key: cord-0727644-oexk8a4b
authors: Ghazal, Taher M.; Said, Raed A.; Taleb, Nasser
title: Internet of vehicles and autonomous systems with AI for medical things
date: 2021-07-22
journal: Soft comput
DOI: 10.1007/s00500-021-06035-2
sha: d91b1540c36c1e3cbeb70587026ded62b243dece
doc_id: 727644
cord_uid: oexk8a4b

The current world faces a considerable traffic rate on roads due to the increase in various types of vehicles. It caused emergency vehicles to delay and increasing the patients' health risk factor. Internet of vehicle-based artificial neural network (IoV-ANN) has been proposed for effective health autonomous system in medical things. The proposed IoV-ANN provides a secure network to monitor and track the vehicle's motion using the global positioning system. It consists of an autonomous system which is enabled with an artificial neural network (ANN). ANN model has three layers. First layers collect the data using IoV sensors. Second or hidden layers process the sensor data, predict the road's traffic condition and reroute the emergency vehicle to an exact route. IoV-ANN helps to reduce road congestion in this article to enhance the timely functioning of an emergency vehicle. ANN categorizes the congestion networks of traffic. Traffic restrictions such as changing the queue gap in the road signals and the alternative roads are altered automatically due to congestion. It allows the government to develop ideas for alternate routes to enhance traffic control. The output layer gives commands to the driver autonomously. The simulation analysis of the proposed method proved that the system could work independently. The IoV-ANN achieves the highest performance rate of (97.89%), with a reduced error rate (9.12%) traffic congestion rate (10.31%), travel period (32 s), vehicle detection rate (93.61%), classification accuracy (95.02%), MAPE (8.4%), throughput rate (93.50%) when compared to other popular methods.

Today, road transport is rising, growing and by 2050 the number is projected to reach 2.5 billion. The difficulty of traffic control is concurrently growing with the increasing rise in traffic (Casado-Sanz et al. 2019 ). The smart transport system is popular in the self-employed sector, with its various road protection and media implementation (Kitchin and Dodge 2019) . The analysis of the linked automobile has become one of the main research areas under the framework of the IoV (Rathee et al. 2019) . IoV focuses mainly on vehicular network with the infrastructure aircraft. Vehicle to anything connectivity is seen as one of the main traffic assisting technologies in an intelligent city addressed by a smart transport system (Mounce and Nelson 2019) . However, multiple vehicles can often yield positive results for essential applications that require high bandwidth and reduced power (Maes and Steppe 2019). Connected vehicles need secure, quick communication with very little delay to and from a network operator (Chen et al. 2020 ). In the meantime, the control of many applications without any road congestion must prove more effective (Li et al. 2020) . A road vehicle often starts a new communication application to the traffic manager (Nanda et al. 2019) . Every time different systems receive a route malfunction query, the traffic coordinator calculates a new course (Simonetto et al. 2019) . The whole way to measure Communicated by Vicente Garcia Diaz.

Taher.ghazal@skylineuniversity.ac.ae a new route causes an additional fee, contributing to slowed traffic (Xu et al. 2020) .

The vehicle-to-all (V2X) system AI provides information from several sources, including vehicles, trains, buses, etc., and makes it possible to boost driver performance and accident avoidance predictions. This growth has enabled us to comprehend the clever conduct, based on the notion of replicating true driving behaviour, reducing human errors and providing drivers with pleasant security. Many services have been established for adjusting traffic, a heritage of autonomous vehicle systems to IoV, from crowds to light roads.

Smart vehicles provide incomparable opportunities for enhancing road safety and have been welcomed for their welfare programs and market opportunity (Chai et al. 2021) . Accurate car position continues to be an important field of emphasis for many automation concepts, including autonomous vehicles, destination resources etc. The primary cause of deaths among many college students among different age groups and the largest mortality time is traffic accidents injury (Shahbazi et al. 2019) . The vehicle-toeverything (V2X) system AI provides information from different sources, including automobiles, trains, buses, etc., and allows drivers to be more realized and forecasted for accidents to be avoided. This advancement has enabled the understanding of smart driving based on the notion of replicating genuine driving behaviour, reducing human blunders and offering drivers comfortable safety. Many services from crowd and light road traffic have been established to convert traffic from self-supporting vehicles to the IoV. A certain study covers the vehicle's rate as one of the major factors contributing to road injuries, drinking, non-use of the belt and distractions (Lomia et al. 2020 ). Around 90% of vehicle incidents are caused by weak road networks and weak traffic control systems in low and medium nations (Heydari et al. 2019) .

