Open Access
Add to BibSonomy
Abstract
Oversampled modulation (OM) and undersampled modulation (UM) are two commonly used optical camera communication (OCC) modulation schemes for high-speed communication in short-range and reliable communication at long distances, respectively. However, the relationship between these two schemes and the tradeoff in performance for arbitrary communication distances have not been thoroughly investigated. In this study, we analyze the impact of distance and modulation parameters on pixel efficiency and packet delivery rate performance, demonstrating the underlying unity of traditional OM and UM schemes. Furthermore, we propose a generalized modulation scheme that allows for achieving predefined link performance at a given distance by adjusting the modulation parameters, such as packet length and repetition counts. Simulation and experimental results show that the proposed generalized modulation scheme provides OCC with a unique distance-aware capability other than the traditional OM and UM schemes, which are two special cases focusing on effectiveness and reliability, respectively. This research enhances our understanding of OCC data modulation and establishes a theoretical foundation for achieving efficient and reliable OCC transmission in complex environments.
© 2024 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Optical camera communication (OCC) is a form of visible light communication (VLC) that utilizes a light-emitting diode (LED) as the transmitter and an image sensor as the receiver [1,2]. Data bits are modulated by controlling the intensity of the emitted light in the transmitter and recognized by an image sensor in the receiver by analyzing the spatial and/or temporal distribution of pixel values in the sequence of the output image frames. Since the frame rate of the image sensor is as low as tens of frames per second, modulating the light intensity as the camera’s frame rate results in perceivable flicker for human eyes, which degrades the illumination quality, and thus violates the coexistence principle [3]. To address the conflict between the camera’s frame rate and non-flicker frequency, especially for cameras with a rolling shutter (RS), two types of on-off keying (OOK) based modulation schemes have been developed, namely, undersampled modulation (UM) and oversampled modulation (OM) [4].
The modulation schemes are "undersampled" because the sampling frequency capturing the intensity of the light source is much lower than the Nyquist frequency of the modulated signal [4–6]. As a result, the bit rate of the UM schemes is extremely low (about 0.5 to 1 bit/frame) limited by the camera’s frame rate, but the required number of pixels in each image frame is minimized. By utilizing the RS mechanism present in most complementary metal oxide semiconductor (CMOS) image sensors, the intensity of the light source can be independently and sequentially sampled by the pixel rows in each image frame, resulting in an increased sampling rate multiplied by the number of pixel rows [7]. This mechanism serves as the basis of the OM schemes. The bit rate of these schemes is thus enhanced compared with UM schemes because more bits can be transmitted during each image frame [8–11].
OCC is considered distance sensitive because the choice of modulation schemes and its performance is significantly influenced by the communication distance [12–14]. Conventionally, OM is preferable for most short distant communication scenarios where the light spot is large enough to cover most pixel rows in the output image [15]. The RS mechanism allows exchanging the redundancy in space or pixel domain for the sampling rate in the time domain, in a manner of "space for time", which supports high-speed data transmission. However, OM schemes are infeasible for most long-range OCCs because the shrinking light spot leads to insufficient space redundancy to compensate for the sampling rate. In this case, UM is a better option because the data reception from a small light spot is guaranteed in UM schemes by minimizing the modulation and data rates.
Conventionally, OM and UM are regarded as distinct modulation modes suitable for short and long distances, respectively. However, a unified view of OCC modulation has not been extensively explored in arbitrary communication distances. Furthermore, a comprehensive framework of performance analysis is needed to investigate the effects of modulation structural parameters and communication distance. In addition, previous research has not addressed a generalized modulation scheme from a system perspective, that guarantees a predefined trade-off between reliability and efficiency for a given communication distance. These limitations undermine the system design and optimization in scenarios with significant variations in communication distances, such as vehicle to everything (V2X) via OCC [16,17].
The length and repetition count of a data packet serve as a pivotal factor for data transmission in OCC, which enables a trade-off between reliability and effectiveness across diverse communication distances. It reveals that the OM and UM are not separate modulation modes, but rather two specific manifestations of the same modulation scheme that emphasizes either effectiveness or reliability at different communication distances. In this paper, we analyze the impacts of packet length and repetition count on the performance of the packet delivery ratio (PDR) and pixel efficiency (PE). Based on this analysis, a generalized modulation scheme is proposed for OCC with variance communication distance. This approach allows for a predefined trade-off between transmission reliability and efficiency by tuning the modulation parameters in communication scenarios with different distance requirements. Theoretical analysis, numerical simulation, and experimental results demonstrate the availability of the proposed generalized modulation via a performance transition from short to long distance. This highlights the intrinsic unity of OM and UM within our proposed approach and reveals the fundamental principle of data modulation for OCC.
