In order to overcome the several shortcomings of Particle Swarm Optimization (PSO) e.g., premature convergence, low accuracy and poor global searching ability, a novel Simple Particle Swarm Optimization based on Random weight and Confidence term (SPSORC) is proposed in this paper. The original two improvements of the algorithm are called Simple Particle Swarm Optimization (SPSO) and Simple Particle Swarm Optimization with Confidence term (SPSOC), respectively. The former has the characteristics of more simple structure and faster convergence speed, and the latter increases particle diversity. SPSORC takes into account the advantages of both and enhances exploitation capability of algorithm. Twenty-two benchmark functions and four state-of-the-art improvement strategies are introduced so as to facilitate more fair comparison. In addition, a t-test is used to analyze the differences in large amounts of data. The stability and the search efficiency of algorithms are evaluated by comparing the success rates and the average iteration times obtained from 50-dimensional benchmark functions. The results show that the SPSO and its improved algorithms perform well comparing with several kinds of improved PSO algorithms according to both search time and computing accuracy. SPSORC, in particular, is more competent for the optimization of complex problems. In all, it has more desirable convergence, stronger stability and higher accuracy.

Obstacle avoidance planning is the critical technology for driverless trackless rubber-tired vehicles (TRTV) in confined spaces of the deep wells. The narrow, long and circuitous characteristics of the confined spaces challenge the conventional obstacle avoidance planning methods for vehicle operation. Therefore, this paper studies the obstacle avoidance planning method for the driverless TRTV. Firstly, a novel environment model is constructed with a basic grid map and a risk one. The basic grid map describes the obstacle information with multiple planes, while the risk grid map characterizes the safety of drivable areas with five levels. Then, an improved obstacle avoidance planning algorithm is proposed by designing a risk grid particle swarm optimization (RGPSO) algorithm. Finally, by resorting to four typical map benchmarks, our proposed RGPSO is comprehensively evaluated. Compared with the existing heuristic algorithm based obstacle avoidance planning algorithms, the proposed obstacle avoidance planning method has significant merits for complex environment with narrow detours and dead zone traps. Furthermore, based on our results, a two-layer environment model is build for the case mine, and RGPSO is used for obstacle avoidance planning. The results meet the practical mining requirement.

Environmental perception is crucial for achieving autonomous driving of auxiliary haulage vehicles in underground coal mines. The complex underground environment and working conditions, such as dust pollution, uneven lighting, and sensor data abnormalities, pose challenges to multimodal fusion perception. These challenges include: (1) the lack of a reasonable and effective method for evaluating the reliability of different modality data; (2) the absence of in-depth fusion methods for different modality data that can handle sensor failures; and (3) the lack of a multimodal dataset for underground coal mines to support model training. To address these issues, this paper proposes a coal mine underground BEV multiscale-enhanced fusion perception model based on dynamic weight adjustment. First, camera and LiDAR modality data are uniformly mapped into BEV space to achieve multimodal feature alignment. Then, a Mixture of Experts-Fuzzy Logic Inference Module (MoE-FLIM) is designed to infer weights for different modality data based on BEV feature dimensions. Next, a Pyramid Multiscale Feature Enhancement and Fusion Module (PMS-FFEM) is introduced to ensure the model’s perception performance in the event of sensor data abnormalities. Lastly, a multimodal dataset for underground coal mines is constructed to provide support for model training and testing in real-world scenarios. Experimental results show that the proposed method demonstrates good accuracy and stability in object-detection tasks in coal mine underground environments, maintaining high detection performance, especially in typical complex scenes such as low light and dust fog.

Monorail cranes (MCs) are the emerging transportation in the downhole and often work in dangerous deep confined spaces. To improve the safety production efficiency, a driving behavior decision-making method is proposed for unmanned MC. First, a multisource perception data layering method is presented. Based on the results of sensor performance testing in the downhole, an onboard perceptual platform is designed, and the sensor information are layered and integrated in a structured manner, considering the significance of different sensor information on safe driving. Second, a driving behavior decision-making framework for unmanned MCs is proposed. The decision-making objective and constraints are derived from the coal mine safety regulations, the MC driving manual, and the downhole testing results. This framework logic is novel in unmanned MC, which can achieve downhole driving by combining rule-based driving behavior decision-making and reactive one. The experiments in the heading area and the upper yard show that the proposed method increases the perception distance by $3.62\times $ compared with that of remote driving and reduces the decision-making response time by 96%. MC has enough distance and time to slow down first until it has to brake. In contrast, this MC driving behavior is more stable and safe in the downhole.

Membrane materials are crucial components that maintain the stability of tensile membrane structures, and damaged membrane material can readily trigger sudden rupture and collapse. To address the strong subjectivity and low efficiency of manual visual inspection methods, this paper proposes an improved dense convolutional classification network framework for damaged membrane material images, namely, DenseNet121-ECA-PTL. Firstly, two tensile membrane test platforms were independently constructed, and 10,000 damaged membrane material images covering five categories (tear, wrinkle, smudge, screen, and scratch) were collected by using smartphones and UAVs to establish a custom dataset. Secondly, an ECA-enhanced DenseNet121 model combined with transfer learning was then developed to improve complex feature extraction, convergence, and generalization, and its performance was validated through comparative and ablation experiments. The proposed model achieved overall testing accuracy, precision, recall, specificity, F1score, and Matthews correlation coefficient of 98.87%, 98.69%, 98.69%, 98.68%, 99.67%, and 98.35%, respectively. The minimum average prediction time of the proposed model was 0.0472 s for each damaged membrane material image. Finally, model interpretability analyses based on T-SNE and Grad-CAM were used to comprehensively examine the significant feature learning and attention-focusing capabilities of the proposed model for multiple damage types of membrane material images.

Actual construction errors and accumulated damage seriously affect the spatial structural form of the tensile membrane structure, making it difficult to characterize the full-life spatial form of the physical tensile membrane structure based on the original theoretical model. There are still significant challenges in achieving the virtual–real correspondence of digital twin results based on the original theoretical design model. It is necessary to propose a high-precision digital twin model construction method that adapts to the full life cycle of the tensile membrane structure. For this reason, a refined digital twin model reconstruction and morphological deviation visualization method for the umbrella-shaped tensile membrane structure combined with 3D point clouds, computational geometry algorithms, and 3D3S finite element simulation modeling is proposed in this paper. Firstly, three-dimensional point cloud data of the umbrella-shaped membrane structure test bench were acquired, and the point cloud data were accurately registered based on the Iterative Closest Point (ICP) algorithm. Secondly, the physical membrane surface reconstruction was obtained by applying the Screened Poisson Surface Reconstruction (SPSR) algorithm. Subsequently, the 3D3S theoretical model of the umbrella-shaped tensile membrane structure was updated according to the contour of the measured membrane surface reconstruction model. Finally, global spatial geometric morphological deviation was compared among the theoretical models and the reconstructed physical model. Furthermore, different environmental load combinations were analyzed to support hazard warning of the refined digital twin model of the physical umbrella-shaped tensile membrane structure. The results show that the maximum spatial form deviation of the original theoretical model was 43.56 mm, whereas the maximum form deviation of the updated theoretical model was only 0.05 mm. The updated theoretical model accurately reflected the spatial form characteristics of the physical membrane structure and satisfied the requirement for digital twin physical–virtual consistency. In addition, the global stress distribution between the original/updated theoretical models differed markedly, and the updated theoretical model is more suitable for identifying weak areas and evaluating safety performance of the actual umbrella-shaped tensile membrane structure under different ultimate environmental loads.