1. Introduction
Cities are considered complex systems with massive numbers of interconnected citizens, transportation, communication network, varieties of services and businesses, and utilities for improving the lifestyle of urban people. Vast numbers of people are coming towards cities, and the city government is pressured to provide the minimum services required for daily life. The excess population and rapid urbanization bring many problems, such as socio-economic, technical, and organizational problems, and risks to urban cities’ environmental or economic sustainability. Several modern cities faced rapid urbanization by meeting the standards and, in the process, generated pollution, traffic congestion, and socio-economic inequality [
1]. Over the past few years, many individuals have migrated towards urban areas, and predictions indicate that by 2030, 60% of the global population is going to be living in urban settings [
2]. Due to the increment of people, varieties of smart applications are introduced to make life easier, contributing to smart cities’ development [
3,
4,
5,
6]. The smart city concept entails intelligent management of valuable components such as transportation, medical services, utility services, residences, agriculture [
7,
8], and environmental building [
9]. Furthermore, in smart cities, different telecommunication or wireless infrastructures are required to provide the services effectively and connect millions of devices with numerous technologies, such as machine-to-machine (M2M) communication, network virtualization, wireless sensor networks, and gateways [
10,
11].
Figure 1 illustrates the data rate, power consumption, establishment cost, and coverage range of available communication technologies [
12].
IoT plays a dynamic role in the new communication paradigm in our everyday objects, which are equipped with microcontrollers, radio modules, and appropriate communication protocols [
13,
14]. IoT enables IoT devices to communicate and become essential to smart cities. Many national governments and private organizations use the IoT concept in Information and Communication Technologies (ICT) solutions to manage the idea of a smart city [
15]. The IoT concept aims to use community resources better, improving the quality of services (QoS) with reduced operational and management costs in smart cities [
16,
17]. IoT technologies are essential in develo** the landscape of present smart cities and steering the smart city standard to the enormous data scale [
18,
19]. According to Statista Research in 2022 [
20], by 2030, the global count of IoT-enabled devices is projected to exceed 29 billion, nearly three times the figure of 9.7 billion in 2020. The figures mentioned earlier demonstrate that IoT is regarded as one of the most valuable emerging technologies, bringing forth fresh avenues for services, opportunities, and challenges in implementing intelligent applications and offerings. IoT’s significance is intrinsically linked to the advancement and progression of diverse smart city applications, where it serves crucial functions in driving sustainable development and fostering distinctive innovations. Numerous strategies, contexts, technological remedies, and application domains have been put forward to mitigate the complexity and administration of smart cities.
Integrating AI with the IoT in smart cities is a growing trend changing how cities are managed and developed. This integration involves using AI algorithms to analyze the vast data IoT sensors generate in smart cities. This integration also facilitates the advancement of development processes, offering novel opportunities and features, all while significantly reducing human interaction [
21]. IoT sensors are devices embedded in different parts of a city’s infrastructure, such as buildings, bridges, roads, and public spaces. These sensors collect and transmit data on various parameters, such as temperature, humidity, traffic flow, energy consumption, and air quality. In general, the use of IoT devices generates vast quantities of data [
22,
23]. These data subsequently enhance city management and elevate residents’ living standards. However, the sheer volume of data generated by IoT sensors can be overwhelming for human operators to analyze and interpret. AI plays a significant role in this aspect by utilizing machine learning (ML) algorithms to analyze vast amounts of data and detect patterns and trends that might be challenging for humans to discern. Analyzing large amounts of complex data can be challenging to accurately determine the most precise and effective course of action [
24,
25]. One key area where AI can enhance IoT in smart cities is predictive maintenance. By analyzing data from IoT sensors, AI algorithms can predict when maintenance is needed for city infrastructure, such as bridges and buildings, before a failure occurs. This can help prevent costly and dangerous failures and ensure the city’s infrastructure is always in good condition. AI has enabled various applications such as smart water supply, energy management, waste management, and mitigating pedestrian and traffic congestion, noise, and environmental pollution. Most smart city programs and technologies have focused on collecting large amounts of data and creating solutions that tackle the complexities and dynamics of specific applications [
26,
27]. AI-powered applications enable the utilization of large volumes of data and knowledge to facilitate decision-making. Around 30% of smart city applications are now significantly integrating AI to enhance urban sustainability, resilience, social welfare, and vitality, including urban transportation solutions. This trend is expected to continue, with an anticipated increase in AI-powered smart city initiatives by 2025 [
28]. The rapid expansion of AI-based smart city concepts can be attributed to the relentless drive of researchers, government officials, and urban dwellers to explore new information and methods for building smart cities. The proliferation of automation and AI is becoming increasingly inevitable, with growing demand for their implementation in smart city development [
22]. Smart sensor nodes [
29] generate large amounts of data associated with various smart city applications and are significantly under-used. Existing Information and Communication Technology (ICT) [
30] infrastructure can generate heterogeneous information, which is essential for consolidation. Finally, AI can improve public safety by analyzing data from security cameras, microphones, and other sensors to detect and prevent crime and other security threats. While integrating AI with IoT in smart cities can revolutionize urban development and management, it also raises concerns about privacy, data security, and potential biases in AI algorithms. As such, smart cities must implement strong ethical and regulatory frameworks to ensure these technologies are used responsibly and transparently. The potential threat and concern are not the discussion point of the current review paper.