The main contributions of this work are as follows.
- (1) Utilizing the imaging mechanism of a rolling shutter camera, the intrinsic unity of traditional OM and UM modulation modes is explored, which develops the fundamental principles of OCC data modulation.
- (2) A concept of equivalent gap interval is utilized to describe the reduction in spot size resulting from increased distance. Based on this concept, the impacts of data frame structure parameters (i.e., packet length and repetition count) on packet delivery ratio and pixel efficiency are quantitatively analyzed and the optimization strategies are investigated in different scenarios.
- (3) A generalized modulation scheme is proposed that provides a theoretical foundation for designing distance-aware OCC systems with rolling shutter cameras. This scheme enables a predefined trade-off between reliability and efficiency across different communication distances by tuning the length and repetition count of transmitted data packets.
The rest of this paper is arranged as follows. The models of the OCC system are presented in Section 2. The proposed generalized modulation scheme and performance analysis are detailed in Section 3. Section 4 includes the simulation and experiment results, and Section 5 concludes this paper.
2. System model
2.1 OCC system with rolling shutter-based camera
A typical OCC system consists of a transmitter and receiver, as shown in Fig. 1. The transmitted bits are organized into a data frame structure and then encoded using a running-length line (RLL) encoder to ensure an equal distribution of binary levels [18]. The encoded waveform is transmitted by controlling the LED light source’s state through a driver circuit. In the receiver, a digital camera equipped with a rolling shutter captures the light source’s states and stores them in a series of image frames. After RLL decoding and de-framing, the received bits are identified based on the patterns of the output image frames [19].
The camera plays a crucial role at the receiver of OCC by converting the intensity of incident light into an electrical signal. A typical camera is modeled as a combination of an optical system and photoconverter, which are commonly realized by optical lenses and an image sensor, respectively [20]. The image sensor usually outputs images at regular intervals, which determines the frequency at which image frames are produced by the sensor, referred to as the frame rate. To ensure that the pixel values in the resulting image accurately represent the spatial distribution of the incident light’s intensity, image sensors utilize a "shutter" mechanism to regulate when and how long each pixel is exposed to the light source in generating each image frame, known as exposure control. Global shutter (GS) and rolling shutter (RS) are two major types of exposure control methods exploited in charge-coupled device (CCD) and CMOS image sensors, respectively. In GS mode, all the pixels in an image are exposed simultaneously, while, in RS mode, the pixels are exposed sequentially in rows with a fixed time delay.
The exposure model depicted in Fig. 2 outlines the operation of the rolling shutter-based image sensor. Here, $T_0$ represents the time interval of output images, which is the inverse of the frame rate. $T_e$ signifies the exposure duration for each row of pixels, dictating how long they are exposed to the light source. $T_r$ accounts for the readout time, including the time required to convert the induced voltage to pixel value by an analog-to-digital converter (ADC). As CMOS image sensors typically share a common set of ADC devices, $T_r$ also serves as the sampling interval for the light source state between adjacent pixel rows. In addition, a non-exposure gap interval $T_g$ appears between adjacent image frames, during which any intensity changes of the light source can not be recorded in the output images.
2.2 Normalization of communication distance and packet length
Within the framework of the rolling shutter mechanism, Fig. 2 describes the process of translating the information from the temporal domain to the pixel space. Consequently, the variables altering the distribution of pixel values and thereby influencing the communication performance of OCC can be investigated by translating their impacts into the temporal domain consistent with the transmitted signal. Leveraging the exposure model of the rolling shutter, this paper translates the impacts of communication distance and packet length on the output images into the temporal domain and provides the normalized metrics by the frame interval for analyzing the performance of the proposed modulation scheme.
