Open Access
Add to BibSonomy
Abstract
Ultraviolet (UV) communication has been the topic of extensive recent research due to its non-line-of-sight (NLOS) communication, anti-interference, and high confidentiality. Given the lack of the UV Media Access Control (MAC) protocol, this paper proposes a UV lossless contention MAC (UVLLC-MAC) protocol creatively. This MAC protocol is based on the superposition logic of UV power, which effectively reduces the collision loss of multi-node contention access. The basic working mechanism and protocol flow are given in detail, and the network performance is simulated and analyzed mathematically based on a four-node UV network. Comparing with the slotted ALOHA protocol, the simulation results indicate that the network has a higher throughput and lower delay under the protocol. The practical networking of four communication terminals is designed and implemented, and the effectiveness of the new UVLLC-MAC protocol is fully verified.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Nowadays, the radio spectrum resources are increasingly limited. And radio communication is vulnerable to electromagnetic interference, which cannot meet the stringent requirements for communication applications in complex environments. Traditional Free Space Optical Communication (FSO) can overcome the above difficulties and realize long-distance, large-capacity, and reliable communication [1]. However, FSO is based on line-of-sight (LOS) communication and relies on the automatic Pointing, Acquisition, and Tracking (PAT) system. It is difficult to realize interconnection among multiple communication terminals, and it is even more difficult for the networking of mobile terminals. Ultraviolet (UV) communication has the characteristics of omnidirectional transmission, anti-interference, and high confidentiality based on atmospheric scattering, which provides a new effective method for wireless secure communication networking [2, 3]. It eliminates the strong directionality of FSO, and the PAT system that restricts networking applications can also be greatly simplified. Multiple UV nodes can easily achieve “communication in motion” within their effective communication range.
At present, a lot of theoretical research has been done on UV communication and its networking [4–12]. Qin et al. studied the interconnection characteristics of UV networks [13]. Arya et al. conducted a preliminary study on the key technologies of UV networking such as design implementation, channel modeling and communication links [14]. They studied the problem of multi-user interference in UV communication networking, and considered a shot-noise limited photon-counting signal model to minimize the multiple access interference laying the foundation for UV networking.
Appropriate and efficient Media Access Control (MAC) protocol is the key to the realization of a UV communication network. A novel contention-based MAC protocol was proposed and outdoor experiments were conducted by Li [15]. Relatively complex collision detection and avoidance mechanisms need to be designed in the contention-based MAC protocol. This method will further consume limited bandwidth resources, and result in unsatisfactory network performance. Chen et al. discussed some main rules in UV networking based on the Time Division Multiple Access (TDMA) mechanism [16]. An improved TDMA networking method was proposed, reducing the network latency [17]. The advantage of this method is simple and reliable, but the waiting time delay in queuing becomes longer and there may be a waste of communication resources when there are more nodes in the network. Limited by the key technology of UV communication, especially the lack of effective networking methods, practical applications of UV networks are still rarely reported. Therefore, it is necessary to comprehensively consider both TDMA and competition-based mechanisms and propose a new MAC protocol that can solve the problem of channel contention in synchronous networks.
Based on the superposition mechanism of optical power, a novel UV lossless contention MAC (UVLLC-MAC) protocol is proposed, and its basic working mechanism and protocol flow are given in detail. This protocol is easy to implement, and its complexity is greatly reduced compared to traditional collision avoidance schemes. A four-node UV communication network model is established with the protocol, and the network performance is simulated and analyzed mathematically. The simulation results indicate that the new MAC protocol has better channel utilization, better network throughput, and lower network delay. Finally, the practical networking of the four terminals is implemented, and the effectiveness of the UVLLC-MAC protocol is fully verified.
2. Preliminaries
The schematic diagram of the UV communication network is shown in Fig. 1. Multiple terminal nodes are constructed into a typical UV communication network. For simplicity, the communication link between any two nodes in the network is reachable in one-hop.
2.1 Working mechanism
Based on the consideration of the scattering and non-line-of-sight (NLOS) characteristics of UV communication, the optical antenna of the terminal nodes has an omnidirectional characteristic. Any node can transmit data and other nodes can receive it, which can be regarded as broadcast data. The channel is set to a single wavelength channel, and multiple network nodes share a single wavelength channel. The communication between nodes needs to obtain channel resources through contention, and the corresponding multi-node contention access collision needs to be solved. It should be noted that the network is based on a physical synchronization mechanism, and the time slots are strictly aligned in our designed protocol and network.
