Internet of Vehicles and Computer Vision Solutions for Smart City Transformations

In the development of smart cities, technology plays a crucial role in shaping civilian life in multiple sectors like agriculture, healthcare, industry, education, transportation, etc. Smart city seeks a wide range of diverse applications, including media, surveillance, package delivery, and search and rescue operations. Unmanned aerial vehicles (UAVs) are visualised as a potential candidate technology for fulfilling smart city requirements. Thus, UAVs are poised to revolutionize global lifestyles shortly. Integrating UAVs into smart city frameworks holds immense potential and promising benefits, such as intelligent traffic management, media services, efficient surveillance, improved farming and healthcare systems, etc. This integration aims to achieve increased levels of efficiency, effectiveness, and innovation, thereby contributing to the advancement and sustainability of urban environments worldwide. The potential of UAVs to transform smart cities is significant, offering cost-effective solutions across numerous domains. These UAV applications require specific communication and spectrum requirements, catering to non-payload and payload services. On the other hand, quality of service (QoS), reliability, scalability, and security are the challenges associated with using UAVs for urban applications. Thus, choosing appropriate wireless communication technologies for UAV operations is essential. This chapter delves into the pivotal role of UAVs in the evolution of smart cities. To ensure seamless connectivity and security, the chapter discusses the technical aspects of integrating UAV communication to enable various smart city applications. The chapter candidly acknowledges the challenges associated with UAV communications, including connectivity, adaptability, safety, privacy, and reliability issues. Despite these challenges, the chapter provides valuable insights into current technological advancements and prospects for addressing these challenges, thereby optimizing UAV communications within smart city frameworks.

The applications of ITS in smart cities are one of the many IT applications resulting from the implementation of IoT applications. Some of the barriers to exploiting the IoT in city mobility management include: Different devices and communication protocols to be fully compatible with each other is a significant challenge when it comes to the IoT. This research presents a new system design in an attempt to enhance the communication and overall compatibility of devices in smart city ITs based on IoT architecture. We then proceed to the identification of the needs and constraints not only of the various stakeholders, infrastructure as well as mobility services that make up smart city transport system but also the individual systems that comprise it as well. The feature of modularity is inbuilt into our system to allow for the integration of various internet connected devices, sensors and communication protocols generally found in urban mobility systems. Its human interface or terminal requires an application interface that can be adapted to connect with other ITS applications The software architecture also comprises a data acquisition layer, a middle layer that handles protocols and data formatting in its entirety, among others. Our primary goal is to reflect changes in the conditions of smart city environments and new trends in IoT standards, so it is relevant to use a standard-based approach prioritizing compatibility and extensibility. With a parallel aim to supporting previously deployed IoT devices and also for forward compatibility the system employs standard IoT protocols such as MQTT, CoAP and HTTP. If we desire better mechanisms to express and understand data in different systems and platforms, namely semantic interoperability, we need to adopt common data models and ontology definitions. To show that their recommended system architecture works, we build a proof of concept in a digital smart city environment. Using their features and behavior, we substantiate the neurological and tender integration of various IoT devices including traffic sensors, vehicle tracking systems and units, public transportation facilities, and stations, and environmental measuring platforms. The current design of our systems can be seen as offering a solid means of enhancing the usage of the Internet of Things as a means of enabling ITMS to exchange information between itself and other systems. In Smart cities, this framework provides ways of developing proper mobility systems that are proper, friendly to the environment, and all geared towards satisfying the users.

The potential impact of generative AI (GenAI) on urban transportation is vast and significant. This chapter explores a future where diverse mobile agents, empowered by GenAI models, collaborate efficiently through ITSs to achieve a variety of objectives. GenAI has the potential to enhance decision-making for vehicles and infrastructure, leading to safer roads and smoother traffic patterns. It has the potential to generate realistic driving and traffic scenarios. Additionally, GenAI can assist agents in predicting outcomes and planning actions without direct real-world interaction, thus improving the reliability and adaptability of machine learning (ML)-based ITS. This chapter also explores large multi-modal models (LMMs) and their potential applications in the ITS represent a promising frontier in this landscape, providing advanced abilities to interpret and generate diverse multimodal data inputs. Rather than seeing GenAI as a substitute for existing ML algorithms, we consider it as a valuable addition, offering advanced capabilities in abstraction, planning, and reasoning, which can evolve conventional systems from autonomous to cognition.