While the intrinsic gap interval $T_g$ remains constant for a specific image sensor, combining the exposure model in Fig. 2 with the pinhole model [21,22] suggests that the increase in communication distance is equivalent to an increase in an "equivalent" gap interval. Denoting $T_g^{\prime }$ the equivalent gap interval, we have
where $\delta$ is the extra gap interval caused by increased communication distance. As shown in Fig. 3, when the camera is close enough to the light source with unit size and the entire image is covered by the light spot, $T_g^{\prime } = T_g$ as $\delta$ approximates to 0. As the camera is moving away from the light source, $\delta$ increases proportionally as the light spot shrinks accordingly. When the transceivers are significantly separated, the light spot in the output image will shrink to a single point. In this case, $\delta$ is large enough to make the equivalent gap interval occupy the entire frame interval $T_0$, i.e., $T_g^{\prime } \approx T_0$.By comparing the equivalent gap interval to the frame interval, a normalized metric of communication distance $\rho$ is defined within the range of $[T_g/T_0,1)$, as
Given that the size of the light spot in the output image, or the equivalent gap interval, is determined by not only the communication distance but also the light source size and image sensor parameters, the normalized communication distance denoted by $\rho$ for a specific image sensor is discussed assuming a unit-sized light source.
The packet length $N_f$ represents the number of symbols in a data packet. Assuming that the frame rate of the image sensor is $F_f = 1/T_0$ and the duration of each data symbol is $T_s$, the time required to transmit a data packet is $N_fT_s$. Therefore, comparing the transmission time of a single packet with $N_f$ symbols to the frame interval $T_0$, a normalized packet length is defined as
Due to the non-exposure gap interval, a data packet lasting longer than $T_0$ will not be correctly received, resulting in $\eta \in (0,1)$.
However, due to the asynchronous transmission characteristics inherent in OCC, even short data packets may not be reliably received. To maintain the integrity of received data, the same packet is transmitted repeatedly. Denoting the repetition count for each packet as $M$, existing literature typically sets $M$ to 2 or 3. However, in this study, $M$ is a variable parameter within the proposed generalized modulation scheme. The paper investigates the impact of different combinations of normalized packet lengths and repetition counts, denoted as ($\eta$, $M$), on communication reliability and effectiveness, given a normalized communication distance, $\rho$. This analysis serves as the fundamental for optimizing the modulation parameters in distance-aware OCC.
3. Generalized modulation for OCC
In this section, we provide a step-by-step implementation of the proposed generalized modulation scheme, along with the corresponding demodulation methods. In addition, the resulting performance trade-offs between reliability and effectiveness are explained in terms of the relationship between diversity and multiplexing, which clarifies the generalization of the traditional OM and UM schemes.
3.1 Symbol modulation and data framing
In optical communication, the simplest mapping scheme from bits to symbols is known as on-off keying (OOK) modulation. The OOK-modulated symbols are packed into packets and frames for efficient transmission. As shown in Fig. 4, a data frame contains several data segments, and each segment is comprised of $M$ repetitions of a data packet with $N_f$ symbols [19]. In addition to $N_p$ payload symbols, a data packet includes $N_h$ header, $N_t$ footer, and $N_c$ control symbols for synchronization, i.e.,
3.2 Structure of generalized modulation
An overview of the proposed generalized modulation and demodulation processes is illustrated in Fig. 5.
3.2.1 Modulation
The generalized modulation is implemented in the following steps:
Step M1: Determination of modulation parameters. The objective of this step is to determine an appropriate combination of modulation parameters ($N_f$ and $M$) given the specified communication distance $d$ that satisfies specific reliability and effectiveness requirements. This is achieved by selecting a combination of $(\eta, M, \rho )$ referring to the performance trade-offs detailed in the next subsection and de-normalizing using Eqs. (2) and (3).
Step M2: Packeterization. An individual data packet is constructed by arranging a specific number of payload bits (i.e., $N_p$) within the packet structure based on the selected parameter (i.e., $N_f$), referring to Fig. 4. To avoid the payloads coinciding with the header, control bits are inserted into the payload sequence every $N_h-2$ successive payload bits, with each state being the opposite of the previous payload bit [18]. For example, the available number of payload symbols is $N_p=7$ with $N_c=2$ control symbols, given $N_f=15$, $N_h=5$, and $N_t=1$.
Step M3: Data framing. The proposed generalized modulation constructs a data frame based on the packet structure through duplication and concatenation. The bit sequence with a length of $N_b$ from the upper layer is equally split into $n=\lceil N_b/N_p\rceil$ segments. Then, each segment is arranged into a packet structure as the payloads. In addition, each packet is duplicated for $M$ times, resulting in $n$ individual data packets with $M$ duplications for each. Finally, these duplications are concatenated one by one to generate the modulated data frame.
