Research on pseudolite signal quality monitoring and availability assessment methods for indoor positioning

Indoor pseudolite positioning systems can provide GNSS-like ranging signals in GNSS-denied environments, but their signal quality is easily degraded by indoor multipath, attenuation, antenna displacement, blockage, and receiver tracking instability. This study proposes a multi-source signal quality assessment method for indoor pseudolite positioning based on monitoring-station observations. An experimental platform was established with pseudolite signal generation, transmitting antenna deployment, receiver-side acquisition, and GNSS-SDR processing. Three manually controlled degradation scenarios were designed: transmitting antenna displacement, signal attenuation, and signal blockage. Signal-, correlation-, tracking-, and observation-domain indicators were extracted, including C/N0, EVM, SCB-related indicators, tracking errors, lock indicators, pseudorange residuals, carrier-phase residuals, Doppler residuals, and observation-domain C/N0 residuals. These indicators were converted into degradation-oriented residuals, robustly normalized, and fused into an Indoor Pseudolite Quality Index (IPQI). Experimental results show that different indicators exhibit distinct responses to different degradation mechanisms. The proposed Indoor Pseudolite Quality Index further distinguishes transient antenna displacement, sustained signal attenuation, and intermittent signal blockage, demonstrating its effectiveness for integrated signal quality monitoring and scenario-level degradation assessment in indoor pseudolite positioning systems.

With the rapid development of intelligent manufacturing, underground-space exploitation, emergency response, warehouse logistics, and public safety applications, the demand for accurate and reliable indoor positioning has increased significantly. In contrast to open-sky outdoor environments, indoor environments are characterized by severe blockage and attenuation of Global Navigation Satellite System (GNSS) signals caused by walls, floors, ceilings, and metallic structures. As a result, conventional GNSS can hardly provide continuous and reliable positioning services indoors. Various indoor positioning technologies, including Wi-Fi, Bluetooth, ultra-wideband (UWB), radio-frequency identification, vision-based positioning, visible light positioning, and inertial navigation, have therefore been extensively investigated (Zafari et al., 2019; Jiménez Ruiz and Seco Granja, 2017; Rahman et al., 2020; Farahsari et al., 2022; Leitch et al., 2023; Al-Okby et al., 2024). However, these technologies differ in coverage, infrastructure cost, accuracy, compatibility, and long-term stability, and a single technology is often insufficient to simultaneously satisfy the requirements of wide-area coverage, high accuracy, and high reliability.

A pseudolite is a ground-based or indoor transmitter that broadcasts GNSS-like ranging signals. By receiving signals from multiple pseudolites, a user receiver can estimate its position using pseudorange, carrier-phase, or Doppler observations. Compared with communication-signal-based indoor positioning technologies such as Wi-Fi and Bluetooth, pseudolite systems have several attractive features, including a clear navigation-signal structure, explicit ranging models, controllable transmitter deployment, and potential compatibility with GNSS receiver architectures. Therefore, pseudolites have been regarded as an important supplementary solution for high-accuracy positioning in GNSS-denied and indoor environments (Xu et al., 2015; Vervisch-Picois and Samama, 2009; Liu et al., 2023). Previous studies have demonstrated that pseudolite-based indoor positioning can achieve improved performance through time-bias estimation, carrier-phase processing, constrained positioning, clock synchronization, and stochastic error modeling (Guo et al., 2018; Wang et al., 2019; Sheng et al., 2020; Wang et al., 2020; Sun et al., 2021; Sun and Wang, 2022; Sheng et al., 2022).

Despite their potential, pseudolite indoor positioning systems still face significant challenges due to complex indoor signal propagation. First, walls, floors, ceilings, equipment, and metallic objects can generate strong multipath effects, which distort the correlation peak and further introduce pseudorange biases and positioning errors. Second, non-line-of-sight propagation may produce systematic ranging errors that cannot be easily mitigated by simple averaging or conventional random-noise models (Sun and Wang, 2022). Third, because pseudolite transmitters are deployed relatively close to the receiver, large received-power differences may cause near-far effects and degrade the acquisition and tracking of weak signals. In addition, transmitter clock synchronization errors, power fluctuations, transmitter coordinate calibration errors, and receiver tracking-loop failures can directly affect measurement quality and final positioning performance (Guo et al., 2018; Vervisch-Picois and Samama, 2009; Wang et al., 2020). Therefore, the performance of a pseudolite indoor positioning system depends not only on the positioning algorithm, but also on signal quality, measurement quality, geometric configuration, and indoor propagation conditions.