Figure 2 illustrates the communication technologies that are currently available and can be utilized in various applications within a smart city.
IoT-based technologies combined with AI in a smart city can be viewed from various perspectives, including smart sensors, communication technologies, and multiple applications. Among the different technologies, the most critical IoT technologies entail the deployment of reliable and robust networking and communication infrastructures to facilitate efficient data and information exchange among the diverse components of smart city services. Smart city IoT services are designed at varying scales, depending on the type of application. They may require diverse networking and communication technologies for their implementation and operation. Several review papers [
16,
25,
31,
32,
33,
34,
35,
36,
37,
38] have previously examined various issues related to big data, network security, and the potential of IoT in smart cities. However, there is a scarcity of resources that comprehensively cover the combination of IoT technologies and the contribution of AI to the smart city concept. This review article discusses the concept of smart cities and the role of IoT in develo** them, explaining how IoT enables devices to communicate and become essential components of smart cities. It also highlights how integrating AI and IoT in smart cities can facilitate the advancement of development processes, offering novel opportunities and features while reducing human interaction. The article delves into the potential of AI in analyzing the vast amount of data produced by IoT sensors in smart cities. It highlights how such technology can enhance city administration and ultimately lead to better living standards for residents.
The article is structured as follows:
Section 1 is going to provide an introduction to the topic of smart cities, highlighting their significance and potential benefits.
Section 2 is going to explain the methodology of searching the relevant article.
Section 3 is going to define the concept of smart cities and discuss their essential characteristics and components.
Section 4 is going to delve into the IoT-based technologies necessary for creating smart cities, including sensors, networks, and data analytics.
Section 5 is going to explore the various AI algorithms suitable for smart cities and discuss their potential impact on urban life.
Section 6 is going to examine the possible future trends of smart cities, such as the integration of 5G communication technology and the impact of AI on various applications. Finally, in
Section 7, we are going to provide conclusions and prospects for the future of smart cities. Overall, this review article aims to provide a comprehensive overview of the current state and future directions of smart city development, highlighting the potential benefits of this emerging field for urban residents, governments, and businesses alike.
3. The Smart City Paradigm
There exist multiple definitions of the smart city [
15,
39,
40,
41]. The term “smart city” is often used interchangeably with other terms such as “intelligent city”, “knowledgeable city”, and “digital city”. This has resulted in diverse theoretical choices for the smart city concept [
4]. The term “smart city” is an ambiguous concept used in multiple ways. There is no specific template for defining the smart city, nor is there a unique definition that is universally accepted [
42,
43]. Smart city concepts emerged as a response to the challenges posed by urbanization and the need for sustainable urban development. Integrating various technologies, such as the IoT, AI, and big data analytics, aims to improve the efficiency of urban services, reduce resource consumption, and elevate the standard of living, which can lead to an improved quality of life for citizens. Technological development has advanced so much that a modern smart city can collect data from various applications, such as smart agriculture, smart industry, smart farming, smart health care, smart traffic, and smart pedestrians, and then analyze and integrate critical data to provide a decision for an improved standard of life [
44]. The rapid development of the IoT [
45,
46], cloud computing technologies [
47,
48], and AI [
49,
50] could be vital factors in improving urban facilities’ performance, quality, and interactivity, which is going to decrease management costs and enhance skills.
While the definition of a smart city can vary, there are generally two distinct approaches based on the key areas a smart city should focus on. In the first trend, smart city definitions emphasize a single urban feature, such as technology or ecology, while leaving the rest of the features of a city included. The definitions do not consider the primary aim of a smart city, which is to introduce a novel framework for managing urban areas and assessing the interconnectedness of various components within an urban ecosystem. Focusing solely on the development of a single aspect of an urban system does not necessarily imply that the problems of the entire system can be overcome. It is important to consider the interconnectedness of all urban characteristics to tackle modern cities’ challenges and ensure a thorough and effective approach [
51,
52,
53]. Under a different set of definitions, the focus is on how the smart-city concept represents a comprehensive approach that considers all urban features integral parts of a system. This approach emphasizes that the smart city is a new methodology for organizing and develo** a city, which considers economic, technological, and social factors to ensure the stability and sustainability of the urban system as a whole. The descriptions suggest a comprehensive approach to urban issues that leverages new technologies, allowing for a redefinition of the smart city model and the relationships between its stakeholders [
54,
55,
56,
57,
58].