This chapter presents a unique application of Large Language Models (LLMs) to enhance human-machine collaboration, specifically focusing on the operation of Unmanned Aerial Vehicles (UAVs) in urban environments. The development of a universal framework embedding LLMs to act as UAV “Co-Pilots” is explored, enabling the UAVs to interpret and execute human intentions accurately while optimizing interaction through a well-defined utilization workflow. A context management mechanism is introduced to systematically manage the information involved in these tasks. Additionally, the chapter outlines a collaborative control framework for large-scale UAV deployments, addressing computational and communication challenges by predicting environmental changes and optimizing information exchange. The potential of hybrid intelligence-combining human insights with machine autonomy-is emphasized, showcasing the practicality and future potential of LLMs in facilitating autonomous UAV operations in complex urban scenarios. Key discussions cover the integration of LLMs into existing UAV systems, potential challenges, and prospective advancements, establishing the groundwork for improved human-machine interactions in real-world applications.

Urban ecosystems in the midst of digital transformation utilize technologies like Internet of Things (IoT), big data, and artificial intelligence (AI) to elevate citizens’ well-being and champion sustainable development. Notably, LLMs such as GPT-3 or GPT-4, crafted by OpenAI, and Google’s AI tools like Gemini and Bard, stand as pivotal components, holding immense potential to revolutionize EVs in smart city initiatives. Generative AI and LLM integration prove influential in optimizing EV functionalities, enhancing security, and fortifying data privacy. This chapter delves into battery technology, battery capacities, energy storage, charging infrastructure optimization, autonomous driving, educational outreach, energy management, and predictive maintenance. In particular, LLMs address EV challenges by optimizing routes, speeds, and behaviors, reducing energy usage and emissions. They facilitate the smart integration of EVs and renewable energy sources, managing charging schedules based on availability and grid demands. Additionally, LLMs contribute to public education on EV benefits, environmental impact, cost savings, and sustainable transportation promotion. The chapter explores synergies between LLMs and EVs, focusing on potential applications and challenges, propelling progress in ITSs. Similarly, Multimodal Large Language Models (MLLMs) and their potential applications in the era of the EV represent a promising frontier in this landscape, offering enhanced capabilities to interpret and generate multimodal data, further enriching the EV ecosystem.

The Internet of Vehicles (IoV) plays a crucial role in the development and implementation of smart city initiatives by enhancing transportation systems, improving road safety, and optimizing urban mobility. The main challenges in smart cities are to minimise interference from Vehicle-to-Vehicle (V2V) pairs and Vehicle-to-Infrastructure (V2I) users to improve the overall throughput and efficiency of IoV systems. In this chapter, the power allocation problem is investigated to minimise the interference and maximise the throughput of V2V pairs for Non-orthogonal multiple access (NOMA)-enabled IoV system. The maximisation problem is non-convex in nature and hence difficult to solve directly. Therefore, the maximisation problem is converted into a convex one by using the Lagrange Multiplier technique. The Hungarian method is applied for reuse pair matching to enhance the spectrum efficiency of the system. Hence, the throughput of V2V pairs is analysed in terms of the Signal-to-interference noise ratio (SINR) of V2V users, and transmit power of V2V users. The effect of the Successive Interference Cancellation (SIC) parameter is also investigated on V2V throughput for different values of transmit power. Furthermore, the proposed scheme achieves a higher throughput improvement of 39.5 \(\%\) in comparison to the existing scheme.

