Introduction

Radar stands for Radio Detection and Ranging, and is one of the most important technologies used for detecting and measuring the distance and speed of targets1,2,3. Since its invention during World War II, radar has made significant progress along with the advancement of technologies such as computers and materials. Today, radar is extensively used across diverse industries, encompassing military, civil, and commercial areas, evolving towards collaboration, intelligence, integration, multi-functionality, streamlining, and cost efficiency.

Conventionally, modern radar systems are made in chain architectures that comprise radio-frequency (RF) transmitters and receivers, as well as digital signal processors (DSP)2,4. The radar signal generation relies on voltage-controlled oscillator (VCO) or direct digital synthesis (DDS)5,6 technologies by using active RF components such as mixers and amplifiers, which are associated with high cost, high power consumption, high complexity, and small flexibility. In addition, the separation between the antennas and the RF front end reduces the integration degree of systems. At the receiver, radar echoes are captured by antennas and subsequently converted from analog to digital formats through high-speed analog-to-digital converters (ADCs)1,3. This conversion process is critically reliant on the performance of the sampling components, which can become a limiting factor, particularly in broadband radar systems. The transition from RF reception to digital processing not only introduces additional time delays but also increases the cost, complexity, and power consumption of the system. Furthermore, this transition can introduce noise, which may affect the quality of the signals. Therefore, it is of great significance to investigate high-efficiency systems with low cost, simplified architecture, and high flexibility.

The rapid progress in metasurface science and technologies7,8 opens up opportunities for new radar designs. As a two-dimensional (2D) version of metamaterial9,10,11,12,13,14, the metasurface consists of sub-wavelength elements arranged periodically or quasi-periodically. Over the past decade, the metasurface has demonstrated strong abilities in tailoring electromagnetic (EM) waves, involving beam steering15,16, polarization conversion17,18, nonreciprocal transmission19,20,21,22, vortex-wave generation23,24, and absorption25,26, to name a few. With the integration of electrically switching and tunable components and field programmable gate array (FPGA), a new kind of metasurfaces, digital coding and programmable metasurfaces, was proposed9, making it highly flexible in manipulating the EM waves in the time, frequency, and spatial domains in real time by properly designing the time-coding, space-coding, and space-time-coding (STC) sequences9,28,29,30. The programmable metasurface was further extended to information metasurface due to its special capability to control the EM waves and modulate digital signals simultaneously27,28. The programmable and information metasurfaces have been extensively investigated in many scenarios, such as harmonic beamforming30, high- efficiency nonlinear frequency conversions31,32,33, and EM imaging34,35. In the field of wireless communication, programmable metasurfaces have been attracting significant attention due to their ability to reshape channel environments36,37. Moreover, with the introduction of time modulation strategies, the information metasurfaces can directly modulate the digital information on the harmonics, eliminating the need for mixers and amplifiers commonly used in traditional transmitters38,39,40,41,42.

Recent years have also witnessed the investigation of metasurfaces in radar scenarios. For instance, phase-controllable metasurfaces were developed to replace the conventional phased arrays43,44,45,46,47 and frequency diverse arrays48,49,50. Some studies have explored to use metasurfaces in integrated sensing and communication systems51,52,53,54,55,56,57,58,59,60 and inverse synthetic aperture radar systems61,62. Apparently, they were focused only on the radiation part. Critical functions such as radar signal generation at the transmitter and dechirp at the receiver still rely on traditional hardware architectures. Hence, these efforts have not achieved deep integration with the radar architectures or delivered transformative innovation to the design of radar systems. Recently, benefiting from the ability to produce nonlinear harmonics, the information metasurfaces were employed to generate frequency-modulated continuous wave (FMCW)63 and micro-Doppler signals64, but these works were only focused on the transmitting end. Till now, very little attention has been given to the use of metasurfaces at the receiving end. This is primarily due to two factors. Firstly, conventional radar receivers primarily perform signal processing in the digital domain, which does not take full advantage of real-time RF processing potentials offered by the information metasurfaces. Secondly, the dechirp operation requires the metasurfaces to possess full-phase control capability, but few metasurfaces are capable of fulfilling this requirement. Recent research has explored the potential to use metamaterial or metasurface for computing, including various operations such as differentiation, integration, convolution, discrete Fourier transform (DFT), and matrix equation solving65,66,67,68,69,70. However, these studies required additional RF circuit components65,66 and complex iterative algorithms67,68 or artificially intelligent networks69,70. To the best of our knowledge, there has been no advancement in the use of metasurfaces for radar receivers. Hence, the research on metasurface-based radar systems remains stagnant.

