Open Access
Add to BibSonomy
Abstract
Orthogonal frequency-division multiplexing (OFDM) waveform is highly preferred as a dual-function candidate for integrated sensing and communication (ISAC) systems. However, the sensitivity to both carrier frequency offset (CFO) and phase noise greatly impedes its applications in millimeter-wave ISAC systems. Here, we propose and experimentally demonstrate a photonic millimeter-wave ISAC system employing the virtual-carrier-aided self-coherent OFDM technique, wherein a digitally-generated local oscillator is transmitted along with the OFDM signal. Then, a compact CFO-immune and phase noise-immune envelope detection method is implemented for down-converting millimeter-wave communication and radar echo signals. In experiments, a V-band ISAC system is successfully implemented with a simplified remote radio unit, using the remote photonic millimeter-wave heterodyning up-conversion for downlink and the envelope detection-assisted down-conversion for uplink (or radar echoes). In the converged transmission link with a 5-km fiber link and 2-m space link, the Kramers–Kronig (KK) receiver supports a communication data rate up to 16-Gbit/s by mitigating signal-signal beat interference (SSBI). More significantly, the SSBI leads to negligible effects on the sensing performance when classic matched filtering is adopted for target identification. Consequently, a 4.8-cm range resolution and a 4-mm range accuracy are obtained for the radar sensing function.
© 2024 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
As one of the essential usage scenarios of sixth generation (6 G) networks endorsed by the International Telecommunication Union Radiocommunication Sector (ITU-R), integrated sensing and communication (ISAC) has attracted tremendous research interests in both academia and industry [1]. It efficiently utilized congested wireless and hardware resources, offering mutual benefits for both functions [2]. Fueled by emerging applications like autonomous driving and augment/virtual reality, which demand ultra-fast communication data rates and ultra-high-resolution sensing capabilities, ISAC systems operating at high-frequency millimeter-wave (MMW) or terahertz (THz) bands have been regarded as one of the cornerstones of future wireless networks [3,4]. Nevertheless, the practical implementation of broadband MMW/THz ISAC systems using traditional pure electronic approaches faces technical challenges, including limited bandwidth, poor reconfigurability, and high complexity. Correspondingly, photonics technologies hold a lot of promise in achieving extremely high-capacity and ultra-high-resolution MMW/THz ISAC systems [5–8], given their competitive advantages of wide bandwidth, low frequency-dependent transmission loss, and flat frequency response.
Among the multitude of reported photonics-assisted ISAC systems, the most straightforward strategy for integrating communication and sensing functions is through orthogonal (time/frequency/code/polarity) wireless resource allocation. The code division multiplexing (CDM) method is harvested in [5] to facilitate a Ka-band photonics ISAC system, which experimentally demonstrated a communication data rate of 1-Gbit/s and a sensing resolution of 3.5-cm. W-band and sub-THz-band photonics ISAC systems are proposed in [6] and [7] based on the time division multiplexing (TDM) scheme, achieving a radar range resolution of 1.58-cm and up to a 38.1-Gbit/s communication data rate in experiments. To simultaneously achieve sensing and communication, the frequency division multiplexing (FDM) strategy is reaped in [8,9] resulting in over 78-Gbit/s data rate and centimeter-level radar resolution. The authors propose a spectral-efficient FDM MMW ISAC system based on super-resolution radar technology, achieving up to an 18-Gbit/s communication data rate and a radar range resolution as high as 2.14-cm. Additionally, electromagnetic polarization multiplexing is investigated in [10], realizing a 15-mm sensing spatial resolution and a 92-Gbit/s data rate in the W band. However, these schemes are plagued by a significant concern of poor wireless resource efficiency.
Alternatively, the integrated waveform approach employs a shared signal for both sensing and communication functions, aiming to maximize the utilization of wireless resources. The design philosophies for integrated waveforms in photonics-assisted ISAC systems can be categorized into communication-centric and sensing-centric approaches. In radar-centric ISAC systems, communication symbols are embedded into the sensing waveforms, typically linear frequency-modulated (LFM) radar signals. The key parameters such as amplitude, phase, and frequency carry communication information data. For instance, amplitude shift keying (ASK)-LFM and phase shift keying (PSK)-LFM integrated waveforms are utilized in [11] and [12] to support photonics-aided ISAC systems operating at microwave and THz bands. The authors experimentally demonstrate an MMW-band photonics-assisted ISAC system using a constant-envelope orthogonal frequency-division multiplexing (OFDM)-LFM integrated waveform, obtaining a range resolution of 1.5-cm and a data rate of 8-Gbit/s. Nonetheless, radar-centric schemes may have limited communication data rate. In contrast, communication-centric waveforms, exemplified by the OFDM-like ones, offer advantages in communication performance and particularly compatibility with 5 G and future 6 G standards [13,14]. Thus, a tunable W/K-band OFDM MMW ISAC system based on the optoelectronic oscillator technique is presented in [15], realizing a 32-Gbit/s communication data rate and 1.5-cm range resolution. MMW-band OFDM radio over fiber (RoF) ISAC systems are also discussed in [16,17], exhibiting intriguing features such as fiber-based low-loss large geographical coverage of high-frequency signals and easy integration with the efficient centralized radio access network.