Existing studies on pseudolite indoor positioning have mainly focused on positioning models, ambiguity resolution, time synchronization, geometric deployment, and multipath mitigation. For example, Sun et al. proposed indoor precise point positioning and indoor PPP-RTK methods using estimated time biases to handle clock-related issues in pseudolite systems (Sun et al., 2021). Sun and Wang investigated stochastic modeling for the mitigation of multipath and non-line-of-sight errors in ground-based indoor positioning (Sun and Wang, 2022). Sheng et al. proposed indoor PPP algorithms for constrained pseudolite environments, improving positioning robustness under poor geometry and nonlinear conditions (Sheng et al., 2020; Sheng et al., 2022). Other studies optimized pseudolite deployment geometry or investigated Doppler-based positioning strategies (Gan et al., 2019; Tang et al., 2021; Liang et al., 2025). These studies have improved pseudolite positioning performance from the perspectives of positioning models, deployment optimization, and error mitigation. However, the monitoring and interpretation of signal-quality degradation in indoor pseudolite systems have received relatively less attention.

In the GNSS field, signal quality monitoring and measurement quality assessment have been widely studied. Signal deformation, correlation-domain distortion, multipath-induced tracking bias, and abnormal waveform monitoring are important for navigation signal integrity and service reliability (Phelts and Akos, 2006; Pagot et al., 2018; Wesson et al., 2018; Wang et al., 2022; Zhang et al., 2023). Measurement quality indicators such as carrier-to-noise density ratio, observation residuals, cycle slips, and observation completeness have also been used to evaluate the reliability of GNSS observations. More recently, signal quality assessment concepts have been introduced into pseudolite monitoring. Zhang et al. developed a pseudolite monitoring station and proposed a signal quality assessment framework for distributed array pseudolite systems (Zhang et al., 2026). Nevertheless, most existing studies mainly evaluate individual signal or measurement indicators, while an integrated and degradation-oriented quality assessment method for indoor pseudolite positioning systems remains insufficiently developed.

Therefore, it is necessary to establish a comprehensive quality assessment method for pseudolite indoor positioning systems by jointly considering signal status, correlation peak distortion, tracking stability, and observation-domain consistency. A single indicator is usually insufficient to characterize the actual system state. For instance, a high C/N0 does not necessarily indicate high observation quality, because strong multipath may increase the received power while simultaneously introducing pseudorange biases. Similarly, favorable positioning geometry cannot guarantee reliable positioning if the measurements are contaminated by non-line-of-sight propagation, signal blockage, or tracking instability. Hence, multi-source quality features should be extracted from different receiver processing stages and integrated into a unified quality assessment framework.

Motivated by the above issues, this paper studies signal quality monitoring and degradation-oriented quality assessment for pseudolite indoor positioning systems. The proposed work focuses on three aspects. First, a multi-source quality feature set is constructed from signal strength, correlation-domain distortion, tracking stability, and observation-domain residuals. Second, controlled degradation experiments are designed to evaluate the response characteristics of different indicators under transmitting antenna displacement, signal attenuation, and signal blockage. Third, an Indoor Pseudolite Quality Index (IPQI) is constructed by converting multi-source indicators into degradation-oriented residuals, applying robust normalization, and fusing them into a compact quality score. The proposed method provides an integrated representation of signal quality degradation and supports scenario-level comparison of different indoor disturbance mechanisms.

The main contributions of this paper are summarized as follows: A multi-source signal quality feature framework is developed for pseudolite indoor positioning by integrating signal strength, correlation peak distortion, tracking stability, and observation-domain residuals.

Controlled indoor degradation experiments are conducted under transmitting antenna displacement, signal attenuation, and signal blockage, providing real measurement data for analyzing the sensitivity of different quality indicators.

A degradation-oriented IPQI is proposed to fuse multi-domain quality indicators into a unified assessment score, enabling integrated comparison of transient, sustained, and intermittent signal degradation patterns in indoor pseudolite positioning systems.

Overall methodology and data processing

The overall workflow of the degradation-oriented signal quality assessment method is illustrated in Figure 1. This study develops a signal quality assessment method for indoor pseudolite positioning systems. Indoor pseudolite signals are strongly affected by multipath propagation, signal attenuation, antenna placement, and temporary blockage. These effects are reflected not only in signal strength but also in correlation-peak shape, tracking-loop stability, pseudorange consistency, and positioning geometry. Therefore, a single indicator such as C/N0 is insufficient to represent positioning service quality.

The proposed method integrates monitoring information from the signal, correlation, tracking, measurement, and geometry layers. Based on these features, an Indoor Pseudolite Quality Index (IPQI) is constructed to provide a compact degradation score for scenario-level quality assessment. The framework quantifies the overall system degradation state rather than predicting positioning error.

The main innovation of the proposed method is that monitoring data are used not only to detect abnormal signal behavior, but also to quantify the overall degradation state of the indoor pseudolite positioning system. By integrating signal-, correlation-, tracking-, and observation-domain indicators, the method shifts signal quality monitoring from isolated indicator inspection to multi-source degradation-oriented quality assessment.

Pseudolite observation model

The receiver position is given by Equation 1. The position of the i -th pseudolite transmitter is given by Equation 2.

The pseudorange observation of the i -th pseudolite is modeled by Equation 3. where ρ i is the pseudorange observation, r i is the geometric distance between the receiver and the transmitter, c is the speed of light, δ t u is the receiver clock offset, δ t i is the transmitter clock offset, m i is the multipath-induced error, n i is the non-line-of-sight error, and ε i is the remaining random noise.