However, several routes are fitted with surveillance and speeding sensors in these increased nations to identify and later recognize those vehicles that have exceeded the allowed speed of traffic (Khan et al. 2020) . While these devices dramatically improved traffic congestion, they did not achieve any successful improvement to avoid or mitigate the occurrences due to accidents (Retallack and Ostendorf 2019) . Besides, over-speed automobile traffic collisions seem to have more occurrences, with other local organizations' participation, including people, motorcycle riders, or other cars (Abdulrazzaq et al. 2020) .

IoV intends to provide modern and creative solutions with integrated traffic control for different transport modes (Guo et al. 2019) . In view of the significance of IoV usage of Artificial Intelligence (AI) since it offers intelligent models for most of its uses, this article presents a quick overview of one of the AI approaches known as machine learning and its possibilities for various particular features of IoV network applications. In IoV networks, the most important obstacles needing an intelligent optimization strategy are edge computing and caching difficulties. Challenges in edge computing and cacheing relate to a variety of elements such as channel status, changing topology of communication and the management of resources.

IoV allows automobiles to transmit communications to enhance road safety and performance. Even then, because of untrusted conditions, the reliability of the communications individuals obtain is challenging for automobiles to determine. The V2X Paradigm bases primarily on the sharing of information, as in Fig. 1 . It consists of the following: vehicle-to-infrastructure (V2I), vehicle-to-pedestrian (V2P) and vehicle -to-vehicle (V2V). The V2X communication system has three primary aspects: traffic efficiency, road safety and energy efficiency. The traffic flow information is a key use case of V2X. This information may be used to carry out smart activities such as correction of traffic congestion, better usage of plug-in electric vehicles (PEV), minimize fuel consumption and improve location-based services.

IoV needs confidence protection, and it is unnecessary to form trustworthy links to modules on the roadside, which could inevitably reach a vehicle throughout its path before time (Feng and Haykin 2019) .

Constant speed sensors' key task is to catch any vehicle that reaches the specific location's posted speed and potentially finish the over-speed automobile/driver on highways in developing countries. Alternatively, the effective stress does not appear to impact the over-speeding problem since drivers may use the propulsion systems or warning lights to know where these cameras. The complete automated vehicles would carry economical and social benefits. The autonomous vehicle field, along with the conventional automotive industry, is controlled by Artificial Intelligence, the rapid growth of the fifth generation's connectivity market, and advanced electronics innovation. The test phase remains, however, for autonomous vehicles. The technology needs are much harder for autonomous driving-related to conventional wireless communication. Level-automated driving has to follow the double standards of extremely low latency and superior efficiency in complicated and unpredictable road conditions when carrying out a dense traffic mission. The main contribution IoV-ANN is described as follows.

• The IoV-ANN proposed provides a reliable network for monitoring and tracking movement via the global positioning system. It comprises an autonomous system activated by an artificial neural network (ANN).

• There are three layers of the ANN model: to collect data, sense data, and the emergency vehicle is diverted to an exact path.

The remaining article is organized as follows: Sect. 2 comprises various background studies concerning the IoV with AI. Section 3 elaborates on proposed IoV-ANN with a secure network to monitor and track the vehicle's motion using the global positioning system. Section 4 constitutes the results that validate the performance and predictability with the corresponding descriptions. Finally, the conclusion with future perspectives is discussed in Sect. 5.