Notwithstanding, challenges remain for the widespread deployment of photonic MMW ISAC systems. These difficulties encompass: (i) the extreme sensitivity of OFDM signal to carrier frequency offset (CFO) and phase noise, which can be more severe in RoF ISAC systems generally leveraging the optical heterodyne beating (of free-running lasers) for flexible MMW signal generation; (ii) the complex and costly MMW coherent receiver requires high-performance mixer and local oscillator (LO) circuits, conflicting with the essential demand for simplified remote radio units (RRUs). Efforts to mitigate the phase noise issue in the photonics-aided OFDM ISAC system are discussed in [18], but it requires the cumbersome optoelectronic oscillator configuration. Conversely, the non-coherent envelope-detection receiver stands out as a simple and practical approach to address the aforementioned challenges [19,20,21], yet only communication function is demonstrated in these systems. A system was implemented by employing the envelope-detection receiver to simultaneously implement both communication and sensing functions, [22], emphasizing the flexible utilization of time-frequency resources rather than achieving higher spectral efficiency.
In this paper, we propose and experimentally demonstrate a photonic communication-centric ISAC system employing the virtual-carrier-aided (VC) self-coherent OFDM technique [23] to deal with the CFO and phase noise issues from laser, wireless channel, or local oscillator. In the proof-of-concept experiments, we establish a photonic V-band ISAC system featuring a simplified RRU. It incorporates a high-speed photodetector responsible for optical heterodyne-based MMW signal generation in the downlink, and a compact directly-modulated laser for delivering the down-converted signals after envelope detection in the uplink. The integration of a digitally synthesized LO carrier sent along with the information OFDM signal enables a simple and CFO/ phase noise-immune envelope-detection-based non-coherent receiver for both MMW communication and radar echo signals. Harnessing the Kramers–Kronig (KK) receiver scheme for alleviating signal-signal beat interference (SSBI), we experimentally demonstrate a communication data rate of 16-Gbit/s over the 5-km fiber and 2-m wireless link. Notably, this work presents a first-of-its-kind investigation into the impacts of SSBI on sensing performances, covering theoretical analyses, simulations, and experiments. Our results indicate that the SSBI has minimal influence on the sensing performance when using the matched filtering processing scheme for sensing information retrieval. As a result, at a 2-m detection distance and utilizing the 5-km fiber uplink for transmitting down-converted radar echoes after envelope detection, our experimental measurements reveal a radar range resolution and accuracy of 4.8-cm and 4-mm, respectively.
2. Principle
Figures 1(a) and (b) depict the architecture and operational principle of the proposed ISAC system. It employs the heterodyne beating of two optical frequencies (denoted as ${f_1}$ and ${f_2}$) within a high-speed photodetector, to generate MMW OFDM ISAC signals. One of the two optical tones (e.g., ${f_1}$) carries the VC-OFDM signal:
where $s(t )$ represents the complex OFDM signal encompassing the spectrum from $0$ to B. ${E_0}$ denotes the direct-current component, which is digitally inserted at the lower boundary of the signal’s spectrum. The generation process of the VC-OFDM signal is implemented through the computer before the arbitrary waveform generator. This direct-current component can be regarded as an LO, essentially acting as a virtual carrier that closely accompanies the OFDM signal throughout the system. In this way, both the frequency and the phase of the LO stay synchronized with those of the OFDM signal, $s(t )$. It is shown in Fig. 1(b) that a narrow frequency interval of a few megahertz is typically required between the LO and the OFDM signal to prevent frequency aliasing between the LO and the out-of-band emission of the signal. The amplitude of ${E_0}$ can be flexibly adjusted to achieve the desired carrier-to-signal power ratio (CSPR), thereby ensuring that the received signal is minimum phase. The CSPR is determined by the power ratio between the LO and the OFDM information signals. Subsequently, this VC-OFDM signal, represented in the form of single sideband, is combined with the unmodulated complex optical tone at ${f_2}$(e.g., ${E_2}(t )= {E_2}\exp ({j2\pi {f_2}t} )$) and sent to the photodetector for heterodyne mixing, yielding:
Fig. 1. Principle and architecture illustrations of proposed photonics-aided MMW ISAC system; Tx: transmitting end; Rx: receiving end; OC: optical coupler; PD: photodetector; ED: envelope detection; SSBI: signal-signal beat interference; KK Demod.: Kramers-Kronig algorithm-aided demodulation; Commu.: communication; MF: matched filtering.
