A Superposition Signaling Scheme for Integrated Sensing and Communication
arXiv:2607.12549
The paper analyzes the capacity of an OFDM-based integrated sensing and communication system under power and kurtosis constraints, characterizes optimal signaling for sensing and communication, and proposes a superposition of constant-modulus and Gaussian inputs that achieves near‑capacity performance.
Abstract
We investigate the capacity of an integrated sensing and communication system operating with orthogonal frequency division multiplexing, where the integrated sidelobe level of the transmit signal is adopted as an input cost. The problem is reduced to the capacity of a complex additive white Gaussian noise channel under power and kurtosis constraints. The sensing-optimal and communication-optimal operating points are characterized by circularly symmetric constant-modulus and complex Gaussian inputs, respectively, while the capacity-achieving input in between has discrete amplitude support with a concentric multi-ring geometry. We propose the superposition of the two extremal inputs as a simple and tunable signaling strategy, whose rate admits an exact expression via successive decoding. Numerical results show that the proposed strategy attains a worst-case gap of approximately 0.043 bits to the capacity.