Abstract
The superposition of various application data streams in smart cities can intensify the load of vehicular networks in intelligent transportation systems, which can have an impact on the popularity of smart cities. To improve the performance of large amount of data transmission in telematics, this study proposes a scheme to determine the network state using congestion parameters and routing parameters and matches different data transmission amounts according to different states of the network. The scheme first considers the possibility of the network congestion state. Once congestion is judged to occur, the amount of data sent is reduced, and reducing the data backlog can further mitigate the possibility of congestion formation on the network. Secondly, after rejecting the possibility of congestion, the routing situation of the network needs to be judged, in such cases, whether a change in vehicles in the multi-hop path leads to a path change or an interruption of the data transmission path. Congestion parameters and routing parameters evaluate the state of the network, and the size of the congestion window is appropriately limited by the routing parameters to prevent excessive data volume from causing backlogs in the vehicular network. Experimental simulations show that the proposed scheme exhibits good performance in both linear and crossover vehicular networking scenarios. The research results provide a useful reference for the data transmission of telematics in smart cities.
1. Introduction
The smart city has become the central vision of current urban development. It includes access to customer needs through data collected from a variety of interconnected sensors, devices, and people. The analysis of data is used to improve the efficiency of people's lives in the city and can even solve urban problems such as transportation and network connectivity. The focus of smart cities is based on the operability of service models and connected scenarios [1]. Such services are expected to reduce capital and infrastructure costs while improving the efficiency of service delivery within the smart city framework and enabling customers to use applications remotely from anywhere in the world via the Internet of things (IoT).
In smart cities, intelligent transportation systems carry a large number of data interactions, but the path of their transmission changes dynamically, which can lead to many problems in the transmission of data. Vehicular ad hoc network (VANET) plays an important role in the intelligent transportation system, and the characteristics of its bearer data transmission are mainly reflected in the large capacity, and multiple applications and applications of wireless networks share the transmission protocol [2–4]. This puts new requirements on the transmission of VANET bearer data, whose protocol must have high compatibility and be able to avoid congestion caused by large-capacity data. The intelligent transportation system is designed to support the smart city vision, and the IoT can be applied to the intelligent transportation system and smart city to form an advanced platform for new applications; however, various issues and challenges have emerged. One of the main problems is the data backlog formed by the overlapping transmission of various data, and this backlog is very likely to cause network congestion; in addition, data are faced with random data loss in the IoT of smart cities, coupled with the fact that changes in the location of vehicles in the path can lead to constant changes in the transmission path, which makes it more difficult to effectively control the transmission of data. [5–7].
Based on this, researchers have conducted a lot of research on the data transmission of the vehicle network formed by intelligent transportation in smart cities [8–10]. However, due to the superposition of various application data in smart cities are very easy to cause backlog loss; in addition, different network states will lead to the sending of data that cannot accurately match the sending volume, resulting in the problem of large data transmission fluctuations and low rates. Therefore, how to accurately determine the state of the vehicular network in smart cities and adjust the data transmission according to different states becomes an urgent problem to be solved.
2. Related Work
Intelligent transportation system occupies a very key position in the construction and application of smart cities, and researchers have conducted a lot of research work on traffic system and data processing and transmission in smart cities.
2.1. Intelligent Transportation System for Smart Cities
Xiong et al. [11] briefly reviewed the progress of intelligent transportation system (ITS) research and discussed the problems encountered in the development of ITS, thus providing a useful reference for subsequent researchers. With the widespread deployment and application of 5G, the use of 5G to support intelligent transportation systems has become a hot research topic. Data are collected through various sensors using 5G for communication, which supports the necessary communication infrastructure required for smart cities. Gohar et al. [12] discuss the technical support of 5G for ITS and key issues for subsequent research in several dimensions based on the characteristics of smart cities and intelligent transportation systems. The VANET suffers from many shortcomings similar to wireless multi-hop networks, such as intermittent connection interruptions, high bit error rates, and intense data access competition. However, VANET has been successfully used in intelligent transportation systems to realize many applications in smart cities. Considering that UAVs have the characteristics of line-of-sight communication, load balancing, flexibility, and cost control, Raza et al. [13] proposed a UAV-assisted vehicular self-organizing network architecture for smart cities, and the test results showed that UAV-assisted vehicular networking can effectively improve the robustness of network architecture for smart cities and provide a useful option for future smart city applications.