The MAC protocol incorporates a novel technique for channel access, which ensures that there is no contention loss for the high-priority node even when access collisions occur among multiple nodes. The protocol is designed based on the principle of optical power superposition and channel status. On-off keying (OOK) modulation mode is adopted in the UV light source. The signal bit “0” (respectively, the signal bit “1”) is used to indicate that the terminal nodes emit light (respectively, don’t emit light). From the perspective of the optical physical layer, there is an “AND” and “OR” superimposition logic of UV power when multiple nodes are connected as shown in Fig. 2. The network channel appears in: (i) a light state, when more than one node emits light, (ii) a non-light state, only when all nodes do not emit light. Note that the superimposed transmission power should be controlled strictly and does not exceed the power threshold of the receiver detector. Therefore, when a high-priority communication node emits light (sending bit information “0”), the existing channel status of the UV network will not be effected by other low-priority competing nodes that need to access the channel regardless of whether they emit light or not (sending bit information “0” or “1”). Eventually, the high-priority node will win the contention and complete the communication process.
2.2 Frame structure
The general frame structure of the UVLLC-MAC protocol is shown in Fig. 3, including the frame header, access arbitration segment, control segment, data segment, and frame end. The frame header and frame end are used to indicate the beginning and ending of the frame data, and the Cyclic Redundancy Check (CRC) code can be added to the frame end to realize data error detection. The control segment is mainly used for negotiation, state control, and corresponding network management among communication nodes. The data segment is used to carry communication data. The access arbitration segment including the node ID, which is a key part of the frame structure designed in this protocol, can be exploited to significantly solve the problem of multi-node contention access collision in the time synchronization network.
2.3 Working process
Figure 3 shows the arbitration process of the UVLLC protocol. Taking three nodes A, B, C as an example, the priority is A > C>B. Assume that all nodes need to access the channel and send the data frame in the above-defined format. Firstly, the 4th-bit information of Node B is “1” (no-light), and the 4th-bit information of Nodes A and C is “0” (light). According to the principle of light power superposition specified in Section 2.1, the state of the optical channel is “0” (light) at this time. Thus, Node B exits the channel contention. Similarly, the 2nd-bit information of Node C is “1” (no-light), and the 2nd-bit information of Node A is “0” (light). At this time, the channel state is “0” (light) after the optical power is superimposed, and Node C exits the channel contention. Therefore, Node A wins the channel access contention among the three nodes. In this contention access process, there is no need to introduce complex collision avoidance and time back-off algorithms, so the process of node contention access is greatly simplified.
Figure 4 gives a typical communication process and steps based on the UVLLC-MAC protocol briefly. The node priority is specified as: A > B>C.
Step 1. All Nodes A, B, and C in the UV communication network are initially in the monitoring state to detect optical channel signals to ensure the strict alignment of the time slot. All nodes that need to access the channel send data frames when they detect that the channel is idle.
Step 2. Among all the nodes participating in the channel access contention, high-priority Node A eventually wins the multi-node contention because Nodes B and C withdraw from the contention according to the UVLLC-MAC protocol.
Step 3. Node A starts to transmit the subsequent control segment and data segment. Nodes B and C suspend the current data transmission and enter the monitoring mode.
Step 4. Once Node A completes the data transmission process, it immediately re-enters the channel monitoring state. And Nodes B and C enter the step 1 again to start a new round of contention for the channel access.
3. Simulation and analysis
A four-node UV communication network model is established to verify the UVLLC-MAC protocol, and numerical calculation and simulation analysis on its protocol performance are performed using the Monte Carlo method. The four-node UV communication network model is constructed as shown in Fig. 1, and the priority of the four nodes is A > B>C > D.
3.1 Network throughput
Throughput is one of the important network performance parameters, which is mainly used to evaluate the carrying and service capabilities of the network. Suppose the frame length is fixed to L bit, and the number of frames successfully transmitted per unit time is n, then the throughput can be expressed as n×L bps.