The geometric distance between the receiver and the i -th pseudolite is given by Equation 4.

Compared with outdoor GNSS positioning, indoor pseudolite positioning has different error characteristics. Atmospheric delays are negligible, while local propagation and hardware-related effects become dominant. Multipath changes the correlation peak shape, blockage reduces signal strength and tracking stability, and transmitter or receiver clock instability introduces systematic ranging errors. Therefore, the observation quality must be evaluated jointly from the signal domain and the measurement domain.

Data preprocessing

The monitoring station records observations by epoch and PRN. Each record contains signal strength, correlator outputs, tracking errors, lock indicators, pseudorange observations, and signal quality monitoring indicators. Since the raw data are dense and may include short-term spikes caused by tracking loop fluctuations or abnormal correlation outputs, direct point-by-point analysis is not suitable for stable quality assessment.

The preprocessing procedure includes four steps. First, empty columns and invalid fields are removed. Second, the time records are converted into a unified time axis starting from the first observation epoch. Third, the data are grouped by PRN to preserve the channel-specific behavior of different pseudolite signals. Fourth, each PRN channel is resampled using a 1 s median window.

For the i -th PRN in the k -th second, the median-resampled value is calculated using Equation 5.

The median operator is selected because it is robust to outliers. This is important for SQM data, where abnormal SCB values, temporary lock loss, or isolated correlator distortions may occur.

For pseudorange observations, the absolute value may contain receiver time-reference terms or internal software receiver offsets. The relative pseudorange variation with respect to the reference period is calculated using Equation 6. where ρ i r e f is the pseudorange sequence of the i -th PRN during a selected reference period. This relative representation is more suitable for analyzing changes caused by antenna displacement, signal attenuation, or blockage.

Multi-source signal quality feature extraction

The multi-source signal quality feature structure used for degradation-oriented IPQI construction is summarized in Figure 2. The framework organizes the monitoring outputs into four complementary feature groups: signal strength, correlation peak, tracking stability, and measurement quality. Geometry-related indices are evaluated separately to characterize the spatial configuration of the pseudolite system. Together, these features characterize signal degradation at different stages of the receiver processing chain and provide the inputs required for subsequent IPQI construction.

Feature design principle

The feature set is designed according to three principles. First, the features should be physically interpretable, so that the source of quality degradation can be traced. Second, the features should be sensitive to typical indoor degradation mechanisms, including multipath, attenuation, blockage, and tracking instability. Third, the features should be computable from monitoring-station outputs without requiring additional hardware.

Based on these principles, five feature groups are constructed: signal strength features, correlation peak features, tracking stability features, measurement quality features, and geometry-related features. These groups describe the signal quality from different stages of the receiver processing chain.

Signal strength features

Signal strength is the most direct indicator of received signal availability. The carrier-to-noise density ratio, used here as the primary signal-strength feature, is defined in Equation 7.

A decrease in C / N 0 may indicate signal attenuation, antenna shielding, poor antenna placement, or partial blockage. However, in indoor environments, high C / N 0 does not always guarantee good positioning accuracy, because strong reflected signals may also increase the received power. Therefore, C / N 0 is used together with other quality features rather than as the only assessment criterion.

The prompt correlator amplitude is calculated using Equation 8. where I P and Q P are the in-phase and quadrature prompt correlator outputs. This value provides an additional representation of correlation strength.

Correlation peak features

Indoor multipath and NLOS propagation mainly affect the shape of the correlation peak. Therefore, correlation-domain features are important for distinguishing between simple signal attenuation and multipath-induced measurement bias.

The early correlator amplitude is calculated using Equation 9. The late correlator amplitude is calculated using Equation 10.

The multipath indicator is defined in Equation 11.

The correlation symmetry indicator is defined in Equation 12.

When the correlation peak is symmetric, the early and late correlator amplitudes are close. A larger MPI or CSI indicates stronger correlation distortion and a higher probability of multipath or NLOS propagation.

In addition, EVM, SCB, and SCB-r are used as correlation-domain SQM indicators. EVM reflects the deviation between the observed signal and the ideal signal state. SCB reflects the bias of the discriminator S-curve and is sensitive to correlation peak distortion. SCB-r reflects the reliability or similarity of the observed S-curve. In general, larger EVM and |SCB| indicate poorer signal quality, while SCB-r closer to 1 indicates a more stable signal shape.

Tracking stability features

Signal degradation also affects the receiver tracking loops. Even when the signal can be acquired, poor tracking stability may lead to unstable pseudorange observations and degraded positioning accuracy. The code tracking error is defined in Equation 13. The carrier tracking error is defined in Equation 14. The carrier lock test is defined in Equation 15.

Smaller e i, k c o d e and e i, k c a r r indicate more stable tracking. A larger L i, k indicates a more reliable carrier lock state. These features are particularly useful for detecting signal attenuation and blockage, which often cause loop instability before complete loss of lock.