Step M4: RLL coding and data transmission. To prevent flickering and support precise dimming control in OCC, it is essential to use a running-length line (RLL) coding scheme to balance the distribution of the high and low levels in the modulated waveform [17]. There are several RLL coding schemes available, including Manchester, 4B6B, 8B10B, etc. In the proposed modulation scheme, Manchester coding is utilized, where "0" and "1" bits in the modulated data frame are mapped to falling and rising edges, respectively. The data frame is then transmitted via an LED light source by feeding the encoded waveform to the driver circuit.
3.2.2 Demodulation
Demodulation processes are performed at the receiver including the following steps:
Step D1: Image preprocessing. Image preprocessing aims to derive a binary sequence from the image frames generated by the image sensor, which represents the varying patterns of light intensity emitted by the light source as much as possible. To accomplish this objective, two tasks are undertaken. Firstly, the pixel lines illuminated by the light source are compressed into a sequence of pixel values. This can be achieved by selecting a single column for simplicity or by averaging all columns to eliminate the noise effect. Secondly, binarization is performed to assign each pixel value a binary value of either "1" or "0" based on a specified threshold. The determination of the optimal threshold is crucial to ensure the demodulation performance, and the choice between fixed or adaptive thresholding algorithms depends on the distribution of the pixel values.
Step D2: RLL-Decoding. In the case of utilizing Manchester code in the RLL coding scheme, the conversion of rising and falling edges of "01" and "10" in the binary sequence is involved in this step as a counterpart by applying the conventional Manchester decoding algorithm.
Step D3: De-packeterization. The inclusion of control bits within the payloads enables clear distinction of the sync header, facilitating easy recognition of data packets within the received binary sequence through a simple sliding window detection. Subsequently, footer verification is conducted for each header to discard any incomplete data packets. For each complete packet, the payload bits therein are restored by removing the control bits.
Step D4: Duplication removal. Since a single packet might be received multiple times because of repetitive transmission strategy, duplication removal, and missing identification are crucial in payload reconstruction. This is accomplished by verifying the sequence number (SN) attached to the payloads in received packets. Only one copy of payloads with the same SN is retained to eliminate duplications while missing data is identified through successive payloads with discontinuous SNs. After that, a bit stream is generated from the payloads in received data frames and submitted to the upper layer.
3.3 Performance analysis
The performance of an OCC link can be evaluated based on reliability and effectiveness in data transmission. Due to the noise and exposure effect, the reliability of an OCC link can be evaluated by measuring the error rate at bit or symbol level using the metric of bit error rate (BER) or symbol error rate (SER) [23–26]. However, this paper specifically focuses on investigating how the modulation structure and distance impact the reliability performance at the frame level. In addition, unlike existing research that primarily analyzes and measures the data rate [27–29], this paper emphasizes the utilization efficiency of pixels in the image sensor, as it plays a crucial role in determining the achievable data rate. Therefore, a novel performance analysis framework is proposed in this paper, which utilizes the metrics of packet delivery ratio and pixel efficiency to assess the reliability and effectiveness of data transmission on OCC links, respectively.
3.3.1 Packet delivery ratio
The packet delivery ratio is defined as the ratio of the number of independent packets that are successfully delivered to the total number of packet transmissions. At the frame level, the successful delivery denotes that at least one of the repetitive transmissions of a data packet is identified at the receiver, and the payloads are successfully decoded following the packet structure as shown in Fig. 4.
From the system model and exposure mechanism, it is evident that the length and repetition count of a data packet are crucial modulation parameters influencing the probability of successful packet delivery, i.e., PDR. The optimal parameter combination maximizing this probability varies across different communication distances. Consequently, we evaluate the reliability of OCC at the frame level via PDR within a normalized parameter space, i.e. ($\eta, M, \rho$), through numerical simulations and experiments.
