Internet of Things (IoT) is enabled by the integration of communication and sensor systems that are used to collect important information from objects around the world. In this article, we proposed the integration of laser-based optical wireless communication (OWC) with a fiber sensor system for IoT applications. OWC can be utilized in a harsh environment to send and receive communication and sensor signals in a wireless means. Moreover, fiber sensors are important to achieve real-time, accurate, and smart monitoring in IoT applications. As compared to mechanical and electrical sensors, optical fiber sensors have numerous benefits, including the tolerance to electromagnetic interference (EMI), small size, lightweight, high bandwidth, low cost, provide distributed sensing, high sensitivity, and electronic devices are not required at the sensor point. Moreover, we proposed a deep variational autoencoder (DVAE) model to estimate the strain changes and peak wavelength of multiple fiber Bragg grating (FBG) sensors using only the spectrum of FBGs obtained from the real experiment. The result showed that our proposed DVAE model can estimate the strain changes and peak wavelength of FBGs with low root mean squared error (RMSE) and high estimation accuracy. Furthermore, the performance of the proposed integration of OWC with the fiber sensor system achieves a clear eye diagram and excellent bit error rate (BER) curve.

This paper verifies the feasibility and effectiveness of utilizing an integration technology of free-space optics (FSO) and satellite communications for an open radio access network (O-RAN) system while simultaneously transmitting fiber-optic sensing signals over the same FSO link. The achievement of O-RAN is to improve the flexibility, openness, and cost-effectiveness of current and next-generation wireless networks. Compared with radio wave-based communication, optical wireless communication (OWC), such as FSO, has many advantages, such as high-speed transmission, data confidentiality, and freedom from electromagnetic interference. Therefore, the use of FSO as a transmission link between ground radar transceivers for satellite communication and O-RAN systems will enhance the communication quality of satellite communication-based O-RAN systems. In particular, FSOs are small and easy to place at ground radar transceiver stations, satellite trucks, and communication target points. This makes it easier to quickly switch the transmission route of the radar transceiver station to different communication transmission and sensing monitoring target scenarios than physical optical fiber cable. In addition, in the fiber optic sensing part, use a machine learning model that combines advanced conditional generative adversarial network (CGAN) and convolutional neural network (CNN) models. This model aims to generate a large amount of sensor data and improve the sensing performance of fiber sensing systems. The experimental findings unambiguously show that the proposed integrated system achieves normal functioning in the communication part and outstanding temperature sensing measurement accuracy in the sensing part.

In this work, we designed a novel hybrid of free-space optics (FSO) communication and intensity and wavelength division multiplexing (IWDM)-fiber Bragg grating (FBG) sensor system over a single shared FSO transmission channel for both optical and sensor signal transmission. In the proposed system, we use one broadband light source and one FSO link for both optical and FBG sensor signal transmission, which significantly reduces the cost of system installation and complexity. Besides, the FSO transmission channel is used to solve the problem of geographical constraints when the environment is difficult to install the fiber. In addition, IWDM is one of the best multiplexing technique to increase the sensor number by more than two times the conventional wavelength division multiplexing (WDM) method. Besides, to solve the overlap spectra problem of IWDM based FBG sensor system, we proposed the stacked gated recurrent unit (SGRU) algorithm with the neural network (NN) profile method. The results prove that the proposed SGRU with the NN profile method improves the accuracy of wavelength detection techniques with the Gaussian profile method. Moreover, the FSO transmission channel achieves a clear eye diagram even the system is a hybrid of optical communication and FBG sensors system.

There is a positive feedback mechanism of desorption, pulverization, and redesorption in the process of coal and gas outburst. The coal pore structure changes to some extent in the process of coal pulverization and has a related impact on the dynamic parameters of coal particle gas desorption. To understand the impact of desorption damage on the dynamic characteristics of coal particle gas desorption, in this paper, a self-developed coal particle gas desorption test device is used to measure the amount of methane desorption of different coal samples repeatedly desorbed under the same adsorption equilibrium pressure. The results show that the methane desorption kinetic curves of the four coal samples based on desorption damage basically have the same trend. Nie's diffusion model can better describe the methane desorption characteristics of coal particles. After desorption, the methane desorption amount, initial desorption velocity, diffusion ability, and ultimate amount of methane desorption of the coal samples are greater than those before desorption. The desorption damage affects the relevant dynamic parameters of the coal particle gas diffusion model, and its impact on the outburst coal is greater than that on raw coal. In addition, the pore size distribution and change characteristics of the coal samples before and after desorption are analyzed quantitatively via a low-temperature liquid nitrogen adsorption test and fractal dimension-related theory. It was found that the pore volume peak area of the coal samples affected by desorption damage within each pore size range was significantly larger than that before desorption. Among them, micropores have the most significant impact on the desorption damage of coal samples, and the peak area of the pore volume of protruding coal is greater than that of raw coal, indicating that the internal pores of coal after desorption damage are more developed than those of raw coal.

This article proposes a novel integration of bidirectional free-space optics (FSO) and fiber-based fiber Bragg gratings (FBG) sensor structure for a long-distance sensing signal transmission system. The proposed FSO transmission technique is crucial to address concerns of topographical limitations that make fiber cable installation complex. According to experiments, as compared the proposed bidirectional scheme with a unidirectional FSO transmission channel achieved higher output power and clear reflected spectra of FBGS after 50 kilometers of SMF signal transmission. Besides, a combination of the neural network (NN) profile, recurrent generative adversarial network (R-GAN), and gated recurrent units (GRU) methods are proposed to address the overlap spectra problem and enhance peak wavelength detection accuracy in the FBG sensor system. The two NN profile and R-GAN methods are used to collect large amounts of strain sensor data, while the GRU method is used for peak wavelength detection in the FBG sensor system. The results demonstrated that the proposed two steps of the NN and R-GAN method are capable to generate a significant amount of realistic spectral training sets to improve the performance of the GRU deep learning model. Moreover, the results indicate that the well-trained GRU model using datasets generated by the NN and R-GAN method achieves superior performance than other previously reported machine learning techniques.

Violent attacks have been one of the hot issues in recent years. In the presence of closed-circuit televisions (CCTVs) in smart cities, there is an emerging challenge in apprehending criminals, leading to a need for innovative solutions. In this paper, the propose a model aimed at enhancing real-time emergency response capabilities and swiftly identifying criminals. This initiative aims to foster a safer environment and better manage criminal activity within smart cities. The proposed architecture combines an image-to-image stable diffusion model with violence detection and pose estimation approaches. The diffusion model generates synthetic data while the object detection approach uses YOLO v7 to identify violent objects like baseball bats, knives, and pistols, complemented by MediaPipe for action detection. Further, a long short-term memory (LSTM) network classifies the action attacks involving violent objects. Subsequently, an ensemble consisting of an edge device and the entire proposed model is deployed onto the edge device for real-time data testing using a dash camera. Thus, this study can handle violent attacks and send alerts in emergencies. As a result, our proposed YOLO model achieves a mean average precision (MAP) of 89.5% for violent attack detection, and the LSTM classifier model achieves an accuracy of 88.33% for violent action classification. The results highlight the model’s enhanced capability to accurately detect violent objects, particularly in effectively identifying violence through the implemented artificial intelligence system.