Measurement and geometry features

Measurement quality describes whether the extracted observations are internally consistent. The pseudorange residual RMS is calculated using Equation 16. where v ρ, i, k is the pseudorange residual of the i -th PRN at epoch k. The observation availability ratio is calculated using Equation 17.

A decrease in R o b s, k indicates that fewer pseudolite observations are available for positioning.

The positioning geometry is evaluated using dilution of precision. The covariance factor matrix based on the design matrix H is given by Equation 18. GDOP is calculated using Equation 19. HDOP is calculated using Equation 20. The number of valid PRNs is defined in Equation 21.

Measurement and geometry features provide a link between signal quality degradation and final positioning performance. This is essential because not all signal anomalies have the same impact on positioning accuracy.

IPQI construction for degradation-oriented quality assessment

The extracted indicators describe signal quality from different processing stages, but they have different units, dynamic ranges, and physical meanings. Some indicators decrease when the signal quality becomes worse, such as C / N 0 and carrier lock, whereas others increase under degraded conditions, such as EVM, multipath-related indicators, tracking errors, and observation residuals. Therefore, before constructing the Indoor Pseudolite Quality Index (IPQI), all indicators are first converted into degradation-oriented residuals. In this study, a larger IPQI value indicates poorer signal quality.

A reference period is selected from the stable observation interval before artificial disturbance is introduced. For the j -th indicator, the reference value is the median over the selected reference period, as defined in Equation 22. where T r e f denotes the reference period. In the experimental analysis, the first 30 s of each scenario is used as the reference period because the pseudolite system is stable and no artificial degradation is introduced during this interval.

For indicators that decrease under degradation, the degradation residual is defined in Equation 23.

This type of indicator includes C / N 0, carrier lock test, observation availability, and other positive-quality indicators.

For indicators that increase under degradation, the degradation residual is defined in Equation 24.

This type of indicator includes EVM, S C B, MPI, code tracking error, carrier tracking error, and other instability-related indicators.

For observation-domain indicators, the absolute deviation from the reference value is defined in Equation 25.

This formulation is used for pseudorange residual, carrier-phase residual, Doppler residual, and other measurement-domain deviations. Through this transformation, all indicators have the same physical direction: a larger d j t indicates stronger degradation.

Because the degradation of residuals still has different units and numerical ranges, robust normalization is applied. The robust scale factor for each indicator is the 95th percentile of its degradation residual, as defined in Equation 26.

The normalized degradation score is calculated using Equation 27. where q j t ∈ 0, 1. A value close to 0 indicates that the indicator remains close to the reference state, whereas a value close to 1 indicates severe degradation. The use of the 95th percentile reduces the influence of extreme outliers while preserving sensitivity to abnormal signal behavior.

For practical real-time monitoring, the reference period is established through stability-gated initialization rather than from future observations. A rolling 30-s window is accepted only when all required PRNs remain available, observation completeness is stable, and no disturbance alarm is active; the median of the accepted window then initializes the reference value. The normalization scale is initialized as the 95th percentile of the degradation residual from a representative commissioning dataset spanning normal operation and the expected operating range. During subsequent operation, both parameters may be recalibrated at scheduled intervals using only verified stable data and are frozen whenever a warning or alarm is active. This procedure prevents degraded samples from contaminating the reference while allowing slow receiver and environmental drift to be accommodated without using future observations.

The IPQI is obtained by fusing the normalized degradation scores, as shown in Equation 28. where M is the number of selected indicators, q j t is the normalized degradation score of the j -th indicator, and w j is the corresponding weight. The indicator weights satisfy the constraint in Equation 29.

In this study, equal weights are used for the selected indicators to avoid introducing subjective preference toward a specific observation domain. The fused indicators include signal-strength degradation, correlation-domain distortion, tracking instability, and observation-domain residuals. Specifically, the IPQI is constructed from C / N 0 degradation, EVM increase, S C B increase, MPI increase, code tracking error increase, carrier tracking error increase, carrier lock degradation, pseudorange residual, carrier-phase residual, Doppler residual, and observation-domain C / N 0 degradation.

The IPQI therefore provides a compact representation of the overall degradation state of the indoor pseudolite positioning system. A low IPQI indicates that the monitored signal and observation features remain close to the reference state. A high IPQI indicates that one or more signal-processing stages have deviated from the normal condition. The quality-state classification is defined in Equation 30. where T 1 and T 2 can be calibrated using reference observations and controlled degradation experiments. For example, they can be determined from the upper percentiles of the IPQI distribution under normal and degraded conditions. This thresholding strategy allows the monitoring system to provide a direct interpretation of the current service-quality state.

IPQI does not directly output or predict positioning error. Instead, it provides an integrated representation of signal and observation degradation that is related to positioning service reliability. A higher IPQI indicates a stronger deviation from the reference condition and a higher risk of degraded positioning performance. Accordingly, this study uses IPQI as a scenario-level quality assessment index to compare degradation mechanisms and identify periods of abnormal signal behavior.

Experimental scenarios and validation

Three typical degradation scenarios are designed to verify the proposed method.

