UWB NLOS Identification and Mitigation Based on Gramian Angular Field and Parallel Deep Learning Model

Ultrawideband (UWB) wireless localization technology has been widely applied in the field of indoor localization due to its good ability of noise resistance, strong penetration, and high measurement accuracy. However, the performance of UWB-based localization technology becomes poor when suffering from nonline-of-sight (NLOS) propagation conditions. Thus, it is necessary to identify NLOS propagation and mitigate the NLOS error. In this article, a novel NLOS identification and mitigation method based on multiinputs parallel deep learning model and Gramian angular field (GAF) is proposed. We utilize GAF to transform 1-D channel impulse response (CIR) signal into 2-D colored images, which adds additional high-level abstract features to the CIR signals. In the model training phase, the convolutional neural network (CNN) is used to extract temporal features from original CIR signals, and the residual network (ResNet) is used to extract visual features from GAF-encoded images. Besides, the received signal strength (RSS) information is also considered as an auxiliary feature to assist in identifying some NLOS scenarios with similar CIR features and further reduce the NLOS error. The experimental results show that our method has good ability in both line of sight (LOS) and NLOS binary classification and NLOS multiclassification, with accuracy over 96%. Additionally, based on the identification results, the proposed method can reduce the mean absolute error (MAE) and root mean square error (RMSE) of the range error from 65.61 and 96.82 cm to 4.19 and 6.95 cm, respectively. In the real indoor localization experiment, the proposed method can improve the localization accuracy by over 80%.

I. Introduction

High-accuracy localization technology plays a more and more important role in the field of Internet-of-Thing (IoT). Different from outdoor environments, the mature localization technologies based on satellite systems such as Beidou and Galileo cannot provide real-time and high-accuracy localization services in indoor environments because of the poor penetration of the localization signal [1]. Thus, using wireless localization technologies such as radio frequency identification (RFID), ZigBee, long range radio (LoRa) and ultrawideband (UWB) to realize high accuracy indoor localization has become a hot research topic recently [2], [3], [4], [5]. However, not all of the indoor positioning technologies can meet the increasingly complex indoor scenarios. For instance, though LoRa can achieve long-distance localization it would not be suitable for large-scale commercial use because of its poor security. For the ZigBee positioning technology, it can be only used in some specific scenarios which do not request high real-time performance such as attendance system. Besides, the performance of the ZigBee positioning system is seriously affected by environmental factors, which means it cannot be used in harsh indoor environments. The RFID positioning technology can be divided into active RFID-based and passive RIFD-based. The former one is more susceptible to the environment and the latter is limited by its valid working distance, only a few meters. Consequently, the positioning technology mentioned above cannot meet the requirement of high-precision indoor positioning especially when the environment is complex. Among common wireless positioning technologies, UWB is considered the most reliable and promising technology which has already been widely used in industrial and challenging indoor environments such as underground mine and large-scale warehouse due to its admirable accuracy in long-distance positioning and good penetration [6]. In spite of the mentioned benefits, the performance of UWB experiences degradation when suffering from nonline-of-sight (NLOS) conditions.