The integration of various transportation technologies in an intelligent transportation system makes it possible to be used in smart city architecture. Zhao and Jia [14] provide a literature review on how intelligent transportation systems can contribute to the environmental sustainability of smart cities, noting that the data required by a large number of vehicle terminals may result in a backlog of large volumes of data that must be processed by intelligent transportation systems in smart cities to ensure efficient transportation by controlling traffic flow and preventing accidents. The concept of smart cities was first proposed in the early 1990s. Subsequently, based on the development of various information technologies, the study of smart cities has achieved certain results [15]. Smart cities are the way to the future development of cities, which are bound to contribute to the improvement of the quality of life of the residents with the development of information and communication technologies (ICT) and transportation technologies. There is no doubt that the requirements for smart cities in modern intelligent transportation systems are getting higher and higher, and various new information technology developments have made the combination of smart cities and intelligent transportation closer and closer. Meanwhile, modern information technologies such as cloud computing [16], blockchain, and Internet of things [17] have their own advantages in facing the processing of large volumes of traffic data, so it is especially important for the above study to use cloud computing to deal with automatic monitoring and management of vehicle flow during traffic congestion. In addition, blockchain and IoT for model evaluation of intelligent transportation systems in smart cities are evident for the improvement of the quality of life of residents in sustainable urban transportation systems.
2.2. Data Transmission Control for Intelligent Transportation Systems
Joseph et al. [18] provide a comprehensive overview of five intelligent transportation system projects (TIME, Sentient Transport, EVT, DynaCHINA, TrafficView) with a focus on the collection, transmission, and analysis of focused traffic and vehicle data. The authors point out that providing intelligent service to drivers is particularly important for future intelligent systems, followed by an introduction to intelligent data identification and data collection in the railroad environment. The data connection of Telematics and various transportation vehicles can promote the effective data transmission in smart cities. In the process of data transmission, real-time transmission control and traffic system will generate a large amount of transmission data, and Sumalee and Ho [19] analyzed the problems that need to be paid attention to data transmission in the intelligent transportation systems, especially in the connected vehicle to build the intelligent transportation system. Considering the recent rise of IoT technologies, Sodhro et al. [20] proposed an IoT-driven intelligent transmission system control scheme aimed at supporting future IoT-driven vehicle-to-vehicle (V2V) multimedia transmission communication. The scheme analyzes the communication performance of the proposed scheme over V2V in terms of QoS metrics such as green (i.e., energy efficiency), sustainability (i.e., less battery charging consumption), reliability (i.e., less packet loss), and availability (i.e., more coverage). Zhang and Lu [21] studied the vehicle communication network in IoT-based intelligent transportation system, and model simulation tests by OPNET modeler showed that AODV protocol outperforms DSR protocol in terms of throughput, average network delay, routing load, packet loss rate, and average routing hops, which is more suitable for the network communication of intelligent transportation system.
The magnitude of traffic flow in intelligent transportation systems is particularly important in the transmission and processing of data, and the prediction of various traffic flows is also the focus of research on intelligent transportation data transmission in smart cities. Zhang et al. [22] proposed a scheme based on a quantum particle swarm optimization strategy. The scheme metaphorically incorporates a genetic simulated annealing algorithm into the initialization of traffic flow data according to the characteristics of the transmission data of traffic flow in smart transportation and applies a radial basis neural network prediction model to optimize the parameters. The simulation results show that the proposed algorithm can reduce the error of data prediction and play a more stable role in the transmission of data in the intelligent transportation system. Similarly, for the traffic congestion problem caused by high traffic volume, Saharan et al. [23] effectively managed the problem of data transmission in intelligent transportation by developing a dynamic pricing strategy. Firstly, the authors provide a literature review and analysis of dynamic pricing techniques, followed by an in-depth discussion of various problems solved by dynamic pricing techniques, evaluation of parameters and their limitations, and, finally, an analysis of future dynamic pricing applications for intelligent transportation systems. Intelligent control of traffic in smart cities has an important impact on enhancing the transmission of data. To improve the efficiency of intersection vehicles, Lv et al. [24] modeled intersection vehicles using artificial intelligence techniques that have developed rapidly in recent years and used dynamic scheduling algorithms to improve the communication network in intelligent transportation systems. Finally, the effectiveness of the proposed model and the improved scheme was evaluated by simulation tests. The results show that the model can be used to predict the passage time of queued vehicles at intersections with a small error and a high success rate of data transmission. Suryadithia et al. [25] provide a detailed discussion of the technical barriers encountered in intelligent transportation systems, starting with a review of the literature on various intelligent transportation technologies in recent years, followed by an example of intelligent transportation research and providing examples of applications that can be used in intelligent transportation systems.
