Ing. Josef Krška

Ultra-wide Band (UWB) positioning systems estimate the user's position from several time-based measurements. While usually achieving decimeter-level accuracy, enhancing the measurements by additional measurements, such as Angle of Arrival (AoA), should improve the positioning performance and robustness. In this work, the design of our custom UWB-AoA board capable of time-based and angle-based measurements is presented. The board contained two DW1000 UWB transceiver chips with a common clock source and with independent communication interfaces, enabling parallel control of the chips. The AoA measurements using Time Difference of Arrival (TDoA) and Phase Difference of Arrival (PDoA) methods are discussed. The AoA estimation accuracy using TDoA and PDoA is evaluated experimentally with turntable experiments, done for several antenna configurations. The results show that with PDoA measurements it is possible to achieve AoA errors with 6.1 degrees standard deviation and +0.73 degrees mean for half-wavelength antenna separation and even better for longer antenna separations.

An integral step in an ultra-wideband localization network installation is deter­mining the positions of the fixed infrastructure nodes, the anchors. This process is time-consuming and usually requires specialized equipment. Additionally, it is difficult to achieve scalability, as any change or addition in the network requires a redetermination of the affected anchors. One can automate this process by utilizing the distance-measuring capabilities of the network infra­structure and employing a distributed position estimation algorithm, such as the consensus subgradient (CSG) algorithm. Yet, the CSG suffers from scalabil­ity issues due to high problem dimensionality and data-sharing bottlenecks in practical applications. Consequently, implementation in embedded devices is difficult. In this article, we propose a modification of this algorithm, the neigh­borhood CSG, which aims toward embedded implementation by local reduction of the problem dimensions without hindering the precision of the original CSG algorithm or its convergence rate.

Ultra-wide Band (UWB) positioning systems are a system of choice for indoor positioning. Though UWB is more resilient against multipath propagation, the positioning accuracy can still be hindered in its presence. The positioning accuracy and availability can be improved by combination with other sensors, typically inertial ones. In the case of pedestrian positioning the use of a dead-reckoning system, which uses the inertial data from a smartphone, may be advantageous. In this paper we propose the use of Extended Kalman filtering for the data fusion of UWB and STEP system, which detects the user making steps and estimate the step-length from the vertical acceleration of the user. For the fusion we assume two models: fusion of UWB and step-length with heading and fusion of UWB and step-length. The performance is evaluated on real data and it is shown that the UWB positioning error of 0.59 m is improved to 0.22 m and 0.36 m when fusion with both the step-length and heading or with only the step-length is used, respectively.

The application of the impulse-radio ultra-wide band technology is already established in indoor localization. Usually, two-way ranging (TWR) or time-difference of arrival (TDoA) approaches are utilized. In this paper it is shown that it is possible to perform time of arrival (ToA) positioning with accuracy similar to the TDoA. The ToA method estimates the bias between the clocks of the anchors (fixed infrastructure nodes) and the tag (localized user equipment). As a direct consequence, any timestamps taken by the tag may be converted to the common system timescale with accuracy of several nanoseconds. The approach is verified by real measurements. We demonstrate that the positioning solution can be obtained by least-squares optimization on epoch-by-epoch basis. We also show that the extended Kalman filter can be used to estimate tag position along with the bias and drift of its clock.

UWB positioning combines time measurements from several devices to estimate the position of a target device. These measurements, however, are expressed in the device's own timescale that is derived from its free running clock source, generally a crystal oscillator. The timescales are inherently offset and with different and time-variant drift, both of which have to be compensated during the estimation. With the coherent IR-UWB chips it is possible to measure the drift either directly (Carrier Frequency Offset) or indirectly with timestamps. This work provides the stability analysis of such measurements as well as the characterization of the noise that affects them. From the results the usability of drift measurement methods for synchronization and positioning is inferred.

High-capacity impulse-radio ultra-wideband (IR-UWB) indoor localization systems are typically based on the time difference of arrival (TDoA) principle. When the fixed and synchronized localization infrastructure, the anchors, transmit precisely timestamped messages, a virtually unlimited number of user receivers (tags) are able to estimate their position based on differences in the time of arrival of those messages. However, the drift of the tag clock causes systematic errors at a sufficiently high magnitude to effectively deny the positioning, if left uncorrected. Previously, the extended Kalman filter (EKF) has been used to track and compensate for the clock drift. In this article, the utilization of a carrier frequency offset (CFO) measurement for suppressing the clock-drift related error in anchor-to-tag positioning is presented and compared to the filtered solution. The CFO is readily available in the coherent UWB transceivers, such as Decawave DW1000. It is inherently related to the clock drift, since both carrier and timestamping frequencies are derived from the identical reference oscillator. The experimental evaluation shows that the CFO-aided solution performs worse than the EKF-based solution in terms of accuracy. Nonetheless, with CFO-aiding it is possible to obtain a solution based on measurements from a single epoch, which is favorable especially for power-constrained applications.

This paper proposes an approach of TDoA positioning in UWB networks, where user tags localize themselves by means of exploitation of the broadcasted synchronization messages of the anchor network. Such approach promises unlimited number of localized devices, moreover, the position is available directly at the user terminal. The key challenge of this method is to eliminate errors caused by tag clock drifts, which render the TDoA measurements useless when left uncorrected. Our method employs extended Kalman filtering (EKF) for the estimation of position and elimination of the drift-induced errors. It is shown that the system performance is similar to the more common TDoA method, where the tags transmit blinks received by the anchors. However, the anchors are still able to receive the blink messages and estimate position of those tags, since the synchronization messages are exploited. Therefore, it is possible to use both directions of the TDoA positioning concurrently; a limited number of tags is tracked by the infrastructure and all tags may compute their positions. The TDoA solutions have achieved RMS horizontal accuracy of 25.9 cm and 33.6 cm, respectively.

The Time Difference of Arrival is a popular method for UWB-based positioning, since it allows high position update rate even for multiple users. However, it requires the network infrastructure (anchors) to be synchronized, preferably with sub-nanosecond accuracy. Herein, an approach for synchronizing multiple anchors in a wireless, line-of-sight manner is described. This method is able to deal with UWB modules equipped with inexpensive drift-prone oscillators, as such impairments are estimated and compensated. By applying the proposed approach the influence of generally variable environment (e.g. temperature) on timing and positioning performance is reduced. Moreover, the presented algorithm is suitable for straightforward chaining of the line-of sight-segments in order to allow synchronization of distant anchors that cannot be synchronized with the master anchor directly.

In the ultra-wide band (UWB) localization networks the time difference of arrival (TDoA) is often used. The TDoA method is advantageous in comparison with two-way ranging approach, however, requires the UWB infrastructure of the network to be synchronized to a sub-nanosecond level. The wireless synchronization is in favor due to practical reason, nonetheless its application is constrained by the need of line-of-sight between the synchronized nodes. This paper focuses on an experimental evaluation of a Kalman-filter-based chained synchronization algorithm, which allows synchronization of a distant, directly unreachable UWB nodes through multiple nodes and the respective line-of-sight path segments.