Antenna displacement: Selected transmitting antennas are physically moved during data collection. This scenario evaluates the effects of transmitter-position changes and local propagation variation. Transmitting antenna displacement may alter received power, correlation-peak shape, tracking stability, and pseudorange consistency.

Signal attenuation: Part of the pseudolite signal is attenuated. This scenario is used to evaluate the effect of received power reduction. Signal attenuation is expected to reduce C / N 0, degrade tracking stability, and increase the probability of invalid observations.

Signal blockage: Part of the signal propagation path is blocked. This scenario is used to evaluate the effect of NLOS propagation and obstruction. Signal blockage may cause power reduction, correlation peak distortion, tracking instability, and pseudorange residual increase.

Observations from all available PRNs were collected in each scenario. For the detailed response analysis in Figures 3–6, PL01 was selected as a representative link so that the same channel could be compared consistently across the three scenarios; no inter-PRN averaging was applied. The indicators were compared before and after each controlled event to evaluate whether the selected features and IPQI captured the associated physical degradation mechanism.

To determine the event boundary objectively, a sliding-window difference method is used. For a candidate change time, the sliding-window change score is defined in Equation 31. where μ j b e f o r e and μ j a f t e r are the mean values of the j-th indicator before and after τ, σ j is the standard deviation of the j -th indicator over the full observation period, and m is the number of indicators.

The estimated change point is determined using Equation 32.

The indicators used for change detection include mean C / N 0, mean EVM, mean |SCB|, mean code tracking error, and mean carrier lock test. If the actual operation time is available, it can be used directly as the event boundary. Otherwise, τ * provides an automatic estimate of the major quality transition.

Experimental setup and data overview

The experimental platform was designed to evaluate the signal quality and positioning performance of a pseudolite indoor positioning system under controlled indoor degradation conditions. The test was carried out in an indoor environment with ceiling-mounted pseudolite antennas, receiver-side equipment, cables, walls, and other structural elements, as shown in Figure 7a. In addition to the inherent multipath and attenuation effects caused by the indoor environment, external manual interventions were introduced to deliberately create representative signal degradation scenarios, including signal attenuation, transmitting antenna displacement, and signal blockage.

The hardware system consists of a pseudolite signal generation subsystem, transmitting antenna groups, a receiver-side acquisition platform, and software-defined receiver processing. The pseudolite transmitter generates GNSS-like spread-spectrum signals, which are distributed to multiple transmitting antennas installed in the test area. The transmitting antennas are organized into three groups, namely PL01-06, PL07-12, and PL13-18, as illustrated in Figure 7b. This grouped deployment provides spatially separated signal sources and makes it possible to observe how local signal disturbances affect different pseudolite links.

The spatial configuration of the transmitting antennas was determined from surveyed coordinates and transformed into a room-aligned two-dimensional coordinate frame. The resulting layout shows the relative positions of the three antenna groups with respect to the room boundary, as presented in Figure 7c. This coordinate-based layout serves as the geometric reference for subsequent analysis, especially for interpreting signal changes caused by antenna displacement, local obstruction, and power attenuation.

The data acquisition chain contains two main functional parts: signal generation and signal reception. In the signal generation part, the pseudolite signal generator produces the GNSS-like signal, which is then converted to RF, adjusted by the power control unit, and radiated by the transmitting antennas. In the signal reception part, the receiving antenna collects the indoor pseudolite signals, and the USRP RF/IF front-end performs down-conversion and sampling. The sampled IQ data are stored by the IQ data recorder and further processed by the GNSS-SDR software receiver, as summarized in Figure 7d. Through acquisition, tracking, and demodulation, the software receiver outputs observables such as pseudorange, carrier phase, Doppler frequency, and carrier-to-noise density ratio.

Before the controlled degradation experiments, a static pseudolite positioning test was conducted to characterize the nominal geometry and observation quality of the indoor deployment. Figure 8 presents the geometry and observation quality characteristics under static pseudolite positioning conditions. The GDOP distribution in Figure 8a indicates that the indoor geometry is generally favorable over most of the test area. In particular, the region surrounded by the three transmitting antenna groups has GDOP values mostly below 2, suggesting that the spatial configuration can provide strong geometric constraints for positioning. A static test point was selected within this well-conditioned region, as marked in Figure 8a, and a 2-h static observation was conducted at this location.

The PRN visibility result in Figure 8b shows that all pseudolite signals from PL01 to PL18 were continuously received during the 2-h observation period. No evident interruption or loss of signal availability was observed, indicating stable tracking conditions for all transmitting links. Figure 8c further summarizes the C / N 0 statistics of the 18 pseudolite signals during the same 2-h period. The overall C / N 0 has a mean value of 56.22 dB-Hz and a median value of 56.20 dB-Hz, indicating that the received signal strength was generally stable under the static condition. The mean C / N 0 values among different PRNs range from 53.29 to 58.85 dB-Hz, showing moderate link-dependent differences caused by antenna position, propagation path, and hardware characteristics. In terms of temporal fluctuation, the overall standard deviation is 7.16 dB-Hz and the interquartile range is 10.94 dB-Hz, suggesting that most C / N 0 observations remain within a relatively concentrated range.

