Mohammed S.M.Gismalla,Asrul I.Azmi,Mohd R.Salim,Farabi Iqbal,Mohammad F.L.Abdullah,Mosab Hamdan,Muzaffar Hamzahand Abu Sahmah M.Supa’at

1Faculty of Electrical Engineering,Faculty of Engineering,Universiti Teknologi,Malaysia,Skudai,81310,Malaysia 2Lightwave Communications Research Group(LCRG),Faculty of Electrical Engineering,Universiti Teknologi,Skudai,Johor,Malaysia

3Department of Communication Engineering,Faculty of Electrical and Electronic Engineering,Universiti Tun Hussein Onn Malaysia(UTHM),Parit Raja,Batu Pahat,Johor,86400,Malaysia

4Faculty of Computing and Informatics,Universiti Malaysia Sabah,Kota Kinabalu,88400,Malaysia

5Department of Computer Science,University of S?o Paulo,S?o Paulo,Brazil

ABSTRACT Visible light communication(VLC)has a paramount role in industrial implementations,especially for better energy efficiency,high speed-data rates,and low susceptibility to interference.However,since studies on VLC for industrial implementations are in scarcity,areas concerning illumination optimisation and communication performances demand further investigation.As such,this paper presents a new modelling of light fixture distribution for a warehouse model to provide acceptable illumination and communication performances.The proposed model was evaluated based on various semi-angles at half power(SAAHP)and different height levels for several parameters,including received power,signal to noise ratio(SNR),and bit error rate(BER).The results revealed improvement in terms of received power and SNR with 30 Mbps data rate.Various modulations were studied to improve the link quality,whereby better average BER values of 5.55×10-15 and 1.06×10-10 had been achieved with 4 PAM and 8 PPM,respectively.The simulation outcomes are indeed viable for the practical warehouse model.

KEYWORDS Visible light communication(VLC);industrial applications;warehouse model;light fixtures;bit error rate(BER)

1 Introduction

Radio frequency (RF) technology suffers from several restrictions that limit its performance in many applications.These restrictions include capacity crunch,electromagnetic interference,bandwidth bottleneck,high power,and cost[1-3].Therefore,visible light communication(VLC)utilising light-emitting diodes(LED)seems to be a promising technology that can be employed to overcome all RF restrictions,besides supporting the upcoming fifth generation and beyond (5 GB) [4,5].A laser diode(LD)based VLC was employed as an alternative technology to RF,mainly because LD is more efficient in operating power and has good reliability,and coherent light with high output power,especially in free space and long-distance communication[6,7].The VLC occupies a high frequency that varies from 400 to 800 THz[8,9],while simultaneously offering low power consumption,no license fee,high data rate,robustness against electromagnetic interference,and many other advantages over the existing RF technology [10-13].Additionally,VLC has dual functionality as it can be used in illumination and fast data communication concurrently[14,15],thus having a pivotal role in the fourth industrial revolution(IR 4.0)framework.

The transition of VLC from theory to industrial standardisation is discussed in[16].Based on a recent market analysis,the universal VLC market has been targeted to hit US$51 billion by 2023 with a compound annual growth rate(CAGR)of 70%[16].The VLC can be used in intelligent manufacturing and smart factory systems[17,18],primarily due to its low power consumption and cost;in comparison to incandescent and fluorescent fixtures[15,19].The VLC has been applied in conference rooms[20],parking lots[21,22],aviation systems[23,24],mining areas[25],smart homes[26],and hospitals[27].It has also been implemented in underwater scenarios[28,29]and other essential applications[30-32].

The rest of this paper is organised as follows: Section 2 describes the related works,while Section 3 illustrates the warehouse system model with illumination analyses,the top view of light fixture distribution,equations of channel model,and modulation techniques.Section 4 describes the simulation results and analyses of the proposed model.Finally,this paper ends with a conclusion in Section 5.