3.3.2 Pixel efficiency
In OCC, the light intensity emitted by the light source is estimated by analyzing the pixel values of the output image captured by the image sensor. Based on this estimation, the transmitted data bits are determined accordingly. If we consider each pixel in the rolling shutter-based image sensor as an independent communication channel, the data rate of the OCC can be described as
where $h$ denotes the number of pixels in each output image that are illuminated by the light source, and $\gamma$ represents the information capacity that can be represented by each pixel in a single image frame, measured in bits per pixel. According to Eq. (5), the transmission rate of the OCC system is theoretically determined by $\gamma$, given the specified image sensor and communication distance. Therefore, similar to the concept of spectrum efficiency in wireless communication, we define $\gamma$ as the pixel efficiency in OCC, which characterizes the communication effectiveness.The pixel efficiency is inversely proportional to the number of pixels needed to represent a modulation symbol in the output image, denoted as $N_{\text {pps}}$. For image sensors in GS mode, $N_{\text {pps}}$ corresponds to the total number of pixels within the illuminated area. Conversely, in RS mode, $N_{\text {pps}}$ equals the number of rows included in each dark or bright stripe, that is
The pixel efficiency for the proposed modulation scheme is derived as
It indicates that duplicated copies of the packet and the overheads therein increase the $N_{\text {pps}}$ equivalently and reduce the pixel efficiency, namely the transmission effectiveness. Using Eqs. (3), (4) and (6) in Eq. (7), we derive
3.3.3 Performance trade-offs
From the perspective of a visual multiple in multiple out (MIMO) system, the reliability and effectiveness of OCC data transmission can be characterized in terms of the array gain or degrees of freedoms (DoFs) provided by the image sensors [30–32]. Since the total DoFs are determined given the image sensor and communication distance, the choice of packet length and repetition count in the generalized modulation scheme determines the allocation of DoFs on obtaining multiplexing or diversity gains, respectively, which in turn determines the link performance with better effectiveness or reliability. Therefore, the proposed generalized modulation provides OCC with a unique capability of distance-awareness since a predefined performance trade-off is achieved at arbitrary communication distance by tuning the modulation parameters.
4. Simulation and experiment
In this section, numerical simulations are conducted on the performance metrics of PDR and PE to demonstrate the trade-offs given communication distance. The simulation results serve as a guideline for determining the appropriate modulation parameters. In addition, experimental results are presented to demonstrate the effects of packet length and repetition counts on the performance of PDR, as well as the distance-aware capability provided by the generalized modulation scheme.
4.1 Numerical simulation
Given the challenge of deriving an analytical expression of PDR for the parameter set $(\eta, M, \rho )$, the numerical simulation results shown in Fig. 6 illustrate the average PDR performance across different distances for combinations of the length and repetition of data packets.
Fig. 6. Numerical simulation of average PDR under various packet lengths and repetition counts in different communication distances, where a smaller value of $\eta$ represents a shorter data packet while a larger value of $\rho$ indicates a longer communication distance.
To illustrate the performance of pixel efficiency, a numerical simulation is conducted under specific conditions, where $T_r=2.7$ $\mu$s, $T_0=1/F_f=1/30$ s, $H=1080$, and $T_s=1$ ms. The results for different repetition counts (i.e., $M=1, 2, 3, 4$) are compared and depicted in Fig. 7.
Fig. 7. Numerical simulation of pixel efficiency under various packet length and repetition counts given a short OCC link, where $T_r=2.7$ $\mu$s, $T_0=1/F_f=1/30$ s, $H=1080$, and $T_s=1$ ms. Note that PE approximates to a minimum as the data packet is short enough to comprise only one symbol, like undersampled phase shift OOK (UPSOOK).
The simulation results depicted in Fig. 6 and Fig. 7 validate the performance trade-offs between reliability and effectiveness. First, at a specified communication distance (or equivalently $\rho$), reducing the value of $N_f$ (or equivalently $\eta$) leads to an increase in PDR and a decrease in PE. This is attributed to the higher probability of shorter packets aligning within the light spot, resulting in complete reception. Nevertheless, shorter packets carry fewer payloads, causing a decrease in transmission efficiency. For the same packet length, more repetitions enhance PDR by introducing redundant transmission, at the cost of reducing PE. Second, as the communication distance increases, the size of the light spot decreases, resulting in a decrease in both PDR and PE. In such a situation, increasing the repetition count might be necessary to meet the PDR performance requirements.
4.2 Experiment setup
An experimental environment is established to demonstrate the performance of the proposed generalized OCC modulation scheme. The experimental setup is depicted in Fig. 8 and the corresponding parameters are listed in Table 1.
