Pedestrian dead reckoning (PDR) is widely used in handheld indoor positioning systems. However, low-cost inertial sensors built into smartphones provide poor-quality measurements, resulting in cumulative error which consists of heading estimation error caused by gyroscope and step length estimation error caused by an accelerometer. Learning more motion features through limited measurements is important to improve positioning accuracy. This paper proposes an improved PDR system using smartphone sensors. Using gyroscope, two motion patterns, walking straight or turning, can be recognised based on dynamic time warp (DTW) and thus improve heading estimation from an extended Kalman filter (EKF). Joint quasi-static field (JQSF) detection is used to avoid bad magnetic measurements due to magnetic disturbances in an indoor environment. In terms of periodicity of angular rate while walking, peak–valley angular velocity detection and zero-cross detection is combined to detect steps. A step-length estimation method based on deep belief network (DBN) is proposed. Experimental results demonstrate that the proposed PDR system can achieve more accurate indoor positioning.

The fingerprint positioning (FP) algorithm has been investigated extensively owing to the fact that it can provide a relatively ideal indoor positioning result. However, the effectiveness of the fingerprint algorithm relies on the size of fingerprint database, which prevents the algorithm from being widely applied in practical applications. In this paper, an improved fingerprint algorithm with access point (AP) selection strategy and reference point (RP) selection strategy is proposed to reduce the size of the fingerprint database and improve the positioning accuracy. The experimental results show that the proposed algorithm can reduce the storage size of the fingerprint database by more than 42·64%. Moreover, compared with the FP algorithm, the fingerprint algorithm with segment characteristic distance (FP-SCD) and the fingerprint algorithm with RP selection strategy (FP-RPSS), the average positioning error of the proposed algorithm is reduced by 20·15%, 10·83% and 11·57%, respectively. Therefore, the proposed algorithm has a good application in real positioning scenarios.

Indoor positioning systems have received increasing attention for supporting location-based services in indoor environments. Wi-Fi based indoor localisation has become attractive due to its extensive distribution and low cost properties. IEEE 802.11-2016 now includes a Wi-Fi Fine Time Measurement (FTM) protocol which can be used for Wi-Fi ranging between intelligent terminal and Wi-Fi access point. This paper introduces a framework of Wi-Fi FTM data acquisition and processing that can be used for indoor localisation. We analyse the main factors that affect the accuracy of Wi-Fi ranging and propose a calibration, filtering and modelling algorithm that can effectively reduce the ranging error caused by clock deviation, non-line-of-sight (NLOS) and multipath propagation. Experimental results show that the proposed calibration and filtering method is able to achieve metre-level ranging accuracy in case of line-of-sight by using large bandwidth. Estimation results also show that the proposed Wi-Fi ranging model provides an accurate ranging performance in NLOS and multipath contained indoor environment; the final positioning error is less than 2·2 m with a stable output frequency of 3 Hz.

Fluctuation of the received signal strength (RSS) is the key performance-limiting factor for Wi-Fi indoor positioning schemes. In this study, the Manhattan distance was used in the weighted K-nearest neighbour (WKNN) algorithm to improve positioning accuracy. Reference point (RP) intervals were optimised to reduce the complexity of the system. Specifically, two new positioning schemes are proposed in this paper. Scheme 1 uses the cellular network to refine the fingerprint database, while Scheme 2 uses the cellular network positioning to locate the node a priori, then uses the Wi-Fi network to further improve accuracy. The experimental results showed that the average positioning error of Scheme 1 was 1·60 m, a reduction of 12% compared with the existing Wi-Fi fingerprinting schemes. In Scheme 2, when double cellular networks were used, RP usage was reduced by 64% and the calculating time was 0·24 s, a reduction of up to 69·5% compared with the Manhattan-WKNN algorithm. These proposed schemes are suitable for high accuracy and real-time positioning situations, respectively.

The demand for accurate indoor positioning continues to grow but the predominant positioning technologies such as Global Navigation Satellite Systems (GNSS) are not suitable for indoor environments due to multipath effects and Non-Line-Of-Sight (NLOS) conditions. This paper presents a new indoor positioning system using artificial encoded magnetic fields, which has good properties for NLOS conditions and fewer multipath effects. The encoded magnetic fields are generated by multiple beacons; each beacon periodically generates unique magnetic field sequences, which consist of a gold code sequence and a beacon location sequence. The position of an object can be determined with measurements from a tri-axial magnetometer using a three-step method: performing time synchronisation between sensor and beacons, identifying the beacon field and the beacon location, and estimating the position of the object. The results of the simulation and experiment show that the proposed system is capable of achieving Two-Dimensional (2D) and Three-Dimensional (3D) accuracy at sub-decimetre and decimetre levels, respectively.