This section discusses several works that various researchers have carried out; Lu et al. (2019) developed The cognitive Internet of vehicles (CIoV). CIoV suggests an inter-strategy for self-driving, focused on a cloud services trend and the IoT service architecture. CIoV approach achieves smart and scalable automated driving tasks with the conceptual Internet of vehicles' aid instead of current studies that rely primarily upon technological developments. The experiment is then performed to demonstrate the effects of the automated driving conceptual IoV. Chang et al. (2019) introduced A deep learning-based Internet of vehicles (DL-IoV). DL-IoV document contains an in-vehicle automotive smart transport system, a personality warning detector for automobiles and a front camera, a deeper cloud training framework and a fog control platform. DL link provides the deeper IoV device DeepCrash. Suppose a face or single-vehicle crash is found. In that case, crash identification is submitted for self-collision car accident identification to the cloud-based web servers, and a subsequent incident warning is issued. The research data suggest that the vehicle accident detection rate can be as high as 96%. Sahraoui et al. (2020) proposed remote sensing to control respiratory viral diseases (RS-CRVD). Vehicles with a wide range of devices like standard monitors that can be substituted with heating monitors are fitted in the omnipresent network topology. RS-CRVD consequently introduces a new model for detecting breathing infectious diseases using IoV to obtain in actual time observations of passenger heart rate and breathing rate. The knowledge can classify regions influenced by potential cases of COVID-19 and introduce effective prevention measures to reduce infection transmission further. Sodhro et al. (2019) developed artificial intelligencebased QoS optimization (AI-QoS). AI-QoSO paper provides two different techniques called quality of experience (QoE) optimization and buffer-aware and correlates the output with the Reference. AI-QoSO provides two new algorithms, and it suggests a mechanism for communication systems. The QoE automation system is presented via mobile devices across multimedia applications in the IoV method. Furthermore, experimental results show that mobile devices' enhanced service lives suggested PQO and BQO implementations are better than the standard with QoE optimizations. Internet of vehicles and autonomous systems with AI for medical… Zacharaki et al. (2019) introduced complex engineering systems (CES). The cybersecurity needs of connected vehicle (CV) are thus increasingly significant. In addition to cyber defence, threatened reaction and car restoration, connectivity and network device level of the entire IoV environment at complicated cases by resolving the protected architecture dimensions of the CV. Therefore, Lin-kedIn and separate motor vehicles are costly, with a new range of critical threats related to higher hacking attacks. A computer threat in a linked vehicle may result in high retrieval costs, financial losses, and even harm human protection. Wazid et al. (2019) proposed authenticated key management IoV (AKM-IoV). AKM-IoV is checked for safety assessment using the structured security assessment of the commonly agreed 'Automated Validation of Internet Security Protocols and Applications (AVISPA)' template, traditional and nontraditional safety tests. AVISPA uses the NS2 simulator to demonstrate the realistic implementation of AKM-IoV. Besides, the performance, usability and protection features provided by AKM-IoV were evaluated to other current protocols undertaken in a comprehensive common context.

The IoV-ANN proposed establishing a safe network to track and monitor the vehicle's motion with the global positioning system based on the survey. IoV-ANN is composed of an autonomous system that ANN supports.

This study deliberated the internet of vehicle-based traffic congestion reduction for emergency vehicles and improved its performance. The emergency vehicles have long stay on the road and raised the health risk factor of their patients. A major problem worldwide is the steadily growing congestion of traffic due to the tremendous growth of vehicles. Shifting from the present route to the smart is a real problem in which no cellular network interfaces are required for both cars and road networks. There is a more congested healthcare region than road situations with more delays and travel time concerns. Different types of routing recommendation protocols were suggested where most consider smart vehicles fitted with sophisticated and intelligent communication systems. Hence, in this paper, IoV-ANN reduces traffic congestion to improve an emergency vehicle's timely performance. ANN classifies the traffic congestion networks. The traffic rules like adjusting the queue gap in the road signals and the alternative routes will be automatically changed based on the traffic congestion. It allows the government to create strategies to establish alternative routes for improved control of traffic. Figure 1 shows the proposed IoV-ANN. One of the major features of intelligent transport infrastructure is the IoV system for traffic control. It is a device that can be used to track and assess the traffic condition of the multiple road segments by the health care or transport department. The computer-primarily collects traffic data in real-time and determines congestion situations for decision-making and traffic management. The proposed system divides road freight loads from heavy congestion, medium congestion and free flow into three groups.

