Method and System for Adaptive Data Forwarding for Comprehensive Infotainment Services in a VNDN Environment Including Autonomous Vehicles

Abstract

The present invention relates to an adaptive data forwarding method and system for comprehensive infotainment services in a VNDN environment including an autonomous vehicle. More specifically, it relates to an intelligent adaptive forwarding method that integrally manages various types of information data (e.g., safety information, traffic information) required by an autonomous vehicle in an autonomous driving environment and entertainment data for passengers, and selectively applies a pull-based or push-based forwarding method according to the characteristics and priority of each data. According to the present invention, first, various types of data traffic (e.g., emergency safety warnings, traffic information, multimedia content) occurring in an autonomous driving environment are clearly distinguished through class attributes, and by selectively applying a push-based APDF or pull-based EPDF forwarding strategy according to the characteristics of each data type, it is possible to simultaneously achieve rapid dissemination of information essential for autonomous driving safety and efficient provision of high-quality entertainment services.

Description

Technology Field

[0001] The present invention relates to an adaptive data forwarding method and system for comprehensive infotainment services in a VNDN environment including an autonomous vehicle. More specifically, it relates to an intelligent adaptive forwarding method that integrally manages various types of information data (e.g., safety information, traffic information) required by an autonomous vehicle in an autonomous driving environment and entertainment data for passengers, and selectively applies a pull-based or push-based forwarding method according to the characteristics and priority of each data. Background Technology

[0002] Autonomous vehicles must process and exchange vast amounts of data in real time for safe and efficient operation. This includes critical safety information essential for the operation of the autonomous driving system itself (e.g., emergency warnings, sensor fusion data), general information that enhances driving convenience (e.g., real-time traffic conditions, guidance on nearby facilities), and various entertainment content for passengers (e.g., video streaming, music). Each of these data types has different QoS requirements regarding transmission methods (push or pull), latency, reliability, and more.

[0003] Existing VNDN environments are primarily based on pull-based data request-response models, and only recently have some push-based techniques been introduced for the rapid dissemination of critical information. However, in environments like autonomous driving where information and entertainment services coexist and the importance and urgency of data vary within each service, it is difficult to expect optimal performance by simply using only one of these methods—either pull or push—or by operating them separately. For instance, processing entertainment data requests may interfere with the dissemination of urgent safety information, or conversely, transmitting all information via the push method may increase network load. To satisfy both the safety and user experience of autonomous vehicles, an integrated and intelligent adaptive forwarding mechanism is essential that can identify various data types and dynamically select and apply forwarding strategies tailored to the characteristics of each data. Prior art literature

[0004] 1. I. Ali and H. Lim, “NameCent: Name centrality-based data broadcast mitigation in vehicular named data networks,” IEEE Access, vol. 9, pp. 162438-.162447, 2021.2. R. Hou et al., “Data forwarding scheme for vehicle tracking in named data networking,” IEEE Trans. Veh. Technol., vol. 70, no. 7, pp. 6684-.6695, Jul. 2021. The problem to be solved

[0005] The present invention was devised to solve such problems, and its main purpose is to provide the first integrated adaptive forwarding method capable of effectively distinguishing and simultaneously servicing information data (critical / non-critical) and entertainment data having different characteristics and QoS requirements, particularly in a complex and dynamic VNDN environment where autonomous vehicles operate, while selectively providing pull-based or push-based forwarding optimized for each data type.

[0006] Specifically, the present invention aims to provide a method for mitigating broadcast surges and maximizing overall data delivery efficiency, QoS satisfaction, and communication stability of an autonomous driving system by identifying the type of data (e.g., critical information for autonomous driving, general information, entertainment) through a Class attribute included in a data packet, and by utilizing a Weight attribute to intelligently determine whether and how to forward based on network parameters (e.g., centrality, direction, speed, hop count, RSSI) that reflect the characteristics of the autonomous driving environment at each intermediary node. means of solving the problem

[0007] To achieve the above objective, in a vehicular named data networking (VNDN) environment including an autonomous vehicle according to the present invention, an adaptive data forwarding method for comprehensive infotainment services, in which an autonomous vehicle or a Road Side Unit (RSU) generates or receives a data packet; (b) classifying the data packet according to a class attribute included in the data packet; (c) calculating a current node weight (CNW) for the data packet from parameters for calculating weights for the data packet including the class attribute, and including the corresponding weight value in the weight attribute of the data packet; (d) determining the forwarding order of the data packet according to the calculated current node weight; and, (e) a step of forwarding the data packet according to the determined order, wherein in step (c), the current node weight (CNW) is calculated by weighted sum of normalized values ​​of a combination of parameters including a centrality (NC) parameter whose definition is converted to node centrality or name centrality according to the class attribute according to the classification of step (b), wherein the weight of the centrality term is set as a coefficient such that the sum of the weights of the other parameters is 1.