Table 1. Parameters of the experiment environment
In the transmitter, data modulation is carried out using MATLAB on a PC, where a random bit stream is generated and modulated into a data frame based on specified packet length and repetition count parameters. The modulated data frame is then converted into a waveform with binary levels using an MCU. This waveform is repetitively fed to a driver circuit through a GPIO, which controls the blinking LED for data transmission. In the receiver located at a distance from the transmitter, a rolling shutter camera is utilized to capture a video clip. The pattern of image frames within the video clip is extracted and analyzed using MATLAB on a PC. Following data demodulation, the performance of the average PDR for each packet can be measured by counting the received independent packets compared with the total number of packet transmissions.
4.3 Results
For short-range communication (i.e., $T_g^{\prime }=T_g$ and $\rho \approx 0.02$), we evaluate the average PDRs through experiment for the normalized packet length $\eta$ ranged from $0.1$ to $0.9$, and repetition count for $M=$1, 2, 3, 4, respectively. The results are shown in Fig. 9 and compared with the simulation results for reference. The average PDR is retained for short packets but drops significantly as the packet length exceeds a certain limit. In addition, increasing the repetition counts for each packet is a feasible way to improve PDR for a long packet. A good match is observed between the simulation and experiment results, which justifies the availability of the proposed modulation scheme. These findings align with the typical OM schemes used in short-distance OCC. In such schemes, due to the large amount of pixels available for the trade-off between effectiveness and reliability, long data packets with a minimum of three repetitions are utilized to maximize pixel efficiency while meeting the minimum PDR requirement.
Fig. 9. Experimental PDR of generalized modulation with different packet length and repetition counts in short communication scenario.
The experimental evaluations of averaged PDR are conducted for short and long packets (i.e., $\eta =0.3$ and $\eta =0.7$, respectively) with varying repetition counts as the communication distance increases (from 0.3 m to 2.0 m). The results in Fig. 10 demonstrate that the average PDR decreases for both packet lengths as the communication distance increases. Moreover, it is observed that, compared to a short packet, a long packet is more susceptible to degradation in long-range communication. Additionally, increasing the repetition counts for each packet helps maintain PDR, but at the expense of reduced PE. This can be attributed to the fact that the number of "data bearing" pixels decreases as the communication distance increases, resulting in a reduction in available DoFs for trading effectiveness with reliability. Consequently, the utilization of limited DoFs to enhance effectiveness or reliability depends on the packet length and repetition counts, respectively. It can be inferred that when the communication distance is long enough for the light spot to occupy only one pixel, a viable approach to establish a reliable communication link is to repeat a one-symbol packet throughout the entire frame interval. This scenario coincides with traditional UM schemes like undersampled phase shift OOK (UPSOOK).
Fig. 10. Experimental PDR of generalized modulation with a short packet length in different communication distances.
4.4 Discussion
The experiment results show that the communication distance plays a pivotal role in determining the maximum number of rows of effective pixels available for data transmission. The packet length and the repetition count employed in the data modulation scheme dictate the allocation of these pixel rows. Some are designated for independent data transmission, contributing to multiplexing gain, while others are utilized for redundancy, enhancing diversity gain. Adjusting the combination of these two parameters for a given communication distance allows us to strike a balance between effectiveness and reliability in communication performance.
When the communication distance is short, employing a long packet (e.g., $\eta = 0.45$) in the data modulation scheme enhances transmission effectiveness. To mitigate the impact of non-exposure time, several repetitions (e.g., $M = 2$ or 3) are utilized to ensure the reliability meets the requirements through moderate diversity (e.g., PDR $\ge 0.99$). In scenarios with long communication distances, on the contrary, the spot size is smaller, and there are fewer rows of effective pixels. To ensure transmission reliability, one has to compromise on effectiveness to minimize the packet length (i.e., multiplexing) and maximize the repetition count (i.e., diversity).
Therefore, the OCC system utilizing the generalized modulation scheme is distance-aware because the trade-offs between effectiveness and reliability at different distances are predefined. This can be accomplished by picking different modulation parameters emphasizing either multiplexing or diversity.
