Summary

Chapter 29 Summary: Integrated Sensing and Communication

Key Points

• 1.

  The ISAC/DFRC paradigm unifies radar sensing and wireless communication into a single system sharing one waveform, one hardware platform, and one frequency band. Key use cases include automotive radar with V2X communication (77 GHz), WiFi-based indoor sensing and gesture recognition (802.11bf), drone detection via 5G NR base stations, and sub-THz imaging for 6G. The fundamental design challenge is managing the trade-off between communication performance (rate, BER) and sensing performance (detection probability, estimation accuracy as bounded by the CRB).

• 2.

  Radar signal processing foundations rest on the matched filter and the ambiguity function $\chi(\tau, f_d)$, which characterises the joint range-Doppler resolution of a waveform. Range resolution is $\Delta r = c/(2B)$ (determined by bandwidth $B$) and velocity resolution is $\Delta v = \lambda/(2T)$ (determined by observation time $T$). Moyal's identity constrains the total ambiguity volume to be constant, so waveform design involves shaping the ambiguity surface rather than shrinking it. Range-Doppler maps are computed efficiently via 2D-FFT processing.

• 3.

  OFDM is a natural ISAC waveform because the communication data modulation can be removed by element-wise division (since the transmitted symbols are known at the ISAC node), leaving pure range-Doppler phase information. An IFFT across subcarriers extracts range; an FFT across OFDM symbols extracts Doppler. Range resolution depends on the total bandwidth $B = K\Delta f$ and velocity resolution on the frame duration $M T_{\text{sym}}$. FMCW chirps remain the dominant waveform for dedicated radar (automotive 77 GHz) due to their constant-envelope property and simple de-chirp processing.

• 4.

  MIMO radar achieves a virtual aperture of $N_T \times N_R$ elements using only $N_T + N_R$ physical antennas by transmitting orthogonal waveforms from each transmit antenna and separating them at the receiver. The angular resolution is determined by the virtual aperture length, providing an $N_T$-fold improvement over a receive-only phased array. Joint communication-sensing beamforming optimises the transmit covariance matrix to simultaneously form communication beams (maximising user SINR) and a sensing beam (maximising beampattern gain at the target angle), with a trade-off parameter $\rho$ controlling the Pareto frontier.

• 5.

  The rate-CRB trade-off characterises the fundamental limit of ISAC systems: the communication rate and the sensing CRB cannot be simultaneously optimised when the communication users and radar targets are in different spatial directions. The trade-off is mild when the target lies in the same direction as a user (the communication beam doubles as a sensing beam) and severe when the directions are orthogonal. Increasing the number of antennas relaxes the trade-off by providing more spatial degrees of freedom.

• 6.

  RF imaging via compressed sensing exploits the sparsity of typical radar scenes to reconstruct spatial reflectivity maps from far fewer measurements than the grid size. The LASSO formulation with $\ell_1$ regularisation is solved by iterative algorithms (ISTA, FISTA) and enjoys RIP-based recovery guarantees requiring only $O(s\log(G/s))$ measurements for an $s$-sparse scene on a $G$-point grid. Deep unfolding approaches (LISTA) accelerate convergence by learning the algorithm parameters from data.