In this work, we present a simplified radar architecture that integrates the STC information metasurfaces as the critical components throughout the entire signal chain, encompassing the signal generation, transmission, reception, and processing. At the transmitter, the STC metasurface directly synthesizes the radar signals at the RF stage under monochromatic wave excitation, enabling real-time reconfiguration of radar parameters via dynamic temporal-spatial encoding. Crucially, this approach eliminates the need for the conventional circuit-level signal synthesis techniques such as VCOs or DDS, as well as RF components such as mixers and transmit/receive (T/R) modules. At the receiver, we establish a theoretical framework wherein the STC metasurface executes the RF-domain dechirp processing, which is conventionally completed in the digital domain. It compresses broadband chirp signals into narrowband counterparts at the RF stage, which can dramatically reduce the requirement for ADC sampling rates and thus eliminate high-speed digitizers and associated storage demands. This marks a fundamental rethinking on the roles of metasurfaces, which are transformed from the auxiliary elements into active processors and waveform generators in next-generation radar architectures. We remark that the roles of the two STC metasurfaces can be interchanged due to their identical structures and reprogrammable nature, thus increasing the flexibility of the proposed radar system.

Several experiments are conducted to validate the detection and ranging capabilities of the STC-metasurface-based radar system. Furthermore, our radar demonstrates its capability for speed measurement. This work demonstrates the use of information metasurfaces as core components in a radar system that integrates STC capabilities, information modulation, beamforming, and RF computing functionalities. This simple STC-metasurface-based radar architecture is expected to offer solutions for emerging fields such as integrated communication, sensing, and smart radar.

Results

STC-metasurface-based radar architecture

In the proposed radar system, two distributed STC information metasurfaces are employed as key components at the transmitting and receiving ends, as shown in Fig. 1. This configuration can effectively prevent the transmitting signals from entering the receivers when the information metasurfaces are used for the RF signal generation and RF dechirp processing in space. From the perspective of information metasurfaces, using multi-metasurface control techniques can facilitate the synchronization of the transmitter and receiver in the presented radar. The two STC metasurfaces are designed with identical structures, including the unit cells and controlling circuits, to ensure the consistency of the chirp signals and the perfect matching filter operation.

Fig. 1: Conceptual diagram to illustrate the space-time-coding (STC)-metasurface-based radar system.
Fig. 1: Conceptual diagram to illustrate the space-time-coding (STC)-metasurface-based radar system.
Full size image

One metasurface is responsible for generating the radio-frequency (RF) chirp signals and transmitting them to targets, while the other STC metasurface receives the echo signals and performs the RF dechirp processing directly on the RF level. The abbreviation FPGA is for field programmable gate array.

The fundamental architecture of the STC metasurface-based radar is shown in Fig. 2a, while the conventional radar system is provided in Fig. 2b for direct comparison. On the transmitter end, the traditional radar relies on a DDS, a digital-to-analog converter (DAC), and an RF mixer with a local oscillator (LO) to generate the chirp signals. In contrast, the transmitter of the propsed radar employs the metasurface to directly convert a monochromatic wave into an RF chirp signal, eliminating those components and substantially simplifying the transmission system. Since the excitation signal for metasurface is monochromatic, the performance requirements of the RF components are substantially relaxed. Distinguished from the conventional beamforming technique that requires the phased arrays and dedicated T/R modules, the STC metasurface achieves the beam steering solely through spatial coding strategies in a digital manner.