2 Related Works

VLC is a promising technology for 6G networks and uses LEDs to generate visible light for illumination and communication.It is expected to be used in many new applications in indoor environments.Practically,VLC has been deployed in room and office models,where various light fixtures are distributed on the ceiling to provide an efficient communication system[33-35].A number of parameters,algorithms,optimisation methods,and modulation techniques have been investigated to improve communication quality[36-38].However,only a few studies have looked into the industrial environment despite the fact that various companies are developing light fixture products with specific characteristics for industrial environments (i.e.,manufacturing,warehouse,transportation,sports facilities,and cold storage)besides mere incandescent and fluorescent fixtures[39,40].

Plenty of problems,potentials,and implementations of VLC in the industrial environment have been discussed in [1,41],where experimental designs were evidenced.The initial actual VLC results in industrial environment were reported by Berenguer et al.in[42],whereby acceptable data rate was achieved.Therefore,the objective of this paper is to investigate the VLC system for a warehouse model and to design a new model that can enhance the quality of the communication system.Different solutions were investigated and proposed to improve the VLC system quality in terms of received power,signal to noise ratio(SNR),and low bit error rate(BER)for industrial environment.One of these conceivable solutions is the distribution of light fixtures on warehouse ceilings[43-45].

Two models of light fixtures are proposed for warehouse implementation.They consist of various high-powered commercial light fixtures made by Philips and OSRAM and designed for industrial regulations with output power and different beam angles [45].Performance evaluation of the two models was carried out,where acceptable communication quality with a data rate of up to 10 Mbps was obtained.A broadband of 8×6 multiple input multiple output(MIMO)VLC system was implemented in the industrial model[46]and the results showed that the line of sight(LOS)link provided acceptable link quality in terms of SNR,while fading was observed as the LOS link was blocked.A new distributed multiuser MIMO is discussed in[41],where improvement in coverage area and link quality had been recorded.

The authors in [47] were analysed the effect of industrial environment,such as dust,artificial light,industrial processes,and other particles,on the VLC system.They prescribed considering these effects when designing VLC systems to avoid attenuation.The distribution of multiple light fixtures for factory automation was elaborated in[48],where advanced modulations and systems were suggested to meet the communication requirement for 5 GB.More wireless network structures for the warehouse model were investigated[49].In[50],an experimental test using three-dimensional(3D)visible light positioning algorithm was conducted to trace a drone in the industrial setting.The authors executed the experiment by tilting the receiver and the multipath reflections[51].

There is an increasing interest in VLC research work to cater to a wide range of indoor and outdoor implementations.However,few studies have been conducted to evaluate the VLC system in industrial settings,such as warehouses and manufacturing plants[52].The suitable distribution and the number of light fixtures needed for the room model are discussed extensively to produce an acceptable performance,especially in terms of uniform distribution in the entire room[33-35,53,54].Notably,the distribution of 13 light fixtures outperformed the others[55].As a result,this number of light fixtures with high power was used in this study for viable distribution in warehouse applications.The proposed warehouse model’s results were compared to the outcomes reported in[45].

The main contributions of this paper are to propose a distribution system of 13 high bay(HB)light fixtures for the warehouse model and identification of the impact of different semi-angle at half power(SAAHP),as well as various height levels,received power,SNR,and average BER.Additionally,this study evaluated the behaviour of average BER against various SAAHP for the first time in a warehouse model,where several modulation techniques are discussed to determine the most suitable one for a practical VLC system.

3 System Model

The general warehouse model is illustrated in Fig.1.The dimension of the warehouse model is 20 m×20 m×h m for length,width,and height,respectively.h refers to the variable height level that depends on the application scenarios included in warehouse,workplace,lab,etc.In this study,120 W AGC HB light fixtures were distributed on the warehouse ceiling[56]instead of the 155 W of power applied in the previous model[45].Therefore,the new model portrayed in Fig.1 can save 22.58%of power;signifying improved power efficiency.Table 1 tabulates the technical specifications of the light fixtures used in the proposed warehouse model.The height level from the light fixtures to the receiver plane varied from 5,7.5,to 10 m,while the height level of the receiver plane or forklift to the ground surface was 2 m.