Stationary sensors in this field gather different road parameters, including vehicle speed and number of vehicles on the road in real-time, and forward them to the ANN-based data processor, which decides the street segments' traffic congestion status. The proposed framework comprises the data collection unit, the transmitting unit and the information processing unit. These are defined in this manner. For decision-making and effective traffic control plans, traffic knowledge is very relevant. In-road stationary sensors are used for traffic control in real-time. On both ends of the road section, the sensors are deployed. The sensors capture the vehicle speed and the number of vehicles through stationary sensors on the road. The sensed traffic data must be moved annually from the field to the remote data processing centre. The local data retrieval system sends aggregate data through mobile or wireless communication. The raw traffic information gathered on the ground is used to calculate vehicles' average speed and road traffic. The research is to deduce traffic congestion in various traffic industries that regulators can timely schedule and take decisions. Figure 2 demonstrates the basic Internet of vehicle diagram. The automobile emblem provides vital details, as it is the classic picture of an automobile. Thus, substitution or misappropriation is complicated. The information gathered from the ANN technology is, therefore, of great importance. ANN technologies can help to reliably distinguish the car if the licence plate is lost, replaced or the vehicle type is changed. This knowledge is critical for traffic control, car park management and identification of vehicle models; Fig. 2 presents a scenario of the implementation of IoV-ANN technology. The first is to obtain traffic information from visual sensor devices used on the lane. The car identification technology is then used to collect information from the traffic data collected to find suspicious vehicles. Finally, if there is a suspicious car, a warning shall be sent to the police vehicle, and the lights shall be controlled at the appropriate intersection.

Furthermore, traffic data can be sent to the cloud repository and sharing centre by the visible sensor and the vehicle information derived by the vehicle detection technology. Cloud data network creation and information sharing strategies are illustrated. The vehicle detection system can easily collect vehicle identification information and can be used for certain situations like car parks, suburban areas and public transport. Vehicles are normally moving at high speed, and the shot pictures of vehicles are often distorted in movement. This paper is used for eliminating movement blur and vehicle emblem detection. However, it cannot be used for a deblurring portion of the vehicle logo identification mission explicitly since it has two big defects. First of all, it cannot determine whether the picture has a whirlwind of motion. Secondly, it cannot concurrently deblur images in multiple resolutions. Blur ratings and image resolutions are both different in large datasets. People must, therefore, recognize photos that involve flowering and flowering. First, the blur of each image can be easily measured and represented in numerical form. Second, ANN will correctly determine whether the picture should be debilitated based on its black meaning. Thirdly, the images of varying resolutions will essentially be blurred by ANN. ANN can be added directly to the vehicle logo identification process with the help of the functions mentioned above. ANN has developed three target detection layers. For identifying big, medium and small objects, three different sizes of detection layers were used. Table 1 Symbols and Description.

If the F vehicle enters a road section, its speed and location, along with its targeted destination, can be passed to the RSU. The RSUs have directional antennas and are fitted for both longer transmission and more simultaneous 

Internet of vehicles and autonomous systems with AI for medical… transmissions. When RSU interacts with cars that drive along a certain highway, it uses directional antennas, which are very important. It will minimize signal interference significantly. If a vehicle has to meet the destination on the ideal path, it sends the application to the RSU. The automobiles follow routes from a given source to a destination according to the RSU recommendation. RSU locations cars in a road segment when determining the congestion index of this road segment. The index congestion is calculated as a share of the total number of vehicles already on the road to the number of vehicles allowed on the road, as seen in Eq. (1)

In Eq. (1), DI is a congestion index that is regarded as the actual number of B vehicle in vehicles for the permitted number of vehicle Q vehicle in the PT j road section. B vehicle includes wireless network interface makes vehicles called equipped or autonomous vehicles or non-wireless interfaces, as illustrated in Eq. (2).

As shown in Eq.

(2), the amount of the actual vehicle has been determined. The total number of vehicles is represented by equation two by adding autonomous and fitted vehicles. DI is measured using F vehicle and MF vehicle and great significance if the total number of vehicles is to be obtained more accurately. RSUs continue to recommend a specific road segment to the F vehicle until they collect details on the traffic congestion exceeding the congestion index value Eq. (3).

