should be noted, however, that this distribution should not affect other users already present in the network. For providing dedicated LEDs to each user, we need to ensure an equal number of LEDs allocated to each user in the LED array on the basis of the user locations. Figure 4(a) shows the flow diagram of the proposed algorithm. If the location of the new user is similar to an old user, then the LED allocation is performed in a circular fashion, otherwise the allocation is made by creating a new sector. Therefore, it is ensured that there is one user per sector, i.e. the LED allocation is performed by creating a sector in a circular or sectorial fashion. Figure 4(b) shows the pictorial representation of the proposed algorithm for up to 4 users that can easily be extended to 10 users. It is important to note that only the number of LEDs allocated to each user varies according to a varying number of users in the indoor environment, while maintaining the downlink connections. Therefore, there is no requirement for reconfiguration as such, only the reallocation of LEDs takes place.
In addition, the proposed scheme supports the movement of the users in an indoor environment. A threshold signal to noise ratio (SNR) is initally defined for receiving the signal at IR uplink receiver. When the user moves from one place to another within the indoor environment, the following will occur. If the received signal at the IR uplink receiver is detected below the predefined threshold SNR level, the control unit will look for the highest SNR value received by other PDs, which indicates the new location of the user moved. Once the PD detecting the highest SNR level is identified, the LEDs for that section of the area will be reallocated to accommodate the user’s movement in a circular or sectorial fashion.
3. Theoretical analysis
3.1. Received power and SNR
First, we discuss the received power and SNR received by the receiver. The LED is assumed to have a Lambertian radiant intensity [1] as in Eq. (1),
where ϕ is irradiance angle, ml is the order of Lambertian emission and it is related to the LED’s semi-angle at half power ϕ1/2, which can be defined asHence, the channel direct current (DC) gain is described by [1] :
where A is the physical area of PD, Dd is the LOS distance between the LED and the PD, Ts(ψ) is the gain of an optical filter, ψ is the angle of incident, and ψFOV is the field of view (FOV) of PD. After modulation, the output signal power P0(t) from LED can be described as: where PLED is the launched optical power by LEDs, mi is the modulation index and x(t) is the non-return-to-zero (NRZ) on-off-keying (OOK) signal. The received optical power detected by the PD is equal to, where R is the rensponsitivity of the PD, H(0) is the channel transfer function described in Eqs. (3a) and (3b). The received power intensity of a user, Puser(t), can be written as a function of the number of users, n, and can be given by:Now, when the detected signal is DC-blocked, the detected electrical signal sout (t) from the PD can be derived as:
The equation of SNR of the detected electrical signal can be given by:
where PN is the power of noise defined in [1].3.2. Illumination
The luminance expresses the brightness of an illuminated surface. It is assumed that the light intensity emitted from the source has a cosine dependence on the angle of emission with respect to the normal surface [1]. The luminous intensity at angle ϕ is given by :
A horizontal illuminance Ehor at point (x,y) is given by
where Dd is the LOS distance between the LED and the PD, ϕ is the angle of irradiance, ψ is the angle of incident and ml is the order of Lambertian emission as mentioned in Eq. (2).3.3. BER and maximum achievable data rate
In the present analysis, the modulation format employed for data transmission is NRZ OOK, which is the simplest modulation scheme mentioned in PHY I of IEEE standards [8]. The BER equation for NRZ OOK modulation scheme can be given as [1]:
where For a given error rate (Pe), the maximum achievable data rate (ϒm) can be defined as: where Rmax represents the maximum data rate for the system concerned and BERm denotes maximum tolerable BERs. It can be observed from Eq. (13a) that the system can achieve the maximum data transmission rate in case of error free transmission, i.e. Pe is equal to 0. Since it is not possible to achieve an error free transmission, the maximum tolerable BERs, BERm, can be defined for applications. As an example, the value of BERm can be set to 10−3 for voice transmission and 10−6 for multimedia transmission.4. Results and analysis
A resource allocation scheme in an indoor VLC environment must fulfill the need of sufficient illumination. For this requirement, we performed simulations in an indoor VLC environment (5m × 5m × 3m) having the proposed LED setup as shown in Fig. 1(a). The transmission was performed by 2520 LEDs with the transmitted power of a single RGB LED being 60mW and with 50° semi-angle at half power. The reception was made at 0.85 m above the floor by PD with an optical filter having optical filter gain, Ts(ψ), equal to 1. PD has a physical area, A, of 1.0 cm2, responsitivity, R, equal to 1 and FOV of 60°.
Figure 5(a) shows the illumination distribution in an indoor environment. It is obvious that the proposed LED setup provides sufficient illuminance of 900 to 1350 lx, defined by ISO [1], over the entire room. We also performed simulations for the distribution of the received power in the same indoor environment. Figure 5(b) shows the power distribution across the room.
Average user SNRs and BERs relative to the number of users are shown in Fig. 6. We used the NRZ OOK modulation scheme, the simplest modulation format described in the IEEE standard [8], and performed simulations to obtain the SNRs. From the obtained SNRs, we calculated the BERs using Eq. (11). It is clear that when the number of users increases, the average user SNRs decrease, while the BERs increase. When the number of users reaches the maximum number, i.e. 10, the average user SNR is reduced to 9.76 dB, whereas a BER performance increases up to 2.27×10−4.
