Processing of signals-of-opportunity (SOP) is often performed on generic processing platforms, such as PCs, using general-purpose signal acquisition systems like USRPs from NI/Ettus Research. Starlink is one of the most promising systems for providing alternative PNT due to its large constellation size. Real-time processing of Starlink signals is challenging because the signals have a relatively large bandwidth (240 MHz) and the system employs eight channels. Fortunately, these signals can be processed in a manner similar to GNSS, and we have shown in previous publications that precise code, Doppler, and, to some extent, carrier-phase observations can be obtained. As with other research groups, all of our previous processing was conducted in post-processing.
To achieve real-time operation, several improvements have been implemented and are presented in this paper: (a) development of a real-time LabView-based interface between the USRPs and the PC via a Thunderbolt connection; (b) synchronization of two USRP-2944 units to provide eight RF channels with a maximum sample rate of 100 MHz; (c) streamlining of memory operations (copy tasks) on the PC to reduce latency; and (d) optimization of CUDA kernels for OFDM replica signal generation and correlation.
A test setup has been installed at UniBwM, consisting of a reference station and a rover mounted in a measurement bus. A conventional upward-pointing LNB is used as the Starlink antenna. Overall, this work summarizes the software and hardware engineering efforts required to achieve real-time Starlink signal processing with two USRP-2944 devices and a high-end gaming PC (AMD Ryzen 9 9950X3D, NVIDIA 5090). Profiling tools such as VTune and NVIDIA Nsight are used to assess processing timing. The focus of this work lies on signal tracking operations.

Critical infrastructure across the globe is dependent on Positioning, Navigation, and Timing (PNT) data and services. Currently, the primary solution to deliver this is the use of Global Navigation Satellite Systems (GNSS) and is integrated into millions of devices across both civil and defence applications. However, GNSS is susceptible to signal obstructions and multipath due to buildings and other infrastructure, as well as interference from space weather, and increasingly from intentional jamming or spoofing. All these effects impact GNSS performance and can cause loss or degradation of PNT information, impacting services that rely on PNT.

One option to improve the resilience of PNT services, by providing an alternative in case of loss of GNSS, is the utilisation of Satellite Signals of Opportunity (SatSOOP). This involves using signals from satellites that are not designed for PNT in an opportunistic manner to generate user level solutions. The surge in launches of satellites and constellations over the past few years (driven by increasing demand in broadband communications) means many candidate satellites have started to be explored in detail as a viable means of generating global PNT.

One advantage of SatSOOP is that there are many different potential satellite constellations with a total of thousands of space vehicles, spanning across various orbital regimes. Additionally, the frequency diversity of the signals – ranging from VHF to Ka-band – greatly benefits PNT resilience to jamming or spoofing, as this makes it very resource intensive for an adversary to disrupt the availability of signals. Finally, the number and diversity of operators means that political and economic factors less likely to impact all satellites in total.

Nevertheless, there are a number of challenges with utilising SatSOOP to resolve PNT. As using the SatSOOP signals are opportunistic and non-cooperative, often little is known about the observed satellite signal structure and ephemerides. These key pieces of information are needed to process captured signals into PNT solutions at competitive accuracies with GNSS. Particularly, the signal structure is different per service provider and can be changeable over time with service demands. Publicly available information on signal structure across different constellations is either limited or unavailable from the service provider or academic research teams. This alone is not enough to resolve PNT solutions at all or to within 1 km of accuracy.

To address this challenge, we have adapted a blind estimation technique to resolve the signal beacon pilot tones for Starlink. Currently, only the synchronisation sequences are available and there is no ICD. Furthermore, insight from literature has shown the signal behaviour has changed over time. By using this technique, this enables us to process the Starlink signals to enable a reasonable Doppler-based positioning solution. This work not only goes over the technique but also how this can be generalised to resolve for signal beacons from other service providers beyond Starlink to enhance alternative PNT capability. As such, this technique moves towards developing a realised SatSOOP receiver system capable of providing PNT solutions to end-users.

The Fused PNT approach, integrating PNT capabilities into K-band broadband Satcom systems, is a critical, rapidly maturing technology. It offers a sustainable, cost-effective, and complementary navigation solution for complex scenarios where GNSS are denied, jammed, or operationally complex. This synergistic concept operates by leveraging specific PNT-driven signals, either intrinsically embedded within or intentionally superimposed upon the main high-throughput Satcom signals. Current research primarily delineates two established implementation paradigms. The Fused Paradigm 1 involves allocating communication link resources for dedicated PNT sequences, such as adapting the 3GPP PRS to NTN. Its advantage lies in protocol transparency and on-demand activation, ensuring compatibility, though performance is inherently limited by dedicated resource allocation. The Fused Paradigm 2 entails systematically superimposing an additional Spread Spectrum (SS) signal onto the main communication carrier across the entire communication bandwidth. The critical challenge is guaranteeing the necessary power budget for reliable PNT ranging while simultaneously limiting interference that may degrade communication service performance.
This paper completes the Fused PNT framework by proposing a possible augmentation via a hybrid, dual-layer constellation structure. The foundational layer consists of the Satcom Constellation, assumed to implement one of the legacy Fused Paradigms. Augmentation uses a dedicated Interoperable Navigation Layer (NavL). The NavL comprises a smaller number of purpose-built satellites or hosted payloads generating dedicated SS signals via a widebeam antenna. The PNT signals are received by the same Satcom terminal but originate from a dedicated satellite transmitter, meaning fusion occurs only at reception level, as the transmission chain is unshared. While this architecture entails increased initial capital expenditure, this cost must be critically evaluated against achieving a truly disruptive boost in the robustness, availability, and overall quality of the navigation service with a reduced number of additional in-orbit transmitters. A crucial feature is signal reception mechanism and geometric diversity. The user terminal can augment baseline Satcom signals with additional PNT signals from the dedicated NavL satellites, which can be received at lower elevation, also from available antenna sidelobes, and are specifically tailored for PNT purposes. To achieve this, signals must be meticulously designed not to interfere with the main communication reception in the same K-band, but to be reliably trackable using the substantial processing gain achieved through de-spreading at the receiver. The core objective of this paper is the comprehensive design of the NavL and its specific signal structure. The design must meet three stringent criteria: minimal unavoidable interference in the communication main lobe, compatibility with spectrum regulatory constraints in the allocated K-band, and the ability to detect the navigation signal under the interference of the data signal. To validate this concept, the paper presents preliminary signal design results and performs a critical assessment of the impact on positioning performance. This evaluation utilizes a Service Volume analysis (SVS) for an illustrative use case centred on augmenting an IRIS-like constellation with the NavL, quantifying the expected performance gains and demonstrating the viability of this advanced Fused PNT approach.