As represented in Eq.

(3), the number of vehicles permitted Q vehicle on a road segment is the allowable number of vehicles named vehicles. Equation 3 indicates the number of Q vehicle calculated from the length of the road sector to the length of the vehicle. The length of each RSU is known, and the vehicle's length is measured as 5 m. The kind of traffic can determine the length of the vehicle on the road section. Its value can be increased by long roads such as roads close to industrial areas or declining cities with large numbers of medium-length vehicles. Since the Q vehicle has wireless interfaces and can communicate with close RSU, it is possible to measure road occupancy by F vehicle . Road occupancy means the usage of the road by the length of the F vehicle and the minimum difference between two vehicles as seen in Eq. (4).

As deliberated in Eq. (4), road occupancy with j and i. PR ji means the main intersection between j' and i of two RSUs. When the PR ji exceeds the length of the PT length the path then is taken by the F vehicle roughly, and MF vehicle are discarded Eq. (5).

As initialized in Eq. (5), the amount of vehicle equipped has been calculated. In the other scenario, both equipped and unfitted vehicles can be used in street segments. It is not enough to create a more reasonable congestion rate when only fitted vehicles are occupied. Therefore, the Internet of vehicles roads, it is very important to recognize the F vehicle along with MF vehicle . When asked to RSU, the F vehicle sends its location. If the distance of two vehicles f vþ1 ð Þ À f v is measured as C and comparing the average length of a vehicle (U length and minimum gap H) the number of vehicles between the vehicles is estimated. To estimate the MF vehicle number in this section is used in Eq. (6).

As expressed in Eq. (6), the amount of vehicle equipped has been determined. In addition, the data centre uses an auxiliary approach to conducting data fusions in historical and real-time. RSUs routinely transmit data to the data centre. In addition, the road network comprises three elements, which are cars, RSUs and road sectors. For all road segments linked to RSU, each RSU keeps traffic congestion information. The Table of Information for Road Segment (RSIT) shown in Eq. 7 displays vehicles heading to destinations following the index value of congestion. RSU updates the total vehicle number at different times of the day to determine the highway segment's high time congestion Eq. (7).

As shown in Eq. (7), a Table of Information for Road Segment has been formulated. This module uses the vehicle details to calculate the congestion index by counting the number of cars against approved vehicles. The index value of congestion ranges between 0 and 1. RSU holds vehicles on the road segment until the value of the congestion index is 1. The index value for the congestion regulated the number of vehicles to prevent the vehicle's congestion in the road sector. RSU arranges all vehicles' destination reports in that road segment at each intersection of this module's route. Figure 3 demonstrates the ANN-based path planning model. The number of vehicles occupying the given length of a road segment is known as the density (C) of traffic. It is possible to express Eq. (8).

As deliberated in Eq. (8), density has been evaluated. Where m is denoted, the number of vehicles and k has explored the road segment length. Average speed T avg is an amount of speed divided by the number of the total vehicle for all vehicles, where Speed T j is defined for each unit length for the entire journey distance travelled by the vehicle. Therefore, the average speed can be expressed (9).

As obtained in Eq. (9), the average speed has been calculated. where, Distanece j explores the distance travelled by j th the vehicle, and s is the period. ANN is an important artificial intelligence branch. It's an incredible way to construct an intelligent processing system. The ANN-based method is based on biological neuronal networks to address various complex issues and categorize congestion status effectively. The ANN has several simple computing machines, such as neurons, related to such weights. A neuron receives input and uses a transfer function to produce output. ANN typically has three component layers: (i)input layer, (ii) hidden layer and (iii) output layer, and several neurons compose of each composite layer. A multi-layer perceptron with complete connection is used for this work. Figure 4 shows clearly that each input-layer node is connected with a certain weight to each hidden layer node. Every hidden layer node with some weight is connected to each output layer node. The weighted matrix is represented by Z 1 from the input layer to the hidden layer and Z 2 from hidden layer to output layer. Matrix Z 1 and Z 2 , where Z j i;l represents the weight of the i th neuron connection from a j th neuron on the l th the layer is shown as follows. Figure 4 illustrates the ANN-based traffic flow prediction. Every neuron (node) in the hidden layer uses a weighted sum method to process information based on inputs and bias and generate output based upon activation Eq. (10).

