Public Safety Applications Research Articles

Artificial Intelligence in Governance: Opportunities, Challenges, and Ethical Implications for Public Administration

The integration of Artificial Intelligence (AI) in the public sector presents both unprecedented opportunities and significant challenges for governments worldwide. This article examines the multifaceted impact of AI on public administration, exploring its applications in service delivery, urban planning, public safety, and administrative efficiency. While AI offers substantial benefits, including enhanced decision-making, cost savings, and improved citizen services, it also raises critical concerns regarding bias, privacy, accountability, and workforce displacement. This article provides a comprehensive analysis of the ethical considerations surrounding AI deployment in governance, emphasizing the need for transparency, inclusivity, and human oversight. By critically evaluating both the promises and perils of AI in public sector operations, this article contributes to the ongoing discourse on responsible AI adoption in government. The findings underscore the importance of developing robust governance frameworks that can harness AI's potential while safeguarding citizens' rights and ensuring equitable service delivery. As governments navigate the AI revolution, this article offers insights into strategies for balancing technological advancement with ethical governance, paving the way for a more efficient, fair, and responsive public sector.

Accurate crowd counting for intelligent video surveillance systems

The paper presents a novel deep learning approach for crowd counting in intelligent video surveillance systems, addressing the growing need for accurate monitoring of public spaces in urban environments. The demand for precise crowd estimation arises from challenges related to security, public safety, and efficiency in urban areas, particularly during large public events. Existing crowd counting techniques, including feature-based object detection and regression-based methods, face limitations in high-density environments due to occlusions, lighting variations, and diverse human figures. To overcome these challenges, the authors propose a new deep encoder-decoder architecture based on VGG16, which incorporates hierarchical feature extraction with spatial and channel attention mechanisms. This architecture enhances the model’s ability to manage variations in crowd density, leveraging adaptive pooling and dilated convolutions to extract meaningful features from dense crowds. The model’s decoder is further refined to handle sparse and crowded scenes through separate density maps, improving its adaptability and accuracy. Evaluations of the proposed model on benchmark datasets, including Shanghai Tech and UCF CC 50, demonstrate superior performance over state-of-the-art methods, with significant improvements in mean absolute error and mean squared error metrics. The paper emphasizes the importance of addressing environmental variability and scale differences in crowded environments and shows that the proposed model is effective in both sparse and dense crowd conditions. This research contributes to the advancement of intelligent video surveillance systems by providing a more accurate and efficient method for crowd counting, with potential applications in public safety, transportation management, and urban planning.

Efficient Violence Recognition in Video Streams using ResDLCNN-GRU Attention Network

Detecting Violence in video streams is essential for public safety and security due to the rising frequency of violent incidents. Despite the extensive deployment of CCTV for surveillance, the available human monitoring resources still need to catch up with the need for vigilant supervision. This research presents a new lightweight model to address this gap by accurately identifying and categorizing violent behaviors in various scenarios, including CCTV footage. The proposed method leverages optical flow and RGB data to capture spatiotemporal features in the Violence data. Built on a Residual DLCNN architecture integrated with the Attention mechanism and GRU components, the model effectively handles high-dimensional video data, enhancing accuracy by prioritizing crucial frames containing violent and nonviolent instances. The proposed model's performance was validated on the Hockey Fights (HF), Movie Fights, and SCVD datasets, achieving impressive accuracies of 98.38%, 99.62%, and 90.57%, respectively. Here, we developed the Extended Automatic Violence Detection Dataset (EAVDD), featuring 1530 videos of violent scenes in movies, public spaces, social media, and sports. Testing the model with top fight scenes in rated movies yielded outstanding results. This research supports surveillance systems and advances short video analysis and understanding with applications in public safety, social media, sports, and law enforcement.

Integrated Image Processing and Gas Sensing for Enhanced Hazard Detection.