5. Conclusion
This paper investigated the distance-aware approach to data modulation of OCC systems in arbitrary communication distance. Furthermore, a generalized modulation scheme was proposed based on a novel link performance analysis considering the modulation parameters and communication distance. The main findings of this paper include: 1) The inherent connection between the traditional OM and UM schemes is revealed in OCC systems utilizing image sensors with rolling shutters. Simulation and experimental results demonstrated that the OM and UM can be seen as special cases of the proposed modulation scheme that emphasize effectiveness through longer packet lengths and reliability through more repetition counts, respectively. 2) From the visual MIMO perspective, the communication distance determines the available DoFs that can be utilized to improve the link performance. In short-range communication, there are sufficient DoFs to achieve high pixel efficiency. However, in long-range communication, most of the available DoFs are utilized to meet the minimum PDR requirement, leaving fewer DoFs for enhancement in pixel efficiency. 3) The generalized modulation scheme introduces a distinctive capability of distance awareness to OCC, which is unavailable in existing modulation schemes. Adjusting the modulation parameters of packet length and repetition count allows for achieving a predefined level of effectiveness and reliability for an arbitrary communication distance.
In addition to the communication distance, the modulation parameters depend on the cameras exploited in OCC receivers. As a result, the challenge of non-uniformity arising in practical scenarios with multiple OCC links using different cameras is an issue that needs to be addressed in future work.
Funding
Talent of Colleges and Universities Service Enterprise of Xi'an (22GXFW0094); Innovation Team of Higher Education Institutions of Shaanxi Province; Key Core Technology Research and Development Project for Key Industrial Chain of Xi'an (103-433023062).
Acknowledgments
The authors gratefully acknowledge the support from the Talent of Colleges and Universities Service Enterprise of Xi’an (No. 22GXFW0094), Innovation Team of Higher Education Institutions of Shaanxi Province, and Key Core Technology Research and Development Project for Key Industrial Chain of Xi’an (No. 103-433023062).
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. T. Yamazato, I. Takai, H. Okada, et al., “Image-sensor-based visible light communication for automotive applications,” IEEE Commun. Mag. 52(7), 88–97 (2014). [CrossRef]
2. M. Z. Chowdhury, M. T. Hossan, A. Islam, et al., “A comparative survey of optical wireless technologies: architectures and applications,” IEEE Access 6, 9819–9840 (2018). [CrossRef]
3. M. K. Hasan, M. O. Ali, M. H. Rahman, et al., “Optical camera communication in vehicular applications: a review,” IEEE Trans. Intell. Transport. Syst. 23(7), 6260–6281 (2022). [CrossRef]
4. P. Luo, M. Zhang, Z. Ghassemlooy, et al., “Undersampled-based modulation schemes for optical camera communications,” IEEE Commun. Mag. 56(2), 204–212 (2018). [CrossRef]
5. T. T. Tsai, C. W. Chow, Y. H. Chang, et al., “130-m image sensor based visible light communication (VLC) using under-sample modulation and spatial modulation,” Opt. Commun. 519, 128405 (2022). [CrossRef]
6. R. D. Roberts, “Undersampled frequency shift ON-OFF keying (UFSOOK) for camera communications (CamCom),” in 2013 22nd Wireless and Optical Communication Conference, (IEEE, 2013), pp. 645–648.
7. W. A. Cahyadi, Y. H. Chung, Z. Ghassemlooy, et al., “Optical camera communications: principles, modulations, potential and challenges,” Electronics 9(9), 1339 (2020). [CrossRef]
8. H.-Y. Lee, H.-M. Lin, Y.-L. Wei, et al., “RollingLight: enabling line-of-sight light-to-camera communications,” in Proceedings of the 13th Annual International Conference on Mobile Systems, Applications, and Services, (ACM, Florence Italy, 2015), pp. 167–180.
9. C. W. Chow, C. Y. Chen, and S. H. Chen, “Visible light communication using mobile-phone camera with data rate higher than frame rate,” Opt. Express 23(20), 26080–26085 (2015). [CrossRef]
10. C. W. Chow, Y. Liu, C. H. Yeh, et al., “Display light panel and rolling shutter image sensor based optical camera communication (OCC) using frame-averaging background removal and neural network,” J. Lightwave Technol. 39(13), 4360–4366 (2021). [CrossRef]
11. C. Chen, W. Huang, L. Zhang, et al., “Robust and unobtrusive display-to-camera communications via blue channel embedding,” IEEE Trans. on Image Process. 28(1), 156–169 (2019). [CrossRef]
12. H. Nguyen, T. L. Pham, H. Nguyen, et al., “Trade-off Communication Distance and Data Rate of Rolling Shutter OCC,” in 2019 Eleventh International Conference on Ubiquitous and Future Networks (ICUFN), (IEEE, Zagreb, Croatia, 2019), pp. 148–151.