Usually, the channel transmission rate R bps is used to normalize the throughput. Therefore, the normalized throughput S is defined as [15]
where T is the average transmission time of a single frame. The normalized throughput is related to multiple factors such as channel utilization, protocol contention mechanism, and collision time of data frame in the access protocol. In the designed MAC protocol, the normalized throughput S of each node can be specifically defined as PA, PB, PC, PD are the probability of frames successfully transmitted of Nodes A, B, C, and D in each time slot, respectively.Since all nodes in the UV network share a channel, and only one node can use channel resources in the same time slot, the total normalized throughput of the network is
The basic model of the UV terminal node without buffer is analyzed firstly, and the network throughput of the UVLLC-MAC protocol is numerically simulated and researched. In this model, when a contention access collision occurs among nodes, the data sent by low-priority nodes cannot be reserved due to the lack of cache space. The network parameter settings are shown in Table 1.
Table 1. Simulation parameter settings
Figure 5(a) shows the normalized throughput of Nodes A, B, C, and D varies with the data generation probability P. For comparison, the normalized throughput of four nodes with the slotted ALOHA protocol was also simulated, as shown in Fig. 5(b).
Fig. 5. Relationship between normalized throughput and data generation probability without buffer: (a) the UVLLC-MAC protocol; (b) the ALOHA protocol.
Figure 5(a) indicates the channel is gradually occupied by the highest-priority Node A, and the normalized throughput of other low-priority nodes gradually degrades as the data generation probability increases. The throughput of lower-priority nodes decreases more obviously.
Figure 5(b) depicts that the total normalized throughput of the slotted ALOHA protocol decreases drastically after a brief increase. There is no difference in the priority of Nodes A, B, C, and D and the throughput variation trend of the four nodes is the same, resulting in the overlap of the change curves. When the data generation probability of each node increases, the contention among nodes becomes more intense, causing the insufficient utilization of channel resources.
However, even as P increases, high-priority nodes can still obtain the right to use channel resources under the UVLLC-MAC protocol, ensuring that the normalized throughput of the entire network continues to increase. When the data generation probability is 0.3, the normalized throughput of the ALOHA protocol is 0.42, which is only 51.2% of the UVLLC protocol. And the peak network throughput of the UVLLC protocol increased by 77.8% compared to the ALOHA protocol.
In the UVLLC-MAC protocol without buffer, low-priority nodes often lose data packets due to contention failures. A certain amount of the data cache space is introduced at each terminal node to alleviate this problem. When the node contention fails, the data is transferred to the buffer and waits for the next time slot to be retransmitted.
Figures 6(a), (b) and (c) respectively show the change of the normalized throughput versus the buffer capacity for three values of P under the UVLLC-MAC protocol. In the simulation process, the data generation probability P of the system is set to 0.2, 0.3, 0.7, corresponding to the light load, medium load, and heavy load of the network service, respectively. The network parameter settings are shown in Table 1.
Fig. 6. Normalized throughput of the UVLLC-MAC protocol with buffer: (a) P=0.2; (b) P=0.3; (c) P=0.7.
Figure 6(a) indicates the normalized throughput of each node and the total change with the cache capacity under light load conditions (P=0.2). We observe that the throughput of each node is effectively increased as the cache increases. Among them, Node A with the high-priority always has a higher data processing capability. Since the probability of data generation P is relatively low and the channel access contention is not fierce, all nodes have better normalized throughput as the cache capacity increases.
Figures 6(b) and 6(c) depict that Node A with the high-priority has always maintained the highest throughput. As the cache capacity increases, the normalized throughput of Node B with the sub-priority gradually increases. For Node D with the lowest-priority, its throughput gradually decreases and more channel resources are occupied by high-priority nodes as the buffer capacity increases.
Besides, we find that the total normalized throughput of the system is close to 1 under medium load and heavy load conditions while the total normalized throughput is stable at about 0.8 under light load conditions, which shows the system channel has been fully used. However, the fairness of the network has deteriorated significantly under heavy load conditions, and there are “starve to death” situations at low priority nodes. This is not in line with the original intention of the network system design. Thus, the excessive load of the system should be avoided as much as possible.