The layout of the organization of the rest of this study is as follows. Section 3 describes the proposed congestion control algorithm. Section 4 goes through the experimental simulations. The last section concludes the study and gives relevant conclusions.
3. Data Transmission in Intelligent Transportation Systems
Data for smart cities come from a variety of interconnected sensors, and sensor data from these devices are critical to intelligent transportation systems. The ITS vehicle network is the core component of the smart city. The current widespread use of a large number of wireless communication technologies can lead to low performance of traditional transmission protocols. Meanwhile, the large-capacity data transmission generated by various sensors poses new challenges to the data transmission system of telematics. In the multi-hop network built by telematics, there will be a large amount of interaction data between the sender and receiver of the vehicle, and these data will experience various network states, such as congestion in telematics, loss of data in the wireless link, change in vehicle building path, or even interruption of data transmission. However, the above situations cannot be directly identified in these interaction data, which create an obstacle to efficient data processing in smart cities.
3.1. Data Interaction Problems in Smart Cities
The amount of data interaction in a smart city is very large. The multi-hop network built by VANET, various sensor data, and application data forms a huge pressure on the transmission system, where the biggest risk faced is the congestion caused by the backlog of transmission data. The judgment of data congestion in the network is round-trip time (RTT) timeout, and the sender of data in the telematics network directly reduces the amount of data transmission, but the guideline of this judgment is exposed to many misjudgments, especially for vehicle-based transmission systems with wireless communication technology. One of the main misjudgment cases that the wireless link caused is data loss, and this data loss is caused by a variety of reasons, such as various interference to the wireless channel, competition caused by the media access, and link bandwidth limitations. The characteristics of this data loss are very different from the case of data loss caused by congestion, and data loss in wireless links often occurs randomly. There should also be a big difference between the data sender's handling of random loss and congestion backlog loss. The former only needs to retransmit the lost random data, and the amount of data sent can still be maintained at a larger amount to improve the transmission efficiency of the network.
In addition, the driving path of vehicles in the traffic system is not a stable path, which adversely affects the construction of multi-hop routing in vehicular networks. When multiple vehicles build multi-hop transmission paths, the data transmission in multi-hop networks can form stable transmission routes. However, when there is an unexpected accident and temporary parking among multi-hop vehicles, the path of data transmission will be interrupted, and the data transmission will be disordered or even lost when looking for a suitable vehicle to build a multi-hop path. At this time, the data sender needs to reconstruct a stable transmission path and reduce or even temporarily stop the transmission of data. In the extreme case, the transmission path of data is irreversibly restored and the data connection of multi-hop network is interrupted. At this time, the sender of data should find a new transmission path as soon as possible to build a stable multi-hop transmission route, and the sender of data should suspend or stop sending data until a new transmission path is established.
Considering that the interaction of data in smart cities is very frequent and various applications will overlay different types of data, making the transmission of data face the risk of congestion caused by backlogs, and congestion is the first network state to be considered in the above two cases. Once congestion is judged to occur, the amount of data sent is reduced. Secondly, in the case of rejecting congestion, the routing situation of the network needs to be judged, and in such cases, whether a change in the multi-hop path occurs due to a change in vehicles or an interruption in the data transmission path.
3.2. Determination of Transmission Status
The sharp increase in the amount of data in smart cities may bring the risk of congestion, and the above analysis shows that in the process of data transmission, it is necessary to first determine whether congestion occurs in the car network in smart transportation and then determine the change in routing. The amount of information for judging the network state in the interaction of data in the transmission protocol is very small. This section first defines two parameters for judging the congestion state, congestion parameter and congestion parameter , which are used to judge the congestion of the network in terms of time and throughput, respectively. In the state of negative congestion, then two parameters are defined to determine the state of routing, routing parameter and routing parameter , which determine the state of routing from the proportion of out-of-order packets and the proportion of packet loss rate in the network, respectively.