Figure 7(a) represents the maximum achievable data rates relative to the number of users with and without the resource allocation scheme. For the performace of the OOK without the resource allocation scheme, we assumed the multiuser data transmission in a time division manner. It can be observed from Fig. 7(a) that the maximum achievable data rate of the proposed resource allocation scheme calculated from Eq. (13) for 10 users is approximately 81 Mbps/user on the basis of a BERm value of 10−3. Therefore, it is found that this maximum achievable data rate is five times higher than the data rate of the OOK without the resource allocation scheme when the number of users is 10.
Fig. 7 Comparison of maximum achievable data rate (a) OOK with and without the proposed scheme. (b) OFDMA based user allocation scheme with the proposed scheme applied to OFDMA.
As the proposed scheme focuses on user or resource allocation, it is interesting to perform a comparison with conventional user/resource allocation schemes. As noted earlier, however, the scheme has distinctive features that differentiate conventional user/resource allocation schemes in terms of structure and complexity in a multiuser VLC. Although it appears to be an indirect comparison, we have undertaken comparative investigation with OFDMA based user allocation scheme [5]. Figure 7(b) shows performance comparison in terms of the maximum achievable data rate per user. It is apparent that when the number of users is 10, an additional gain of approximately 420 Mbps per user is achieved with the proposed scheme applied to OFDMA.
The results exhibit robustness and significance of the proposed resource allocation scheme. It should be noted that although the present scheme is presented with 10 users, the scheme can support any number of users. For good link quality and performance, however, the maximum number of users can be determined on the basis of the number of LEDs, power of each LED installed, required BER/user and required achievable data rate/user. The BER performance and data rate of the proposed scheme can further be enhanced by using other modulation schemes and error detection and correction techniques described in [8].
5. Conclusion
We have proposed a novel smart LED allocation scheme for efficient multiuser visible light communications in an indoor VLC network. Every user is allocated with a particular sector on the basis of its location within the room. The location of the user can be identified from the transmitting sector and the receiving IR PD. An algorithm is proposed for the dynamic allocation of the LEDs to a particular sector. The performed simulations show that the proposed scheme offers an efficient resource allocation for VLC networks and also significantly increased data rate compared with the VLC network without the resource allocation scheme.
Acknowledgments
This research is funded by the Research Grant of BB (Brain Busan) 21 project of 2015.
References and links
1. T. Komine and M. Nakagawa, “Fundamental analysis for visible-light communication system using LED lights,” IEEE Trans. Consum. Electron. 50(1), 100–107 (2004). [CrossRef]
2. K. Bandara and Y. H. Chung, “Novel color-clustered multiuser visible light communication,” Trans. Emerg. Telecommun. Technol. 25(6), 579–590 (2014). [CrossRef]
3. Y. F. Liu, C. H. Yeh, C. W. Chow, Y. Liu, Y. L. Liu, and H. K. Tsang, “Demonstration of bi-directional LED visible light communication using TDD traffic with mitigation of reflection interference,” Opt. Express 20(21), 23019–23024 (2012). [CrossRef] [PubMed]
4. D. Bykhovsky and S. Arnon, “Multiple access resource allocation in visible light communication systems,” J. Lightwave Technol. 32(8), 1594–1600 (2014). [CrossRef]
5. A. Sewaiwar, S.V. Tiwari, and Y.H. Chung, “Novel user allocation scheme for full duplex multiuser bidirectional Li-Fi network,” Opt. Commun. 339, 153–156 (2015). [CrossRef]
6. C. H. Yeh, C. W. Chow, S. P. Huang, J. Y. Sung, Y. L. Liu, and C. L. Pan, “Ring-based WDM access network providing both Rayleigh backscattering noise mitigation and fiber-fault protection,” J. Lightwave Technol. 30(20), 3211–3218 (2012). [CrossRef]
7. S. M. Kim and S. M. Kim, “Performance improvement of visible light communications using optical beamforming,” in International Conference Ubiquitous and Future Networks (ICUFN, 2013), pp. 362–365.
8. S. Rajagopal, R.D. Roberts, and S.K. Lim, “IEEE 802.15.7 visible light communication: modulation schemes and dimming support,” IEEE Commun. Mag. 50(3), 72–82 (2012). [CrossRef]
9. Y. Jang, K. Choi, F. Rawshan, S. Dan, M. Ju, and Y. Park, “Bi-directional visible light communication using performance-based selection of IR-LEDs in upstream transmission,” in International Conference on Ubiquitous and Future Networks (ICUFN, 2012), pp. 8–9.
10. J. M. Luna-Rivera, R. Perez-Jimenez, J. Rabadan-Borjes, J. Rufo-Torres, V. Guerra, and C. Suarez-Rodriguez, “Multiuser CSK scheme for indoor visible light communications,” Opt. Express 22(20), 24256–24267 (2014). [CrossRef] [PubMed]
11. J. Fakidis, D. Tsonev, and H. Haas, “A comparison between DCO-OFDMA and synchronous one-dimensional OCDMA for optical wireless communications,” in IEEE 24th Int. Symp. on Personal Indoor and Mobile Radio Comm. (PIMRC), London, UK, 2013, pp. 3605–3609.
