Recent Advances in Ubiquitous Positioning Systems for Mobility Applications

For many years, global navigation satellite systems (GNSSs) have been used in professional navigation, land surveying and transport. The last decade has been a period of extending the area of their use to social applications, including those related to sports and recreation. Nowadays, both handheld recorders and mobile phones equipped with GNSS modules are widely used in fitness. Moreover, due to the popularisation of jogging, a mobile phone has become one of the essential navigational instruments which enables, inter alia, the determination of the distance covered by an amateur athlete, the duration of completing the route and the athlete's speed. However, each of satellite-based solutions is characterised by a different positioning accuracy and positioning availability, which have a decisive influence on the reliability of the parameters being measured. The article proposes a new measure of positioning accuracy, the stadium cross track error (SXTE) with a confidence level of 95% and 68%, as a universal comparison criterion for satellite sport receivers. The application of the method is presented with the example of receivers used in nine Samsung Galaxy series mobile phones. The method uses an athletics stadium that meets the International Association of Athletics Federation's standard for the running track length (400 m). The study demonstrated significant differences in positioning accuracies of particular GNSS receivers in Samsung Galaxy series mobile phones while allowing their accuracy to be unambiguously assessed.

The development of a global navigation satellite system (GNSS) brings the benefit of positioning, navigation and timing (PNT) services with three or even more available frequency signals. This paper developed five-system multi-frequency precise point positioning (PPP) models based on mathematical and stochastic models. Static positioning performances were evaluated and analyzed with multi-GNSS experiment (MGEX) network datasets and a vehicle-borne kinematic experiment was conducted to verify the kinematic PPP performances. In addition, the receiver clock, zenith tropospheric delay (ZTD), inter-frequency bias (IFB) and differential code bias (DCB) estimates were discussed. Results show that the triple-frequency PPP performances perform slightly better than the dual-frequency solutions, apart from the GPS-related PPP models based on a single ionosphere-free (IF) combined measurement. By introducing the external ionospheric products, the mean convergence time is reduced. For instance, the mean convergence time of ionosphere-constrained (IC) multi-frequency PPP is reduced by 7.4% from 35.7 to 33.1 min and by 19.0% from 7.8 to 6.3 min, for Galileo-only and five-constellation solutions, respectively, compared with dual-frequency IF PPP models. Similarly, the kinematic PPP can also achieve improved performances with more frequency signals and multi-GNSS observations.

In high-precision navigation applications, a well-designed self-calibration method is a convenient approach to ensuring the positioning performance of a rotational inertial navigation system (RINS). Benefiting from the gimbal structure, traditional inertial measurement unit (IMU) sensor errors, including gyro drifts, accelerometer biases, scale factor errors and installation errors, could be estimated through a filter process under a proper rotation scheme. However, when the IMU rotates, inner lever-arm effects may bring additional errors to the observations, which may reduce the self-calibration accuracy. In this paper, an improved self-calibration method that includes consideration of the inner lever-arm effect is proposed for a dual-axis RINS. Based on analysis of the error propagation characteristics, a novel rotation scheme with variable angular rate is designed. By adopting the proposed self-calibration method, traditional IMU sensor errors can achieve much higher accuracy, and the inner lever-arm parameters can also be well calibrated simultaneously. Long-term vehicle navigation indicates that the positioning accuracy was significantly enhanced after the compensation of the calibration results, fully illustrating the effectiveness of the proposed method in ameliorating navigation performance for the dual-axis RINS.

Nowadays, global navigation satellite systems (GNSSs) are widely used in location-based services (LBSs) as they can provide high-accuracy position services continuously. However, their performance deteriorates in indoor scenarios, in which GNSS signal reception is limited or completely impossible. In this paper, an enhanced constrained Kalman filter is presented to enhance the indoor positioning performance of a LBS and for use in a WiFi/pedestrian dead reckoning (PDR) integrated navigation algorithm. A robust scheme for computing the gross error in constrained conditions is suggested to make the performance of the constrained condition model in the WiFi/PDR integrated system more robust. The results of simulation analysis indicate that the robustly constrained Kalman filter can reliably determine gross errors in constrained conditions. An indoor field experiment was conducted to test the performance of the proposed filter algorithm, and the results show that the improved filter can eliminate the effect of gross error from constrained conditions in the WiFi/PDR system.