In industrial and public safety applications, detecting hazardous situations in a timely manner is critical for averting accidents and maintaining human safety. This work offers a novel method for identifying dangerous environments by using cutting-edge image processing algorithms. Using sophisticated image processing algorithms, the technology extracts pertinent information from visual data that is recorded by cameras or other sensors. Gas sensors are simultaneously used to track the chemical composition of the surrounding air, with a special emphasis on chemicals that may pose a threat. The integration of gas and optical sensor data allows for a more thorough analysis, which enhances the system's precision in identifying and categorizing dangerous situations.

Impact of gamma rays on DNA quality and genetic traits in human blood samples

This research investigates the effects of gamma rays on DNA quality and genetic traits in eight participants from Saudi Arabia (4 males and 4 females). The study aims to understand changes in genetic fingerprint sites resulting from varying levels of radiation exposure. It addresses three key research questions related to DNA quantity, obtaining complete genetic fingerprint sites, and the reliability of genetic trait results from cancer patients undergoing radiation therapy. The study's significance lies in its potential applications in forensics, medical diagnostics, and public safety, providing insights into identifying victims and perpetrators in radiation-related incidents and enhancing treatment planning for cancer patients exposed to radiation therapies. The research contributes to a broader understanding of radiation's impact on genetic material, promotes scientific knowledge, and safeguards public health. This research investigated the impact of gamma radiation on DNA quality and genetic traits in human blood samples from Saudi Arabian volunteers. Exposure to gamma radiation at doses similar to therapeutic levels used in cancer treatment was found to cause DNA fragmentation, resulting in decreased concentrations of detectable genetic loci. The study revealed that higher radiation doses had a more pronounced effect on DNA content in blood cells.

A naked-eye colorimetric assay for highly sensitive detection of CL-20 explosive

2,4,6,8,10,12-hexanitro-2,4,6,8,10,12-hexaazaisowurtzitane (CL-20) stands as the most energetically condensed explosive ever found, and poses a great security threat to humans. Here we report a naked-eye colorimetric assay for highly sensitive detection of CL-20. Through effective cascading of the hydrolysis process of CL-20 with nitrite-involved Griess reaction to yield a pink-colored azo product, the visual quantification of CL-20 is realized by taking advantage of corresponding colorimetric signals. Under optimal experimental conditions, the present colorimetric assay is successfully used for highly selective determination of CL-20 in deionized water and environmental water samples with a nanomolar sensitivity. The resulting limit of detection is much lower than conventional methods. Moreover, benefiting from the distinct color change of the detection solution, direct naked-eye semi-quantitative analysis of CL-20 is achieved with a threshold concentration of 1 μM. Such simple, cost-effective, selective, and highly sensitive visual colorimetric assay will have valuable applications in public safety, environmental monitoring, and other related fields.

Four port MIMO antenna for IoT applications in public safety band and sub-6 GHz TDD 5G band

This paper proposes a Multi Input Multi Output (MIMO) antenna specifically developed for Internet of Things (IoT) Applications in Public Safety 4.9 GHz (4940–4990 MHz) band and Sub-6 GHz TDD 5G band (4800–4990 MHz). The proposed MIMO antenna consists of Koch fractal element, and it is introduced in four port configuration. HFSS is used to design, analyze, and optimize the performance parameters of the introduced MIMO antenna. To achieve optimum performance in terms of isolation, initially all two radiating elements of 1 × 2 MIMO configuration with varying orientations have been evaluated, which are positioned on top of a cost-effective FR4 substrate. Later on 2 × 2 and 1 × 4 MIMO configurations are evaluated; and 1 × 4 configuration is fabricated and tested in lab for the verification of simulated results. The MIMO antenna demonstrates resonance at 4.97 GHz, offering a bandwidth of 320 MHz (4795–5115 MHz). It exhibits excellent performance with Envelope Correlation Coefficient (EC-C) < 0.4, Diversity Gain (D-G) > 9.8, Mean Effective Gain (ME-G) < 3 dB, Total Active Refection Coefficient (TAR-C) < −14 dB, Channel Capacity Loss (CC-L) < 0.4 bits/s/Hz, and Isolation < −20 dB across the entire operating frequency range. The proposed MIMO antenna has a maximum gain of 1.2 dBi and radiation efficiency (η) of 80%. This proposed antenna offers several advantages over traditional ones, including high efficiency (η), low CC-L, and significant isolation, even without the use of special techniques to achieve isolation.