Table 1 summarizes the definitions of the smart city.
The characteristics of a smart city can be defined as follows:
The integration of advanced technologies such as IoT, AI, and big data analytics to enhance sustainability, efficiency, and the overall well-being of citizens.
Implementing renewable energy sources, eco-friendly buildings, intelligent transportation systems, electric vehicles, and efficient traffic management. Moreover, the platform provides insights into various aspects of a smart city, including energy management, smart homes, optimized transportation, smart grid systems, water monitoring, waste management, and streamlined administration. These initiatives aim to monitor existing infrastructure and enhance the quality of life for urban residents.
Emphasis on citizen engagement and active participation through open data initiatives and digital platforms, fostering transparency and collaboration between citizens and the government.
4. The Architecture of IoT-Enabled Smart City
The advancement in technologies could bring millions of intelligent sensors to instrument the infrastructure of the cities, working with speed and network structure equivalent to 1000 IoT devices. Effective management of the data collected from sensors is another crucial aspect. This data’s availability and intelligent management play a vital role in the operation of complex smart city ecosystems, ensuring the accurate functioning of city services. A classification based on the application type is essential to establish a solid foundation for smart cities. As previously mentioned, the degree of smartness in various areas such as governance, economy, healthcare, and others must be evaluated within the context of the smart city application.
The environment of a smart city is characterized by advanced communication technologies and an IoT infrastructure, which requires specific communication technologies and systems architecture to enable all the applications and advantages of smart cities. The architecture of IoT-enabled smart cities typically requires an ICT infrastructure that facilitates information exchange among the various stakeholders within the urban environment, regardless of the specific application or service being used. Communication is essential to transmit data generated from various sensors in different applications between devices and information sinks in both directions. To accomplish this, three commonly used communication patterns are utilized: (i) utilizing Cellular Mobile Networks for communication, (ii) using IoT-Dedicated Cellular Networks for communication, and (iii) employing Multi-Tier Networks for communication.
Figure 5 illustrates the main layout for the three different architectures. Therefore, to achieve a proper architectural ground for smart cities, classification based on the type of application is necessary, as the smartness of entities such as governance, economy, healthcare, etc., must be measured in the smart city application.
The following sections consist of the architectural components with the leading communication technologies available in all three architectural classes in a smart city scenario.
4.1. Architectural Components
The literature has discussed five or three components of IoT [
32,
66,
67,
68]. However, four major components, such as data sensing/actuating, networking and communication components, vital components, and cloud/fog computing for various services and applications, can be summarized and discussed in the following sections.
4.2. Perception or Sensing Layers Components
The components referred to here are the physical objects of electronic devices such as sensors and actuators. These devices interact with the physical world by sending and receiving data, utilizing wireless networks [
6,
7,
8,
69,
70]. Various smart city applications include sensors, actuators, and smart device technologies in this context. There is a vast array of commercial devices available that can measure different physical quantities and external variables, i.e., sensors for measuring humidity [
71,
72], temperature [
73], environment monitoring [
74], water nutrient monitoring [
75], etc. Actuators are used for controlling/moving other objects or systems through physical interaction or virtual control [
76]. They are usually categorized into pneumatic, electrical, and hydraulic categories [
77]. Several software applications and solutions can be used for deploying low-level IoT applications [
78].
4.3. Networking and Communication
In the context of smart cities, various communication protocols are available to connect different devices and components of the infrastructure. These protocols have additional features and specifications that make them suitable for specific applications. For example, some protocols such as Wi-Fi, ZigBee, and Z-Wave are ideal for short distances where devices and their coverage areas are limited, while other protocols such as LoRaWAN, NB-IoT, Sigfox, and Long-Term Evolution (LTE) are more suitable for long-range applications. Each of these protocols has unique features that enable them to support various smart city applications. For instance, ZigBee is commonly used for low-power applications with a low data rate, offering a secure network and longer battery life. It also provides network topologies suitable for various applications, such as mesh, star, and tree. On the other hand, LoRaWAN, Narrowband-Internet of Things (NB-IoT), Sigfox, and LTE require devices to be connected through a central gateway to collect information and sink the data, and they operate on licensed and unlicensed spectrums, depending on the applications. Therefore, it is essential to choose the appropriate communication protocol based on the specific requirements of each application and the range of coverage needed. Understanding the features and capabilities of different protocols can help smart city planners and developers choose the most suitable protocol for their specific use case and integrate various IoT-enabled devices and artificial intelligence algorithms to create a more efficient, sustainable, and livable city. The subsequent sections are going to discuss some of the significant protocols.
