Dual Functions in One Device: A Portable 220-GHz-Band System for Terahertz Integrated Sensing and Communications
The concept of terahertz integrated sensing and communications (THz-ISAC) has been discussed for years, especially since ITU’s 6G framework began explicitly positioning ISAC as a core capability. Yet, across the research community, progress has largely unfolded along several disconnected tracks. Device-level studies typically focus on either high-speed communications or high-resolution imaging, while algorithm-level research often assumes idealized hardware.
Modern THz applications therefore face significant challenges. High-speed communication prototypes remain short-range, static, and fragile under mobility. High-resolution imaging systems rely on bulky optics and slow mechanical scanning. Most importantly, the two functions almost always reside on entirely independent hardware chains with different frontends and waveforms. “Integration” becomes an afterthought, achieved through mode switching or partial reuse rather than through a unified architectural philosophy. Unsurprisingly, this has fueled growing criticism in recent years: Can THz communications and sensing truly be integrated in practice?
A System-Level Overview from an Operator Perspective
From an operator’s perspective, practical deployment demands far more than an elegant ISAC theory. A deployable device must support mobile terminals, function outdoors, and operate under uncertain sensing conditions. Under these constraints, the traditional separation between sensing and communications increasingly appears not as a historical habit, but as a fundamental obstacle to real 6G-class THz systems. Therefore, our core research problem extends beyond merely asking: “How do we optimize THz communications?” or “How do we improve THz imaging?”, and instead becomes something far more foundational:
We want to build a single THz device, with one hardware chain, one waveform and one signal-processing framework, that performs both tasks naturally and without compromise.
In essence, the community needs a system-level blueprint that anchors future research in physical reality. This became the conceptual core of our work: a portable THz-ISAC prototype with 100% hardware reuse, implemented in a compact 20×20×10 cm³ device designed for practical deployment. As shown in Fig.1, the device supports high-speed mobile communications and high-accuracy free-trajectory imaging using the same RF chain, the same OFDM waveform, and the same processing framework.
Contribution 1 – 100% Hardware Reuse Enabled by HTCC SiP
A practical ISAC system must begin with practical hardware. Today’s THz systems still rely heavily on cascaded waveguide modules linked by bulky mechanical interfaces. Their size, fragility, and complexity make real-world deployment nearly impossible. Our first design question was therefore straightforward:
Can we build a compact integrated THz frontend without sacrificing performance?
Using high-temperature co-fired ceramic (HTCC) system-in-package technology, we developed a 3-D heterogeneous integration platform combining multiple chips and materials into a single structure. The result is one of the most compact THz-ISAC devices to date: an 8-Tx / 4-Rx phased array with strong radiation efficiency and robust beamforming capability. Measured results show an EIRP up to 43 dBm and ±30° electronic beam scanning, placing its performance and integration level among the leading THz systems worldwide.
Most importantly, the device is built upon:
- one RF chain
- one OFDM waveform
- one baseband & processing framework
There are no sensing-only or communication-only modules, no duplicated circuits, and no mode switching. Everything is shared. This transforms the hardware from a set of subsystems into a single physical-layer device capable of expressing both communications and sensing behaviors.
Contribution 2 – Making THz Mobile Communications Real
On top of the unified hardware, we developed a communications system explicitly targeting mobility, something almost never validated at THz frequencies. A long-standing question in the THz community has been:
“THz is great for point-to-point links, but can it actually support mobility?”
Our phased-array frondend supports ±30° beam scanning. While scanning measurement have appeared in literature, no prior work has verified a fully mobile THz communications link in outdoor conditions. Achieving outdoor mobility forced us to rethink beam alignment two fundamental aspects, which it demands extracting geometric cues from real-time reflected signals. This led us to a hierarchical tracking framework:
- STARE algorithm: Performs spatial-temporal clustering to extract trajectory patterns, with special attention to beam-switching regions.
- BRIGHT algorithm: Builds on this with a hidden Markov model that infers motion tendencies and predicts future beam transitions.
In outdoor experiments, with a UE moving at 2 m/s at a 38-m link distance, the system achieved stable 64-QAM transmission over a 12.8-GHz IF bandwidth, yielding a measured throughput of 64 Gbps. The resulting distance–rate product of 2432 m·Gbps, which is over six times higher than the best prior published result, successfully demonstrating that THz mobile communication can be reliable in complex outdoor environments.
Contribution 3 – Making THz Sensing Efficient and Precise
Given that the device adopts a fully shared hardware architecture, communication and sensing can operate simultaneously. However, THz nondestructive evaluation has been studied for years, yet commercial adoption has remained limited by a fundamental bottleneck: THz sensing is too slow. Raster scanning and mechanical stages constrain speed and require bulky optical setups. This naturally led us to ask:
Can we achieve high-efficiency, high-accuracy sensing while simultaneously supporting high-speed communications?
Inspired by this idea, we adopted a free-trajectory scanning approach. A robotic arm moves the device along arbitrary 2-D paths. Using the shared OFDM waveform, the device records reflections along these trajectories.
For reconstruction, we designed a compressed-sensing algorithm tailored for OFDM reflection channels. Using a spherical-wave model and AMP-like residual refinement, the algorithm converges in ~35 iterations and reconstructs millimeter-level structures with only 16% sampling density. A full acquisition and reconstruction cycle completes in roughly 10 seconds. Experiments clearly resolve ~5-mm features, with efficiency improved by more than five times compared to traditional raster scanning.
One Device, Dual Functions — An integrated THz System Ready for future industry
We consider this 220-GHz portable THz-ISAC device marks a transition from laboratory-scale prototypes toward systems that can realistically enter the world for future industry. Rather than treating THz communications and sensing as two separate domains, this device demonstrates that both functions can coexist within a single, compact hardware. Its core breakthroughs include:
- A fully-unified architecture: The first portable device to achieve 100% hardware and waveform reuse for THz communication and sensing, establishing an architectural foundation for THz-ISAC deployment.
- A truly mobile THz communications: Demonstrating stable 2 m/s outdoor mobility, proving engineering viability beyond static laboratory setups.
- State-of-art mobile-communications performance: A distance–rate product of 2432 m·Gbps, exceeding the previous best record by more than six times.
- High-efficiency, high-resolution free-trajectory sensing: Achieving millimeter-level detail with a five times efficiency improvement over traditional scanning methods.
Taken together, these advances shift THz-ISAC from “two worlds patched together” to a single unified system. We hope this device serves not only as a practical milestone but also as a blueprint for future terahertz platforms, pushing THz technology toward broader real-world deployment and making THz-ISAC a fundamental building block of future information infrastructure.