During the experiment, three types of manually controlled disturbances were introduced to emulate typical indoor signal degradation mechanisms, as summarized in Table 1. The experimental datasets include antenna displacement, signal attenuation, and signal blockage, each consisting of three repeated groups with a duration of 10 min per group. The nominal sampling interval was 0.02 s, resulting in 30,000 expected epochs for each degradation scenario. All experiments used pseudolite signals in the BeiDou B1 frequency band.

For the antenna displacement scenario, selected transmitting antennas were physically moved to modify the propagation geometry and received signal characteristics. For the signal attenuation scenario, the effective received signal strength was reduced by adjusting the transmitter configuration. For the signal blockage scenario, obstacles were manually placed to obstruct selected signal paths, leading to reduced signal availability and degraded tracking performance. These controlled disturbances provide real measurement data for evaluating the sensitivity of the proposed signal quality indicators to different degradation sources.

Signal and receiver-level response analysis

Signal- and receiver-level responses were analyzed using C / N 0, correlation-domain indicators, and tracking stability indicators. These indicators describe different stages of the receiver processing chain: C / N 0 reflects received signal strength, EVM/SCB/MPI describe correlation peak distortion, and tracking indicators reflect the stability of code and carrier tracking.

Unless otherwise stated, the statistics in Figures 3–6 were calculated from PL01 at the nominal 0.02 s sampling interval, without inter-PRN aggregation. In Figure 3, C/N0 means and standard deviations were calculated over complete intervals bounded by the controlled operations: [0, 199.6), [199.6, 383.7), and [383.7, 600] s for antenna displacement; [0, 49.6), [49.6, 194.6), [194.6, 389.6), [389.6, 516.0), and [516.0, 600] s for signal attenuation; and [51.90, 181.90), [229.60, 241.48), [292.82, 300.10), and [375.48, 518.08) s for signal blockage. Figures 4, 5 use the same PL01 time series and controlled-event markers.

Figure 3 shows the temporal C / N 0 response under the three controlled degradation scenarios. For antenna displacement, C / N 0 changed from 49.78 dB-Hz before the first movement to 52.52 dB-Hz after the first event and then returned to 50.01 dB-Hz after the second event. The standard deviation also increased from 0.40 dB-Hz to 3.00 dB-Hz and 2.54 dB-Hz, indicating that antenna displacement not only changed the received signal level but also increased short-term signal fluctuation. For signal attenuation, C / N 0 decreased stepwise from 51.32 dB-Hz to 45.92, 42.29, and 38.25 dB-Hz, before recovering to 49.53 dB-Hz after the final adjustment. This confirms that the transmitter power configuration directly affected the received signal strength. For signal blockage, the C / N 0 response showed stronger interval-dependent fluctuation, with mean values of 39.24, 38.67, 44.81, and 42.43 dB-Hz in the four blockage periods. The corresponding standard deviations were 3.25, 2.90, 3.08, and 2.70 dB-Hz, suggesting that blockage produced more irregular signal variation than simple attenuation.

The interval-wise results in Figure 3 should not be interpreted as immediate event responses. For the first antenna movement, the mean over the complete post-event segment is 52.52 dB-Hz, whereas the local median used in Figure 6 decreases from 49.55 dB-Hz in [169.6, 199.6) s to 40.00 dB-Hz in [199.6, 209.6) s, yielding −19.27%. For the second movement, the corresponding local medians decrease from 53.39 to 44.63 dB-Hz, yielding −16.42%. Thus, Figure 3 characterizes longer-term segment levels, whereas Figure 6 quantifies short-term transients around each movement event.

Figure 4 presents the correlation-domain indicators, including EVM, |SCB|, SCB-r, and MPI. These indicators are sensitive to changes in the correlation peak shape. Under antenna displacement, clear responses appeared near the event times, showing that changes in antenna position affected the signal correlation structure. Under signal attenuation, EVM, |SCB|, and MPI generally increased during stronger attenuation periods, indicating degradation of correlation peak stability. Under signal blockage, the correlation-domain indicators showed local fluctuations around blockage onset, especially for |SCB| and MPI, suggesting that obstruction affected not only signal strength but also the symmetry and stability of the correlation function.

Figure 5 shows the tracking stability indicators. The carrier lock test, absolute code tracking error, and absolute carrier phase error provide receiver-level evidence of signal degradation. Compared with C / N 0 and correlation indicators, these metrics are closer to the final observation quality because they directly describe whether the receiver can maintain stable tracking. For example, during antenna displacement, the event-oriented analysis showed that carrier lock decreased by 11.57% and 4.92% for the two movement events, while |code error| increased by 98.50% and 143.66%. This indicates that even short antenna displacement can propagate into tracking instability.

