In this paper, we present a simple-structure Pedestrian Dead Reckon (PDR) system based on commercial IMU sensor and UWB ranging system. In PDR system, the accuracy of step and heading angle estimation completely decide the precision of location result. In order to select proper zero-velocity intervals for step and heading estimations in the foot stance/still stage, a modified zero-velocity detection way called Heuristic Step Detection (HSD) has been designed based on zero-velocity update algorithm (ZUPT). Based on the modified zero-velocity detection algorithm, i.e., HSD, an Kalman-type filter is used to get rough heading angle by fusing zero-velocity information and single UWB anchor ranging results. After that, a constrained sigma point based filter is used to further constrain heading angle range. Moreover, the range measures, provided by a UWB localization system with only one reference anchor, are used to correct the pedestrian step length. Trough UWB ranging measures analysis between two consecutive steps, the step length and pedestrian heading direction correcting processes are related to each other, and as a result, a more stable positioning result can be gotten. The corresponding practical experiments are conducted in real indoor environment over 2000 meters, and the results show that, compared with only INS-aided PDR, our scheme can reduce the average position error by more than 80%, and it can achieve long-term high accuracy and robust localization results.

A fingerprint-based localization system is an economic way to solve an indoor positioning problem. However, the traditional off-line fingerprint collection stage is a time-consuming and laborious process which limits the use of fingerprint-based localization systems. In this paper, based on ubiquitous Wireless Fidelity (Wi-Fi) equipment and a low-cost Ultra-Wideband (UWB) ranging system (with only one UWB anchor), a ready-to-use indoor localization system is proposed to realize long-term and high-accuracy indoor positioning. More specifically, in this system, it is divided into two stages: (1) an initial stage, and (2) a positioning stage. In the initial stage, an Inertial Measure Unit (IMU) is used to calculate the position using Pedestrian Dead Reckon (PDR) algorithm within a preset number of steps, and the location-related fingerprints are collected to train a Convolutional Neural Network (CNN) regression model; simultaneously, in order to make the UWB ranging system adapt to the Non-Line-of-Sight (NLoS) environment, the increments of acceleration and angular velocity in IMU and the increments of single UWB ranging measures are correlated to pre-train a Supported Vector Regression (SVR). After reaching the threshold of time or step number, the system is changed into a positioning stage, and the CNN predicts the position calibrated by corrected UWB ranging. At last, a series of practical experiments are conducted in the real environment; the experiment results show that, due to the corrected UWB ranging measures calibrating the CNN parameters in every positioning period, this system has stable localization results in a comparative long-term range. Additionally, it has the advantages of stability, low cost, anti-noise, etc.

Indoor localization service is an indispensable part of modern intelligent life, among which Wi-Fi based fingerprint localization system is popular in indoor positioning researches due to its advantages of low cost and widely deployment. However, Wi-Fi based localization system is susceptible to dynamic environment, and fingerprint collection and updating are time-consuming and labor-intensive. To address this problem, we propose a novel positioning framework based on multiple transfer learning fusion using Generalized Policy Iteration (GPI). Firstly, a 1-Dimension Convolutional Autoencoder (1-D CAE) is designed to extract features from one-dimensional fingerprint data; similar to Convolutional Neural Network (CNN), it can not only pay more attention to the information of different dimensions of fingerprints, but also compress redundant information and reduce noise. After that, Domain Adversarial Neural Network (DANN) and Passive Aggressive (PA) algorithm are fused to train localization model based on unlabeled fingerprint of target domain using the theory of GPI in offline stage. Finally, the model is fine-tuned with unlabeled fingerprints and few labeled fingerprints in daily online predictions to improve the performance of the localization system. Various evaluations in five typical scenarios validate the effectiveness of proposed algorithm in dynamic environment, with low tendency, easy recalibration, long-term stabilization high accuracy and so on.

The sound velocity profile is the base of various underwater acoustic equipment. In this paper, a low-cost sound velocity profiler is designed based on the time difference method. It mainly includes three parts: the control unit, the storage module and the ultrasonic measurement module. Its overall volume is small, and the standby power consumption is low. It can be integrated into various underwater measurement platforms and profilers to realize the sound velocity measurement, and it also could be used as a self-contained sound velocity sensor. Furthermore, according to the sound velocity measurement principle and response characteristics, a calibration algorithm based on Recurrent Neural Network (RNN) and Discrete Wavelet Transformation (DWT) is proposed, which can improve the accuracy and adapt to the nonlinear response of the system by using multiple sets of time data obtained from the measurements. It is verified by calibration experiments that the neural network calibration algorithm can effectively reduce the nonlinear system error in the measurement, and its effect is better than the traditional linear regression method. The designed system prototype can achieve measurement accuracy of 0.05m/s after calibration, which can meet the needs of low-cost and high-precision underwater sound velocity measurement.

In indoor positioning systems, Wi-Fi fingerprint positioning often attracts the attention of researchers due to its advantages such as popularity and low cost. However, Wi-Fi fingerprint positioning relies on a large number of labeled fingerprints, and these fingerprints are prone to becoming invalid in dynamic environments, resulting in very frequent Wi-Fi fingerprint collection, which is unfavorable for the promotion of Wi-Fi fingerprint positioning. To address this issue, this work proposes a fingerprint augmentation method based on wavelet transform, which reduces the need for extensive fingerprint data collection while maintaining localization accuracy. Specifically, the method first utilizes the virtual positions of Wi-Fi access points (APs) to embed spatial information into the fingerprint sequence, transforming the one-dimensional fingerprint sequence into a multidimensional fingerprint grayscale image; Furthermore, various fingerprint augmentation schemes based on Discrete Wavelet Transform (DWT) are proposed to expand the fingerprint database and further enhance the robustness of the positioning model. Finally, a series of schemes are designed to verify the effectiveness of the fingerprint augmentation schemes, and comparisons are made with similar works. The results indicate that the DWT-based fingerprint augmentation scheme proposed in this paper effectively reduces the reliance on labeled fingerprints without a significant increase in time complexity. Compared to similar works, it demonstrates greater adaptability to dynamic environments.