The weighted value (m 1 1 ) at the first node in the hidden layer Internet of vehicles and autonomous systems with AI for medical…

As found in Eq. (10), the weighted value has been computed. A weighted sum (m 1 1 ) is used to generate the net output (R 1 1 ) at first nodes in the hidden layer and can be expressed by the activation function Eq. (11) Figure 5 shows the net output and where E is the function of activation. This paper has the function of log sigmoid activation, which maps output between 0 and 1 and is shown by Eq. (12)

As explored in Eq. (12), activation function has been described. Thus, at the first nodes in the hidden layer weighted number (m 1 1 ) can be shown as a dot product Eq. (13):

As expressed in Eq. (13), total weight has been found. The weighted total (m 1 1 ) is supplied by the log sigmoid activation function to get the net output (R 1 1 ) in the first nodes in the hidden layer, as follows Eq. (14).

As shown in Eq. (14) net output of the first node has been expressed. Likewise, net output R 1 1;i ¼ 5 for, it is possible to quantify all five nodes in the secret layer. In the output node for each output layer node, the concealed layer output is now entered to decide the final output. The weighted number (m 2 1 ) at second node (class 1node) can be seen in dot product form in the output layer Eq. (15)

As explored in Eq. (15), second node weighted value has been obtained. The weighted sum (m 2 1 ) is then applied to sigmoid activation to obtain the net output R 2 l of the first nodes in the output layer.

In dot product form, net R 2 1 at the second node in the output, the layer is displayed Eq. (16).

As displayed in Eq. (16), net output in the second node has been expressed. Initializing weight Z 1 and Z 2 by random values in the neural network. The weights must, therefore, be adjusted. Backpropagation is a technique for changing the weight (s), by computing the error gradient function, of each node in an artificial neural network.

In dot product form, net output R 2 l at the second node in the output layer can be shown as Eq. (17): Figure 6 demonstrates the total error. The target is the actual output, and output is the value generated by the node. Equally, node two-and three-node error functions F 02 and F 03 of the output, the layer should be determined, respectively.

Total error F total is network is represented by Eq. (18)

As calculated in Eq. (18), the total error has been computed. The network is now expanding back to reduce the error by changing weight values. For this reason, the rate of error change must be determined by F total about the weight change represented by lF total =lz 2 1;1 Eq. (19)

Modification weight gas has been determined in Eq. (19) . Thus, the weight z 2 1;1 is modified lF total =lz 2 1;1 is calculated, and the weight z 2 1;1 is changed by the following Eq. (20):

All other weights changed from the hidden layer into the output layer.

The input layer weights must be matched to the hidden layer. For the weight change z 1 1;1 , the following equivalents are used for measuring lF total =lz 1 1;1 and the weight z 1 1;1

Eq. (21) Equivalently, all input layer weights need to be adjusted to the hidden layer. The entire backpacking phase is replicated five thousand times to minimize network error. Neural networks have to be trained after they have been developed with predefined or known traffic datasets. As an input to the neural network, the document takes travel density and average speed and classifies traffic congestion Internet of vehicles and autonomous systems with AI for medical… on a neural network basis. The section of results explained why input is based on average speed and density. This paper takes three levels of congestion into account: high congestions, low congestions and free flows. Figure 7 signifies the actual traffic status prediction.

The proposed IoV-ANN has reduced traffic and improved healthcare performance to achieve performance, error rate, traffic congestion rate, travel period, vehicle detection rate, classification accuracy, MAPE, and throughput rate.

The constantly expanding traffic congestion due to the huge expansion of cars is a big concern worldwide. Shifting from the current route to the smart route is a serious difficulty when there is no need for both road and automotive cellular network interfaces. There is a more crowded area of health than roads with greater delays and problems with travel time. Various forms of route suggestion algorithms where smart cars equipped with sophisticated, intelligent communication systems are the major consideration have been offered. IoV-ANN is therefore reducing road congestion in this article to enhance the timely functioning of an emergency vehicle.