4.3.1. Evolutions of GSM and LTE
The cellular network working groups handle the functions of administration of GSM/GPRS and Edge Radio Access Networks (GERAN). Their efforts are focused on introducing alternative methods to enhance the efficiency of GSM/GPRS for M2M communication. Various methods are available to improve M2M (machine-to-machine) communication efficiency. These approaches comprise upgrading the uplink capacity, extending the coverage for control and data channels, and reducing the power consumption and complexity of M2M devices [
79]. Extended Coverage GSM (EC-GSM) is another progressive technique that utilizes Frequency Division Multiple Access (FDMA) in conjunction with Code Division Multiple Access (CDMA) on the uplink to accommodate a more significant number of IoT devices transmitting on the same frequency band. Bind repetition is achieved for all transport channels and extended the coverage regions. It is a technique where the transmitter repeats the same data block several times to achieve higher receiving gains. Different devices are used for different repetition levels in the blind repetition technique. Another approach is to support the narrowband GSM spectrum with channelization of 200 kHz. Narrowband Cellular IoT (NB-IoT) accommodates narrow asymmetric bands in downlink and uplink, where 200 kHz is divided into 48 narrowband channels are downlink channels, and 36 narrowband channels are divided into uplink channels. The Orthogonal Frequency Domain Multiple Access (OFDMA) is employed for the downlink, while the uplink uses Frequency Division Multiple Access (FDMA) to comply with the specific requirements of each device.
One of the evolutions of LTE is LTE Rel-11 [
80], which has focused on improving the overload functionalities of RAN for handling the access of many IoT devices with reduced capabilities. The complexity and the cost are reduced by reducing the radio transceiver and limiting the transport block sizes by allowing half-duplex FDD during the operation mode. Moreover, device power saving was also introduced, a significant advantage for IoT applications. Another substantial evolution is the LTE Rel-13 [
80] which is improved at the physical layer by including the characteristics of narrowband transmission and provides a better response for M2M requirements. EC-GSM enables the use of cost-effective and low-power hardware while simultaneously enhancing coverage areas. Additionally, Rel-12 has undergone further enhancements and now features an Enhanced Power Saving mode (EPS) and Extended Discontinuous Reception (DRX) functionality.
Cimmino et al. have reported a framework for various applications from the perception of different communication technologies in a smart city scenario [
81]. LTE technology has concentrated on leveraging small-cell technology to augment bandwidth within this framework. Small cell technology allows for inexpensive and low-power intelligent devices for various applications. LTE service can be deployed over a larger geographical area and can fulfil all communication necessities, including interoperability, low power consumption, resilient communication, and multi-modal access, which can enhance the overall quality of user experience. Thus, LTE service can be utilized for many applications and may be a superior option for smart cities.
4.3.2. Cellular Mobile Networks
Cellular mobile networks play a significant role in facilitating various services in smart cities. These networks are primarily designed to enable communication between human-to-human and human-to-machine interactions, such as telephony, texting services, multimedia downloading, and streaming [
81]. The devices ecosystem also encompasses environment monitoring devices that gather data and require data exchange capabilities with the backend to receive feedback on the monitored data.
The Radio Access Network (RAN) operates on a licensed network, while the Core Network (CN) includes various entities, such as user mobility and registration. For instance, the CN of the LTE system consists of several components, including the Serving Gateway (SGW), the Mobility Management Entity (MME), the Packet Data Network Gateway (PGW), the Home Subscriber Server (HSS), and the Policy and Charging Rules Function (PCRF) server [
82]. In smart cities, the Internet of Things (IoT) has ushered in a new communication paradigm known as Machine-to-Machine (M2M) or Machine-Type Communications (MTCs). This paradigm relies on minimal or no human interaction, enabling devices to communicate directly with each other and exchange information without human intervention [
83,
84].
The characterization of M2M communications are featured distinctively concerning the “legacy” of human-to-human communications, and the end-user human is often replaced by a more significant number of devices, accessing the network periodically or irregularly [
85,
86,
87]. Although 2G/3G (GSM, GPRS) cellular technologies are used for most M2M communications, the RAN and CN of mobile cellular networks face challenges due to the massive M2M data traffic [
88]. Hence, extensive research is being conducted to effectively enhance cellular architectures to address the obstacles in M2M communications [
87].