Fig. 2: Hardware architectures of the space-time-coding (STC) metasurface-based and conventional radar systems.
Fig. 2: Hardware architectures of the space-time-coding (STC) metasurface-based and conventional radar systems.
Full size image

a The STC-metasurface-based radar system. b The conventional radar system. The abbreviations LO, FPGA, ADC, RF, DSP, DDS, and DAC refer to local oscillator, field-programmable gate array, analog-to-digital converter, radio frequency, digital signal processor, direct digital synthesis, and digital-to-analog converter, respectively.

The advantage of STC metasurface is also extended to signal reception and processing. The conventional system, as plotted in Fig. 2b, includes an RF mixer, a high-speed ADC, a digital mixer, and a DSP module. The STC metasurface-based receiver, as shown in Fig. 2a, adopts a dechirp processing strategy. Owing to the time-varying modulation feature, the metasurface performs the signal demodulation directly at the RF level by compressing the broadband echoes into RF beat signals. By doing so, narrowband analog beat signals are obtained after the RF mixer. For these narrowband signals, the conventional high-speed ADC and large data storage are not required, and a low-speed one is enough to yield the digital beat signals. After that, the conventional operations including spectrum analysis via 2D fast Fourier transform (FFT) and constant false alarm rate (CFAR) detection are performed, which are identical to those in the conventional radar systems. It can be seen that the metasurface advances the signal processing to the RF front end, thus introducing a process-upon-reception paradigm that enables real-time analog-domain computations with enhanced operational efficiency. As a result, it significantly simplifies the digital chain required for the signal demodulation processing.

The STC information metasurface consists of a lattice of unit cells, in which each unit is embedded with electrically tunable varactor diodes, as illustrated in Fig. 1. By controlling the reverse biasing voltages across these diodes, the reflection phase of the unit cell can be adjusted with a range greater than 2π. Details on the cell structure and its full-phase control are introduced in Methods and Supplementary Note 1. When connected to a digital module and controlled by time-coding signals, the STC metasurface can tailor the EM waves in the frequency domain. In other words, it can produce multiple frequencies and manipulate their phases precisely, meeting the crucial prerequisites for chirp generation and dechirp behavior in this work.

According to the principle of STC metasurface, harmonic frequencies can be generated by altering the reflection phase of the unit cell over time31. To further introduce a phase shift at the mth-order harmonic, a time shift \({t}_{d}\) is produced to the modulation waveform. Then the phase shift becomes \(\varphi=-2\pi m{f}_{m}{t}_{d}\), where \({f}_{m}\) is the modulation frequency. To demonstrate the STC metasurface’s abilities to generate harmonics and tune their phases, the scattering characteristics of the 1st-order harmonic are studied as an example. Based on the array theory, a phase gradient should be formed across the metasurface to realize a beam deflection angle \({\theta }_{{des}}\) in the far-field region, and the harmonic phase difference between the neighboring meta-columns should be

$$\Delta \varphi=\frac{2\pi {f}_{o}d}{c}\cdot \sin {\theta }_{{des}},$$
(1)

where \(c\) denotes the light speed in free space, \({f}_{o}\) represents the operational frequency, and \(d\) accounts for the distance between the adjacent meta-columns.

The metasurface under test in this section has eight meta-columns, each spaced 24 mm apart, and the incident frequency is 4.12 GHz. Firstly, we set the modulation frequency of the controlling signals for the STC metasurface to 200 kHz. To realize two steering angles of −20° and 30°, respectively, space phase gradients are purposely designed by time-shifting the controlling waveform. Figure 3a, b display the calculated and measured scattering beams at 4.1202 GHz, which agree well with each other. In the same manner, we also demonstrate the first-harmonic beams steered to angles 10° and −40° with good performance when the modulation frequency is 100 kHz, as shown in Fig. 3c, d. Please refer to Supplementary Note 2 for more details on the designs of harmonic space phase gradients in these examples.

Fig. 3: The first-harmonic beamforming properties realized by the STC metasurface, in which the incident frequency is 4.12 GHz.
Fig. 3: The first-harmonic beamforming properties realized by the STC metasurface, in which the incident frequency is 4.12 GHz.
Full size image

a, b The beam deflection angles of −20° (a) and 30° (b) when the modulation frequency is 200 kHz. c, d The beam deflection angles of 10° (a) and −40° (b) when the modulation frequency is 100 kHz.