There is a more crowded area of health than roads with greater delays and problems with travel time. Various forms of route suggestion algorithms where smart cars equipped with sophisticated, intelligent communication systems are the major consideration have been offered. IoV-ANN is therefore reducing road congestion in this article to enhance the timely functioning of an emergency vehicle. ANN categorizes the congestion networks of traffic. Traffic restrictions such as changing the queue gap in the road signals and the alternative roads are altered automatically due to congestion.

IoV-ANN has been validated based on the performance and the error rate. The actual performance is the phase in the quantification of the task. The expert recognizes the performance of the IoV-ANN and documents that evaluating position concerning the normal performance framework of the vehicles. Vehicles have to function efficiently with an effective traffic reduction rate. The steady-station sensors collect multiple road metrics such as vehicle speed and number of vehicles on the road in real time, and pass them to the traffic congestion processor based on ANN which has been discussed in the results and discussion. The structure presented includes the data gathering unit, the transmission unit and the information handling unit. This is how they are defined based on experimental analysis section. Traffic knowledge is particularly important for decision-making and successful traffic control measures. Stationary on-road sensors are utilized for real-time traffic control. The sensors are installed on both ends of the road.

Whereas with the performances ratings, there are several sources of error, the error can be minimized by rank teaching and component assessment scale. The ANN structures, performance conditions and the requirements of vehicles place on traffic control are different tasks. IoV-ANN examined the traffic-based data collection obtained by the global positioning system. The performance rate of IoV-ANN is shown in Fig. 8 .

The monitoring and tracking focus mainly on the intelligent vehicles traffic control system, performed by the IoV-ANN method. Moreover, the vehicles have its limitation for traffic congestion, and there is a low level of error in the road condition level. The secured IoV-ANN framework is implemented to reduce traffic congestion attacks with a reduced error rate. The attacks on numerous roads on IoV platforms are compatible, and a protected level is built to intercept these traffic congestion using the primitive system of data transfer. The error rate of IoV-ANN is shown in Fig. 9 . Each vehicle's travel time is used with fitted and unfitted vehicles reflected in the traffic congestion criteria. As the main period of the morning is higher, there are significantly higher traffic congestions. Each vehicle's total travel output must be reduced than the traffic condition and reroute the emergency vehicle to an exact route. Travel period is not feasible during traffic congestion periods and better accuracy at nighttime if a limited number of vehicles are driving along the road. The traffic congestion rate of IoV-ANN is shown in Fig. 10 .

The changing amount of vehicles progressively increases and influences the final delay in vehicles' route direction; when there is a rise in the number of vehicles, their travel time increases. If the number of vehicles increases, the congestion on the road increases. IoV-ANN is proposed to reduce road traffic and the travel time reduces automatically. The automobiles are diverted on various paths based on the traffic index score to prevent a shortage of one-lane section. The interruption at the final moment would be the same in non-congested periods, demonstrates traffic road congestion variations. The travel time of vehicles is shown in Fig. 11 .

In the proper driving support systems, vehicle identification performs a significant function. The enhanced template is commonly used in vehicle identification due to its high accuracy and good performance. The vehicle detection technologies are focused on computer vision; the issue associated with reductions in slightly blurred vehicles' Fig. 9 The error rate of IoV-ANN Internet of vehicles and autonomous systems with AI for medical… false positive rate is really difficult. The vehicle detection rate in IoV-ANN is shown in Table 2 . The classification results in extremely applicationspecific impacts of errors in the vehicle detection and traffic control mechanism. While traffic enforcement mechanisms could be durable to misclassify road visibility in relative car concentrations, smart payment systems may necessitate accurate distinctions among trucks and passenger vehicles. Table 3 illustrates an aggregate summary of the resulting classification accuracies for various research models and the categories.

The mean absolute percentage (MAPE) error is often recognized as a calculation of the predictive reliability in data, e.g. in the impact development, employed even as a gradient descent for deep learning dimensionality reduction. MAPE and its conceptual meaning of absolute errors are often used as a loss variable for regression tasks and model assessment. A statistic calculation of a reliable forecast system is based on the MAPE. It calculates the exact value as a fraction, and it can be determined by dividing true values as estimated total percentage errors for each period. The MAPE estimation is illustrated in Table 4 .