Efficient CNN-based low-resolution facial detection from UAVs

AbstractFace detection in UAV imagery requires high accuracy and low execution time for real-time mission-critical operations in public safety, emergency management, disaster relief and other applications. This study presents UWS-YOLO, a new convolutional neural network (CNN)-based machine learning algorithm designed to address these demanding requirements. UWS-YOLO’s key strengths lie in its exceptional speed, remarkable accuracy and ability to handle complex UAV operations. This algorithm presents a balanced and portable solution for real-time face detection in UAV applications. Evaluation and comparison with the state-of-the-art algorithms using standard and UAV-specific datasets demonstrate UWS-YOLO’s superiority. It achieves 59.29% of accuracy compared with 27.43% in a state-of-the-art solution RetinaFace and 46.59% with YOLOv7. Additionally, UWS-YOLO operates at 11 milliseconds, which is 345% faster than RetinaFace and 373% than YOLOv7.

Sustainable Abnormal Events Detection and Tracking in Surveillance System

With the proliferation of surveillance cameras, managing and analyzing vast amounts of video data have become challenging. This paper introduces a sustainable automated approach to detect abnormal events in surveillance footage. Leveraging Convolutional Neural Networks (CNNs) and deep learning techniques, our system identifies unusual activities by analyzing video frames. By automating this process, we reduce the burden of manual monitoring and enable timely responses to security threats. This sustainable solution has broad applications in public safety, security, and crime prevention.

Enabling Metaheuristics with Deep Learning based Resource Allocation in Unmanned Aerial Vehicles Wireless Networks

Unmanned aerial vehicle (UAV) network offers a variety of applications in public safety, disaster management, advertising and broadcasting, overload situation, etc. Due to the dynamic characteristics of MU, it is challenging to provide robust transmission services to mobile users (MU). Resource allocation (RA), including sub-channel, serving user, and transmit power, is a crucial problem; also, it is critical to enhance the coverage and energy efficiency of UAV-enabled communication protocol. Furthermore, system resources are limited (for example, spectrum, and transmission power) and UAV transmission coverage and on-board energy are limited. In order to meet the QoE of any user with limited UAV energy and limited resource system, we jointly enhance UAV trajectory, user communication scheduling, and bandwidth allocation and transmit power to satisfy user QoE requirements and increase energy efficiency. Thus, the study proposes a new mud ring optimization with deep belief network-based resource allocation scheme (MRODBN-RAS) technique for UAV-enabled wireless networks. The proposed MRODBN-RAS approach focuses on the effectual accomplishment of the computational and energy-effective decision. Besides, the MRODBN-RAS technique assumed the UAV as a learning agent by forming RA decisions as actions. In addition, the MRODBN-RAS technique designed a reward function to reduce the weighted resource utilization. The MRODBN-RAS technique uses DBN model with hyperparameter tuning using MRO algorithm to allocate the resources. The design of the MRO algorithm helps in the optimal selection of the hyperparameter related to the DBN model. The simulation results of the MRODBN-RAS method are examined under various measures. The extensive comparison study highlighted the better performance of the MRODBN-RAS approach over existing techniques.

Dynamic-Distance-Based Thresholding for UAV-Based Face Verification Algorithms.