With this feature, the proposed metasurface-based radar can be further optimized to measure the spatial angles of targets with a high resolution. Unlike recent studies on the direction of arrival (DOA) estimations using metasurfaces71,72, which primarily rely on passive detection, our approach actively emits the EM waves toward the target. This active emission enables precise control over the transmitting waveform, frequency, and timing, thereby enhancing the target detection capabilities across diverse environments. In the next section, we will concentrate on target detection and range/velocity measurements.

The STC-metasurface-based receiver

The first advancement achieved in this work is the theory for radar signal dechirp processing. As illustrated in Fig. 1, the primary function of the STC metasurface situated in the receiving section is to directly execute the RF dechirp processing and deliver the resulting multi-tone signals to post-stage circuits to acquire the target’s range and velocity. When the scattered echoes from targets arrive at the receiving metasurface, the signals are reflected again, and the electric field \({E}_{r1}\) is written as

$${{E}_{r1}=\varGamma }_{1}{\cdot E}_{i1},$$
(2)

where \({\varGamma }_{1}\) is the time-varying reflectivity of the STC metasurface that can be regulated by the controlling voltages. \({E}_{i1}\) describes the electric field of the signal impinging on the receiving metasurface, which is the echo from the target. In the field of radar engineering, it can be presented as1,4

$${E}_{i1}\left(t,\,k\right)=\mathop{\sum }\limits_{n=1}^{N}{\sigma }_{n}{e}^{j\left(2\pi {f}_{c}\left(t-\frac{({R}_{n}-2(k-1){v}_{n}{T}_{p})}{c}\right)+j\pi \frac{B}{{T}_{p}}{\left(t-\frac{({R}_{n}-2(k-1){v}_{n}{T}_{p})}{c}\right)}^{2}\right)}$$
(3)

where \(t\) represents the fast time, while \((k-1){T}_{p}\) represents the slow time, in which \(k\) is the kth chirp signal in one frame; \(N\) represents the number of targets; \({\sigma }_{n}\) is proportional to the radar cross section (RCS) of the nth target, range attenuation, and so forth; \({f}_{c}\) is the carrier frequency; \({R}_{n}\) represents the transceiver range, that is, \({{R}_{n}=R}_{t}+{R}_{r}\), where \({R}_{t}\) and \({R}_{r}\) indicate the distances from the transmitter to the nth target and from the nth target to the receiver, respectively; \({v}_{n}\) denotes the radial velocity of the nth target; and \(B\) and \({T}_{p}\) account for the bandwidth and duration of the FMCW signal, respectively. It should be mentioned that Eq. (3) is entirely distinct from the monochromatic plane-wave incidences considered in the previous research on metasurfaces, and it leads to the unique wave-signal-matter interaction on the STC metasurface for radar signal processing.

To realize the dechirp processing on the STC metasurface, we introduce a time-varying reflection phase to cancel out the phase term \(\pi B/{T}_{p}\) in Eq. (3), while preserving the phase terms that contain the range/speed information. Specifically, the reflectivity \({\varGamma }_{1}\) is expressed as

$${{\varGamma }_{1}\left(t\right)=\,e}^{-j\pi \frac{B}{{T}_{p}}{t}^{2}}$$
(4)

From Eq. (4), it can be seen that the amplitude of \({\varGamma }_{1}\) is a constant, and the phase of \({\varGamma }_{1}\) is a quadratic function of time. More importantly, the reflection phase of \({\varGamma }_{1}\) should undergo \(2\pi\) variation in each cycle, which raises an essential requirement for the metasurface to be used. With the mapping relationship between the biasing voltage and reflection phase of the STC metasurface, \({\varGamma }_{1}\) can be accurately regulated in real time. By substituting Eq. (3) and Eq. (4) into Eq. (2), and by neglecting part of constant term and the residual video phase, \({E}_{r1}\) can be deduced as

$${E}_{r1}(t,k)=\mathop{\sum }\limits_{n=1}^{N}{\sigma }_{n}{e}^{\; j2\pi {f}_{c}t}{e}^{-j\left[\frac{2\pi B{R}_{n}}{{{cT}}_{p}}t+\frac{2\pi ({R}_{n}-2(k-1){v}_{n}{T}_{p})}{{\lambda }_{c}}\right]}$$
(5)

