To resolve the issue of inaccurate global mapping in coal mines, which impedes reliable support for autonomous driving, the paper proposes an enhanced SLAM global mapping method for coal mines based on multimodal data, which consists of LiDAR-inertial-wheel odometry frontend and a multi-factor motivated optimization backend, to realize the pose estimation and map establishment for global mapping of underground roadways. The frontend fuses inertial measurements from the M-SINS, which are compensated for the Earth rotation, with wheel encoder data and LiDAR points using an iterated error-state Kalman filter, mitigating the LiDAR degeneration. The backend utilizes a series of identified spherical targets, sparsely deployed in the tunnel, to introduce global position constrains and unambiguous loop closure detection under a multi-factor pose graph of length distance. Meanwhile, Since the algorithm operates within a globally consistent navigation coordinate system, point-based corrections using GNSS/UWB signals or spherical targets can enhance the accuracy of the global map. The proposed method is evaluated and compared to state-of-the-art approaches through simulation tests in a virtual subterranean roadway and field experiments at an active underground coal mine using our own robot platform. These evaluations demonstrate the method’s performance and practicability in large-scale underground coal mine environments.

A multiparameter control strategy and method for the cutting arm of a roadheader is proposed through the operation analysis of roadheader. The method can address the problems of low intelligence and low cutting efficiency faced by the roadheader in the cutting process. The control strategy is divided into two parts: the cutting load identification part and the swing speed control part. The former part is designed using a backpropagation neural network that is optimized by an improved particle swarm optimization algorithm. The latter part is optimally designed using a fuzzy PID controller with improved simulated annealing particle swarm optimization. The simulation analysis in SIMULINK showed that the response time was reduced, proving the robustness of the method. In addition, experimental studies verified the good control effect of the method under different cutting states. The proposed method uses multiparameter to intelligently change the swing speed, providing a theoretical and practical basis for the realization of intelligent and unmanned cutting of roadheader.

In response to the current situation of backward automation levels, heavy labor intensities, and high accident rates in the underground coal mine auxiliary transportation system, the mining trackless auxiliary transportation robot (MTATBOT) is presented in this paper. The MTATBOT is specially designed for long-range, space-constrained, and explosion-proof underground coal mine environments. With an onboard perception and autopilot system, the MTATBOT can perform automated and unmanned subterranean material transportation. This paper proposes an integrated odometry-based method to improve position estimation and mitigate location ambiguities for simultaneous localization and mapping (SLAM) in large-scale, GNSS-denied, and perceptually degraded subterranean transport roadway scenarios. Additionally, this paper analyzes the robot dynamic model and presents a nonlinear control strategy for the robot to autonomously track a planned trajectory based on the path-following error dynamic model. Finally, the proposed algorithm and control strategy are tested and validated both in a virtual transport roadway environment and in an active underground coal mine. The test results indicate that the proposed algorithm can obtain more accurate and robust robot odometry and better large-scale underground roadway mapping results compared with other SLAM solutions.

The working environment at coal mining faces is harsh, leading to high failure rates and significant maintenance issues with roadheaders. This study explores multi-layer dimensionality reduction of vibration signal features in complex environments to enhance the differentiation of different operational states of a roadheader, thereby achieving fault recognition of key components. Concurrently, reducing dimensionality in manifold spaces positively influences operational state differentiation. Therefore, this paper integrates manifold learning to conduct multi-sensor and multi-layer data mining to enhance the differential phenotypes between faults of key components of the roadheader. Initially, we constructed multiple status-reference sample sets for each sensor individually, forming multiple manifolds at different spatial points, and utilizing locality-preserving projections (LPP) to extract low-dimensional manifold features. Further fusion of low-dimensional features from multiple sensors was used to elevate samples, constructing an enhanced spatial pseudo-manifold. Finally, we used LPP to re-reduce the enhanced sensitive feature set from multiple vibration sensors, establishing a dual-layer sensitive feature enhancement learning model. Conducting fault recognition analysis on experimental vibration signals, using k-nearest neighbors (KNN) to classify the enhanced feature set, we achieved a recognition success rate of 98.75% for samples, proving the method’s feasibility in fault recognition under complex loads.

Underground coal mines belong to the GNSS-denied environment, and the Strapdown Inertial Navigation System (SINS) has a significant advantage in the precise positioning of equipment in this environment because of its operation without requiring interaction with external information and strong anti-interference capabilities. Nonetheless, the vibrations of the installation platform adversely affect the positioning accuracy of SINS. This article focuses on the monorail crane in coal mines as the subject of research, developing a dynamic model for the motion unit consisting of the “track + drive unit + driver’s cab”, while analyzing the relationship between track roughness conditions and the vibration excitation of this unit. Subsequently, utilizing the dynamic model, the study calculated the angular and linear vibration characteristics and formulated models to address coning error and sculling error specific to the SINS in this vibration condition. Lastly, by employing a multi-sample compensation algorithm, this article compensated for positioning errors in the SINS caused by track roughness-induced vibrations during uniform straight-line motion of the motion unit, thus achieving optimal positioning information for the monorail crane. The simulation results demonstrated that employing a four-sample compensation algorithm reduces the coning error in SINS positioning calculations by a minimum of 50% and decreases the sculling error by at least 31%, satisfying the positioning accuracy requirements for precise parking of the monorail crane during the transportation phase, while establishing the foundation for autonomous precise positioning and integrated navigation of underground track transport equipment in coal mines.

The safe and stable operation of roadheader is of great significance to the efficient and rapid production of a coal mine. Health diagnosis based on vibration signals has been studied in bearings and motors. Complex geological conditions and bad working environment lead to the characteristics of nonlinear and time‐varying vibration signals of a roadheader. In this paper, a health state analysis method based on reference manifold (RM) learning and improved K‐means clustering analysis was proposed; the method was verified by using the real‐time collected roadheader cutting reducer fault signal. Firstly, the comparison signal and analysis signal were extracted from the actual collected vibration data of the roadheader, and the referential analysis samples were constructed through time domain and wavelet packet energy analysis. Then, the characteristic structure of the low‐dimensional space of the referential analysis samples is obtained by Locally Linear Embedding (LLE), which is a method of manifold learning. Through the improved K‐means clustering analysis method, the low‐dimensional structure parameters were analyzed and the clustering effect index was obtained, which was used as the health evaluation index (HEI). Finally, the normal distribution model of the health evaluation index is established, and the confidence interval of the health evaluation index is determined, so as to realize the health state analysis of the roadheader and realize the fault warning function. Through the analysis of data of three sensors, the results show that the roadheader failed on the 15th day, which is consistent with the actual working condition. Through practical analysis, the effectiveness of the method was verified and provided a kind of fault analysis idea and method for equipment working under complex working conditions and the theoretical basis for fault type analysis.