This chapter explores the intersection of artificial intelligence (AI) and real-time traffic monitoring in the context of smart cities. Focusing on the imperative need for swift, data-driven decisions in urban transportation, the discussion unravels the core components of real-time traffic systems. It scrutinizes the deployment of AI, particularly machine learning algorithms and computer vision, to enhance the speed and precision of traffic analysis. The integration of Internet of Things (IoT) sensors emerges as a linchpin, ensuring comprehensive data collection. The chapter navigates the symbiotic relationship between AI and IoT, emphasizing communication protocols that underpin seamless connectivity. Real-world case studies amplify the exploration, distilling lessons from cities adept at leveraging AI for dynamic traffic surveillance. Challenges and ethical considerations inherent in real-time monitoring are confronted, encompassing technical hurdles and privacy issues. Looking forward, the chapter extrapolates the trajectory of AI and IoT in traffic management, envisioning novel methodologies and technologies on the horizon. In essence, this chapter serves as a compact yet comprehensive guide, unveiling the transformative potential of AI in real-time traffic monitoring for the sustainable evolution of smart cities.

Autonomous Vehicles (AVs) equipped with camera systems have emerged as a pivotal solution for smart urban mobility. The escalating demand for AVs emphasizes the need to prioritize driving safety, especially in challenging weather conditions like heavy rain. In this context, the accurate perception of environmental features, notably lane markers, becomes imperative for effective autonomous navigation. Severe weather can lead to camera image degradation, including blur and loss of details, impacting the accuracy of subsequent image processing. Despite the prevalence of camera-based methods, sensitivity to environmental noise, such as rain streaks, poses a challenge, necessitating preprocessing mechanisms like rain removal to enhance lane detection accuracy. This chapter focuses on the development of a vision-based algorithm dedicated to detecting and tracking lane markers, coupled with an efficient rain streak removal algorithm. A progressive approach to lane detection on city roads is presented, incorporating sliding windows and Kalman filter methodologies into a model-based method. Integration of the Kalman filter has yielded a notable improvement in video processing speeds, from 1.67 to 2.72 frames/s, enhancing overall operational efficiency. Furthermore, a novel neural network structure, amalgamating convolutional neural networks (CNNs) and long short-term memory (LSTM), is introduced for rain streak removal before performing lane marker detection. Comparative analysis against existing methods demonstrates an average 2.3% improvement in peak signal-to-noise ratio (PSNR) for rain removal and an 8% enhancement in Google Vision test results.

Road information plays a vital role in the development of rural and urban areas. Road information is considered to be an important for providing better connectivity in the areas. Mapping the road can be useful for planning and developing rural and urban areas. Moreover, road extraction from aerial and Unmanned Aerial Vehicle (UAV) imagery has been an active research topic over a few years due to many key applications like Transportation system, Geographic information system (GIS) application, etc. According to the real challenges like complex road structure usually consist of a form of heterogeneity. Some type of obstacles during road class extraction either in the form of shadow occlusion or visual occlusion occurs in satellite images. Dataset preparation is very important and also their relevance with the objective of road extraction. Due to recent advancements in aerial and UAV technology, High-Resolution UAV images are widely available which is quite impactful to extract the road networks. Extracted features from the UAV images are being utilized to detect the road class which utilizes the road properties. Road extraction methods are classified in some major categories viz. traditional methods, computer vision, ensemble methods and deep learning based methods. In the traditional methods, road extraction is evaluated based on feature level, template matching, edge and filter methods. Road monitoring is also an important factor for the transportation system and urban planning. Thus, this chapter aims to provide detail review on extraction and monitoring of the roads from aerial and UAV images using different type of approaches.

Smart parking solutions are transforming urban mobility by utilizing technology to tackle the challenges of parking in crowded cities. These solutions incorporate sensors, data analytics, mobile applications, and automation to optimize parking space usage, alleviate congestion, enhance traffic flow, and improve the overall parking experience for drivers. This chapter explores the use of computer vision in smart parking systems, revolutionizing traditional parking practices. It commences with an examination of the fundamental components, advantages, challenges, and future directions of smart parking solutions. The chapter then delves into the principles of computer vision, elucidating how cameras and image processing algorithms detect and analyze parking spaces, vehicles, and movement patterns. It discusses critical elements of a smart parking system, such as camera positioning, image processing methods, and data integration with parking management systems. Emphasis is placed on the significance of real-time data processing in dynamic parking environments, enabling functionalities such as parking availability alerts and occupancy forecasts. Overall, this chapter offers a comprehensive overview with case studies of how computer vision is reshaping parking management, fostering smarter, more efficient, and sustainable cities.

