Visible light technology shows great potential in the underground mine positioning today. Most visible light-based underground positioning technology deploy the receiver, such as photodiodes (PDs), on the mine tunnel ground and consider the miner’s lamp on the ceiling as the transmitter. This incurs the unreliable reception of visible light signals and the difficult implementation of VLP systems due to the complex underground mine environment, such as the receiver being obscured or trampled by objects (e.g., Minecarts and Ore). We present a reliable underground positioning algorithm with visible light considering the actual mine environment. The key innovation is that we design a novel reverse mechanism, i.e., utilizing the existing miner’s headlamp as the transmitter for broadcasting the specific light beacon representing the mine’s identity information, and the PDs deployed on the ceiling of the mine tunnel as the receiver for optical signal reception. We can infer the current locations of all miners using a series of base stations constructed by PDs. If some special mine environment limits the large-scale deployment of base stations and causes the positioning blind spots, the inertial navigation is combined to achieve all miners’ real-time positioning. We overcome some technical challenges for this new mechanism and the underground mine environment, such as optimal frequency selection, robust frequency identification, and accurate base station positioning. We conduct extensive experiments on our algorithm in a simulated mine tunnel, and the experimental results verify that our algorithm exhibits good performance.

As an essential part of intelligent monitoring, behavior recognition, automatic driving, and others, the challenge of multi-object tracking is still to ensure tracking accuracy and robustness, especially in complex occlusion environments. Aiming at the issues of the occlusion, background noise, and motion state violent change for multi-object in a complex scene, an improved DeepSORT algorithm based on YOLOv5 is proposed for multi-object tracking to enhance the speed and accuracy of tracking. Firstly, a general object motion model is devised, which is similar to the variable acceleration motion model, and a multi-object tracking framework with the general motion model is established. Then, the latest YOLOv5 algorithm, which has satisfactory detection accuracy, is utilized to obtain the object information as the input of multi-object tracking. An unscented Kalman filter (UKF) is proposed to estimate the motion state of multi-object to solve nonlinear errors. In addition, the adaptive factor is introduced to evaluate observation noise and detect abnormal observations so as to adaptively adjust the innovation covariance matrix. Finally, an improved DeepSORT algorithm for multi-object tracking is formed to promote robustness and accuracy. Extensive experiments are carried out on the MOT16 data set, and we compare the proposed algorithm with the DeepSORT algorithm. The results indicate that the speed and precision of the improved DeepSORT are increased by 4.75% and 2.30%, respectively. Especially in the MOT16 of the dynamic camera, the improved DeepSORT shows better performance.

The precise tracking of unmanned aerial vehicle (UAV) in coal mines remains a challenging task that faces strict energy constraints. The existing approaches mainly focus on achieving better accuracy but ignore energy consumption. This article proposes an on-demand precise tracking (OPT) framework based on the fusion of the ultrawideband (UWB) and inertial measurement unit (IMU). The fusion unscented Kalman filter (fusion UKF) is devised for energy-efficient tracking with customized performance. First, OPT provides an adaptive adjustment mechanism of UWB signal transmission to trade off between accuracy and energy consumption for UAV local positioning in coal mines. Second, we propose an on-demand trigger algorithm for UAV remote tracking to determine whether the data are transmitted to the ground server. Especially, the on-demand trigger condition is ameliorated due to the existence of packet drops in complicated coal mines. Finally, extensive simulations in the Gazebo Platform and field experiments utilizing P440 UWB nodes in the indoor and coal mine laboratories have been conducted to verify the feasibility. Results show that OPT is efficient and practical and balances the performance and energy consumption under appropriate parameters. Compared with the existing event-triggered extended Kalman filter (ET-EKF) scheme, the accuracy is improved by 10.3%, and the communication rate is contracted by 11.4% in the coal mine environment.

The localization of coal mine robots (CMRs) serves as the foundation for intelligent mines. Despite various approaches proposed by academia and industry, existing range-based localization methods generally encounter the non-line-ofsight (NLOS) problem, leading to severe accuracy degradation in complicated underground mines. Facing the above challenge, this paper proposes an edge-assisted cross-domain NLOS localization (CrossDNL) framework based on semi-supervised learning. Specifically, we analyze the channel impulse responses (CIRs) of ultra-wideband (UWB) to identify the multi-channel NLOS conditions and further mitigate ranging errors via the deep neural network (DNN). To achieve reliable localization across various scenarios, CrossDNL adopts an edge-assisted semi-supervised learning architecture that enables it to optimize DNN models by self-training on the edge server and achieve the real-time service on the CMR. Besides, we propose a range-based error state Kalman filter (ESKF) scheme to improve the localization performance. We implement CrossDNL on the CMR and the edge server and evaluate it in various scenarios. Results show that CrossDNL can achieve a 14cm localization error in real time, outperforming the latest solutions by more than 20%.

Device-free passive indoor human tracking based on radio frequency (RF) signals has prompted extensive research in academia and industry. Most existing approaches track a single target due to the coarse-grained spatial resolution. Multi-person tracking methods face challenges in complex and dynamic scenes, leading to a sharp decline in accuracy and an exponential increase in computation. In this paper, we employ the impulse radio ultra-wideband (IR-UWB), known for its high spatial resolution and range accuracy, to break down these limitations. We design, implement, and evaluate UWTrack, a clustering assisted device-free tracking system that can achieve the trajectories of multiple persons in real-time with two key components. First, a multi-person detection scheme with the adaptive motion filter and the range compensation is devised, which significantly improves the ranging accuracy and reduces false detections. Second, we propose a clustering-based multi-person tracking method to remove noise points and decline extra updating operations in the Gaussian mixture probability hypothesis density (GM-PHD) filter. It makes UWTrack enhance the real-time tracking performance and decrease the computational cost. We conduct experiments to evaluate UWTrack under various scenarios and conditions. The results reveal that UWTrack can achieve 35.3cm average tracking accuracy and 90.25% detection accuracy in real-time, outperforming the existing solutions by more than 36%.