For the radar detection function, the matched filtering is executed by computing the cross-correlation between the signal after envelope detection and the reference signal $s(t )$. Neglecting the scale factor $2{E_0}$ and focusing solely on the useful signal and SSBI term, the output of the matched filter is given by:
This can be interpreted as the sum of the reference OFDM signal’s auto-correlation function (ACF) and the cross-correlation function (CCF) between the reference and SSBI. Being compared with ACF, the CCF between the reference and the unmatched channel tends to approach zero, resulting in a correlation peak only for the target channel. For given conditions, the output amplitude of the matched filter in Eq. (4) can be further simplified as:
This simplification illustrates that SSBI can be effectively mitigated by the matched filtering, reducing the need for digital signal processing (DSP) at the radar receiver. It indicates that the KK algorithm is not necessary at the radar receiver, unlike at the communication receiver. The relationship between ${R_u}(t )$ and ${R_{SSBI}}(t )$ will be elucidated in section 3, providing additional details for feasible simplification.
Considering $s(t )$ as the complex OFDM waveform with m symbols and n subcarriers within a single pulse, the general expression, excluding the cyclic prefix, can be mathematically formulated as follows:
The matched filtering corresponds to the time-based displacement of ACF or the ambiguity function, displaying an impulse peak around the delay value $t = {t_0}$. According to Rayleigh width (peak to first null width), the range resolution $\Delta R$ in time is approximately calculated as $|{t - {t_0}} |\approx 1/B = N\Delta f$ seconds. In some works, $\Delta R$ can be defined in terms of the full width at half-maximum (FWHM) of the ACF [27]:
The resolution defined by FWHM is typically slightly larger than that defined by the Rayleigh criterion, but it is more convenient for practical measurements.
3. Simulation results
A preliminary simulation concerning OFDM radar detection was initially conducted using the commercially available simulation software, VPItransmissionMaker (VPI), in conjunction with MATLAB. This simulation aimed to validate the conclusions derived from Eqs. (3)-(5).
Here we define the signal-to-interference ratio (SIR) as the ratio between the maximum power value of ${R_u}(t )$ and that of ${R_{SSBI}}(t )$:
Since ${R_u}(t )$ is associated with full coherent integration [28], while ${R_{SSBI}}(t )$ pertains to the noncoherent integration, the SIR exhibits a linear dependency on ${T_s}$, the OFDM symbol duration time. For an OFDM signal with a 4-GHz bandwidth but variable duration, the simulation result is shown in Fig. 2.
In most ISAC systems utilizing OFDM waveform, the symbol time duration ${T_s}$ typically falls within the order of microseconds. As a result, the SIR typically reaches levels of approximately 30-dB. This implies that during such a time scale, the interference term ${R_{SSBI}}(t )$ can effectively be disregarded. The impact of full coherent integration through matched filtering essentially depends on the size of data involved in DSP, and this size is directly proportional to the time-bandwidth product of the signal. Therefore, the greater the time-bandwidth product of the signal, the more pronounced the effect of matched filtering, leading to a larger SIR. It is noteworthy that at the communication receiver, a conventional matched filter is typically combined with a sender-side pulse shaping filter to act on each symbol or bit with a small time-bandwidth product, resulting in a relatively small SIR. In such cases, the influence of SSBI cannot be overlooked. Conversely, the matched filter directly acts on the echo signal with a large time-bandwidth product at the radar receiver, resulting in a substantial SIR. For this case, the impact of SSBI can be disregarded.
In subsequent simulations, the generated MMW VC-OFDM signal, employing 16QAM modulation format, is centered at 53-GHz and possesses a 4-GHz bandwidth, with one symbol duration approximately equal to 1.6 microseconds. The CSPR of the VC-OFDM signal is set at 10-dB to satisfy the minimum phase condition (MPC), ensuring that its time trajectory does not encircle the origin and thus supports KK processing.