3.2 Network delay
When the access protocol introduces the cache setting, the network delay will inevitably increase. The delay includes transmission delay, propagation delay, and queuing delay. Transmission delay τ1 depends on the length of the transmitted data L bit and the channel transmission rate R bps,
Propagation delay τ2 is the time required for an optical signal to travel a certain distance d in the channel at the speed of light c, Because the speed of light c is extremely high, the propagation delay is generally negligible. Queuing delay includes sending waiting time and receiving processing time. The former is also called access delay, which represents the time interval from when data needs to be sent to when data is sent. The access delay can be used to reflect the access efficiency of a single node. And the network delay is mainly caused by the access delay [16]. We only consider the access delay part when discussing the impact of the cache capacity on the network delay.Figures 7(a) and 7(b) indicate the variation of network delay with the buffer capacity under the condition that the data generation probability P is 0.2 and 0.3 respectively. The simulation results show that the average delay of the light-load network is stable at about 25ms. The average delay of the network node is inversely proportional to the node priority. That is, the higher the node priority is, the lower its average delay is. For low-priority nodes, the data will wait longer in the cache when the cache area is larger, and the average delay will increase accordingly.
Through the comparison, we notice that the delay of low-priority nodes will increase faster when the system load becomes larger. In Fig. 7(b), network delays of Nodes A, B, and C are all less than 200ms when the cache capacity is relatively large. While average delay of Node D reaches nearly 1s, which can only complete communication services with low real-time requirements.
4. UV experiment network
To verify the effectiveness of the UVLLC-MAC protocol, UV physical networking experiments were carried out and a UV communication network with four terminal nodes was implemented.
As is showed in Fig. 8, the hexahedron structure is used in the UV communication terminal. Each side has separate transmitting and receiving devices. A full-duplex communication method is used, and the UV terminal node can send and receive data at the same time. Therefore, the direction of the nodes need not be adjusted deliberately.
The optical power of a single UV source is 102mw and the wavelength is 265nm which is in the “solar blind” region of UV. A 1×6 array of light sources is adopted at the transmitting end, and the optical power of a single plane of a single node is 6×102 = 612mw. The communication radius of the nodes can reach tens to hundreds of meters and the communication link between any two nodes in the network is reachable in one-hop. In the communication process, the photodetector needs to be selected reasonably to ensure that the superimposed optical power value does not exceed the threshold of the photodetector when all independent light sources emit light. That is, the superimposed optical power is within the dynamic range of the photodetector. The rise time of the photodetector is 0.57ns and its response time can even be on the order of nanoseconds. The four UV Nodes A, B, C and D are placed as shown in Fig. 9. In the experiments, the data generation probability P is 0.3 and the cache capacity is 5.
Figure 10 illustrates intuitively that data can be transmitted normally between any two nodes, and high-priority nodes have better data processing capability. The effective communication among four UV terminal nodes is completely realized under the UVLLC-MAC protocol. The experimental results denote that the effective communication rate is about 115kbps, and the average delay of the high-priority nodes in the network is about 100 ms, while the average delay of the lowest-priority node is about 500 ms. In the simulation section, Fig. 7(b) shows the delay of the lowest-priority node is 410 ms under the condition that P is 0.3 and the cache capacity is 5. By comparing the simulation results and the experimental results, we find that the communication rate is roughly the same (about 100kbps), but the delay of each node in the experimental results is higher than that in the simulation results. In the simulation process, only the access delay is considered because the access delay is the main factor affecting the network delay [16]. However, the network delay includes transmission delay, propagation delay, and queuing delay, so the delay in the experimental results is relatively high.
5. Conclusion
A novel UVLLC-MAC protocol is proposed in this paper. The basic working mechanism and protocol flow are given in detail, and the performance of the UV communication network is simulated and analyzed mathematically. The simulation illustrates that high-priority nodes have ideal throughput and relatively short delay, and can be used as key nodes to complete the transmission of important information. Finally, the actual networking of the four communication terminals is designed and implemented, and the effectiveness of the lossless contention access mechanism is verified through experiments. Furthermore, the comparison between the simulation results and the experimental results is completed and the differences are analyzed.
Given the fact that high-priority nodes always have priority to use the channel, the communication scheduling and flow balance of nodes with different priorities need to be considered as a whole in the upper application to ensure certain network fairness. The directions for future research is dynamic adjustment of node priority to enhance networking performance and to apply the UVLLC protocol to the low rate communication of mobile terminals in the complex electromagnetic environment as a supplement to the traditional communication means.
