Experimental Validation of SBFD ISAC in an FR3 Distributed SIMO Testbed

Integrated sensing and communication (ISAC) is a key enabler for future radio networks. This paper presents a sub-band full-duplex (SBFD) ISAC system that assigns non-overlapping OFDM subbands to sensing and communication, enabling simultaneous operation with minimal interference. A distributed testbed with three SIMO nodes is implemented using USRP X410 devices operating at 6.8 GHz with 20 MHz bandwidth per channel. A total of 2048 OFDM subcarriers are partitioned into three subbands: two for sensing using Zadoff-Chu sequences and one for communication using QPSK. Each USRP transmits one subband while receiving signals across all three, forming a 1 x 3 SIMO node. Time synchronization is achieved through host-server coordination without external clock distribution. Indoor measurements, validated against MOCAP ground truth, confirm the feasibility of the SBFD ISAC system. The results demonstrate monostatic sensing with a velocity resolution of 0.145 m/s, and communication under NLoS conditions with a BER of 3.63e-3. Compared with a multiband benchmark requiring three times more spectrum, the SBFD configuration achieves comparable velocity estimation accuracy while conserving resources. The sensing and communication performance trade-off is determined by subcarrier allocation strategy rather than mutual interference.

💡 Research Summary

This paper presents a practical implementation and experimental validation of a sub‑band full‑duplex (SBFD) integrated sensing and communication (ISAC) system operating in the FR3 band at 6.8 GHz. The authors build a distributed testbed consisting of three USRP‑X410 software‑defined radios, each equipped with one transmit (Tx) chain and three receive (Rx) chains, forming a 1 × 3 single‑input‑multiple‑output (SIMO) node. All devices share a single 20 MHz channel; the total OFDM bandwidth is divided into three non‑overlapping sub‑bands. Two sub‑bands (≈598 active sub‑carriers each) are dedicated to radar‑type sensing and are populated with Zadoff‑Chu (ZC) sequences, while the third sub‑band carries a QPSK‑modulated communication payload. Guard sub‑carriers are inserted between sub‑bands to provide spectral isolation and suppress leakage.

The signal model defines the transmit waveform for each USRP, the allocation of data and pilot sub‑carriers (pilots every 20 sub‑carriers), and the use of a cyclic prefix. At the receiver, each Rx chain captures the full 20 MHz band, enabling simultaneous extraction of the sensing and communication components. Channel impulse responses are estimated from the known pilots and subsequently used for both radar processing (range‑velocity estimation) and data demodulation.

Hardware architecture includes 12 dBi directional patch antennas with a measured 10 dB return‑loss bandwidth of 200 MHz (6.7–6.9 GHz), 100 GbE interconnects, and a host server equipped with 512 GB RAM for real‑time buffering of raw baseband samples. Time synchronization across the three USRPs is achieved without external clock distribution: each SDR’s internal clock is reset, and a common timestamp (t′1) is broadcast from the host, limiting start‑time offsets to <0.1 ms.

Three experimental configurations are evaluated:

1. SBFD mode (proposed) – all three USRPs operate at the same 6.8 GHz center frequency, each transmitting its assigned sub‑band. This configuration uses only 20 MHz of spectrum while enabling simultaneous sensing and communication.
2. Multiband mode – each USRP transmits on a distinct center frequency (6.74 GHz, 6.80 GHz, 6.86 GHz) with 20 MHz bandwidth, all dedicated to sensing. The total spectrum consumption is 60 MHz, providing a benchmark for performance when more bandwidth is available.
3. Same‑band mode – all USRPs transmit and receive within the same 20 MHz band without sub‑band separation, representing a worst‑case scenario with severe mutual interference.

A motion‑capture (MOCAP) system with four cameras provides ground‑truth trajectories for a human subject walking within the measurement area. The radar processing pipeline uses the ZC sequences to estimate velocity via standard OFDM‑based range‑Doppler processing. Communication performance is assessed by measuring the bit‑error‑rate (BER) of the QPSK stream under non‑line‑of‑sight (NLoS) conditions.

Key results:

• Velocity resolution of 0.145 m/s is achieved using 1 216 OFDM symbols, sufficient for typical human gait analysis. The root‑mean‑square error (RMSE) of velocity estimates in SBFD mode is comparable to the multiband benchmark, despite using one‑third of the spectrum.
• BER for the communication link in SBFD mode is 3.63 × 10⁻³ under NLoS, demonstrating reliable data delivery while the radar sub‑bands operate concurrently.
• The same‑band configuration suffers from pronounced inter‑sub‑band interference, leading to degraded radar accuracy and higher BER, confirming the necessity of guard sub‑carriers and non‑overlapping sub‑band allocation.
• The trade‑off between sensing and communication performance is governed primarily by the proportion of sub‑carriers allocated to each function, not by mutual interference, thanks to the spectral isolation provided by the SBFD design.

The authors conclude that SBFD ISAC can deliver continuous, simultaneous sensing and communication with a markedly reduced spectral footprint, making it attractive for future 5G‑NR and 6G deployments where spectrum is scarce and multi‑service operation is required (e.g., vehicular radar, drone navigation, and ultra‑reliable low‑latency communication). The paper also outlines future work directions, including scaling the testbed to more nodes, dynamic sub‑band reallocation in mobile scenarios, and integration of machine‑learning‑based channel and radar processing to further improve real‑time performance.