To separate the SSBI term, an 8-GHz frequency guard interval is reserved between the VC and OFDM signal, as Fig. 3(a1). By introducing another radar delay channel, echoes may undergo multipath fading. As depicted in Fig. 3(a2), SSBI falls in the baseband and is denoted in green, while the useful OFDM echo remains in the intermediate frequency, facilitating effective separation. Before the matched filter, the blue echo must be down converted to the baseband, whereas the green SSBI does not require this conversion. The outcomes of the two filtering operations are illustrated in Fig. 3(a3), with the signal to noise ratio (SNR) of ${R_u}(t )$ being roughly 40-dB higher than that of ${R_{SSBI}}(t )$, rendering ${R_{SSBI}}(t )$ negligible. Additionally, spurious peaks caused by cyclic prefix were disregarded in this ranging scenario.
Fig. 3. Simulation of SSBI elimination by matched filtering. The electrical spectrum after the radar channel of integrated signals with (a1) and without (b1) frequency guard interval; (a2), (b2) the baseband spectrum after envelope detection; (a3), (b3) ranging peaks after matched filtering.
In cases where no guard interval is present, as shown in Fig. 3(b1), the useful echo signal overlaps with SSBI after envelope detection as in Fig. 3(b2). Processing them together in a single matched filtering operation still yields excellent ranging performance. As illustrated in Fig. 3(b3), no ghost signals are observed, and SSBI does not introduce any adverse effects. Therefore, as Eq. (5) elucidates, matched filtering can effectively eliminate the SSBI term.
In pursuit of enhanced matched filtering outcomes, we initially explore the application of the KK algorithm to alleviate SSBI before engaging in the filtering process. We extract the integrated signal from Fig. 3(b1) and present its time trajectory in Fig. 4(a). When the phase of the VC is not fixed, the time trajectory of the entire signal will pass through the four quadrants of the complex plane. The absence of the trajectory around the origin indicates that the integrated signal satisfies the MPC. Upon comparing Fig. 4(b) with Fig. 3(b3), it is apparent that applying the KK algorithm followed by matched filtering only results in a marginal SNR improvement of approximately 0.1-dB.
Fig. 4. Simulation of SSBI elimination by KK algorithm; (a) time trajectory of the SSB signal at a 10-dB CSPR; (b) applying KK algorithm before matched filtering and the resulting output.
In other words, in this ranging scenario employing matched filtering, the KK algorithm yields only negligible gains:
where $KK({\cdot} )$ represents the operation of the KK algorithm. For instance, the floating-point operations per second (FLOPs) [29] amount to 37,324,809 when utilizing the KK algorithm, whereas using a matched filtering alone requires only 7,464,969 FLOPs, resulting in a reduction of 29,859,840 FLOPs. This reduction is approximately four times that of matched filtering. Considering that the KK algorithm introduces additional DSP hardware, it is deemed sufficient to utilize matched filter alone at the radar receiver end.4. Experimental results
An experimental setup for the MMW VC-OFDM ISAC system was established based on the outlined principle. At the central office, the emitted light wave from the laser diode underwent modulation using a Mach-Zehnder modulator (MZM) employing carrier suppressed single sideband modulation. This modulation process resulted in the generation of two optical tones adjacent to 60-GHz, which were then separated into individual tones using a wavelength division multiplexer (WDM). One tone traversed through an in-phase/quadrature (IQ)-MZM is modulated by the VC-OFDM signal generated by the arbitrary waveform generator, while the other tone remained unchanged. After combining the two optical paths, the output of the optical coupler was transmitted to the RRU through a 5-km single-mode fiber. Through heterodyne beating, the MMW integrated signal was generated in a photodetector and amplified by a power amplifier, subsequently transmitted to both communication and radar channels. At the user end, the MMW signal is down-converted to the baseband using an envelope detector and sampled for implementing DSP of communications. Simultaneously, echoes from detected targets were captured by the RRU, down converted to baseband signals, and transmitted back to the central office for implementing the DSP of radar. The spectral evolution is visually depicted in Fig. 5(b). It is crucial to note that the DSP procedures at the communication end and the radar end differ, as explained in the preceding section.
Fig. 5. Experimental setup of the proposed ISAC system based on virtual-carrier-assisted self-coherent OFDM scheme and spectrum evolution diagram; LD: laser diode; SG: signal generator; CS-DSB: carrier suppressed double sideband modulation; MZM: Mach-Zehnder modulator; WDM: wavelength division multiplexer; AWG: arbitrary waveform generator; CS-SSB: carrier suppressed single sideband modulation; IQ-MZM: in-phase/quadrature Mach-Zehnder modulator. OC: optical coupler; VOA: variable optical attenuator; PD: photodetector; ED: envelope detector; PA: power amplifier; LNA: low noise amplifier; DML: directly modulated laser; Txa: transmitting antenna; Rxa-R: receiving antenna of radar; Rxa-C: receiving antenna of communication; DSP: digital signal processing.