Funding
Research Center of Optical Communications Engineering & Technology ((ZXF201901).); National Natural Science Foundation of China (62171463).
Acknowledgments
This work was supported by the National Natural Science Foundation of China (62171463).
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. V. W. S. Chan, “Free-Space Optical Communications,” J. Lightwave Technol. 24(12), 4750–4762 (2006). [CrossRef]
2. T. Y. Aung, S. Arya, and Y. H. Chung, “Performance dependence of non-line-of-sight ultraviolet communications on atmospheric parameters of the ultraviolet channel,” Opt. Commun. 443, 7–11 (2019). [CrossRef]
3. W. Hou, C. Liu, F. Lu, J. Kang, Z. Mao, and B. Li, “Non-line-of-sight ultraviolet single-scatter path loss model,” Photon. Netw. Commun. 35(2), 251–257 (2018). [CrossRef]
4. R. J. Drost and B. M. Sadler, “Survey of ultraviolet non-line-of-sight communications,” Semicond. Sci. Technol. 29(8), 084006 (2014). [CrossRef]
5. H. Xiao, Y. Zuo, J. Wu, Y. Li, and J. Lin, “Non-line-of-sight ultraviolet single-scatter propagation model in random turbulent medium,” Opt. Lett. 38(17), 3366–3369 (2013). [CrossRef]
6. T. F. Zhao and X. Z. Ke, “Monte Carlo simulations for non-line-of-sight ultraviolet scattering coverage area,” Acta Phys. Sin. 61(11), 114208 (2012). [CrossRef]
7. D. Han, X. Fan, K. Zhang, and R. Zhu, “Research on multiple-scattering channel with Monte Carlo model in UV atmosphere communication,” Appl. Opt. 52(22), 5516–5522 (2013). [CrossRef]
8. P. Song, X. Z. Ke, and F. Song, “Multi-user interference in a non-line-of-sight ultraviolet communication network,” IET Commun. 10(13), 1640–1645 (2016). [CrossRef]
9. X. Sun, Z. Zhang, and A. Chaaban, “71-Mbit/s ultraviolet-B LED communication link based on 8-QAM-OFDM modulation,” Opt. Express 25(19), 23267–23274 (2017). [CrossRef]
10. C. Li, J. H. Li, Z. Y. Xu, and J. Y. Wang, “Study on the k-Connectivity of Ultraviolet Communication Network under Uniform Distribution of Nodes in a Circular Region,” 2019 IEEE ICCC, 797–802 (2019).
11. C. Li, J. H. Li, Z. Y. Xu, and J. Y. Wang, “Study on the k-Connectivity of UV Communication Network Under the Node Distribution of RWP Mobility Model in the Arbitrary Polygon Area,” IEEE Photonics J. 12(4), 9–12 (2020). [CrossRef]
12. T. Zhao, Y. Xie, S. Xu, J. Wang, and H. Wang, “Flocking of UAV Formation with Wireless Ultraviolet Communication,” IEEE Wirel. Commun. 25(3), 58–64 (2018). [CrossRef]
13. H. Qin, Y. Zuo, F. Li, R. Cong, L. Meng, and J. Wu, “Analytical link bandwidth model based square array reception for non-line-of-sight ultraviolet communication,” Opt. Express 25(19), 22693–22703 (2017). [CrossRef]
14. S. Arya and Y. H. Chung, “Novel Indoor Ultraviolet Wireless Communication: Design Implementation, Channel Modeling, and Challenges,” IEEE Syst. J. 5(99), 2349–2360 (2021). [CrossRef]
15. Y. Li, J. Ning, Z. Y. Xu, S. V. Krishnamurthy, and G. Chen, “UVOC-MAC: a MAC protocol for outdoor ultraviolet networks,” Wirel. Netw. 19(6), 1101–1120 (2013). [CrossRef]
16. M. Chen, Y. L. Xiao, S. T. Wang, and H. T. Li, “Analysis of TDMA time slot allocation algorithm in ultraviolet Ad Hoc network,” Optical Communication Technology7(8) (2016).
17. J. B. Li, X. G. Liu, H. X. Wang, and X. J. Wu, “Improved Ultraviolet Communication Time Division Multiplexing Network Method,” Chinese Journal of Lasers 41(11), 132–137 (2014).