The congestion parameter is defined as follows:
The congestion parameter is used to determine the state of the network easily by the difference between the timestamp of the sender and the receiver, where is the timestamp included in the packet sent by the server in telematics, and is the timestamp in the packet returned by the end device after receiving the data. There is no doubt that the difference between these two reflects all the time that the data have experienced. The congestion parameter is the difference between the time experienced by i+1 packets and i packets. If the difference between the two fluctuates widely, it indicates that the time experienced by i+1 packets and i packets varies widely. Equation (1) can not only calculate the difference between the time stamps but also calculate the time difference between the sender’s continuous data transmission and the receiver’s continuous data reception; the above equation can be organized as follows:
The above equation shows that the initial value of the congestion parameter is the initial difference between the timestamps of the sender and the receiver, which reflects the initial network condition of the vehicular network. However, only the difference in the timestamp intelligently reflects the change in time fluctuation, and it cannot directly reflect the congestion state value of the network. Therefore, the congestion parameter is introduced and defined as follows:
is the number of packets received in two consecutive packet sending intervals, and the value reflects the data transmission rate in a very short time interval. In the case of stable data transmission, the value varies very little. As with the congestion parameter , the congestion state of the network is not directly reflected by the amount of data transmitted in a short time only. However, combining these two parameters together can give a good indication of the congestion status of the network. Firstly, when the congestion parameter is larger and the congestion parameter is smaller, it means that the packet has experienced a long time on the transmission path and the data transmission is very small, so it is obvious that a serious data backlog has occurred in the data, and then, it can be directly judged that the network is congested. On the contrary, when the congestion parameter is larger and the congestion parameter is larger, it cannot be directly determined that the network is sending congestion. In this case, there is a possibility of random data loss resulting in an increase in the time taken for the data to be transmitted. When the congestion parameter is small and the congestion parameter is large, the path of data transmission may change, resulting in a sharp increase in the amount of data sent, and it is not possible to determine whether congestion has occurred. When the congestion parameter is small and the congestion parameter is small, there is a possibility of serious data congestion or a large amount of random data loss or even path interruption, which requires further judgment, so it cannot be directly determined whether the network is congested. When congestion is determined, it is necessary to reduce the amount of data sent. Other cases can jump to the routing parametric feedback after determining the congestion parametric feedback. The above process is shown in Figure 1.
After negating the state of congestion, the data may experience a change in transmission path or interruption, so other parameters need to be introduced to determine the state of the network. The proposed scheme uses two parameters for the evaluation of the routing state, the routing parameter and the routing parameter . These two parameters are the proportion of out-of-order packets and the proportion of packet loss rate in the vehicular network, respectively.
The routing parameter is defined as follows:
In the above equation, is the number of out-of-order packets in the interval time, and is the maximum value of out-of-order packets received by the receiver. The amount of data transmission carried in the intelligent transportation system in a smart city is relatively large, and congestion caused by data overload can lead to a large amount of data loss, but the interference of the wireless link can also lead to random loss of data due to errors in the transmission process. Therefore, the ratio of out-of-order packets within a certain time does not directly reflect a certain state of the network, such as path changes or congestion. It needs to be considered with the help of other network state judgments and parameters. Therefore, it is necessary to introduce a second routing parametric , defined as follows:
In the above equation, is the number of received packets in the interval time and is the maximum value of out-of-order packets received by the receiver. The routing parameter reflects the probability of data loss in the network, and the value reflects the state of the wireless link error. Again, a single routing parameter does not directly determine that a link error has occurred in the network.
Congestion occurs in the network, and link disruptions may also cause fluctuations in this value. If two congestion parameters are combined to negate the congestion state, routing parameter and routing parameter can each reflect a different state of the network. When routing parameter increases, it can be determined that a path change has occurred in the network, while when routing parameter increases, it can be determined that a link error has occurred. The former case does not need to change the sending state of data, and the latter case only needs to resend the lost data, and the sending state of the network can still be maintained. When all the above states are excluded, if the congestion parameter is extremely low, it can be determined that the data transmission path is interrupted and the sending state of the network should be temporarily stopped or even suspended. The above process is shown in Figure 2.