The experimental results, encompassing both the communication performance at the user end and the radar performance at the central office, are illustrated in Figs. 6–10. In evaluating the influence of OFDM waveform parameters on communication and radar performance, we adopted the setup outlined in Table 1, following the parameters specified in the 5 G NR FR2 R17 standard [30]. In our experiment, the subcarrier spacing of the OFDM signal is 980-KHz. Due to limited memory depth in our arbitrary waveform generator and oscilloscope, an OFDM pulse consisting of ten symbols is transmitted, lasting for 10.96 microseconds. Regarding the cyclic prefix, while it is necessary for the communication scenario to prevent inter symbol interference, it can introduce pseudo peaks on both sides of the main peak in the radar detection scenario. Therefore, we selected a cyclic prefix length as short as possible according to the R17 standard to strike a balance between communication and radar performance.
Fig. 6. (a) BER versus the received optical power with/without utilizing the KK algorithm; (b) BER versus CSPR with/without utilizing the KK algorithm under two different modulation schemes, 4-GHz QPSK and 16QAM, respectively; (c) Wireless communication at about 2 m.
Fig. 7. (a) Experimental setup of the two-lasers transmitter; (b) BER versus the received optical power with/without utilizing the KK algorithm; (c) BER versus CSPR with/without utilizing the KK algorithm under two different modulation schemes, 4-GHz QPSK and 16QAM, respectively.
Fig. 8. Experiment results of SSBI elimination; the electrical spectrum of echoes after envelope detection with (a1) and without (b1) frequency guard interval; (a2), (b2) the output of matched filter; (b3) the output of matched filter after using KK algorithm without frequency guard interval; (c) the impact of KK on SNR performance versus the received optical power.
Fig. 9. (a) Theoretical and measured values of range resolution versus bandwidth; (b) multi-targets detection; (c) five targets arranged at about 6-cm intervals.
Fig. 10. (a) SNR performance versus CSPR with two different distances; (b) error of measured distances versus CSPR; (c) Measured accuracy and precision with relative distance.
Table 1. OFDM parameter
For communication function demonstrations employing 4-GBaud QPSK modulation, the evaluation of the bit error rate (BER) with respect to various the received optical powers is presented in Fig. 6(a). When the characteristics of the envelope-detection receiver can be approximated by a square root relationship, SSBI can be eliminated, and the advantages of KK processing are most pronounced at high received optical power. We optimized the received optical power at a constant level of about 5-dBm and explored the optimal CSPR. In Fig. 6(b), for QPSK modulation, an optimal 12-dB CSPR is observed in the conventional SSB scheme, while 9-dB in the proposed KK scheme, resulting in a 3-dB reduction attributed to the improved field reconstruction accuracy of the KK scheme. When the CSPR exceeds the optimum value, BER degrades with the increase of CSPR, mainly due to the energy conservation law. In a system with certain limited transmission power, the greater the power of the VC, the smaller the power of the useful OFDM signal, leading to a smaller received SNR. Additionally, in the case of lower CSPR, decreasing the CSPR also degrades the BER performance due to the violation of MPC and the increase of SSBI. The same set of measurements is repeated with 16QAM signaling, which exhibits a behavior similar to that for QPSK. In both QPSK and 16QAM cases, the KK processing reduced both the BER performance and optimum CSPR compared to the reception without algorithm processing. In the absence of the KK algorithm in the 16QAM scheme, the BER consistently exceeds the forward error correction (FEC) limit by 20% overhead. Thus, for the QPSK or 16QAM modulation scheme, the valid interval of CSPR under limited FEC was 16-dB or 6-dB, respectively. Furthermore, we transmitted this VC-OFDM signal at a center frequency of 52.1 GHz over a wireless distance of 2 m, as shown in Fig. 6(c).
Besides the optical frequency comb scheme illustrated in Fig. 5, we employed two distinct free-running lasers as the heterodyne millimeter-wave source. The transmitter diagram and communication results are detailed in Fig. 7. The communication performances between the optical frequency comb scheme and the two-lasers scheme exhibit remarkable similarity. Hence, no matter which scheme was employed, the experimental results exhibit minimal impact from CFO and phase noise when using a virtual carrier synchronized with the signal.