Emerging Third-Generation Partnership Project (3GPP) - compliant systems in non-terrestrial networks (NTN) for Positioning, Navigation, and Timing (PNT) offer a promising complement to existing Global Navigation Satellite Systems (GNSS) and upcoming low-Earth-orbit (LEO) PNT constellations such as Xona. In particular, the Positioning Reference Signal (PRS) defined for 5G provides deterministic and repeatable structures within the orthogonal frequency-division multiplexing (OFDM) frame. These properties enable PRS tracking using matched-filter techniques such as those already applied in GNSS-centric software-defined receivers (SDRs) and recently adopted for tracking the Primary and Secondary Synchronization Signals (PSS/SSS) of Starlink.
To evaluate the potential of future 5G-based PNT systems, the SDR developed at the Institute of Space Technology and Space Applications (ISTA) of the Bundeswehr University Munich (MuSNAT) was extended to support the processing of such 3GPP-compliant reference signals. Assuming phase coherence over an entire 5G frame and employing an FR2-conform numerology, a data-free downlink from a hypothetical NTN constellation to a moving user equipment (UE) was simulated and investigated.
Tracking was performed using an advanced hybrid architecture consisting of a third-order phase-locked loop (PLL) assisted by a second-order frequency-locked loop (FLL) and a carrier-aided second-order delay-locked loop (DLL). The results show stable carrier-phase tracking with highly accurate observables and very low Code-minus-Carrier (CMC), proving the validity of the framework. These findings indicate that maintaining phase coherence in future 5G NTN signals would be highly beneficial for precise PNT, potentially enabling complementary or standalone navigation capabilities independent of traditional GNSS.

Leveraging 5G Non-Terrestrial Networks (NTN) for positioning can complement GNSS(Global Navigation Satellite System) in environments where satellite navigation signals are weak or unavailable. Accurate Time-of-Arrival (TOA) estimation is critical, as the propagation delay directly translates into range measurements through the speed of light, even a nanosecond-level TOA error results in tens of centimetres of ranging error, which fundamentally limits the achievable positioning accuracy. In 5G systems, TOA is commonly obtained by correlating the received signal with a known positioning reference signal (PRS). As PRS bandwidth is configurable, allocating a sufficiently wide bandwidth can, in principle, enable very fine timing resolution. However, in 5G NTN deployments, configuring wide bandwidth PRS incurs a significant trade-off with communication capacity and overall resource efficiency. Moreover, correlation-based TOA remains limited to sample-level accuracy and typically relies on continuous tracking loops for refinement. In NTN scenarios, large Doppler shifts from LEO satellites, combined with the intermittent transmission of PRS, challenge the operation of conventional tracking loops. This motivates TOA estimation techniques that operate in a snapshot mode without relying on a continuous tracking loop. This paper proposes a three-stage snapshot TOA estimation method that combines Doppler estimation using the Demodulation Reference signal (DMRS), coarse TOA estimation via PRS correlation, and refinement of the fractional TOA using phase-slope analysis across PRS subcarriers.
High Doppler shifts can significantly distort the phase of the received signal, causing the correlation peak of PRS to spread and become ambiguous. Thus, the Doppler shift must be compensated before PRS correlation. DMRS is highly suitable for Doppler estimation due to its deterministic sequence and frequent appearance. The predictable and regularly spaced DMRS symbols enable accurate tracking of phase rotation across consecutive symbols. By analysing the inter-symbol phase rotation of the received DMRS, both the instantaneous Doppler frequency and its variation can be precisely estimated. This process facilitates effective Doppler compensation prior to the TOA estimation. To achieve a fractional-sample level accuracy of TOA after PRS correlation process, the proposed method analyses the phase rotation of each PRS subcarrier. Propagation delay produces a linear phase trend across frequencies. By fitting a straight line to the subcarrier phases, the slope directly provides the fractional TOA component, achieving precision beyond the correlation peak resolution. This approach provides a delay resolution comparable to tracking loop-based techniques, without requiring continuous synchronization, thereby enabling snapshot TOA estimation for PRS-based positioning in high Doppler NTN environments.
The proposed method is evaluated in a simulation that emulates a realistic 5G NTN scenario, including LEO satellite Doppler profiles, fading channels and bursty reference signal transmission. Various Doppler shifts, reference signal configuration and SNR levels are tested to demonstrate the method's robustness. The simulation results show that the proposed method achieves robust and highly accurate TOA estimation under high Doppler, wideband reference signal configurations and NTN channel conditions. The proposed approach offers a low complexity, snapshot-capable TOA estimation method that addresses merging 5G NTN positioning needs and complements GNSS in challenging environments.