[0009] The above class attributes can be classified into critical, non-critical, and entertainment.

[0010] In the above step (c), the current node weight (CNW) is the weight (W) of vehicle n for packet i. in As, mathematical formula Calculated by, where RSSI in represents the Received Signal Strength Indication of vehicle n for packet i, and D in represents the directionality of vehicle n with respect to packet i, and V in represents the speed of vehicle n for packet i, and HC in is the number of hops packet i passed through vehicle n, α1, α2, α3, and α4 are coefficients that adjust the importance of each parameter, and NC in ...indicates the Node Centrality of Vehicle n for information packet i when the above class attribute is important or non-important, and the Name Centrality when the above class attribute is entertainment, which represents the number of interest packets received by Vehicle n for entertainment data packet i, and RSSI max , D max , V max , HC max , NC max is, each of the above parameters RSSI in , D in , V in , HC in , NC in It refers to each maximum value for normalization.

[0011] delete

[0012] The above adaptive data forwarding method may further include: (f1) a step of obtaining a signal strength (RSS), GPS coordinates of the current node autonomous vehicle, and a speed value; (f2) a step of calculating the estimated position of the current node autonomous vehicle using an extended Kalman filter (EKF); (f3) a step of obtaining a signal strength value from a new RSU if the estimated position is not in the same category as the existing RSU (Road Side Unit); and (f4) a step of handing over vehicle information and PIT / FIB tables to the new RSU if the signal strength value of the new RSU is greater than the signal strength value of the previous RSU.

[0014] delete Effects of the invention

[0015] According to the present invention, first, various types of data traffic (e.g., emergency safety warnings, traffic information, multimedia content) occurring in an autonomous driving environment are clearly distinguished through class attributes, and by selectively applying a push-based APDF or pull-based EPDF forwarding strategy according to the characteristics of each data type, it is possible to simultaneously achieve rapid dissemination of information essential for autonomous driving safety and efficient provision of high-quality entertainment services.

[0016] Second, through a weight-based forwarder selection method that reflects the dynamic characteristics of the autonomous driving environment within the APDF and EPDF mechanisms, it effectively mitigates the broadcast surge of information and entertainment data packets and minimizes unnecessary redundant data transmission, thereby maximizing the efficiency of network resources.

[0017] Third, compared to existing single forwarding methods or separated service delivery methods, the proposed integrated adaptive forwarding method significantly improves overall network performance, such as reducing data redundancy, reducing network overhead, enhancing scalability, shortening data packet latency, and improving the data delivery ratio and interest satisfaction rate, thereby enhancing the communication reliability and responsiveness of the autonomous driving system and providing an enhanced infotainment experience to passengers.

[0018] In conclusion, the present invention can establish itself as a core technology for future intelligent transportation systems and autonomous driving communication platforms by proposing, for the first time, a key adaptive forwarding method that integrally manages information and entertainment services and provides an optimal forwarding path in a VNDN environment to satisfy the complex requirements for autonomous vehicles to operate safely and efficiently while simultaneously providing rich infotainment services. Brief explanation of the drawing