3.3. Window Optimization
Even if the original congestion window adjustment strategy is used when congestion occurs, it still faces the problem of excessive data transmission. After all, in a smart city where wireless channel transmission is dominant, excessive data competition for the available medium interface will only aggravate the data loss, so this scheme optimizes the transmission window appropriately.
The main parameter affected by the wireless link is the routing parameter , which reflects the degree of data loss in the vehicular network. The probability of data loss in the link is common for wireless channels, and the routing parameter measures the probability of error in the wireless link. The proposed scheme gives the optimized congestion window by combining the routing parameter and the upper bound on the bandwidth delay product of the vehicular network. The conventional transmission scheme follows the following rule in maintaining the increase in the congestion window.
For vehicular networks where wireless links often lose data, the transmission of large amounts of data can add to the transmission pressure of multi-hop networks, so it is necessary to restrict the above windows appropriately. The proposed scheme uses an interval that limits the congestion window.
is the window fading factor, is the upper limit of the bandwidth delay product within a multi-hop vehicular network, which generally does not exceed , P and Q are the size of the forward and reverse data transmission in the vehicular network, respectively. The k and l are the corresponding number of hops, respectively. The size of the congestion window is limited in the above equation to prevent excessive data volume from stressing the transmission of the vehicular network. This is extremely beneficial for the network transmission system of intelligent transportation systems in smart cities.
4. Result Analyses and Discussion
The routes formed by vehicles in the connected vehicle transportation system of a smart city are mainly in the form of straight lines and intersections. As shown in Figures 3 and 4, in the linear telematic system, vehicles mainly form traffic flow in two opposite directions, and these vehicles can build stable multi-hop transmission relay points. In particular, the lanes in the same direction can form a relatively stationary vehicular network topology map. The intersection form of vehicular networking is more complicated than the linear form of vehicular networking, and the traffic flows in different directions have different transmission paths in different lanes, which is extremely easy to cause interruptions of transmission paths and changes in routes for vehicular networking with wireless links. The simulation of the experiment is firstly carried out in the linear vehicular network, as shown in Figure 5, the topology diagram of the vehicle network by 12 vehicles statically constructing the simulation.
At the MAC layer based on the commonly used IEEE 802.11 scheme with a distance of 200 m between nodes, simulations are performed to test the transmission performance of the classical scheme 1 [26], scheme 2 [27], and scheme 3 [28] and improved algorithms.
The data flow formed by the dynamic overlay of various applications in the smart city builds a huge traffic volume, especially in the car network built by multiple sensors, and the multipoint sensing of data on multiple application data again forms the overlay of data flow that will inevitably intensify the transmission pressure of the network. Figure 6 shows the transmission rate of scheme 1 at six hops, which has been a great success in wired networks, but the transmission rate of this scheme fluctuates greatly in the vehicular network of smart cities, as seen in the figure, the transmission rate fluctuates between 120kps and 170kps, scheme 1 mainly adopts additive increase and multiplicative decrease (AIMD) scheme, the random packet loss is more serious in the vehicle network, the scheme will encounter serious performance degradation and the random loss of data, the scheme is considered to be caused by congestion, and in the transmit window will directly adjust the transmit window to 1. This will greatly reduce the efficiency of the network data transmission.
Figure 7 shows the data transmission efficiency of the proposed scheme at six hops. As seen in the figure, the data transmission rate improves significantly and fluctuates around 210kps. Overall, the data fluctuate less around 200kps. This has an important role in the stable data transmission of the telematics. Considering the proposed scheme in combination with different congestion parameters and routing parameters in the network can fully reflect the different states of the network, so that the data sending rate can be reasonably adjusted. Considering the demand for different application data from terminal devices in smart cities, congestion is bound to form in the network, and once the data backlog is formed in the network, the possibility of the congestion state of the network is considered first. Once congestion is judged to occur, the amount of data sent is reduced to reduce the probability of further formation of data backlog on the network. Secondly, after rejecting the possibility of congestion, the routing situation of the network needs to be judged, and in such cases, whether a change in vehicles in the multi-hop path leads to a path change or an interruption of the data transmission path occurs. The proposed scheme fully considers the different parameters of congestion and routing, which plays an important role in balancing the backlog of data in the network. Also, a stable data transmission environment reduces external disturbances and is friendly to various applications.