Table 1: 120 W AGC HB light fixture specifications[56]

The LOS link was considered to examine the proposed model.Various modulation techniques were investigated to minimise the BER value.Table 2 lists the main simulation parameters mentioned in[8]and[39].

Figure 1:The proposed VLC system:(a)warehouse model,(b)top view of light fixtures distribution and coordinates on the ceiling

3.1 Horizontal Illuminance

The light fixtures should fulfil the illumination requirements while providing an acceptable communication performance[1].Therefore,the illumination intensity level at angleφwas computed as follows:

whereI(0) represents the centre luminous intensity from a group of LEDs,whileφandmdenote the angle of irradiance and the order of Lambertian emission,respectively;wheremis related to light fixture semi-angle of half illuminationand it is expressed in(2):

The illumination of 3D points(x,y,&z)is defined byEhor,as follows:

whereψanddrefer to the angle of incident and distance between light fixture and receiver,respectively.Light fixtures can be distributed on the ceiling of the warehouse model by using several configuration methods illustrated in[45].

Fig.2 shows the contour of illumination distribution at a small 10°(a)and wide 70°(b)SAAHP for 5 m height level.From the illumination results,the wide SAAHP should be considered to improve illumination distribution,while many blinds or uncovered zones were observed at the small SAAHP.The SAAHP is the angle from an axis perpendicular to the light fixture by which the illuminating intensity is just 50%.It is also called the‘viewing angle’and may vary from 10°at no diffusion to 70°for maximum diffusion[57].

3.2 Received Power and SNR Calculations

Upon considering the LOS link,the channel DC gain was computed as follows[45,58]:

whereArepresents the physical area of the photodiode (PD) detector,ψis the FOV angle of the receiver,Ts(ψ) andg(ψ) refer to the optical filter gain and the gain of an optical concentrator,respectively,whileψcis the SAAHP.

Figure 2:Contour of illumination distribution for the proposed warehouse model

For the 3D coordinates,as shown in Fig.1,the light fixtures were installed at the position of(xi,yi,hi),while the receiver or forklift at the position of (x,y,h).Therefore,cos(φ) was computed as follows:

Moreover,it can be expressed that g(ψ)≈g and cos(ψ)=cos(φ).sincTs(ψ)≈Ts.By substituting Eqs.(5),(4)is written as follows:

The received power(Pr)at photodetector(PD)can be calculated from the transmitted power of light fixture(PT)and the channel DC gain,as shown in(7):

The SNR can be further computed by:

whereRrefers to PD responsivity and it can be measured in amperes per watt.Prwas defined and calculated in(7),whererefer to the variances of thermal and shot noises,respectively,expressed as follows:

whereI3=0.0868,whilekandqare constants of Boltzmann and electronic charge,respectively,Tkrefers to the absolute temperature,ηis a fixed capacitance of PD per unit area,B denotes the equivalent noise bandwidth,I2indicates the noise bandwidth factor,G signifies an open-loop voltage gain,Γrefers to the channel noise factor of the field-effect transistor(FET),gmis the FET transconductance,andIbgrepresents the current of background light.

3.3 Modulation Techniques

Upon further investigation,several modulation techniques were discussed to assess the performances of BER and data rate.The simulation work performed in[45]had applied the on-off keying(OOK)modulation,which gave a data rate of 10 Mbps.Turning to this present study,it looked into the modulation techniques in the warehouse model and the effects of various SAAHP aspects on the average BER performance for the first time.The modulation techniques included binary phase shift keying(BPSK),quadrature phase shift keying(QPSK),pulse position modulation(L-PPM),and pulse amplitude modulation(M-PAM).Eqs.(11)to(15)show the BER calculations for the mentioned modulation techniques,respectively,where L and M refer to the order of modulation[59].The BER performance was executed in the presence of Gaussian noise.