Equation (5) is the theoretical foundation of this work, which indicates that the dechirp processing is directly performed on the RF level. It also implies that the RF dechirp processing converts the broadband chirp signals into narrowband signals, substantially reducing the high sampling rates and thereby easing the demands on the entire receiving chain. Specifically, the baseband signal in Eq. (5) manifests a set of discrete frequency tones, where each frequency tone corresponds to a specific range based on the time delay of the reflected signal. When this baseband signal is processed using FFT, the resulting frequency spectrum represents the one-dimensional (1D) range profile, in which the amplitude at each frequency indicates the strength of the reflection from targets at that particular range. Hence, the FFT operation at this stage is commonly referred to as the range FFT.

After detecting the nth target and measuring its distance, we proceed to calculate its velocity. To achieve this goal, the baseband signals from a complete frame are collected. Specifically, Eq. (5) is adjusted as

$${E}_{r1}(t)=\mathop{\sum }\limits_{k=1}^{K}{\sigma }_{n}{e}^{\;-j2\pi \frac{B{R}_{n}}{{{cT}}_{p}}t}{e}^{\;j\frac{4\pi (k-1){v}_{n}{T}_{p}}{{\lambda }_{c}}}$$
(6)

Equation (6) indicates that K returns are obtained from the nth target. These K signals share the same frequency but have different phases. The measured phase difference \(\Delta \varphi=4\pi {v}_{n}{T}_{p}/{\lambda }_{c}\) is the function of the velocity of the nth target. Subsequently, a second FFT, referred to as the Doppler FFT, is applied to the K phasors. This allows the velocity of the nth target to be calculated as \({\Delta \varphi \lambda }_{c}/4\pi {T}_{p}\).

In conclusion, the STC-metasurface-based radar receiver directly performs the dechirp signal processing on the physical platform of the metasurface, and the signals processed by the STC metasurface can be further processed using a 2D FFT to determine the ranges and speeds. Please refer to Supplementary Note 3 for more details. The feasibility of this method will be validated in the experimental section.

The STC-metasurface-based chirp signal transmitter

The ability to transmit the chirp signals into space is the key to an FMCW radar system. As shown in Fig. 1, the transmitting STC metasurface can generate the desired FMCW signals directly. Here, the monochromatic EM waves at the frequency of \({f}_{c}\) are normally incident to the metasurface, and the reflected electric field \({E}_{r2}\) is described as

$${{E}_{r2}=\varGamma }_{2}{\cdot e}^{\;j2\pi {f}_{c}t}$$
(7)

in which \({\varGamma }_{2}\) represents the time-varying reflection coefficient of the STC metasurface. For the chirp generation, \({\varGamma }_{2}\) should be regulated as

$${{\varGamma }_{2}\left(t\right)=\,e}^{\;j\pi \frac{B}{{T}_{p}}{t}^{2}}$$
(8)

in which \(B\) and \({T}_{p}\) represent the bandwidth and duration of the FMCW signal, respectively. It can be seen from Eq. (8) that the reflection phase of \({\varGamma }_{2}\) experiences a 360-degree variance in a single period, which poses a fundamental criterion for the effective functionality of the STC metasurface.

To validate the ability of the STC metasurface to generate the chirp signals, a series of experiments are conducted. The bandwidth \(B\) and duration \({T}_{p}\) of the signals are 100 kHz and \(100\) us, respectively. The experimental setup includes an STC metasurface controlled by a digital module, which converts the incident monochromatic signals into the chirp ones, and a universal software radio peripheral (USRP-2974) that down-converts the chirp signals to the baseband. More details can be found in “Methods”.

