21923 - SCTE Broadband Aug2023 COMPLETE v1 (1)

101 Vol. 45 No. 3 - September 2023 Issue technical 100 and 275GHz are 32.5GHz (116−148.5GHz) and 18.5GHz (231.5−250GHz), but with only 12.25GHz (non contiguous) earmarked for unrestricted xed or mobile terrestrial usage 6 . Transmissions are strictly forbidden in 33.5GHz of spectrum, and conditioned to the protection of coexisting passive users in the remaining spectrum. These conservative regulations apply even without passive users exploiting the spectrum for sensing. This prevents multiplexing of unused resources, making this spectrum less appealing for wireless backhaul. To foster wireless innovation in years to come, it becomes necessary to develop spectrum sharing solutions between communications and passive sensing systems – which is the main contribution of this paper. Moreover, while terrestrial sensing stations can be protected through geographical separation, orbiting satellite systems call for dynamic sharing solutions based on self-adaptive wireless links. Indeed, we show—through an accurate link budget based on ITU channel models—that active transmitters may actually generate harmful interference to highly sensitive sensors on satellites orbiting over communication systems. Shared access between sensing and communications has never been demonstrated experimentally in the spectrum above 100GHz. This is not without reason. Current spectrum policies prohibit transmissions in passive bands, even for experimental purposes. Additionally, a self-adaptive wireless link requires the design of dynamic sub-terahertz RF front- ends and protocol stacks that need to be controlled and recon gured in real time. However, existing transceivers designed for this portion of the spectrum are limited to channel sounding or static baseband processing capabilities, with limited or no recon gurability, and are thus un t to demonstrate dynamic spectrum access. This article presents a dynamic spectrum access, dual- band backhaul system in the above-100GHz spectrum, speci cally, at 123.5–140GHz (referred to as Lower Band (LB) throughout the paper) and 210–225GHz (Upper Band (UB)). This is the rst above-100GHz wireless system that (i) operates a real-time protocol stack in two frequency bands; (ii) performs automated and dynamic spectrum sharing; and (iii) adapts link parameters (e.g., the frequency band) in real time. The proposed system tracks satellite orbits and automatically switches frequency bands, thus preventing backhaul transmissions from interfering with passive incumbents, while enabling the allocation of large, contiguous bands for communication services on shared spectrum. Prototyping this system has required us to address challenges that span from RF design for the dual-band setup, to dynamic control and integration of multiple independent systems through open interfaces. We evaluate the backhaul prototype in an experimental setting with transmitter and receiver deployed outdoor in an urban environment, on the roofs of two buildings, under the limits of an FCC experimental license. We believe that the results shown in this paper can pave the way for more exible regulations, protecting scienti c uses of the spectrum while allowing more reasonable allocations for future generation of communication systems. Results This section introduces the proposed self-adaptive dual- band backhaul link. We discuss the challenges and lessons learned in the development of the RF and communication components, the dynamic switching framework, and present its experimental performance evaluation. RF and communications components Figure 1 shows the prototype hardware, with Transmitter (TX)/Receiver (RX) real-time signal processing back-ends and dual-band sub-terahertz front-ends. The NI (formerly known as National Instruments) mmWave back-end features a hardware-accelerated protocol stack that generates an Orthogonal Frequency Division Multi-plexing (OFDM) waveform inspired by the 5G New Radio (NR) physical (PHY) layer, aggregating eight component carriers with 100MHz each. The dual-band front-ends, supporting 123.5–140GHz and 210–225GHz, accept an Intermediate Frequency (IF) input at the up-converters (TX-side) and generate an IF output at the down-converters (RX-side). They are connected to the baseband-to-IF conversion modules of the NI system through a Single-Pole Double-Throw (SPDT) RF switch, which supports DC–18GHz. Four Keysight E8257D Performance Signal Generators (PSGs) drive the Local Oscillators (LOs) of the dual-band front-ends. The overall system design was hindered by two main challenges. First, the same IF output, with high peak-to- average power ratio, needs to be interfaced with dual-band front-ends with different and limited input power speci cations. Second, phase noise and dual sideband front-ends introduce amplitude impairments in the received signal. The link budget planning aims at having a 10 decibel (dB) SNR margin at the IF input of the receiver back-end, independently on the LB or UB front-ends. This is challenged by two factors.