Indoor localisation has always been a challenging problem due to poor Global Navigation Satellite System (GNSS) availability in such environments. While inertial measurement sensors have become popular solutions for indoor positioning, they suffer large drifts after initialisation. Collaborative positioning enhances positioning robustness by integrating multiple localisation information, especially relative ranging measurements between local users and transmitters. However, not all ranging measurements are useful throughout the whole positioning process and integrating too much data will increase the computation cost. To enable a more reliable positioning system, an adaptive collaborative positioning algorithm is proposed which selects units for the collaborative network and integrates ranging measurement to constrain inertial measurement errors. The algorithm selects the network adaptively from three perspectives: the network geometry, the network size and the accuracy level of the ranging measurements between the units. The collaborative relative constraint is then defined according to the selected network geometry and anticipated measurement quality. In the case of trials with real data, the positioning accuracy is improved by 60% by adjusting the range constraint adaptively according to the selected network situation, while also improving the system robustness.

In this paper, a Three Dimensional Pedestrian Movement Model (3D-MM) capable of probabilistically representing pedestrian movement in challenging indoor and outdoor localization environments is developed, implemented and evaluated. In the scope of this paper, the model is used to generate a ‘movement’ or a transition for dynamic positioning systems that are based on sequential Bayesian filtering techniques, such as particle filtering. It can also be used to assign weights for particles' movements proposed by sensors in Likelihood Particle Filters implementations. Alternatively, the developed model can be applied to other applications domains such as infrastructure design, evacuation planning, robot-human interaction and pervasive computing. The novelty of the model is in its ability to characterize both random and goals-oriented pedestrian motions and additionally use the a priori knowledge of maps and floor plans. It will be shown that an appropriate pedestrian movement model not only improves the positioning accuracy, but is also essential for a robust positioning estimator. Additionally, this work shows that maps and floor plans can improve pedestrian movement models but do not replace them, as several authors suggest.

A version of this paper was first presented at the Royal Institute of Navigation NAV 08 Conference held at Church House, Westminster, London in October 2008.

There exist various applications for indoor positioning, amongst which indoor positioning and tracking in urban environments has gained significant attention. Some user communities, like fire fighters, ideally require indoor accuracy of less than one metre, with accuracies of less than six metres acceptable by some other user communities. Achieving this level of accuracy requires a detailed profiling of error sources so that they can be better understood so that, in turn, indoor positioning accuracy in the presence of these errors can be further improved. Some well known error sources like multipath, NLOS (non line of sight), oscillator drift, dilution of precision and others have been studied and can be found in the literature. A less well known error source that can substantially affect indoor positioning accuracy are the effects of the dielectric properties of building materials on propagation delay.

Various RF and non-RF based prototypes that claim to be suitable for indoor positioning can be found in the literature. Most of the existing literature discusses algorithms and summarizes the positioning results that were achieved during field tests using a prototype system or, more commonly, simulations. Little of this existing literature provides a breakdown of the total navigation system errors observed with the objective of analyzing the contribution of each error source independently.

The paper will first provide a brief overview of the precision personnel locator system developed at the Worcester Polytechnic Institute. The field tests and observed indoor positioning results using this RF prototype will then be summarized and used to provide a baseline to establish a system error budget. The total observed error will be broken down and a detailed analysis of each of the error sources will be presented based on actual measured data in a variety of indoor environments. This leads to a better understanding of how each error source affects indoor positioning accuracy. Each of the error sources can then be independently optimized to minimize the observed errors. Specifically, the interplay between the dielectric properties and multipath profiles will be highlighted. This paper will conclude by presenting an error budget which can be used as a practical lower bound when designing precise indoor positioning systems.

Thanks to its physical characteristics, Ultra-wideband (UWB) is one of the most promising technologies for indoor pedestrian navigation. UWB radio localisation systems however experience multipath phenomena that decrease the precision and the reliability of the user's location. To cope with complex indoor environments, UWB radio signals are coupled with inertial measurements from Micro Electro Mechanical Sensors (MEMS) in an extended Kalman filter. Improved performances of the filter are presented and compared with reference trajectories and with pure inertial solutions. Only specific selection methods enable this improvement by detecting and removing outliers in the raw localisation data.