As per the guidelines set by the Third Generation Partnership Project (3GPP) [
89], the main distinctive features of M2M concerning human-based communication include the following [
80]:
The various market fields, such as smart grid or smart metering, environmental monitoring, and crowd monitoring applications, require different market scenarios to support their diverse applications.
The end-user device should lower costs with powerful capabilities of handling energy efficiently.
The number of communication terminals should be large enough to accommodate many devices.
Every terminal should be able to efficiently handle traffic of varying sizes, ranging from small to large, that originates from the field devices or active smart devices and is transmitted to the network.
In the IoT-based cellular ecosystems, the end-users play an essential role, and the main issues from the user’s side are dealing with cost reduction and network assistive power-saving functionalities, which can improve the lifetime of the devices. From a network perspective, the most critical issues are improving the coverage area and defining lightweight data handling procedures for M2M devices to avoid data overloading difficulties [
90]. It is imperative to take action to tackle traffic congestion in both the radio and core network layers [
55] and manage the interplay between M2M communication by effectively allocating radio resources. The GSM [
91] and LTE standards [
92] are some of the initiatives launched by 3GPP to address these challenges [
93]. An outline of the primary characteristics of upcoming standardization endeavors in cellular IoT is provided in
Table 2.
4.3.3. IoT-Dedicated Cellular Networks
This cellular network tries to design low-cost, low-energy, and low-traffic congestion IoT devices with reduced traffic requirements. IoT-dedicated cellular networks provide many advantages, including reduced power consumption and Total Cost of Ownership (TCO) compared to traditional cellular networks, as well as worldwide coverage and hassle-free plug-and-play connectivity. The network architecture follows a common star topology with extensive coverage areas and can accommodate numerous dedicated devices for applications such as smart lighting, smoke alarm systems, environmental monitoring, smart parking systems, and smart pedestrian counting. These applications may not be feasible with traditional GPRS/GSM networks due to their higher costs, subscription fees, and power consumption [
96]. The key distinguishing factors among IoT-specific cellular networks are their bandwidth, ability to handle bi-directional traffic, and network availability.
Table 3 provides a comparison of the primary technologies and their respective features.
Sigfox Protocol
Sigfox [
97] adopts an ultra-narrowband (UNB) technology where devices communicate directly to the base station in a star-like topology. Multiple base stations are available for data collection, making it a reliable network. Most devices are used for uplink communication, and tiny downlink channels control the device from gateways. Sigfox network supports a message payload capacity of up to 12 bytes and employs a technique where the same payload message can be transmitted multiple times for robust reception. The frequency of messages a device can transmit daily can be customized based on the application’s specific requirements. The network coverage of Sigfox is estimated to reach up to 10 km in urban areas and 50 km in rural areas. Building, installing, and maintaining the low-powered network lies with Sigfox Network Operators (SNO). They are also responsible for business development and maintaining/upgrading the reference architecture if necessary [
97].
Weightless Protocol
The weightless system is controlled by the Weightless Special Interest Group (SIG) [
98], a nonprofit organization for implementing the standard architecture. The reference architecture of the weightless system is illustrated in
Figure 6. It includes the uplink and downlink, which are part of the physical layer. The uplink transmission adopts FDMA to enable concurrent transmissions and prevent interference. The base station also uses a time scheduling method to notify the devices of their time allocation.
The data link layer comprises three sub-layers: the baseband, lower link layer, and upper link layer. The management of radio resources at the Medium Access Control (MAC) layer and configuration of network arrangements are handled by the Radio Resource Manager (RRM). Additionally, the system supports transmission acknowledgement, fragmentation, and multi-cast and can interrupt the base station.
Low Power, Wide Area Networking (LoRaWAN)
The LoRa Alliance [
1] has developed a standard for connecting low-power IoT devices covering extensive areas. The standard uses different ISM bands according to the regulations in specific regions. Various ISM bands comply with specific regional regulations in the standard. The physical layer incorporates Chirp Spread Spectrum (CSS) modulation, facilitating two-way communication between base stations/gateways and end devices [
41]. Three modes are available in the Medium Access Control, enabling access control while balancing the uplink and downlink capabilities.
Figure 7 illustrates the reference architecture of the LoRaWAN. The Class A device requires low power and limited downlink channels for communication from the base station, and Class B shares the same characteristics. However, they have extra time windows not available on Class A devices. Class C device has continuous receiving time windows. It can support multicasts, such as Class A and Class B. During communication between the base station and the end devices, all classes of devices in LoRaWAN utilize the ALOHA access protocol.