In the dynamic landscape of urbanization, Smart Cities emerge as a transformative approach to city development. By leveraging technology, they aim to improve citizens’ lives and streamline urban operations. This chapter offers an exploration of AI-powered surveillance for smart cities, delving into foundational concepts, and the pivotal role of computer vision techniques in shaping urban environments. While navigating through the context of urban development, it contrasts traditional approaches with the innovative integration of cutting-edge technologies for surveillance. The chapter focuses on the evolution of surveillance methods, from basic to advanced smart systems, and how these systems can aid in public safety, urban planning, and crisis management in the context of smart cities. Various uses of smart surveillance in traffic management and law enforcement are discussed, highlighting challenges such as privacy and biases and emphasizing the need for ethical guidelines. Cutting-edge technologies like facial recognition and person re-identification systems are analyzed, along with acknowledging the challenges they present. The exploration extends to complex systems like multi-camera person tracking and drone-based surveillance, which can enhance city awareness and the efficiency of surveillance. Finally, it looks ahead to a future of data-driven decision-making and citizen-focused services in smart cities, aiming to drive innovation and sustainable urban development.

The convergence of the Internet of Things (IoT) and Computer Vision (CV) technologies is revolutionizing public safety and Quality of Life (QoL) in smart cities. This chapter explores how data-driven event detection and dispatch management powered by Artificial Intelligence (AI) and optimization can transform public safety. Moving beyond traditional methods like manual reporting, we propose a data-driven approach utilizing IoT sensors (i.e. optical cameras) to capture real-time data. This data feeds into advanced AI models capable of identifying and anticipating potential events, ranging from security threat detection to traffic management. Upon event detection, a robust dispatch management system integrated with AI utilizes optimization algorithms to efficiently allocate resources and route first responders. An example of a specific Automated Event Detection and Dispatch Management System (AEDDMS), known as Remote Area Management Systems (ReAMS), is presented to demonstrate the concept. ReAMS employs AI to monitor remote areas for security threats. While focusing on this example, the chapter emphasizes the broader applicability of this approach to various smart city domains. The chapter also acknowledges the challenges, future work, and societal considerations like infrastructure planning, safety, ethics, and privacy surrounding these smart city solutions. This chapter paves the way for a data and AI-driven future where technology empowers us to build safer, more resilient communities.

Efficient and secure transportation systems are crucial for the development of smart cities. This chapter introduces Total Road Control (TRC), a software solution by COM-IoT Technologies that revolutionizes road monitoring and management. TRC utilizes LiDAR sensors for high-precision data collection and convolutional neural network (CNN) models for comprehensive data analysis, enhancing road safety, traffic flow, and urban planning. LiDAR sensors in TRC provide unparalleled accuracy in diverse environmental conditions, ensuring complete and precise road data acquisition. CNN models facilitate vehicle detection, tracking, counting, and classification, offering real-time traffic insights. TRC’s advanced features include vehicle speed measurement and nuanced classification into categories such as cars, trucks, buses, and motorcycles, with specific subclassifications for trucks based on weight. TRC also measures vehicle dimensions, aiding in road planning and infrastructure development. By analyzing vehicle distances, TRC enhances road safety by detecting tailgating incidents and near-miss scenarios, mitigating collision risks. Additionally, TRC serves as a predictive tool for accident prevention, identifying high-risk areas and underlying factors contributing to road incidents and empowering authorities to implement targeted safety interventions.