Considering the remarkable progress of the Galileo constellation in recent years, the main objective of this study is to evaluate the performance of dual- and single-frequency Galileo-based precise point positioning (PPP), and its contribution to GPS and Galileo integration with different precise products generated by four analysis centers (ACs) within the context of the Multi-GNSS Experiment (MGEX) of the International GNSS Service (IGS). For this purpose, the daily observation dataset collected at ten IGS stations over one month was processed in both static and kinematic modes for Galileo-only, GPS-only, and GPS/Galileo PPP scenarios. For dual-frequency PPP, the results demonstrate that while the Galileo-only solutions are highly comparable with GPS-only PPP for the static mode, the mean 3D positioning errors for Galileo-only processes are approximately 1-cm higher than those obtained from GPS-only solutions for all agencies. The analysis to evaluate the influence of Galileo satellites with highly eccentric orbits, namely E14 and E18, on dual-frequency Galileo-only PPP performance indicates that including or excluding these satellites has no significant effect on the results. For single-frequency PPP, which is dependent on the GRAPHIC combination, Galileo-only PPP performs significantly better, approximately 40%, than GPS-only solutions in the static mode, whereas kinematic Galileo-only and GPS-only PPP solutions are quite similar for all agencies except for WHU. In addition, the RMS of observation residuals for Galileo in single-frequency PPP was noticeably lower than that for GPS, demonstrating that the observation quality of Galileo code measurements is better than those of GPS. Among the ACs, Galileo-based PPP solutions applying CODE products produced slightly better results than those obtained for GFZ and CNES/CLS in general, whereas processes using WHU products resulted in a worse performance, both in terms of positioning accuracy and of convergence time. The integration of Galileo with GPS was shown to enhance PPP performance significantly in both static and kinematic modes.

In 2018, China established the primary constellation of the BeiDou-3 global navigation satellite system (BDS-3), and the construction of this satellite system is due to be completed in 2020. The signal emitted by BDS-3 will provide global positioning, navigation and timing services. This study assessed the performance of network real-time kinematic (NRTK) in reference to the new BDS-3 signals. The ambiguity resolution, retrieval, and interpolation of the double-differenced (DD) observation corrections and positioning precision are assessed based on datasets collected using four SR480 receivers that track the open signals of BDS-3. Subsequently, the NRTK, using the combined BDS-2 and BDS-3 measurements, is compared with the NRTK using only BDS-2 measurements. In addition, the positioning results of NRTK are compared with those of RTK. The results show that the DD atmospheric delay corrections with centimeter-level accuracy can be derived from the network of regional reference stations, and the positioning accuracy of NRTK is improved by incorporating the BDS-3 measurements.

Ultra-rapid products, which do not require any external connection unlike real-time services, are an important alternative for real-time global navigation satellite system (GNSS) applications. Especially, the inclusion of newly-emerged satellite systems in ultra-rapid products opens up considerable opportunities to improve the positioning performance. In this regard, this study concentrates on the employment of the most recent ultra-rapid products besides traditional ones for real-time single-frequency multi-GNSS positioning using code and carrier phase measurements. In the study, experimental tests were conducted for the ionosphere-free code-carrier combination to evaluate the performance of single-receiver single-frequency positioning. The results reveal that single-frequency positioning is able to provide sub-meter level positioning accuracy with ultra-rapid products despite its performance alters depending on the applied product. Also, the performance of single-frequency positioning which based on code-carrier combination is not influenced significantly from the possible decline in the precision of ultra-rapid products over time due to the convergence of phase ambiguities. On the other hand, the results demonstrate that the accuracy of pseudorange positioning can significantly be improved with the integration of multi-constellation and the improvement ratio can reach 30% compared with the GPS-only solutions. Furthermore, the convergence time of GPS-only solution can be decreased by a ratio of 37% on average with multi-GNSS combinations. Finally, the results show that for the multi-constellation analyses, the solutions which utilize the ultra-rapid product of Wuhan University provide the best performance in terms of positioning accuracy and convergence time.