For radar function in ranging, the impact of SSBI on ranging performance is relatively small. We adjusted CSPR to 10-dB to enable the utilization of KK processing at the receiving end. As shown in Figs. 8(a1) and (b1), the electrical spectrum of echoes after envelope detection exhibited some irregularities, attributed to the multipath fading and non-ideal transmission characteristics of either the IQ-MZM or other devices. When applying the KK algorithm before matched filtering, the SNR would slightly increase by approximately 0.1-dB, as Figs. 8(b2) and (b3). The experimental results shown in Fig. 8 closely resembled those obtained in the simulation. ${R_u}(t )$ consistently far exceeded ${R_{SSBI}}(t )$ in both cases, using only matched filtering and applying the KK algorithm before matched filtering.
To further illustrate the influence of the received optical power on the accuracy and precision of ranging, a variable optical attenuator is placed before the photodetector at the RRU, and the results are presented in Fig. 8(c). While employing the KK algorithm, SNR performance exhibited a slight improvement. At around 29-dB, there exists an SNR cross point. Below this threshold, the KK algorithm does not work effectively, indicating that the phase retrieval algorithm requires the SNR of the received signal to be greater than a certain value.
The measured accuracy is directly related to the range resolution $\Delta R$, defined by FWHM of the output of ACF or matched filter. All the experimental resolution results indicated that the measured values are slightly higher than their theoretical counterparts. In Fig. 9(a), the horizontal dashed line represents the theoretical value of range resolution, and the bar chart illustrates the measured values. The measurements were conducted for bandwidths 1 to 4-GHz in the 16QAM scheme. Taking 4-GHz bandwidth as an example, the measured value shown in the inset was 4.8-cm, while its theoretical value was 4.5-cm, slightly smaller. For multi-target detection considerations, we set five targets with a spacing of around 6-cm, as shown in Fig. 9(c). It is evident in Fig. 9(b) that there are five distinct detection peaks respectively.
To obtain more stable measurement results, we selected QPSK as the modulation format instead of 16QAM [31]. To validate the ranging performance of the system, we positioned the targets at distances of 6-cm and 10-cm, respectively, and the obtained SNR is depicted in Fig. 10(a). Similar to the communication results, within a specific CSPR interval, the SNR can reach its optimal value. The radar function can operate effectively above the 3-dB detection threshold in the CSPR interval from -20 to 30-dB, which is significantly wider than the communication interval represented by the orange area. To explore the impact of SNR on ranging performance, we adjusted the CSPR to obtain different SNR at 6-cm and 10-cm respectively. The SNR-dependent error typically dominates the radar range error, and a standard deviation is given by:
As observed in Fig. 10(b), the ranging error of the target at 10-cm was 4-mm after calibration, smaller than 7-mm for the target at 6-cm, indicating that higher SNR results in better ranging accuracy. Moreover, outside the communication CSPR-restricted area, the radar continues to work well. Finally, the CSPR working interval at the radar end was also constrained by the communication interval from 4 to 20-dB, resulting in a total reduction of 34-dB. By adjusting the distance between the two targets in 2-cm steps, the ranging performance carried out at the central office is illustrated in Fig. 10(c). Under optimal SNR condition, a ranging accuracy of less than 7-mm and a ranging precision of less than 1-mm could be observed.
Applying the ranging principle and utilizing a turntable, a 2D range-azimuth scanning imaging experiment was conducted. Three targets to be measured were strategically placed within a fan-shaped range with an azimuth angle of 50°, all within the 1-m radius. Subsequently, the turntable was rotated to measure targets at various angles. Following digital calibration and smoothing filtering, the resulting image is presented in Fig. 11(a). Notably, the range-azimuth image of two iron cubes is relatively clearer, as their reflection areas are larger compared to that of the corner reflector.
5. Discussion
We conducted a confirmatory simulation to explore the effectiveness of KK algorithm for SSBI in multipath channels, and the results are shown in Fig. 12. To visually identify SSBI, we introduced ample frequency guard intervals between the virtual carrier and the signal. This ensures that, after envelope detection, the SSBI is distinctly separated from the signal in the frequency domain.
Fig. 12. (a1) The electrical spectrum of the received signal by the antenna after the one path channel; (a2) time trajectory of the antenna signal; (a3) the baseband spectrum after envelope detection; (a4) eliminating SSBI by KK algorithm; (b1)-(b4) multipath channels, 0.249 nanoseconds delay; (c1)-(c4) multipath channels, 0.32 nanoseconds delay.
For a single-path channel, when the received signal satisfies the minimum phase condition, its time trajectory does not pass through the coordinate origin, as illustrated in Fig. 12(a). After envelope detection, the KK algorithm can eliminate the SSBI effectively.