The baseband signals are shown in Fig. 4a. In addition, Fig. 4b illustrates a segment of the chirp signals transmitted by the STC metasurface in the form of time-frequency curves. Figure 4a, b indicate that the chirp signals are successfully produced by the metasurface. Besides the chirp signals, some undesired frequency components, such as the fundamental and high-order harmonics, are also reflected by the STC metasurface, as seen in Fig. 4b. The primary cause of the fundamental harmonic is the spatial multi-path effect; the high-order harmonics such as the 1st- and 2nd-order harmonics are mainly the result of fluctuations in the reflection amplitude of the metasurface and the nonlinear characteristics of the controlling components. Nevertheless, due to the identical structure and performance of the receiving and transmitting metasurfaces, the undesired frequency components generated by them are the same. The chirp signals produced by one metasurface can be effectively processed by the other as a matched filter while preserving the ranging and velocity measurement capabilities of the system. It is also highlighted that the reprogrammable feature of the metasurface allows for dynamic regulations of the chirp signal parameters, including bandwidth, making it potentially useful for modern radar anti-jamming techniques.

Fig. 4: The baseband chirp signal generated by the STC metasurface.
Fig. 4: The baseband chirp signal generated by the STC metasurface.
Full size image

a The phase of the generated chirp signal. b Segment of the time-frequency curve of the generated baseband signal.

Experimental validation for the STC-metasurface-based radar system

We conduct a series of experiments to validate the performance of the STC-metasurface-based radar system. The picture of the radar system is presented in Fig. 5. The transmitting section is shown on the left side. A single-tone signal at 4.25 GHz with an instantaneous bandwidth of 1 GHz is generated by a vector signal transceiver (VST, PXIe-5841), which is radiated by a horn antenna (ANT 1) to illuminate the transmitting metasurface. The chirp signal directly generated by the metasurface is transmitted into space. The receiving section is shown on the right side. The radar echoes experience the dechirp processing on the receiving metasurface, and the processed signals are then re-scattered and captured by a horn antenna (ANT 4). ANT 4 is connected to VST, where the signals are converted into baseband signals. Finally, the range and/or speed information of the targets is extracted. The controlling signals for the metasurfaces are provided by a 16-bit waveform generator (WG, PXIe-5433). The VST and WG modules are installed in a multi-slot chassis (PXIe-1092) along with an embedded controller (PXIe-8881), ensuring that the entire system is operated on the same oscillator and maintains coherence.

Fig. 5: The STC-metasurface-based radar system in an indoor scenario.
Fig. 5: The STC-metasurface-based radar system in an indoor scenario.
Full size image

The abbreviations WG, VST, and ANT refer to waveform generator, vector signal transceiver, and antenna, respectively.

In this indoor radar testing environment configuration, a radar echo simulator (details are provided in Supplementary Note 4) receives the signals from the transmitting metasurface via a horn antenna (ANT 2) and re-radiates the echoes encoded with specific motion information of the targets through another horn antenna (ANT 3). In the experiments, the transmitting metasurface emits chirp signals with a bandwidth of 1 MHz, resulting in a range resolution of 0.15 km. Besides that, the pulse width is 100 us, and each coherent processing interval (CPI) consists of 128 chirps. In modern radar systems, it is essential for the transmitting and receiving RF signals to be coherent. Therefore, the low-frequency controlling signals for the two STC metasurfaces are synchronized with a common LO.

For comparison, we construct a conventional software-defined radar (SDR) system based on the National Instruments (NI) platform, whose hardware architecture is shown in Fig. 2b. At the transmitter, digitally modulated signals are generated using a DDS controlled by an FPGA. These signals are then converted to analog signals via a DAC and upconverted to the RF domain using an RF in-phase and quadrature (I/Q) mixer and an LO. The resulting RF FMCW signals are radiated into free space by the antenna. At the receiver, the echo signals are captured by the receiving antenna and coherently downconverted to the baseband by an RF mixer synchronized with the same LO. The resulting analog signals are wideband, so they should be digitized by a high-speed ADC and demodulated via a digital mixer to extract the narrowband beat signals. A DSP module is then used to process these signals using standard algorithms such as spectral analysis and CFAR method for detecting the target and measuring the range-velocity information. The SDR transmits 128 chirps within a single CPI, each with a 1 MHz bandwidth and a 100 us pulse width. Theoretically, this configuration yields the same range and velocity resolutions as the STC-metasurface-based radar.