The development of indoor localization has been advanced by the rapid development of intelligent devices. The well-known methods used for indoor localization such as Wi-Fi fingerprint database positioning and pedestrian dead reckoning (PDR) can be implemented in a self-contained smartphone. However, the existing Wi-Fi fingerprint database positioning method can be easily influenced by the dynamic environment while PDR will generate a cumulative error with an increase in walking steps. In this paper, we propose a new hybrid method using PDR and Wi-Fi information. We divide the localization area into several subareas to improve the accuracy of the Wi-Fi fingerprint matching phase and introduce an enhanced particle filter (PF) algorithm which includes subarea information in the state vector and adopts a clonal selection algorithm (CSA) to improve resampling. We conduct a series of experiments in real-world environments, and the experimental results validate that the proposed algorithm is much better than ordinary PF algorithms and standalone methods.

Fingerprinting localization based on Wi-Fi received signal strength (RSS) is the most widely used indoor localization method. It typically includes offline training and online matching phases. The selection of the RSS characteristic value is a key step. The weighted K nearest neighbor (WKNN) algorithm is the most commonly used position-determination algorithm. The mean value of the RSS data collected over a time interval is usually taken as its characteristic value. However, the RSS measurements contain Gaussian and non-Gaussian noise, which cannot be filtered out effectively by the mean value method. The traditional WKNN algorithm adopts a fixed $K$. However, reference points far away from the test point (TP) may be selected as the nearest neighbors to participate in the position calculation, which may result in accuracy degradation. This paper proposes the weighted dynamic K nearest neighbor algorithm (WDKNN-HF), which utilizes a hybrid of particle filtering and Kalman filtering to extract the RSS characteristic value. In the online matching phase, a dynamic K matching algorithm based on Euclidean distances is developed to determine the coordinates of TPs. Two experiments are conducted in two different indoor scenes. Experimental results demonstrate that the proposed algorithm can obtain better positioning accuracy than existing algorithms, such as KNN, WKNN, enhanced-WKNN (EWKNN) and self-adaptive weighted K nearest neighbor (SAWKNN).

Although the tightly integrated inertial navigation system/ultra wide band (INS/UWB) improves the localization accuracy, it suffers from UWB distance outage. In order to reduce the outage effect on the position accuracy, in this paper we propose using a novel decision tree (DT)-extended finite impulse response (EFIR) filter. When all the UWB distances are available, the EFIR filter tightly fuses the INS and UWB data. Otherwise, the DT builds a relationship between the INS and UWB position errors. Once at least one UWB distance is unavailable, the DT bridges a gap over unavailable measurements. It is shown experimentally that the DT-EFIR filter is an efficient tool to reduce the effect of the UWB distance outage on the INS/UWB system operation, which can improve the localization error by about 40% as compared with the UWB solution.

To solve the accurate positioning problem of mobile robots, simultaneous localization and mapping (SLAM) or visual odometry (VO) based on visual information are widely used. However, most visual SLAM or VO cannot meet the accuracy requirements in dynamic indoor environments. This paper proposes a robust visual odometry based on deep learning to eliminate feature points matching error. However, when a camera and dynamic objects are in relative motion, the frames of camera will produce ghosting, especially in high-dynamic environments, which bring additional positioning error; in view of this problem, a novel method based on the average optical flow value of the dynamic region is proposed to identify feature points of the ghosting, and then the feature points of the ghosting and dynamic region are removed. After the remaining feature points are matched, we use a non-linear optimization method to calculate the pose. The proposed algorithm is tested on TUM RGB-D dataset, and the results show that our VO improves the positioning accuracy than other robust SLAM or VO and is strongly robust especially in high-dynamic environments.