Steadily, the findings reflect the throughput's optimum value instead of subsequent decreases with time as the congestion rises. The output is about the same during the uncongested time; it indicates traffic congestion variations. Therefore, even under congestion conditions, greater efficiency outcomes can be claimed because of the highest number of data packs obtained at IoV-ANN. The throughput rate of IoV-ANN is shown in Table 5 .

The proposed method achieves the highest performance and accuracy when compared with other existing complex engineering systems (CES), a deep learning-based Internet of vehicles (DL-IoV), the cognitive Internet of vehicles (CIoV).

This paper presents IoV-ANN with a secure network to monitor and track the vehicle's motion using the global positioning system. In today's world, road traffic is growing dramatically on account of several automated vehicles. IoV-ANN consists of an autonomous system which is enabled with an ANN. ANN model has three layers. First layers collect the data using IoV sensors. Second or hidden layers process the data collected from sensors, predict the road's traffic condition and reroute the emergency vehicle to an exact route. The output layer gives commands to the driver autonomously. The simulation analysis of the proposed method proved that the system could work independently. The upcoming solutions must consider the communication coverage limitation in a complex healthcare environment.

The blur of each image may first be measured and displayed numerically using ANN which determines its dark connotation. Further, ANN will effectively validate the resolutions in identification process. Three target detection layers have been built by ANN in which sizes of detection layers were employed to detect large, medium and tiny items.

Sophisticated solutions can be established to improve communication ability utilizing diverse technologies such as IoV, AI and medical things for effective healthcare solutions. The IoV-ANN achieves the highest performance rate of (97.89%) with a reduced error rate (9.12%), traffic congestion rate (10.31%), travel period (32 s), vehicle detection rate (93.61%), classification accuracy (95.02%), MAPE (8.4%), throughput rate (93.50%) when compared to other popular methods. 

Conflict of interest The authors declare that they have no conflict of interest.

Ethical approval This article does not contain any studies with human participants or animals performed by any of the authors.

Traffic congestion: shift from private car to public transportation

Population ageing and rural road accidents: Analysis of accident severity in traffic crashes with older pedestrians on Spanish crosstown roads

Infrastructure: echoing with autonomous driving

DeepCrash: a deep learningbased Internet of vehicles system for head-on and single-vehicle accident detection with emergency notification

Intelligent resource allocation management for vehicles network: an A3C learning approach

Cognitive risk control for anti-jamming V2V communications in autonomous vehicle networks

Urban traffic signal control with connected and automated vehicles: a survey

Road safety in low-income countries: state of knowledge and future directions

Smart traffic monitoring system using unmanned aerial vehicles (UAVs)

The (in) security of smart cities: vulnerabilities, risks, mitigation, and prevention

The reliable congestion control mechanism for safety applications in urban VANETs

Socio-demographic determinants of road traffic fatalities in women of reproductive age in the Republic of Georgia: evidence from the National Reproductive Age Mortality Study (2014)

The cognitive internet of vehicles for autonomous driving

Perspectives for remote sensing with unmanned aerial vehicles in precision agriculture

On the potential for one-way electric vehicle car-sharing in future mobility systems

Internet of autonomous vehicles communications security: overview, issues, and directions

A blockchain framework for securing connected and autonomous vehicles

Current understanding of the effects of congestion on traffic accidents

Remote sensing to control respiratory viral diseases outbreaks using the Internet of Vehicles

Analysis of the mortality rate of road traffic accidents and its trend in 11 years in Iran

Real-time city-scale ridesharing via linear assignment problems

Artificial Intelligence-based QoS optimization for multimedia communication in IoV systems

AKM-IoV: authenticated key management protocol in fog computing-based Internet of vehicles deployment

A framework for collaborative air traffic flow management minimizing costs for airspace users: Enabling trajectory options and flexible pre-tactical delay management

September) Complex Engineering Systems as an enabler for security on the Internet of Vehicles: the nIoVe approach

Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations

Internet of vehicles and autonomous systems with AI for medical…