To summarize the event-level response, the median-based response index is calculated using Equation 33. where M_ref and M_resp denote the median values of an indicator in the reference and response windows, respectively. All values in Figure 6 were calculated for PL01 using median aggregation. For antenna displacement, events E1 and E2 occur at 199.6 and 383.7 s; each reference window covers the preceding 30 s and each response window covers the following 10 s. For signal attenuation, events E1-E4 occur at 49.6, 194.6, 389.6, and 516.0 s; the corresponding windows are 30 s before and 2 s after each adjustment. For signal blockage, B1-B4 are defined by onset times of 51.90, 229.60, 292.82, and 375.48 s; the windows are 30 s before and 10 s after each onset. The constant epsilon prevents numerical instability when the reference median is close to zero.

Figure 6 summarizes the median-based response index for the selected signal- and receiver-level indicators. For antenna displacement, both movement events produced clear short-term degradation, with C / N 0 decreasing by 19.27% and 16.42%, while EVM, MPI, and carrier-related errors increased markedly. For signal attenuation, the response varied with the direction of power adjustment: attenuation events reduced C / N 0 by up to 10.39% and increased EVM by up to 82.66%, whereas the recovery event increased C / N 0 by 22.39% and reduced EVM by 68.84%. For signal blockage, the onset-based analysis captured localized obstruction effects, with the strongest response observed at B2, where C / N 0 decreased by 8.69%, EVM increased by 42.88%, and carrier lock decreased by 17.30%. These results confirm that the proposed event-oriented index can compactly describe both degradation and recovery responses across different disturbance mechanisms.

Figures 3–6 demonstrate that different degradation mechanisms affect different parts of the receiver processing chain. C / N 0 is effective for identifying received power changes, correlation-domain indicators reveal signal deformation, and tracking indicators reflect receiver stability. The event-oriented response index further provides a compact quantitative comparison across different disturbance types.

Measurement-domain response analysis

Figure 9 presents the temporal responses of four observation-domain residuals, including pseudorange, carrier phase, Doppler, and C / N 0. All residuals were computed relative to the median value of the reference period, so the curves mainly reflect the deviation of each observable from its initial stable state. For the antenna displacement scenario, all four indicators show evident changes around the displacement events, indicating that antenna displacement affects not only the received signal strength but also the ranging and carrier-tracking observables. The pseudorange residual, carrier-phase residual, Doppler residual, and C / N 0 residual all exhibit distinguishable responses before and after the displacement, confirming that antenna displacement can be effectively detected in the observation domain.

For signal attenuation and signal blockage, the sensitivity of different observables is not identical. The C / N 0 residual shows clear responses in all three scenarios, which is consistent with the C / N 0 -based analysis in Section 3.2. In the statistical results shown in Figure 10, the RMS of the C / N 0 residual is 2.71 dB-Hz for antenna displacement, 8.49 dB-Hz for signal attenuation, and 5.45 dB-Hz for signal blockage. The corresponding 95th percentiles of the absolute residual are 3.87, 13.43, and 10.05 dB-Hz, respectively. These results indicate that C / N 0 is the most direct and stable indicator for detecting signal strength degradation.

The pseudorange residual also shows sensitivity to different degradation conditions, although its response is less direct than C / N 0. The RMS values of the pseudorange residual are 5.85 m, 11.16 m, and 7.22 m for antenna displacement, signal attenuation, and signal blockage, respectively. The signal attenuation scenario produces the largest pseudorange dispersion, with an IQR of 10.43 m and a 95th absolute residual of 20.88 m. This suggests that transmitter power changes can introduce noticeable ranging-domain instability, probably through degraded tracking quality and reduced measurement consistency.

The carrier-phase residual provides additional information, but its response is less consistent for signal attenuation and blockage. For antenna displacement, the carrier-phase residual has an RMS of 0.77 cycles and an IQR of 1.18 cycles, showing a clear response to geometry and propagation-path changes. In contrast, signal attenuation produces a much smaller carrier-phase residual distribution, with an RMS of 0.11 cycles and an IQR of 0.11 cycles. For signal blockage, the robust statistics remain relatively small, with a median of 0.04 cycles and an IQR of 0.12 cycles, although a few abnormal samples increase the RMS. Therefore, carrier-phase residuals can indicate abrupt or local tracking disturbances, but they are less stable as a standalone degradation indicator.

The Doppler residual is particularly useful for identifying signal attenuation and blockage. Its RMS increases from 1.30 Hz under antenna displacement to 2.19 Hz under signal attenuation and 2.88 Hz under signal blockage. The 95th percentile of the absolute Doppler residual also increases from 2.07 Hz to 4.71 Hz and 5.82 Hz, respectively. In Figure 9, the Doppler residual increases more evidently during the lower-power stages of the signal attenuation scenario, suggesting that reduced signal power can weaken carrier tracking and lead to larger frequency-domain fluctuations.

The results show that C / N 0 is the most sensitive and consistent indicator across all degradation scenarios, Doppler residual is effective for attenuation and blockage, and pseudorange and carrier-phase residuals provide complementary measurement-domain evidence, especially for antenna displacement and severe tracking disturbances.