Face verification, crucial for identity authentication and access control in our digital society, faces significant challenges when comparing images taken in diverse environments, which vary in terms of distance, angle, and lighting conditions. These disparities often lead to decreased accuracy due to significant resolution changes. This paper introduces an adaptive face verification solution tailored for diverse conditions, particularly focusing on Unmanned Aerial Vehicle (UAV)-based public safety applications. Our approach features an innovative adaptive verification threshold algorithm and an optimised operation pipeline, specifically designed to accommodate varying distances between the UAV and the human subject. The proposed solution is implemented based on a UAV platform and empirically compared with several state-of-the-art solutions. Empirical results have shown that an improvement of 15% in accuracy can be achieved.

Enhanced Hazardous Material Management through Advanced Robotic Solutions

Abstract: This abstract introduces an advanced explosive disposal robot built to handle dangerous tasks traditionally done by bomb disposal experts. It comes with cutting-edge technology, a strong chassis for tough terrains, and a precise mechanical arm for delicate tasks and controlled detonations. The robot includes high-res cameras, thermal sensors, and chemical detectors for real-time feedback. It's designed for remote control, ensuring operator safety and precise movements. Safety, reliability, and ease of use are focal points, validated through rigorous testing. This robot reduces risks, improves efficiency, and finds applications in military, law enforcement, and public safety.

A Review of GAN-Based Super-Resolution Reconstruction for Optical Remote Sensing Images

High-resolution images have a wide range of applications in image compression, remote sensing, medical imaging, public safety, and other fields. The primary objective of super-resolution reconstruction of images is to reconstruct a given low-resolution image into a corresponding high-resolution image by a specific algorithm. With the emergence and swift advancement of generative adversarial networks (GANs), image super-resolution reconstruction is experiencing a new era of progress. Unfortunately, there has been a lack of comprehensive efforts to bring together the advancements made in the field of super-resolution reconstruction using generative adversarial networks. Hence, this paper presents a comprehensive overview of the super-resolution image reconstruction technique that utilizes generative adversarial networks. Initially, we examine the operational principles of generative adversarial networks, followed by an overview of the relevant research and background information on reconstructing remote sensing images through super-resolution techniques. Next, we discuss significant research on generative adversarial networks in high-resolution image reconstruction. We cover various aspects, such as datasets, evaluation criteria, and conventional models used for image reconstruction. Subsequently, the super-resolution reconstruction models based on generative adversarial networks are categorized based on whether the kernel blurring function is recognized and utilized during training. We provide a brief overview of the utilization of generative adversarial network models in analyzing remote sensing imagery. In conclusion, we present a prospective analysis of forthcoming research directions pertaining to super-resolution reconstruction methods that rely on generative adversarial networks.

A lightweight license plate detection algorithm based on deep learning

AbstractLicense plate detection is an important task in Intelligent Transportation Systems (ITS) and has a wide range of applications in vehicle management, traffic control, and public safety. In order to improve the accuracy and speed of mobile recognition, an improved lightweight YOLOv5s model is proposed for license plate detection. First, an improved Stemblock network is used to replace the original Focus layer in the network, which ensures strong feature expression capability and reduces a large number of parameters to lower the computational complexity; then, an improved lightweight network, ShuffleNetv2, is used to replace the backbone network of the YOLOv5s, which makes the model lighter and ensures the detection accuracy at the same time. Then, a feature enhancement module is designed to reduce the information loss caused by the rearrangement of the backbone network channels, which facilitates the information interaction in the feature fusion process; finally, the low‐, medium‐ and high‐level features in the Shufflenetv2 network structure are fused to form the final high‐level output features. Experimental results on the CCPD dataset show that compared to other methods this paper obtains better performance and faster speed in the license plate detection task, in which the average precision mean value reaches 96.6%, and can achieve a detection speed of 43.86 frame/s, and the parameter volume is reduced to 5.07 M.