When calibrating inter-system biases (ISB), especially the fractional part of inter-system phase biases (F-ISPB), a multi-GNSS inter-system model can effectively improve positioning performance under a complex environment. Usually, the F-ISPB is estimated after fixing the intra-system ambiguities. However, this approach seems inapplicable when it is difficult to obtain intra-system ambiguities under a complex environment. A multi-dimensional particle filter (PF)-based F-ISPB estimate method has been proposed to overcome the problem. Nevertheless, the multi-dimensional PF involves a great quantity of computations. In this contribution, four state optimal estimate-based F-ISPB handling schemes are proposed: step-by-step PF, step-by-step particle swarm optimization (PSO), multi-dimensional PF, and multi-dimensional PSO-based F-ISPB estimate methods. Two baselines were selected to investigate the F-ISPB estimate performance in both open and complex environments. The results show that due to the potential of the wrong F-ISPB to bring about the maximum ratio for a long time during the initial stage, the step-by-step PF method can achieve better performance than step-by-step PSO. Besides, the two-dimensional results show that all of the F-ISPB still cannot be extracted under complex environments by multi-dimensional PSO. Furthermore, compared with step-by-step PF, the multi-dimensional PF method costs too much to obtain the right value. For example, in the two-dimensional case, the step-by-step PF searches 200 times for each epoch, while the two-dimensional PF requires 40 000 times for each epoch, so it is difficult for receivers to provide hardware support for this method. In addition, the step-by-step PF can obtain the right F-ISPB with about 100 epochs no matter what scenario. Thus, under challenging observation scenarios, a step-by-step PF method is recommended to extract the F-ISPB.

Visible light positioning (VLP) is a promising technique to bring location-based service for numerous Internet of Things applications. Recent advances in VLP have shown that machine learning (ML)-based positioning algorithms show satisfying performance in physical environments under highly noisy and interference-rich conditions. With so many ML methods proposed, one major concern is that trained models could fail due to environmental heterogeneity. In this paper, we propose AdVLP, a novel adversarial training method for VLP based on deep neural networks, to address the issue of the vulnerability of data-driven approaches, which happens when channel parameters change. Our proposed method, which is inspired by generative adversarial networks, manages to adapt the two domains of the source dataset and the target dataset. The channel parameters, e.g. the Lambertian order of light-emitting diode (LED) lights and a photodiode, the LED power as well as the location error of the lights, were discussed. We observed in experiments that the accuracy of a neural network method would decrease as the bias of the parameters became larger. Secondly, a higher dimension of the parameter change would make the neural network method more vulnerable. However, our proposed approach could achieve substantial improvement in positioning accuracy.

The release of raw global navigation satellite system (GNSS) observations by Google Android makes high-precision positioning possible with low-cost smart devices. This study contributes to this research trend by constructing a stochastic model based on raw GNSS observation characteristics from Android smartphones and verifying the feasibility of smartphone-based ambiguity fixing in the short-baseline real-time kinematic (RTK) case. This study uses the raw observation standard deviations (ROSTDs) delivered by the Android application programming interface (API) as a stochastic model and takes advantage of the multipath index from the API to rule out unusable observations. As well as these, the ambiguity integer property is investigated by analyzing the residuals of double-differenced carrier-phase observations associated with one smartphone and one geodetic-grade receiver. Furthermore, we note that the carrier-phase observations collected by tested smartphones do not have the integer property but for the Huawei P30 and Xiaomi 8 devices, such an integer property can be successfully recovered by means of detrending. With the use of ROSTD-dependent weighting, we first perform single-point positioning (SPP) and real-time differentition (RTD) using pseudorange observations delivered by the Huawei P30 and Xiaomi 8 devices. The results show that the stochastic model is applicable to the Xiaomi 8. Moreover, the three-dimensional root-mean-square (3D-RMS) errors of the two smartphones for SPP are 1.28 m and 1.96 m, and the 3D-RMS errors for RTD are 0.79 m and 1.64 m, respectively. We next test the RTK positioning performance based on a short-baseline of 882 m using carrier-phase observations with recovered integer ambiguities. For the Huawei P30, the positioning errors achieved were 7.8, 2.4, 1.1 mm for the east, north, and up (ENU) components at the time of first fix while for the Xiaomi 8, the positioning errors achieved were 4.3, 4.2, 4.2 mm for the ENU components at the time of first fix.