IPQI-based quality assessment and scenario comparison

Figure 11 shows the temporal response of IPQI under the three controlled degradation scenarios. For antenna displacement, IPQI remains close to the low degradation level for most of the experiment, but sharp peaks appear around the disturbance moments. This behavior indicates that antenna displacement mainly produces transient signal and measurement perturbations rather than long-duration degradation. The median IPQI is only 0.090, while the maximum value reaches 0.962, confirming that the antenna displacement scenario is characterized by short-term but severe quality changes.

For signal attenuation, IPQI shows a more persistent increase after the power adjustment events. Compared with antenna displacement, the IPQI curve remains at a higher level over a longer duration, indicating that transmitter power reduction affects the received signal quality continuously. The median IPQI reaches 0.409, and the RMS value is 0.479, both of which are the highest among the three scenarios. This result is consistent with the behavior observed in the C/N0 and Doppler residual responses, where stronger attenuation produced more evident measurement instability.

For signal blockage, IPQI exhibits intermittent high-value segments corresponding to the manually introduced obstruction intervals. The response is not completely uniform within each blockage interval, because the signal degradation depends on the obstruction geometry, the affected propagation path, and the receiver tracking state. The median IPQI is 0.299, and the 95th percentile reaches 0.736, indicating that blockage produces stronger quality degradation than antenna displacement in the overall distribution, but its persistence is weaker than that of signal attenuation.

The scenario-level statistics of IPQI are summarized in Table 2. Signal attenuation has the highest median and mean IPQI values, with a median of 0.409 and a mean of 0.404. Signal blockage ranks second, with a median of 0.299 and a mean of 0.326. Antenna displacement has the lowest median and mean values, 0.090 and 0.094, respectively, but its maximum value is the highest among the three scenarios. This confirms that the three degradation mechanisms have different temporal characteristics: antenna displacement is dominated by abrupt transient disturbance, signal attenuation produces sustained degradation, and signal blockage presents intermittent obstruction-related degradation.

Figure 12 further compares the IPQI distributions among the three scenarios. The distribution of antenna displacement is concentrated at low IPQI values, with only a small number of high-value outliers caused by the movement events. In contrast, signal attenuation shows a broader and higher distribution, reflecting a more continuous deterioration of signal quality. Signal blockage presents an intermediate distribution with a relatively broad spread, which is consistent with the repeated obstruction and recovery processes. These results demonstrate that IPQI can not only detect quality degradation but also distinguish the statistical characteristics of different degradation sources.

Overall, the IPQI-based assessment provides a compact representation of multi-domain signal quality. Compared with individual indicators, IPQI reduces the dependence on a single observation type and improves the interpretability of scenario-level quality comparison. The results show that the proposed index is sensitive to abrupt antenna displacement, sustained signal attenuation, and intermittent signal blockage, making it suitable for integrated quality monitoring in indoor pseudolite positioning systems.

This study developed a signal quality assessment method for an indoor pseudolite positioning system based on monitoring-station observations. A complete experimental framework was established, including pseudolite signal generation, indoor transmitting antenna deployment, receiver-side signal acquisition, and software-defined receiver processing. Three manually controlled degradation scenarios, namely antenna displacement, signal attenuation, and signal blockage, were introduced to evaluate the response characteristics of different signal quality indicators.

For the controlled degradation experiments, multi-source indicators from the signal, correlation, tracking, and observation domains showed different sensitivities to the three disturbance types. C/N0 responded clearly to antenna displacement, signal attenuation, and blockage, making it an effective general indicator of received signal strength degradation. Correlation-domain indicators, such as EVM and MPI, were sensitive to signal structure distortion, while tracking-domain indicators reflected loop stability changes during disturbed periods. Observation-domain residuals, including pseudorange, carrier-phase, Doppler, and observation-domain C/N0 residuals, further revealed how signal degradation propagated into receiver measurements. Antenna displacement caused abrupt variations in all observation indicators, while signal attenuation and blockage produced more continuous or intermittent measurement-domain changes.

Based on these indicators, a degradation-oriented Indoor Pseudolite Quality Index (IPQI) was constructed using reference-period residuals, robust normalization, and multi-source fusion. The IPQI results showed clear scenario-dependent characteristics. Antenna displacement produced short-term but severe IPQI peaks, with a relatively low median value of 0.090 but a maximum value of 0.962. Signal attenuation showed the strongest sustained degradation, with the highest median IPQI of 0.409 and RMS value of 0.479. Signal blockage exhibited intermittent degradation, with a median IPQI of 0.299 and a 95th percentile value of 0.736. These results demonstrate that IPQI can distinguish transient, sustained, and intermittent degradation patterns in indoor pseudolite positioning environments.

Overall, the proposed method provides a practical framework for monitoring and evaluating signal quality in indoor pseudolite positioning systems. By combining physically interpretable indicators from multiple receiver processing stages, the method can identify different degradation sources and provide an integrated quality assessment result. The experimental results show that the proposed IPQI is suitable for comparing signal quality under controlled indoor disturbances and can support future real-time monitoring of pseudolite positioning service reliability.