This paper presents an approach for joint radar sensing and communications based on the OTFS (Orthogonal Time Frequency Space) modulation, applied to a moving target detection during regular digital wireless transmission. Most communication systems estimate and correct channel influence using specially designed reference signals or pilots. Pilots are usually known to the receiver before the transmission, and the receiver estimates the channel impulse response (CIR) by analyzing the input (pilots) and output of the channel. Because of the Doppler effect, the movement of reflection sources (or targets in radar terminology) in multi-path channels causes oscillations of CIR coefficients. Based on the frequency analysis of time-varying CIR, one can extract the distances and velocities of moving targets when data is transmitted through the wireless channel. This paper describes the application of OTFS with random pilots using both simulations and real-life field tests. It is shown that applied time-varying CIR estimation based on random pilot signals overcomes the OTFS with zero padding during vehicle velocity estimation in very noisy conditions. The proposed OTFS-based sensing method with the CIR-based channel estimation and vehicle detection is experimentally compared with the cross-ambiguity function (CAF) approach commonly used in radars. A thorough analysis shows that the proposed CIR-based approach outperforms the classical CAF-based radar processing in terms of target detection. Finally, the proposed integrated sensing and communication concept, including bit transmission, is experimentally verified.

In order to recover transmitted bits most receivers have to estimate the channel impulse or frequency response and equalize the channel influence. In multi-path fading channels with Doppler effect the samples of channel impulse response (CIR) oscillates with Doppler frequencies. It is shown in the paper that CIR estimates calculated by means of a 5G NR waveform inside the 5G NR receiver can be used for remote sensing purposes, e.g. finding distances and velocities of object/vehicles moving in the receiver neighborhood. It is demonstrated that appropriately configured standard Channel State Information (CSI) Reference Signal (RS), used in 5G, allows such environment sensing. In order to increase radar functionality some OFDM symbols in 5G frames can be entirely or partly used as additional pilots, i.e. as object illuminators offered in future 5G services. This gives possibility of creating a completely new 5G NR utilization, which can be called a radar-on-demand (RoD). In the paper results from simulational testing of this idea are presented for different pilot configurations and different frequency estimation methods.

In this paper, the possibilities of using the signal of 5G cellular networks for passive radar are investigated. In contrast to the traditional approach, i.e. the passive coherent location (PCL), in which the cross-ambiguity function (CAF) between the transmitted and received signal is calculated, in the presented method, known from the automotive industry, the channel frequency response (CFR) is first estimated, then the channel impulse response (CIR) is computed and spectrally analyzed to obtain a range-velocity map. It is shown that CFR/CIR-based 5G radar, known as an OFDM-based radar, outperforms CAF-based 5G radar in some aspects, e.g. ease of implementation and lower complexity, while maintaining comparable measurement accuracy. In this paper, CFR/CIR is estimated using standard 5G channel state information reference signals (CSI-RS) or some additional radar-on-demand (RoD) OFDM symbols that could be offered by mobile network operators as an extra paid service. Different time and frequency densities of RoD OFDM symbols are tested. The results are compared with the application of CAF and 5G positioning reference-like signals (PRS). The paper shows that even rare CSI-RS pilots can make a low-cost radar device from a 5G receiver. It is demonstrated that slightly irregular sampling of the CIR taps, caused by using cyclic prefixes of different lengths in 5G, does not lead to major velocity estimation errors.

As 5G networks evolve into integrated sensing and communication (ISAC) platforms, OFDM‐based passive radars can significantly enhance sensing by utilising dense user data payloads alongside traditional sparse reference signals. To address the vulnerability of data symbols to decoding errors, we evaluate a data‐driven 5G passive radar under dynamic modulation order, and low‐density parity‐check (LDPC) coding. Our simulations reveal a direct correlation between communication reliability and radar accuracy: Maintaining the standard 10% block error rate (BLER) limit via LDPC coding minimises signal reconstruction errors. Consequently, the peak‐to‐noise‐floor ratio (PNFR) and probability of detection (POD) remain nearly identical to an ideal and error‐free reference. Whereas higher‐order modulations cause a marginal 1–2 dB PNFR degradation compared to QPSK, the POD is unaffected. Finally, under partial resource occupancy, leveraging just 20% of user data guarantees a POD above 90% at signal‐to‐noise ratios (SNRs) exceeding 10 dB.