In multi-path scenarios, the carrier-to-signal power ratio increases if the multi-path effect enhances the virtual carrier, as shown in Fig. 12(b). In this case, the received signal still satisfies the minimum phase condition, such that the time trajectory is further away from the coordinate origin, allowing the KK algorithm to function effectively. However, when multipath fading suppresses the virtual carrier, the minimum phase relationship is disrupted, as illustrated in Fig. 12(c). Then the time trajectory passes through the origin, and the KK algorithm cannot eliminate the SSBI. Therefore, we should guarantee a sufficiently large carrier-to-signal power ratio, before employing the KK algorithm for multi-path channels.
6. Conclusion
We proposed and demonstrated a photonic MMW VC-OFDM ISAC system for the fiber-wireless integrated network. Employing both the KK algorithm and the matched filtering method, this self-coherent system is fully immune to both CFO and phase noise, resulting in greatly simplified system, improved communication performance, and enhanced range performance. In experiments, we have achieved a sensing resolution of 4.8-cm, a ranging accuracy of 4-mm, a ranging precision of 1-mm and a data rate of 16-Gbit/s, through a 5-km fiber link and a 2-m space link. Additionally, the simple and robust structure of this system highlights its potential for the development future ISAC system.
Funding
National Key Research and Development Program of China (2023YFB2804900); National Natural Science Foundation of China (62271422, U21A20507); Sichuan Province Science and Technology Support Program (2022JDTD0013); Fundamental Research Funds for the Central Universities (2682021CX045).
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. ITU, “Draft new Recommendation ITU-R M.[IMT.FRAMEWORK FOR 2030 AND BEYOND] - Framework and overall objectives of the future development of IMT for 2030 and beyond,” (2023). https://www.itu.int/md/meetingdoc.asp?lang = en&parent = R19-WP5D-230612-TD-0905.
2. F. Liu, Y. Cui, C. Masouros, et al., “Integrated Sensing and Communications: Toward Dual-Functional Wireless Networks for 6 G and Beyond,” IEEE J. Select. Areas Commun. 40(6), 1728–1767 (2022). [CrossRef]
3. T. Mao, J. Chen, Q. Wang, et al., “Waveform Design for Joint Sensing and Communications in Millimeter-Wave and Low Terahertz Bands,” IEEE Trans. Commun. 70(10), 7023–7039 (2022). [CrossRef]
4. Y. Wu, C. Han, and Z. Chen, “DFT-spread orthogonal time frequency space system with superimposed pilots for terahertz integrated sensing and communication,” IEEE Trans. Wireless Commun. 22(11), 7361 (2023). [CrossRef]
5. W. Bai, X. Zou, P. Li, et al., “Photonics Millimeter-Wave Joint Radar Communication System Using Spectrum-Spreading Phase-Coding,” IEEE Trans. Microwave Theory Techn. 70(3), 1552–1561 (2022). [CrossRef]
6. Y. Wang, Z. Dong, J. Ding, et al., “Photonics-assisted joint high-speed communication and high-resolution radar detection system,” Opt. Lett. 46(24), 6103–6106 (2021). [CrossRef]
7. Y. Wang, W. Li, J. Ding, et al., “Integrated High-Resolution Radar and Long-Distance Communication Based-on Photonics in Terahertz Band,” J. Lightwave Technol. 40(9), 2731–2738 (2022). [CrossRef]
8. Y. Wang, J. Liu, J. Ding, et al., “Joint communication and radar sensing functions system based on photonics at the W-band,” Opt. Express 30(8), 13404–13415 (2022). [CrossRef]
9. N. Zhong, P. Li, W. Bai, et al., “Spectral-Efficient Frequency-Division Photonics Millimeter-Wave Integrated Sensing and Communication System Using Improved Sparse LFM Sub-bands Fusion,” J. Lightwave Technol. 41(23), 7105–7114 (2023). [CrossRef]
10. M. Lei, M. Zhu, Y. Cai, et al., “Integration of Sensing and Communication in a W-Band Fiber-Wireless Link Enabled by Electromagnetic Polarization Multiplexing,” J. Lightwave Technol. 41(23), 7128–7138 (2023). [CrossRef]
11. H. Nie, F. Zhang, Y. Yang, et al., “Photonics-based integrated communication and radar system,” in Proc. Int. Topical Meeting Microw. Photon. (MWP), 42–45 (2019).
