Communication-Assisted Sensing
Understanding Communication-Assisted Sensing
Communication-Assisted Sensing grew out of the observation that a modern wireless transmitter and a radar share most of the same building blocks: a high-frequency front end, wideband data converters, phased-array antennas, and substantial digital signal processing. When a base station or access point radiates a communication signal, some of that energy reflects off buildings, vehicles, drones, and people before returning to a receiver. A conventional communication system treats those reflections as multipath interference to be equalized away. A communication-assisted sensing system instead interprets the same reflections as deliberate sensing returns and extracts geometry from them. The transmitted symbols are already known to the network, so the receiver can correlate the echo against the known waveform, estimate propagation delay and frequency shift, and build a picture of the surrounding scene.
Monostatic, Bistatic, and Network Sensing
Implementations fall into three broad classes. In monostatic sensing the same node transmits and receives, which requires strong self-interference cancellation because the leakage from the transmitter typically exceeds the weak target echo by 100 dB or more; full-duplex isolation and digital cancellation are therefore central design problems. In bistatic sensing one node transmits and a physically separated node receives, removing the self-interference problem but requiring accurate time and phase synchronization between nodes. Network-level sensing extends the bistatic idea across many cooperating base stations, fusing measurements to localize and track targets over a wide area. The choice depends on the available synchronization, the duplexing scheme, and whether the same site must both talk and listen at once.
Waveform Reuse and Its Penalties
The dominant waveform in cellular and Wi-Fi systems is OFDM, and it can be processed for sensing very efficiently. Because each subcarrier carries a known complex symbol, the receiver can divide the received frequency-domain grid by the transmitted grid, which removes the random data modulation and leaves a clean channel response. An inverse transform along subcarriers then yields a range profile, and a transform across successive OFDM symbols yields a Doppler profile, producing a range-Doppler map without the data-dependent ambiguity sidelobes that would otherwise corrupt the result. The penalty is that communication waveforms are not optimized for sensing: they have high peak-to-average power ratio, their bandwidth and frame timing are chosen for throughput rather than range resolution, and pilot density limits the usable sensing aperture. Designers accept these compromises because reusing licensed spectrum and existing radio sites is far cheaper than fielding a parallel radar network.
Why It Matters for 6G
Communication-Assisted Sensing is widely expected to be a defining feature of sixth-generation networks, where the same millimeter-wave and sub-terahertz infrastructure that delivers high data rates can also map traffic flow, detect drones and intruders, support assisted driving, and feed digital-twin models of a coverage area. Higher carrier frequencies and wider channel bandwidths, common at millimeter-wave, directly improve both range and angular resolution, so the move toward mmWave and beyond makes the network a more capable sensor. Standards bodies including 3GPP and IEEE are actively defining the reference signals, frame structures, and measurement reporting needed to make sensing a first-class network service alongside data transport.
Sensing Equations
R = c · τ / 2
Range Resolution (bandwidth-limited):
ΔR = c / (2B)
Doppler Shift from Radial Velocity:
fD = 2 · v · fc / c
Where R = target range (m), c = 2.998 × 108 m/s, τ = round-trip delay (s), B = occupied bandwidth (Hz), ΔR = range resolution (m), fD = Doppler shift (Hz), v = radial velocity (m/s), fc = carrier frequency (Hz). Example: B = 400 MHz gives ΔR ≈ 0.37 m; a target closing at 30 m/s at fc = 28 GHz produces fD ≈ 5.6 kHz.
Typical Sensing Performance by Band
| Carrier / Band | Typical Bandwidth | Range Resolution | Doppler @ 30 m/s | Representative Use |
|---|---|---|---|---|
| Sub-6 GHz (FR1) | 100 MHz | ~1.5 m | ~0.7 kHz @ 3.5 GHz | Wide-area presence, traffic flow |
| Wi-Fi 6E (6 GHz) | 160 MHz | ~0.94 m | ~1.2 kHz @ 6 GHz | Indoor motion, gesture, occupancy |
| mmWave (FR2, 28 GHz) | 400 MHz | ~0.37 m | ~5.6 kHz @ 28 GHz | Drone detect, V2X, short-range mapping |
| mmWave (FR2, 60 GHz) | 2 GHz | ~0.075 m | ~12 kHz @ 60 GHz | High-resolution imaging, vital signs |
| Sub-THz (6G, >100 GHz) | 5 to 10 GHz | ~0.015 to 0.03 m | >20 kHz | cm-scale sensing, digital twins |
Frequently Asked Questions
What is communication-assisted sensing?
Communication-Assisted Sensing is the reuse of communication waveforms, transceivers, and network infrastructure to perform environmental sensing, such as detecting, ranging, and tracking objects. Instead of deploying a separate radar, the system processes echoes of its own communication signals (for example OFDM symbols) as sensing returns. Range is inferred from echo delay and velocity from Doppler shift, so a base station or access point can sense the scene while still carrying data traffic. It is a practical realization of integrated sensing and communication and a foundational capability of 6G perceptive networks.
How does communication-assisted sensing measure range and velocity?
Range is derived from the round-trip propagation delay of the reflected communication signal. Because radio waves travel at the speed of light c, an object at distance R produces a delay τ = 2R/c, so R = c·τ/2. For a monostatic OFDM system the range resolution is set by the occupied bandwidth, ΔR = c/(2B); a 400 MHz millimeter-wave channel yields about 0.37 m resolution. Velocity is estimated from the Doppler frequency shift across successive symbols, fD = 2vfc/c, where v is the radial velocity and fc the carrier frequency. The maximum unambiguous velocity and range are bounded by the symbol spacing and the cyclic prefix duration.
How is communication-assisted sensing different from a dedicated radar?
A dedicated radar transmits a purpose-built waveform optimized only for sensing, while communication-assisted sensing reuses a waveform whose primary job is carrying data, accepting a sensing performance penalty in exchange for spectral and hardware efficiency. Communication waveforms such as OFDM have high peak-to-average power ratio and a random data-dependent ambiguity function, so processing must remove the data modulation (for example by spectral division of the received symbols by the known transmitted symbols) before forming range-Doppler maps. The benefits are reusing licensed spectrum, existing antennas, and deployed base stations, which lowers cost and avoids the spectrum congestion of running separate radar bands.