The first experiment is conducted to illustrate the single-target range measurement using the metasurface-based radar. In the radar echo simulator, a stationary target is set at a distance of 1 km for testing. The result from the proposed radar (in green) is depicted in Fig. 6a, showing the measured distance of 1.09 km, with a measuring error of 0.09 km that falls below the distance resolution of 0.15 km. The measured result from the conventional radar (in green) is 1.09 km too, as shown in Fig. 6b, which is quite close to the result of the proposed radar.

Fig. 6: The ranges (R) and velocities (V) measured by the STC-metasurface-based radar (displayed in green) and the conventional radar (displayed in blue), respectively.
Fig. 6: The ranges (R) and velocities (V) measured by the STC-metasurface-based radar (displayed in green) and the conventional radar (displayed in blue), respectively.
Full size image

a, c, e, g depict the results of the metasurface-based radar, while (b, d, f, h) showcase the results of the conventional radar. The measurement settings: (a)(b). (1 km, 0 m/s); (c, d). (1 km, 10 m/s); (e)(f). (3 km, 0 m/s), (1 km, 0 m/s); (g, h). (3 km, 0 m/s), (1 km, 10 m/s). The measured results: (a). (1.09 km, 0.03 m/s); (b). (1.09 km, 0.03 m/s); (c). (1.02 km, 10.04 m/s); (d). (1.02 km, 10.04 m/s); (e). (3.01 km, −0.03 m/s), (1.04 km, 0.03 m/s); (f). (3.01 km, -0.03 m/s), (1.04 km, 0.03 m/s); (g). (3.01 km, 0.03 m/s), (1.06 km, 9.97 m/s); (h). (3.01 km, 0.1 m/s), (1.02 km, 10.04 m/s).

Then, to validate the joint range-velocity measurement capability of the presented radar system, we set a target with a range of 1 km and a velocity of 10 m/s. In Fig. 6(c, d), the measured results of the proposed radar and the conventional radar are compared. As described in Fig. 6c, the target’s range and velocity are set to 1 km and 10 m/s, respectively, and the measured values from the proposed radar are 1.02 km and 10.04 m/s. The resulting errors in range and velocity, 0.02 km and 0.04 m/s, are both below their respective resolutions, which demonstrates the capability of the STC-metasurface-based radar for accurate measurement of the range and velocity for single target. As for the conventional radar, the measured results are the same as those from the proposed radar, as shown in Fig. 6d.

In practical applications, it is common to have multiple targets that need to be detected simultaneously. Here, we consider two dual-target cases. In Case I shown in Fig. 6e, f, the two targets are set as (3 km, 0 m/s), (1 km, 0 m/s); while in Case II shown in Fig. 6g, h, the two targets are set as (3 km, 0 m/s), (1 km, 10 m/s). Figure 6e, g depict the results measured by the metasurface-based radar, and Fig. 6f, h show the results obtained by the conventional radar. It can be found that the presented radar has successfully detected all preset targets. Specifically, the measured results in Case I are (3.01 km, −0.03 m/s) and (1.04 km, 0.03 m/s), with the errors below 0.04 km and 0.03 m/s, respectively; and the results in Case II are (3.01 km, 0.03 m/s) and (1.06 km, 9.97 m/s), with the errors below 0.06 km and 0.03 m/s, respectively. These results are almost the same as those from the conventional radar. These experiments have demonstrated the satisfactory capabilities of the proposed radar system for target detection and measurement. To enhance the transmission signal bandwidth, a more advanced metasurface prototype can be designed to accommodate significantly higher modulation speeds. This will be studied in future work.