12. Z. Lyu, L. Zhang, H. Zhang, et al., “Radar-Centric Photonics Terahertz Integrated Sensing and Communication System Based on LFM-PSK Waveform,” IEEE Trans. Microwave Theory Techn. 71(11), 5019–5027 (2023). [CrossRef]
13. S. D. Liyanaarachchi, T. Riihonen, C. B. Barneto, et al., “Optimized Waveforms for 5G-6 G Communication with Sensing: Theory, Simulations and Experiments,” IEEE Trans. Wireless Commun. 20(12), 8301–8315 (2021). [CrossRef]
14. M. F. Keskin, V. Koivunen, and H. Wymeersch, “Limited Feedforward Waveform Design for OFDM Dual-Functional Radar-Communications,” IEEE Trans. Signal Process. 69, 2955–2970 (2021). [CrossRef]
15. Z. Xue, S. Li, X. Xue, et al., “Tunable K/W-band OFDM integrated radar and communication system based on optoelectronic oscillator for intelligent transportation,” Opt. Express 30(20), 35270–35281 (2022). [CrossRef]
16. B. T. Brandão, J. H. Silva, L. S. Leitão, et al., “5G-NR based Joint RADAR and Communication System Using Low-Cost Photonics Fronthaul,” in Proc. Int. Topical Meeting Microw. Photon. (MWP) (2021), pp. 1–4.
17. L. Huang, R. Li, S. Liu, et al., “Centralized Fiber-Distributed Data Communication and Sensing Convergence System Based on Microwave Photonics,” J. Lightwave Technol. 37(21), 5406–5416 (2019). [CrossRef]
18. Z. Xue, S. Li, J. Li, et al., “OFDM Radar and Communication Joint System Using Opto-Electronic Oscillator with Phase Noise Degradation Analysis and Mitigation,” J. Lightwave Technol. 40(13), 4101–4109 (2022). [CrossRef]
19. P. Li, L. Zhu, X. Zou, et al., “Constant-Envelope OFDM for Power-Efficient and Nonlinearity-Tolerant Heterodyne MMW-RoF System with Envelope Detection,” J. Lightwave Technol. 40(20), 6882–6890 (2022). [CrossRef]
20. M. Steeg, F. Exner, J. Tebart, et al., “100 Gbit/s V-band Transmission Enabled by Coherent Radio-over-Fiber System with IF-OFDM Envelope Detection and SSBI Suppression,” in 2020 XXXIIIrd General Assembly and Scientific Symposium of the International Union of Radio Science (URSI GASS) (2020), pp. 1–4.
21. L. Gonzalez-Guerrero, H. Shams, I. Fatadin, et al., “Single Sideband Signals for Phase Noise Mitigation in Wireless THz-Over-Fibre Systems,” J. Lightwave Technol. 36(19), 4527–4534 (2018). [CrossRef]
22. B. Dong, J. Jia, G. Li, et al., “Demonstration of photonics-based flexible integration of sensing and communication with adaptive waveforms for a W-band fiber-wireless integrated network,” Opt. Express 30(22), 40936–40950 (2022). [CrossRef]
23. Q. Jin and Y. Hong, “Self-Coherent OFDM With Undersampling Downconversion for Wireless Communications,” IEEE Trans. Signal Process. 15(10), 6979–6991 (2016). [CrossRef]
24. T. Harter, C. Füllner, J. N. Kemal, et al., “Generalized Kramers–Kronig receiver for coherent terahertz communications,” Nat. Photonics 14(10), 601–606 (2020). [CrossRef]
25. R. K. Patel, I. A. Alimi, N. J. Muga, et al., “Optical Signal Phase Retrieval with Low Complexity DC-Value Method,” J. Lightwave Technol. 38(16), 4205–4212 (2020). [CrossRef]
26. A. Mecozzi, C. Antonelli, and M. Shtaif, “Kramers–Kronig coherent receiver,” Optica 3(11), 1220–1227 (2016). [CrossRef]
27. R. Chen, H. Shu, B. Shen, et al., “Breaking the temporal and frequency congestion of LiDAR by parallel chaos,” Nat. Photon. 17(4), 306–314 (2023). [CrossRef]
28. M. A. Richards, Fundamentals of Radar Signal Processing, 2nd ed. (McGraw-Hill Education, 2014), Chap. 1.
29. H. Qian, “Counting the Floating Point Operations (FLOPS),” (MATLAB Central File Exchange, 2023). https://www.mathworks.com/matlabcentral/fileexchange/50608-counting-the-floating-point-operations-flops.
30. 3GPP, “NR physical channels and modulation,” (3GPP TR38.211 V17.4.0, Sophia Antipolis, France, Tech. Rep., 2022).
31. Y. Xiong, F. Liu, Y. Cui, et al., “On the Fundamental Tradeoff of Integrated Sensing and Communications Under Gaussian Channels,” IEEE Trans. Inform. Theory 69(9), 5723–5751 (2023). [CrossRef]