4.4. Communications in Multi-Tier Architectures
Multi-tier architectures [
99] rely on a layered approach to collect and process sensing data and establish a network infrastructure in a multi-hop manner. The sensor data are typically transmitted to central gateways, which then transmit the data to the internet using various communication technologies. The multi-hop transmission technique [
100,
101] compensates for the limited communication ranges by introducing various radio technologies. This technique also helps to use an extremely low-powered device, which is essential for specific applications. On the other hand, a number of devices can be utilized due to the limited radio technologies needed to ensure connectivity.
Table 4 compares several short-range communication standards for their use in multi-tier architectures [
102]. Such architectures are based on a layered design and utilize multi-hop routing to establish a network infrastructure.
The IEEE 802.15.4 standard [
103] is designed for low-cost, low-data rate, and low-power consumption personal wireless networks. The protocol stack comprises PHY and MAC layers, and it supports operation on three unlicensed frequency bands, namely 868 MHz, 915 MHz, and 2.4 GHz. The standard uses the direct sequence spread spectrum (DSSS) modulation scheme, enabling 20, 40, and 250 kbps data rates on the three frequency bands. Over time, the IEEE 802.15.4 standard [
103] has undergone various enhancements to improve its performance, especially in the PHY layers. The MAC layers employ carrier-sense multiple access with collision avoidance (CSMA-CA) mechanisms to facilitate channel access, synchronization, PAN association, and beacon transmission of devices.
Numerous solutions have emerged from the IEEE 802.15.4 standard [
103], such as ZigBee, 6LoWPAN, WI-SUN, and ULP, which can be used for diverse applications in personal wireless networks. The Z-Wave Alliance developed the Z-Wave protocol stack for home automation applications. It utilizes Gaussian frequency-shift keying (GFSK) modulation and mesh network architecture for routing and security. The Z-Wave protocol stack defines all the available layers, and one company operates it to enable interoperability between companies. Variable data rates of 9.6/40/100 kbps are possible, and its communication range is similar to IEEE 802.15.4-based solutions.
The Metering Bus (M-Bus) is a specialized bus-based transmission technique for metering utility services such as gas, electricity, and water. It is essential for develo** smart utility services and can operate in three frequency bands (169/433/868 MHz) with star or mesh network topologies. It employs synchronized time-division multiple access (TDMA) routing protocols and can communicate up to a range of 300 m with high data rates of 100 kbps.
Bluetooth Low Energy and Wi-Fi Low Power
Bluetooth Low Energy (BLE) has recently been very popular in audio streaming and is suitable for IoT applications. It is a low-powered protocol stack where the BLE devices consume low energy and can be used for a long time. The first modulation scheme is GFSK, with a 1 Mbps data rate. There are 40 different channels; 3 are used for advertising, and 37 are data channels. The BLE can integrate with IPv6 and supports the packet fragmentation technique with basic security features. BLE’s major disadvantage is that it only supports mesh topologies, which limits the applications to real-life scenarios. Wi-Fi Low Power increases Wi-Fi range by using less power and operating in the 900 MHz frequency spectrum. As a result, it is appropriate for IoT applications. The data rate is increased from 150 kbps to 8 Mbps by using the OFDM waveform scheme in conjunction with Binary Phase Shift Keying (BPSK), Quadrature Phase-Shift Keying (QPSK), and Quadrature Amplitude Modulation (QAM) modulation. It is also possible to add more devices by using the MAC layer. It is a better protocol for various applications because it has a power conservation mode, a data optimization method, and an extended slee** state.
Khorov et al. [
104] have reported a work that is based on Wi-Fi technology (IEEE 802.11 [
105]). They found that Wi-Fi technology [
106] has been extensively exploited in applications for smart cities. Due to the abundance of gadgets connecting, there is a severe interference issue. An anti-interference mechanism is essential to reduce the interference effect, and Wi-Fi can be more useful in IoT-based smart cities [
107]. Duarte et al. [
108] have reported a mechanism to measure Wi-Fi signal strength and developed smart medical health care for smart city applications. Every patient in a medical hospital carries a smartphone with Wi-Fi connectivity. The number of patients waiting in the waiting area, where they are located, and their real-time movements are all determined by the signal strength data gathered from their cell phones. The system’s main benefit is that it can be implemented in the current infrastructure without any new infrastructure. The information gathered can be used to spot future medical problems, which is helpful for doctors.
6. Future Trends of Smart City
To manage different aspects of urban life, including but not limited to traffic, energy consumption, waste management, and public safety, a smart city requires extensive use of IoT devices for monitoring and control. These devices generate massive amounts of data that must be transmitted over the internet to the central control system for processing and analysis. Therefore, a gigantic IoT connection is essential for successfully implementing a smart city. Implementing a 5G-based [
216] IoT network is a potential solution for managing a massive network in a smart city scenario. This network [
217,
218], based on the wireless software-defined networking (WSDN) paradigm, offers a flexible and rapid network infrastructure to facilitate the deployment of IoT-based smart city networks [
219]. It has increased speed and bandwidth where the data transfer rate reaches up to 20 Gbps. The 5G network can manage the enormous volume of data IoT devices generate. Its higher bandwidth can accommodate more connected devices, enabling various new applications for enhancing urban life in smart cities.