On the devolvement of fractal antenna for IoT applications

A fractal patch antenna is designed and fabricated on FR4-epoxy substrate for Internet of Things (IoT) applications. Minkowski fractal has been used in the design of radiating element of antenna and slots are incorporated in the partial ground structure. The proposed antenna is miniaturized by 65% in terms of resonating frequency using Minkowski fractal. The proposed structure of antenna has an overall size of 60× 40 ×1.6 mm3. The antenna resonates at 1.223, 1.58, 2.69, and at 3.54 GHz. It exhibits an impedance bandwidth of 0.11 (1.16–1.27) GHz, 0.23 (1.47–1.70) GHz, and 4.5 (1.99–6.49) GHz. The proposed antenna has a peak gain of 3.4 dBi. The key benefits of the proposed antenna are: simple structure, improved performance characteristics, and its usefulness for applications such as Global Positioning System (GPS), Radar Altimeter, Public Safety Applications, and 5G communications in sub-6 GHz band. The simulated results of the proposed fractal antenna have been verified experimentally. The results are highly convincing.

One-Step Synthesis of Self-Stratification Core-Shell Latex for Antimicrobial Coating

Herein, we describe a one-step method for synthesizing cationic acrylate-based core-shell latex (CACS latex), which is used to prepare architectural coatings with excellent antimicrobial properties. Firstly, a polymerizable water-soluble quaternary ammonium salt (QAS-BN) was synthesized using 2-(Dimethylamine) ethyl methacrylate (DMAEMA) and benzyl bromide by the Hoffman alkylation reaction. Then QAS-BN, butyl acrylate (BA), methyl methacrylate (MMA), and vinyltriethoxysilane (VTES) as reactants and 2,2'-azobis(2-methylpropionamidine) dihydrochloride (AIBA) as a water-soluble initiator were used to synthesize the CACS latex. The effect of the QAS-BN dosage on the properties of the emulsion and latex film was systematically investigated. The TGA results showed that using QAS-BN reduced the latex film's initial degradation temperature but improved its thermal stability. In the transmission electron microscopy (TEM) photographs, the self-stratification of latex particles with a high dosage of QAS-BN was observed, forming a core-shell structure of latex particles. The DSC, TGA, XPS, SEM, and performance tests confirmed the core-shell structure of the latex particles. The relationship between the formation of the core-shell structure and the content of QAS-BN was proved. The formation of the core-shell structure was due to the preferential reaction of water-soluble monomers in the aqueous phase, which led to the aggregation of hydrophilic groups, resulting in the formation of soft-core and hard-shell latex particles. However, the water resistance of the films formed by CACS latex was greatly reduced. We introduced a p-chloromethyl styrene and n-hexane diamine (p-CMS/EDA) crosslinking system, effectively improving the water resistance in this study. Finally, the antimicrobial coating was prepared with a CACS emulsion of 7 wt.% QAS-BN and 2 wt.% p-CMS/EDA. The antibacterial activity rates of this antimicrobial coating against E. coli and S. aureus were 99.99%. The antiviral activity rates against H3N2, HCoV-229E, and EV71 were 99.4%, 99.2%, and 97.9%, respectively. This study provides a novel idea for the morphological design of latex particles. A new architectural coating with broad-spectrum antimicrobial properties was obtained, which has important public health and safety applications.

Status of Law Enforcement Drone Education at Two-Year Community College

Small unmanned aircraft systems (drones) have only been widely available for civilian application for less than 10 years. Despite the newness, drones are being used to support a broad range of public safety applications. The purpose of this study is to evaluate if law enforcement programs at two-year community colleges are providing adequate drone education. The course catalogs and websites of 255 community colleges in the Southeast United States were searched for drone-related keywords. Only 16% of the schools had at least one drone-related course available. A follow-up survey was sent to these schools’ law enforcement program coordinators. The survey revealed that the program coordinators know that law enforcement is using drones, felt drones reduced injuries, and desired to expand their UAS curriculum. However, the survey also found that current faculty are unfamiliar with UAS regulations and have a general lack of professional development in this area.