13. P. Chavez-Burbano, J. Rabadan, V. Guerra, et al., “Flickering-free distance-independent modulation scheme for OCC,” Electronics 10(9), 1103 (2021). [CrossRef]
14. K. Eöllős-Jarošíková, V. Neuman, C. M. Jurado-Verdú, et al., “Long-distance indoor optical camera communication using side-emitting fibers as distributed transmitters,” Opt. Express 31(16), 26980–26989 (2023). [CrossRef]
15. D. C. Tsai, Y. H. Chang, C. W. Chow, et al., “Optical camera communication (OCC) using a laser-diode coupled optical-diffusing fiber (ODF) and rolling shutter image sensor,” Opt. Express 30(10), 16069–16077 (2022). [CrossRef]
16. A. R. Ndjiongue and H. C. Ferreira, “An overview of outdoor visible light communications,” Trans. Emerging Tel. Tech. 29(7), 1 (2018). [CrossRef]
17. M. D. Thieu, T. L. Pham, T. Nguyen, et al., “Optical-rol-signaling for vehicular communications,” IEEE Access 7, 69873–69891 (2019). [CrossRef]
18. C. Jurado-Verdu, V. Guerra, V. Matus, et al., “Convolutional autoencoder for exposure effects equalization and noise mitigation in optical camera communication,” Opt. Express 29(15), 22973–22991 (2021). [CrossRef]
19. H. Nguyen, M. D. Thieu, T. Nguyen, et al., “Rolling OFDM for image sensor based optical wireless communication,” IEEE Photonics J. 11(4), 1–17 (2019). [CrossRef]
20. X. Li, N. B. Hassan, A. Burton, et al., “A simplified model for the rolling shutter based camera in optical camera communications,” in 2019 15th International Conference on Telecommunications (ConTEL), (IEEE, 2019), pp. 1–5.
21. T. Yamazato, M. Kinoshita, S. Arai, et al., “Vehicle motion and pixel illumination modeling for image sensor based visible light communication,” IEEE J. Select. Areas Commun. 33(9), 1793–1805 (2015). [CrossRef]
22. A. Q. Liu, W. X. Shi, M. Ouyang, et al., “Characterization of optical camera communication based on a comprehensive system model,” J. Lightwave Technol. 40(18), 6087–6100 (2022). [CrossRef]
23. K. Dong, M. Kong, R. Chen, et al., “BER analysis on exposure effect for optical camera communication,” Opt. Lett. 48(13), 3419–3422 (2023). [CrossRef]
24. C. Jurado-Verdu, V. Guerra, J. Rabadan, et al., “Deep learning for signal clock and exposure estimation in rolling shutter optical camera communication,” Opt. Express 30(12), 20261–20277 (2022). [CrossRef]
25. M. Rátosi and G. Simon, “Robust VLC beacon identification for indoor camera-based localization systems,” Sensors 20(9), 2522 (2020). [CrossRef]
26. K. Dong, X. Ke, and H. Li, “Camera-based channel modeling and symbol error rate analysis of CSK modulation for outdoor OCC systems,” IEEE Access 10, 50254–50264 (2022). [CrossRef]
27. Y. Goto, I. Takai, T. Yamazato, et al., “A new automotive VLC system using optical communication image sensor,” IEEE Photonics J. 8(3), 1–17 (2016). [CrossRef]
28. M. K. Hasan, M. Z. Chowdhury, Md. Shahjalal, et al., “Performance analysis and improvement of optical camera communication,” Appl. Sci. 8(12), 2527 (2018). [CrossRef]
29. L. Q. Liu, R. Deng, and L. K. Chen, “47-Kbit/s RGB-LED-based optical camera communication based on 2D-CNN and XOR-based data loss compensation,” Opt. Express 27(23), 33840–33847 (2019). [CrossRef]
30. A. Ashok, M. Gruteser, N. Mandayam, et al., “Characterizing multiplexing and diversity in visual MIMO,” in 2011 45th Annual Conference on Information Sciences and Systems, (IEEEBaltimore, MD, USA, 2011), pp. 1–6.
31. T. Nguyen, M. D. Thieu, and Y. M. Jang, “2D-OFDM for optical camera communication: principle and implementation,” IEEE Access 7, 29405–29424 (2019). [CrossRef]
32. H. Zhang and F. Yang, “Push the limit of light-to-camera communication,” IEEE Access 8, 55969–55979 (2020). [CrossRef]