Additionally, 5G networks offer low latency and improved reliability, making them desirable for powering IoT-based smart city networks. With latency as low as 1 ms, 5G technology can enable the development of more real-time applications. The reduced latency ensures minimal delay between IoT devices, leading to faster and more efficient communication. This feature is crucial for develo** applications such as smart drone transport, autonomous vehicles, real-time remote surgery, or patient monitoring for smart healthcare in a smart city. Devices with lower latency can send and receive data much more quickly, enabling a range of innovative and time-sensitive use cases. Compared to previous generations (i.e., 2G, 3G, or 4G), 5G networks are considered more reliable, offering features such as network slicing and multi-access edge computing (MEC). These features ensure consistent and uninterrupted connectivity for IoT devices in congested areas of a smart city. As the development of smart cities continues, 5G networks are expected to be increasingly used due to their ability to provide enhanced security features, such as secure boot, network slicing, and secure element, which can help prevent cyber-attacks and data breaches, ensuring the security of IoT devices and the massive data they generate (as depicted in
Figure 8).
The impact of 5G on connecting billions of devices in a single communication technology is significant, especially in creating a massive IoT-enabled smart city. To achieve this, smart devices must be able to interact with each other and share data without the need for human intervention [
220]. It is also vital for these smart devices to support real-time, demand-based events that can be coordinated between end-to-end devices. Additionally, they should be equipped with automatic and intelligent algorithms that can operate seamlessly during each phase of the smart city development, from planning to deployment and maintenance. This is going to ensure that the smart city infrastructure is efficient, scalable, and can adapt to the changing needs and demands of the citizens [
221]. IoT-enabled 5G architectures are designed to provide a common platform that can fulfil the following requirements:
5G networks should be adaptable and independent based on the specific needs of the applications.
Cloud-based Radio Access Network (CloudRAN) should be used to support massive connections of devices for various applications (as shown in
Figure 6).
The core network architecture should be easy to implement and provide on-demand functions for network configurations.
In the future, smart cities are going to rely heavily on 5G communications and advanced IoT technologies, which involve storing and uploading massive amounts of data to an IoT cloud server to extract useful information using data analysis methods. These smart cities are going to incorporate various aspects such as smart traffic, smart homes, smart agriculture, smart grids, and smart management. 5G-enabled IoT networks are going to provide multiple features, including direct device-to-device communication, typical architecture, advanced spectrum sharing, and interference management, which are going to be vital in building a complete smart city. Integrating AI with 5G networks is going to be integral to the future smart city. AI is going to handle massive amounts of data from IoT-based applications, such as public transportation, traffic lights, and citizens’ daily activities, for more accurate and precise decision-making. This is going to enable finding insights into data patterns that can be used to boost the efficiency and productivity of municipal government operations while reducing related expenses. Additionally, AI is going to help reduce human errors, make automated data-driven decisions, implement efficient urban management, and explore new commercial possibilities.
The role of AI in a smart city is to manage and analyze the vast amounts of data generated by IoT devices with accuracy and precision to aid decision-making. For example, public transportation, traffic lights, and citizens’ daily activities generate enormous amounts of data, and AI can identify insights that can increase the efficiency and productivity of municipal government operations while reducing costs. Furthermore, AI can minimize human errors, make automated data-driven decisions, implement efficient urban management, and explore new commercial possibilities. Using AI and 5G technology, energy providers can optimize renewable energy sources such as solar and wind power. AI can predict when the sun is going to shine or the wind is going to blow, and 5G can instantly communicate this information to energy providers, allowing them to adjust their energy supply accordingly. AI and 5G can also help reduce traffic congestion and emissions in the transportation sector by optimizing traffic flow and identifying the most efficient vehicle routes. This can decrease time spent on the road, which in turn can reduce emissions and save energy. Integrating AI and 5G technology can revolutionize how we manage energy and transportation, reducing carbon emissions and moving us towards a more sustainable future.