4 Simulation Results and Discussion

This section demonstrates the simulation results and the analyses of the proposed warehouse model in terms of received power,SNR,and BER,so as to enhance communication performance.Fig.3 shows the received power and the contour of received power distribution at 5 m height level.The average received power of 0.0753 dBm was obtained at wide SAAHP 70°,which fluctuated from-4.4898 to 1.7734 dBm.Next,the average received power of -0.6065 and -1.3652 dBm had been achieved at height levels 7.5 m and 10 m,respectively.The maximum,minimum,and average received power for height levels 7.5 and 10 m are presented in Table 3.

Table 3: Received power and SNR against various height levels for the proposed model

Fig.4 portrays the SNR distribution and its contour at wide SAAHP 70°for 5 m height level.The maximum,minimum,and average SNR values of 73.2095,60.6838,and 69.8135 dB were produced.The average SNR values of 68.45 and 66.9328 dB were obtained at 7.5 and 10 m height levels,respectively (see Table 3).Referring to Table 3,the short height level yielded better average received power and SNR,which are adequate for practical VLC systems.However,the performance decreased as the height level increased.

Figs.5 and 6 demonstrate the behaviour of average received power and SNR performances at different height levels,where various SAAHP values were considered.The received power of 0.0753,-0.6065,and-1.3652 dBm had been recorded at wide SAAHP 70°for height levels 5,7.5,and 10 m,respectively.Next,the SNR values of 69.8135,68.45,and 66.9328 dB were obtained at height levels 5,7.5,and 10 m,respectively.Based on these results,better performance in terms of received power and SNR was observed at 5 m height level,but the performance deteriorated with increasing height level and SAAHP.

Figure 3:(a)Received power and(b)contour of power of the proposed model at 5 m height level

Figure 4:(a)SNR and(b)contour of SNR of the proposed model at 5 m height level

The results of the proposed model at wide SAAHP and height levels of 7.5 and 10 m were acceptable and better,when compared to previous models[45].The variation of SAAHP led to a tradeoff between received power and SNR with the coverage area in the entire warehouse model.Referring to Figs.5 and 6,small SAAHP produced the lowest average received power and SNR distribution at 5 m height level.This is ascribed to the LOS link that bounded the light distribution downward from the light fixture,which failed to cover all areas in the warehouse model.However,both received power and SNR distribution were increased by increasing the SAAHP.Although higher SAAHP can degrade both received power and SNR,it can still yield an acceptable level that guaranteed the link quality.This trade-off was diminished with increased height level,mainly because the coverage area increased when the height of light fixture increased both light diffuses and angular range.

Figure 5:Average received power against various SAAHP for different height levels

Figure 6:Average SNR against various SAAHP for different height levels

Table 4 summarises the received power and SNR results for the past models that used 9 Philips HB and 12 OSRAM HB light fixtures with the proposed model when SAAHP was set at 70°.From Table 4,better maximum,minimum,and average received power and SNR were obtained from the new model of 13 AGC HB at various height levels.

Table 4: Parameters of performance for the warehouse models

For further investigation on improving the link quality in terms of BER,different modulation techniques were assessed.Apparently,the lowest BER yielded the best communication quality.A trade-off was noted when various modulation techniques were applied for communication systems,which successfully improved both power and bandwidth efficiencies,apart from minimising BER[60].Fig.7 displays the BER performance for OOK-NRZ,BPSK,QPSK,4 PPM,8 PPM,and 4 PAM at 5 m height level.

Both 8 PPM and 4 PAM produced minimum BER with a data rate of up to 30 Mbps,which met the required BER of 10-6for acceptable communication links.However,OOK-NRZ,BPSK,and QPSK yielded high BER that exceeded 10-6(see Table 5).Therefore,better average BER at 5.55×10-15was achieved by using 4 PAM.Table 5 tabulates the performance of average BER for all the studied modulations at various height levels.