The escalating climate crisis necessitates innovative, technology-driven solutions for environmental monitoring and resilience. This chapter explores the transformative potential of integrating Long Range (LoRa) technology with advanced Artificial Intelligence (AI) algorithms to address pressing environmental challenges. LoRa technology’s extensive coverage, low power consumption, and scalability provide a robust foundation for comprehensive environmental sensing. When coupled with AI, these sensors evolve into powerful analytical tools capable of deciphering complex environmental patterns and predicting trends with remarkable precision. We delve into the capabilities of LoRa-enabled devices and their integration with AI-driven analytics, illustrating their application through real-world case studies. These examples demonstrate the effectiveness of LoRa and AI in improving air quality monitoring, flood prediction, and other critical areas, highlighting the tangible benefits for communities and ecosystems. The chapter underscores the importance of precision, timeliness, and comprehensiveness in environmental monitoring, offering a roadmap for achieving climate resilience. This chapter not only contributes to the academic discourse on environmental monitoring but also serves as a call to action for governments, organizations, and communities to adopt and implement these innovative technologies. Through collaborative efforts and the harnessing of LoRa and AI, we can pave the way toward a sustainable and resilient future amidst the challenges of climate change.

This chapter explores the evolution of power consumption scheduling in smart cities, focusing on smart homes and electric vehicle charging. It discusses the transition from classical optimization techniques to heuristic methods like genetic algorithms and swarm optimization for addressing complex energy management problems. The chapter also highlights the growing importance of machine learning, particularly artificial neural networks and reinforcement learning, in predicting energy demand and optimizing scheduling decisions. Additionally, it delves into the potential of generative AI, including generative adversarial networks and large language models, to revolutionize power scheduling by generating realistic scenarios, improving user interaction, and enabling more personalized and efficient energy management strategies.

Integrating Deep Learning (DL) techniques in the Internet of Vehicles (IoV) introduces many security challenges and issues that require thorough examination. This literature review delves into the inherent vulnerabilities and risks associated with DL in IoV systems, shedding light on the multifaceted nature of security threats. Through an extensive analysis of existing research, we explore potential threats posed by DL algorithms, including adversarial attacks, data privacy breaches, and model poisoning. Additionally, we investigate the impact of DL on critical aspects of IoV security, such as intrusion detection, anomaly detection, and secure communication protocols. Our review emphasizes the complexities of ensuring the robustness, reliability, and trustworthiness of DL-based IoV systems, given the dynamic and interconnected nature of vehicular networks. Furthermore, we discuss the need for novel security solutions tailored to address these challenges effectively and enhance the security posture of DL-enabled IoV environments. By offering insights into these critical issues, this chapter aims to stimulate further research, innovation, and collaboration in securing DL techniques within the context of the IoV, thereby fostering a safer and more resilient future for vehicular communication and connectivity.

Intelligent Transport Systems (ITS) represent a burgeoning and transformative concept aimed at reshaping the landscape of mobility both within and outside cities. The Internet of Vehicles (IoV) serves as a networked ecosystem that integrates infrastructure, pedestrians, fog, cloud, and vehicles to enhance the capabilities of ITS. While IoV holds tremendous promise for advancing transportation systems, its networked and data-centric nature raises numerous security concerns. Several solutions have recently been proposed to address these IoV-related challenges; however, many of them involve significant computational overhead and exhibit security flaws. Moreover, there is concern about malicious vehicles infiltrating the network and potentially gaining unauthorized access to services. To tackle these challenges, we present SEATS, a ground breaking solution. The system aims to ensure the secure exchange of information, authentication by both parties, and effective key management among vehicles, roadside units (RSU), and cloud servers. We conduct extensive security and privacy assessments on the proposed approach using the Real-or-Random (ROR) oracle model and Scyther tools, supplemented by an informal security study. The framework is simulated using the Objective Modular Network Testbed in C++ (OMNet++). To demonstrate the efficacy of our approach, we compare it to existing methods, evaluating computation and communication costs.