In conclusion, the success of a smart city relies heavily on its ability to integrate cutting-edge technologies such as 5G communication, IoT devices, and AI algorithms into its infrastructure. By working together, these technologies can create a more efficient, sustainable, and livable city. By integrating AI algorithms with renewable energy sources such as solar and wind power, cities can optimize energy usage and reduce carbon emissions, making the city more environmentally sustainable while reducing energy costs for residents and businesses. By leveraging these technologies, smart cities can become more efficient, productive, and ecologically sustainable, ultimately enhancing the quality of life for citizens. By reducing congestion, improving air quality, and optimizing the use of resources, smart cities can create a better living environment for all, making them an exciting prospect for the future of urban development.
Figure 9 illustrates a futuristic smart city that combines IoT, 5G, and AI.
7. Conclusions
Communication technologies are essential to ensure reliable and continuous connectivity in smart cities. However, the proliferation of smart devices can present difficulties. Therefore, current research explores the potential of communication technologies enabled by the Internet of Things (IoT) and their possible applications in smart cities. Wi-Fi, ZigBee, and Z-Wave are suitable for short distances where devices and their coverage areas are limited, as they offer higher data throughput rates than long-distance technologies such as LoRaWAN. However, the choice of communication technology depends on the specific requirements of each application and the range of coverage needed. ZigBee is commonly used for low-power applications with a low data rate, providing a secure network and longer battery life. It offers network topologies suitable for various applications, such as mesh, star, and tree. On the other hand, some long-range technologies require devices to be connected through a central gateway to collect information and sink the data. These technologies include LoRaWAN, NB-IoT, Sigfox, and LTE, and they operate on licensed and unlicensed spectrums, depending on the applications. When choosing a communication technology, not only the low-power spectrum is essential, but scalability, network capacity, security, and regularity must also be considered.
Cellular networks can be a valuable resource for deploying new applications in a smart city. Narrowband IoT (NB-IoT) is a technology that focuses on low-cost devices while ensuring high network security and reliability. It also takes advantage of existing cellular networks, which can reduce the cost of installing a network for new smart city applications. On the other hand, LoRaWAN and Sigfox are unlicensed spectrum technologies that can help overcome the cost issues associated with licensed technologies. Depending on the application and region, these technologies typically use ISM bands, including 2.4 GHz, 868 MHz, 915 MHz, and 433 MHz. Sigfox is particularly useful for low-data applications such as smart environments, smart agriculture, and smart retailing. It claims to connect a million devices within a large coverage area.
Utilizing several AI algorithms in smart cities can enhance efficiency, sustainability, and quality of life. The algorithms discussed include ML, DL, NLP, CV, and RL. ML algorithms enable machines to identify patterns, learn from data, and make predictions based on input. On the other hand, DL models utilize artificial neural networks to understand hierarchical representations of data, allowing them to detect complex features and patterns. NLP algorithms enable computers to comprehend, interpret, and create human language, while CV algorithms enable machines to examine and understand visual data, recognize image patterns and features, and leverage this information to make informed decisions and predictions. RL is a machine learning algorithm in which an agent interacts with an environment through trial and error to maximize a cumulative reward. These algorithms can be used in various applications in smart cities, such as traffic management, energy optimization, waste management, and public safety. They can make smart cities more efficient, sustainable, and safer for residents.
The future of the smart city is going to depend on the development of every sector, such as a smart environment, smart agriculture, smart economy, smart business, and smart governance. The ability to incorporate the latest 5G communication technology, IoT devices, and AI algorithms into a city’s infrastructure is going to play a crucial role in the success of a smart city. Integrating these technologies can harmonize to produce a more efficient, sustainable, and livable city. Combining all the sectors can help build a smart city; this review explains the requirements. The available data transfer standard developed for IoT networks is not compatible. Therefore, lots of work must be conducted to enable intercommunication between the sensor nodes using different communication protocols while operating under low power. Another area the researchers or stakeholders must focus on is develo** efficient storage techniques and low-power hardware design to reduce operational costs. Heterogenous networks for individual applications should be processed in one giant smart city network, and 5G can play a vital role in future smart city concepts. AI also has an enormous possibility for future work, including develo** data fusion techniques for making heterogeneous data sources more accessible and intelligent data reduction for ensuring that surplus or ’uninteresting’ data are not part of the AI development pipeline. In summary, the current communication technologies are inadequate in providing uninterrupted connectivity in smart cities as they were initially created to accommodate a restricted number of devices and possess limited communication capabilities. Thus, there is a pressing need to develop intelligent and standardized protocols, such as the Internet of Things enabled by 5G technology, to cater to the needs of future smart cities effectively.
The current review paper has not thoroughly examined the ethical, social, and political ramifications of integrating AI with IoT in smart cities, including its potential impact on privacy, security, and social justice. Additionally, it lacks a comprehensive critical analysis of the advantages and drawbacks of this integration, leaving unanswered questions regarding the long-term sustainability, scalability, and efficacy of such approaches in smart cities.