Figure 8 shows the transmission rates of the four schemes at different hop counts. From the figure, it can be seen that at 2 hops, different schemes have higher transmission rates than other hops. Due to the less number of hops, there is less interference between various vehicles, and to avoid random data loss caused by data access at the same time, the access mechanism will use a silent waiting scheme to adjust the competing interference of different data access, which will reduce the waiting time in the transmission of data and improve the efficiency of data transmission per unit of time. In addition, in the range of 4 hops to 12 hops, the transmission rate of data of all four schemes decreases to different degrees. It can be seen that the interference between vehicles increases when the number of hops increases, which will wait longer for the access of large amount of data. Although the waiting time increases additionally, the data reduce the data loss caused by competition when multiple data streams are accessed and avoid data retransmission. Similarly, scheme 1 uses the additive increase and multiplicative decrease (AIMD) scheme, and in wireless links where random packet loss is serious, the scheme determines that congestion is sent in the network when data are lost randomly, and the sending window is adjusted to 1 directly in the sending window, which greatly reduces the network This will greatly reduce the efficiency of data transmission in the network. Scheme 2 can handle the random data loss well, but still adopts the strategy of scheme 1 in the congestion window adjustment, which obviously also reduces the data sending rate. Scheme 3 provides good feedback on the state of the network with reasonable judgment, so the data transmission rate is significantly improved based on scheme 2. The proposed scheme makes full use of the advantages of the previous three schemes not only to accurately determine the state of the data in the network but also to reasonably adjust the size of the congestion window. Therefore, the transmission rate is the highest among these schemes.
Table 1 shows the tests for different data streams, as the data transmission path of telematics in the traffic system is relatively fixed. As shown in Figures 3 and 4, two cross-streams and four parallel data streams are used in the simulation for testing. The proposed scheme shows the best transmission rate in two cases as seen in the table. The proposed scheme can fully reflect the different states of the network when combining different congestion parameters and routing parameters in the network, so that the data sending rate can be reasonably adjusted. The medium access control of data in two cross-flow scenario will be more intense. The proposed scheme is still able to judge the state of the network well enough to reasonably adjust the amount of data sent. In four parallel data streams, although the transmission paths of data and cross-streams are not the same, more data streams will likewise aggravate the competition of data transmission, thus leading to the occurrence of random data loss. This makes it particularly important to determine the different states of the network, and the proposed scheme still shows a high performance in this scenario.
5. Conclusion
The vehicular network in smart cities plays an important role in carrying data transmission for various applications. In this study, we try to start from the judgment of the network state and try to judge the network state more accurately through the very little feedback information in the network, to provide a basis for the adjustment of the congestion window. The proposed scheme first defines two parameters to determine the congestion state, which are time and throughput to determine the congestion of the network. In the state of negative congestion, two parameters are then defined to determine the state of routing, which are judged from the proportion of out-of-order packets and the proportion of packet loss rate in the network, respectively. The size of the congestion window is adjusted according to the four different parameters mentioned above to match different network states. To distinguish between congestion and random loss caused by data backlog, this study combines the above different parameters to solve the data transmission problem according to the characteristics of smart cities, and the proposed scheme is simulated and compared with other three schemes, and the results show that the proposed scheme has certain advantages under different hop counts and data flows. The data from VANET in smart cities will increase dramatically, which poses new challenges for high-capacity data analysis, especially today when cloud computing and big data analysis are increasingly important. Future research will not only focus on data transmission enhancement but also combine the analysis of big data to refine the required feature data and better provide personalized services.
Data Availability
The data used to support the findings of this study are available from the corresponding author upon reasonable request and with the permission of funders.
Conflicts of Interest
The authors declare that they have no conflicts of interest.
Acknowledgments
This work was supported by the Jiangxi Provincial Natural Science Foundation (20212BAB202029) and the Science and Technology Foundation of Jiangxi Provincial Education Department (grant no. GJJ201601).