This chapter explores how the Internet of Things (IoT), Artificial Intelligence (AI), and LPWANS-Long-Range (LoRa) communications, are pivotal in advancing the frameworks of smart cities to foster sustainable urban ecosystems. It highlights how the synergistic application of these technologies significantly enhances connectivity, optimizes the allocation of resources, and improves communication systems within urban environments. Through a comprehensive analysis, the chapter investigates the various challenges that include infrastructure compatibility, data privacy, and the digital divide that impede the deployment of these technologies. Additionally, it explores the transformative effects of integrating AI and IoT on urban operations, from predictive maintenance to anomaly detection, thereby enhancing operational efficiency and public safety. Case studies and real-life examples are discussed to illustrate the successful implementation of these technologies in smart cities, emphasizing their alignment with the United Nations’ Sustainable Development Goals (SDGs) and their contribution to creating resilient, equitable, and thriving urban communities. This discourse not only underscores the benefits but also delves into the ethical and technical considerations necessary for leveraging AI, LoRa, and IoT in the evolution of smart urban landscapes.

Vehicle-to-everything (V2X) communication technology is changing the way we move. It allows vehicles, devices, and infrastructures to interact, overcoming traditional limitations, and enabling smart mobility. V2X technologies aim to enhance road safety, transportation efficiency, energy savings, and driver assistance systems, thus being an important milestone in the development of smart cities. To improve the reliability and efficiency of these technologies, researchers and practitioners are increasingly turning to Deep Reinforcement Learning (DRL). This chapter offers an introduction to DRL in V2X, and its synergy with Federated Learning (FL). It starts by explaining the principles of DRL, where vehicles learn themselves which behavior to follow. A strong focus is put on deep policy gradient and actor-critic methods. These methods are crucial in reinforcement learning and rely on using deep neural networks to find good policies and evaluate them. FL, a collaborative machine learning paradigm that promotes collective learning, is also introduced. The fusion of FL and DRL leads to Federated Deep Reinforcement Learning (FedDRL), offering scalable solutions to modern V2X challenges. Federated Deep Reinforcement Learning (FedDRL) is then applied to the use-case of access point selection for communication in Vehicle-to-Everything (V2X) technologies. These experiments demonstrate the potential of combining Deep Reinforcement Learning (DRL) and Federated Learning (FL) to advance V2X technology. This offers intelligent, adaptable, and collaborative mobility solutions for the future.

In recent years, the demand for high-speed wireless communication has experienced exponential growth due to increasing reliance on mobile devices and the proliferation of high-data-rate applications. To meet this demand, the millimeter wave (mmWave) and terahertz (THz) frequency bands have emerged as a promising solutions, offering significant data transmission capacity and bandwidth. Relying on line-of-sight (LOS) connections to ensure adequate reception power, They play a crucial role in fulfilling the data rate requirements for 5G and future 6G networks. Nevertheless, this necessitates the deployment of large antenna arrays using narrow beams at the transmission as well as the reception antennas to secure reliable received signal strength. Therefore, we are facing two types of problems that must be overcome: On one hand, selecting optimal beams for these large antenna arrays usually requires extensive training, challenging mmWave/THz communication systems to promote high consuming 5G/6G applications, such as connected and intelligent vehicle, video streaming, augmented and virtual reality, etc. Thus, there is a significant need for finding new ways to overcome this beam training effort and realize highly mobile mmWave/THz communication systems. On the other hand, deploying mmWave communication systems presents another challenge which is he susceptibility of mmWave signals to occlusion. Obstacles, such as moving objects like cars, trucks, buses, or even human bodies, can have a significant impact on the performance of mmWave communication systems. These impacts can result in a notable degradation of connection performance, increased disturbances, and reduced available data transmission capacity. This situation can lead to decreased download speeds, increased latency, and an unsatisfactory user experience. This phenomenon poses a significant obstacle to achieving reliable and uninterrupted communication in mmWave networks. In this book chapter, we propose addressing these two issues by utilizing computer vision and artificial intelligence (AI) tools within a multi-source data fusion framework. This framework integrates data collected from various sensors, including cameras, LiDAR, and GPS, to advance smart city transformation.