Forget less, count better: a domain-incremental self-distillation learning benchmark for lifelong crowd counting

Crowd counting has important applications in public safety and pandemic control. A robust and practical crowd counting system has to be capable of continuously learning with the new incoming domain data in real-world scenarios instead of fitting one domain only. Off-the-shelf methods have some drawbacks when handling multiple domains: (1) the models will achieve limited performance (even drop dramatically) among old domains after training images from new domains due to the discrepancies of intrinsic data distributions from various domains, which is called catastrophic forgetting; (2) the well-trained model in a specific domain achieves imperfect performance among other unseen domains because of the domain shift; and (3) it leads to linearly increasing storage overhead, either mixing all the data for training or simply training dozens of separate models for different domains when new ones are available. To overcome these issues, we investigated a new crowd counting task in the incremental domains training setting called Lifelong Crowd Counting. Its goal is to alleviate the catastrophic forgetting and improve the generalization ability using a single model updated by the incremental domains. Specifically, we propose a self-distillation learning framework as a benchmark (Forget Less, Count Better, or FLCB) for lifelong crowd counting, which helps the model sustainably leverage previous meaningful knowledge for better crowd counting to mitigate the forgetting when the new data arrive. In addition, a new quantitative metric, normalized backward transfer (nBwT), is developed to evaluate the forgetting degree of the model in the lifelong learning process. Extensive experimental results demonstrate the superiority of our proposed benchmark in achieving a low catastrophic forgetting degree and strong generalization ability.

Blockchain-Driven Intelligent Scheme for IoT-Based Public Safety System beyond 5G Networks.

Mobile applications have rapidly grown over the past few decades to offer futuristic applications, such as autonomous vehicles, smart farming, and smart city. Such applications require ubiquitous, real-time, and secure communications to deliver services quickly. Toward this aim, sixth-generation (6G) wireless technology offers superior performance with high reliability, enhanced transmission rate, and low latency. However, managing the resources of the aforementioned applications is highly complex in the precarious network. An adversary can perform various network-related attacks (i.e., data injection or modification) to jeopardize the regular operation of the smart applications. Therefore, incorporating blockchain technology in the smart application can be a prominent solution to tackle security, reliability, and data-sharing privacy concerns. Motivated by the same, we presented a case study on public safety applications that utilizes the essential characteristics of artificial intelligence (AI), blockchain, and a 6G network to handle data integrity attacks on the crime data. The case study is assessed using various performance parameters by considering blockchain scalability, packet drop ratio, and training accuracy. Lastly, we explored different research challenges of adopting blockchain in the 6G wireless network.

CASTLE: A CONTEXT-AWARE SPATIAL-TEMPORAL LOCATION EMBEDDING PRE-TRAINING MODEL FOR NEXT LOCATION PREDICTION

Abstract. Next location prediction is helpful for service recommendation, public safety, intelligent transportation, and other location-based applications. Existing location prediction methods usually use sparse check-in trajectories and require massive historical data to capture complex spatial-temporal correlations. High spatial-temporal resolution trajectories have rich information. However, obtaining personal trajectories with long time series and high spatio-temporal resolution usually proves challenging. Herein, this paper proposes a two-stage Context-Aware Spatial-Temporal Location Embedding (CASTLE) model, a multi-modal pre-training model for sequence-to- sequence prediction tasks. The method is built in two steps. First, large-scale location datasets, which are sparse but easier to be acquired (i.e., check-in and anomalous navigation data), are used for pre-training location embedding to capture the multi-functional properties under different contexts. After that, the learned contextual embedding is used for downstream location prediction in small-scale but higher spatio-temporal resolution trajectory datasets. Specifically, the CASTLE model combines Bidirectional and Auto-Regressive Transformers to generate contextual embedding vectors rather than a fixed vector for each location. Furthermore, we introduce a location and time-aware encoder to reflect the spatial distances between locations and visit times. Experiments are conducted on two real trajectory datasets. The results show that the CASTLE model can pre-train beneficial location embedding and outperforms the model without pre-training by 4.6–7.1%. The proposed method is expected to improve the next location prediction accuracy without massive historical data, which will greatly drive the use of trajectory data.