Table 5: Average BER performance of different modulation techniques at SAAHP 70°

Fig.8 presents the behaviour of average BER performance against various SAAHP and different height levels.At small SAAHP of 10°or less,the average BER performance had deteriorated (i.e.,BER exceeded 10-6),except for higher orders of L-PPM and M-PAM modulation techniques,while the average BER minimised as the SAAHP increased(see Figs.8a-8c).Relative stability in the average BER performance was observed for SAAHP above 30°.Better average BER of 5.55×10-15was attained at 5 m height level,while the average BER values of 7.5 × 10-15and 1.27 × 10-14were recorded at height levels 7.5 and 10 m,respectively,as shown in Table 5.Notably,the average BER degraded with increased height level,as portrayed in Table 5 and Figs.8b and 8c.

Figure 7: BER performance of the proposed model for 5 m height level: (a) OOK-NRZ,(b) BPSK,(c)QPSK,(d)4 PPM,(e)8 PPM,and(f)4 PAM

Figure 8:Evaluation of average BER performance against different SAAHP for height levels(a)5 m,(b)7.5 m,and(c)10 m

The proposed model of 13 AGC HB light fixtures for the warehouse displayed improvements in terms of received power,SNR,and data rate.Better BER values were obtained when compared with the results reported by Almadani et al.Referring to the BER performance,higher order of L-PPM and M-PAM generated better average BER than other modulations.As depicted in[60],L-PPM was more efficient in terms of power,while M-PAM had better efficiency in terms of bandwidth.Hence,both modulations are suitable for the design of VLC system in a practical warehouse model.

5 Conclusion

In this paper,a new modelling of light fixtures is proposed for a warehouse to investigate the performance of communication system in terms of received power,SNR,and BER.The LOS link was considered and different height levels(5,7.5,&10 m)were used in the evaluation.The average received power values of 0.0753,-0.6065,and-1.3652 dBm had been obtained at height levels of 5,7.5,and 10 m,respectively.Next,the average SNR of 69.8135 dB was recorded at 5 m height level,whereas SNR values of 68.45 dB and 66.9328 dB were achieved at 7.5 and 10 m,respectively.The behaviour of average received power and SNR over various SAAHP and height levels had been assessed.The new model showcased better performance than that reported by Almadani et al.Several modulation techniques,including OOK-NRZ,BPSK,QPSK,4 PPM,8 PPM,and 4 PAM,were investigated to evaluate the BER performance over various SAAHP and height levels.The average BER values of 1.06×10-10,1.44×10-10,and 2.24×10-10had been achieved by using 8 PPM for height levels of 5,7.5,and 10 m,respectively.Next,4 PAM produced average BER values of 5.55×10-15,7.52×10-15,and 1.27×10-14at 5,7.5,and 10 m,respectively.Overall,the new model displayed exceptional performance that is suitable for warehouse application.The higher order of both L-PPM and M-PAM yielded better link quality in terms of BER performance.Therefore,more modelling and analyses are required in future endeavour to produce a general model for warehouses and other industrial applications.

Acknowledgement:The authors acknowledge the funding of this project by Research Management Centre (RMC),under the Professional Development Research University Grant (UTM Vot No.06E59),Universiti Teknologi Malaysia (UTM),Malaysia.This project was also funded by Universiti Malaysia Sabah,Jalan UMS,88400,Kota 599 Kinabalu,Malaysia,under the name of grant“Smart Vertical Farming Technology for Temperate Vegetable Cultivation in Sabah:Practicing Smart Automation System Using IR and AI Technology in Agriculture 4.0”.

Funding Statement:The manuscript APC is supported by the grant names Smart Vertical farming Technology for Temperate Vegetable Cultivation in Sabah: Practicing Smart Automation System Using IR and AI Technology in Agriculture 4.0”.It was also supported by Professional Development Research University Grant(UTM Vot No.06E59).

Conflicts of Interest:The authors declare that they have no conflicts of interest to report regarding the present study.