Discussion

We propose a metasurface-based radar system that leverages the STC metasurfaces in both the transmitter and receiver. The STC metasurface in the transmitter generates the FMCW signals from monochromatic EM wave incidences and sends them into space, eliminating the need for conventional VCO or DDS technology. Meanwhile, the STC metasurface in the receiver captures the echo signal and performs the RF dechirp calculations without using the traditional RF circuits or algorithms. It offers a metasurface-based approach for processing the radar signals at the RF front end, which holds several advantages, including improved signal quality, increased processing speed, and reduced power consumption. Several experiments are carried out for the detection and measurement of single and dual targets, and their distances and velocities are successfully acquired with controllable range and velocity resolutions. The results are on par with those from the conventional radar systems, yet they are achieved with significantly simplified hardware and reduced costs. This methodology has demonstrated promising potential for future applications in radar, integrated communication, and sensing technologies.

Despite its demonstrated capabilities, the proposed STC metasurface-based radar faces several limitations that should be resolved. Firstly, the bandwidth of transmitted signals is constrained by the metasurface’s modulation speed. Overcoming this challenge will require advances in meta-atom design, including the optimized unit-cell layouts, improved diode performance, and enhanced high-speed control circuitry. Secondly, environmental robustness presents a significant challenge. The exposed active diodes and control circuits are susceptible to temperature variations and humidity, potentially compromising long-term reliability. Future research should prioritize to develop real-time environmental monitoring and advanced encapsulation techniques to ensure stable operations under diverse conditions. Thirdly, scaling the metasurface leads to substantial power consumption from the integrated diodes. This could be mitigated by employing low-power alternatives like Schottky diodes or developing customized diode designs specifically for metasurface integration.

Methods

Details on the metasurface and measurement

The metasurface in this study is operated at approximately 4.25 GHz and consists of 8 × 16 meta-atoms. This configuration was chosen as a balance between the radar performance and fabrication complexity. While a larger aperture would enhance the beam directivity, angular resolution, and signal-to-noise ratio (SNR), it would also require more diodes and complicate the controlling network. After evaluating these trade-offs, we adopted the layout with 8×16 elements. The meta-atom features a three-layered structure, as depicted in Supplementary Fig. S1a, with the top layer incorporating four chip capacitors and four varactor diodes that bridge adjacent metallic strips. By varying the bias voltage applied to the varactor diodes from 0 to 13 V, the reflection phase was measured. The relationship between the biasing voltage and the reflection phase is illustrated in Supplementary Fig. S1c. It can be seen that the reflection phase range exceeds 2π as the biasing voltage increases, facilitating the generation of arbitrary signals based on the measured mapping relationship. It is apparent that the meta-atom’s full-phase control capability meets the requirements for chirp signal generation and dechirp calculation, as indicated in Eq. (8) and Eq. (4). Simulation details of the meta-atom are provided in Supplementary Note 1.

Experimental setup to validate the chirp signal generation

The experiment was conducted in an indoor scenario. The setup can be categorized into two sections: the transmitting end and receiving end. At the transmitting end, the STC metasurface was controlled by an NI chassis consisting of a high-speed I/O bus controller, an FPGA module, a DAC module, a DC power supply module, etc. A horn antenna is linked to a microwave signal generator (Keysight E8267D) to send monochromatic excitation signals at 4.25 GHz to the metasurface. At the receiving end, another horn antenna is connected to a universal software radio peripheral (USRP-2974) to capture the chirp signals reflected by the STC metasurface and inject them into USRP, which down-converts the chirp signals to the baseband.

Experimental setup for the reflection phase measurement

The experiment was conducted in a microwave anechoic chamber, with the metasurface being controlled by the NI custom chassis (PXIe-1092) comprising a high-speed I/O bus controller, an FPGA module (PXIe-7976R), a timing (PXIe-6674T), an I/O module (PXIe-5783), etc. On the other side, a vector network analyzer was used to receive the reflected signals via a horn antenna and measure the corresponding reflection coefficient S11. As the biasing voltage increases from 0 V to 13 V, the phases of S11 are recorded. The mapping relationship between them is the reflection phase-voltage mapping relationship.