[0019] FIG. 1 is a diagram showing an example of a VNDN (Vehicular Named Data Networking) environment for infotainment in a smart city. FIG. 2 is a diagram showing an EWFQ (Enhanced Weighted Fair Queuing) scheduling method for infotainment services in a VNDN environment. FIG. 3 is a diagram showing an example of the basic processes of APDF (Advanced Push-based Data Forwarding) and EPDF (Enhanced Pull-based Data Forwarding) in the adaptive forwarding policy of the present invention. FIG. 4 is a diagram illustrating an EKF (Extended Kalman Filter)-based mobility management process in the framework of the present invention. FIG. 5 is a flowchart for providing comprehensive infotainment services in the VNDN of the present invention. FIG. 6 is a diagram illustrating a scenario as an embodiment used to simulate the framework of the present invention in a dense VNDN environment. Figure 7 is a diagram showing the relationship between network load and queuing delay for various schedulers and priorities. FIGS. 8a to 8d are drawings showing a performance comparison between the framework of the present invention applying adaptive forwarding including extended Kalman filter (EKF)-based mobility management, the framework of the present invention applying only adaptive forwarding without EKF-based mobility management, and a simple VNDN scheme. FIGS. 9a to 9d are drawings showing a performance comparison between EPDF and a conventional fragment method in a vehicle speed range of 70 to 100 km / h. FIGS. 10a and 10b are drawings showing a comparison of data forwarding performance between the APDF of the present invention and a conventional method at an average speed of 80 km / h. Figures 11a and 11b show a comparison of performance between the EKF and the conventional method at an average speed of 80 km / h. FIG. 12 is a diagram showing the configuration of a comprehensive infotainment service data forwarding system in a VNDN environment including an autonomous vehicle. Specific details for implementing the invention

[0020] Preferred embodiments of the present invention will be described in detail below with reference to the attached drawings. Prior to this, terms and words used in this specification and claims should not be interpreted as being limited to their ordinary or dictionary meanings, but should be interpreted in a meaning and concept consistent with the technical spirit of the present invention, based on the principle that the inventor can appropriately define the concept of the terms to best describe his invention. Accordingly, the embodiments described in this specification and the configurations illustrated in the drawings are merely one preferred embodiment of the present invention and do not represent all aspects of the technical spirit of the present invention; therefore, it should be understood that various equivalents and modifications capable of replacing them may exist at the time of filing this application.

[0022] FIG. 1 is a diagram showing an example of a VNDN (Vehicular Named Data Networking) environment for infotainment in a smart city, FIG. 2 is a diagram showing an EWFQ (Enhanced Weighted Fair Queuing) scheduling method for infotainment services in a VNDN environment, FIG. 3 is a diagram showing an example of the basic processes of APDF (Advanced Push-based Data Forwarding) and EPDF (Enhanced Pull-based Data Forwarding) in the adaptive forwarding policy of the present invention, and FIG. 4 is a diagram showing an EKF (Extended Kalman Filter)-based mobility management process in the framework of the present invention.

[0023] Hereinafter, with reference to FIGS. 1 to 4, the framework of the present invention for providing a full infotainment service in a VNDN environment will be described. The present invention proposes a comprehensive framework that includes a data packet format, service type classification, an enhanced queuing policy, an adaptive forwarding strategy, and an efficient mobility management method, which are essential for supporting various infotainment services in a VNDN environment, particularly one that includes autonomous vehicles. In particular, the core components include an adaptive delivery strategy that applies push-based and pull-based delivery methods optimized for information data and entertainment data, respectively, in an autonomous driving environment, and an EKF-based mobility management strategy that provides an enhanced handover delay compared to existing vehicle tracking-based mobility management methods.

[0025] 1) VNDN environment

[0026] The VNDN environment consists of Consumers, Producers, Intermediate Vehicles, and Roadside Units (RSUs). A description of each component is given in [Table 1] below.

[0027] individual explanation Consumer In a VNDN environment, a consumer refers to a vehicle that requests data content and generates interest packets to receive that data. Producer A producer is an entity that generates information and entertainment data and distributes it to a network. The producer transmits the generated data packets to consumer vehicles through intermediate vehicles and RSUs. Intermediate Vehicles The intermediate vehicle acts as a forwarder that facilitates multi-hop data exchange between consumers and producers. It receives packets of interest from nearby consumers, forwards them to potential data producers according to the forwarding method, and delivers the data packets to the consumers who requested them. Roadside Unit (RSU) RSUs are connected to the core network backbone and provide direct connectivity to vehicles passing through their coverage area. RSUs serve to transmit data packets from producers to consumers or packets of interest from consumers to producers.

[0028] Figure 1 shows an example of a VNDN environment in an urban smart city. This smart city consists of various vehicles such as buses, police cars, ambulances, taxis, and private vehicles, and is integrated with major facility infrastructure such as two-way roads, highways, hospitals, police stations, schools, offices, and residential buildings. In the smart city, vehicles including autonomous vehicles are interconnected into the VNDN environment through V2X (Vehicle-to-Everything) communication.

[0029] 2) Data packet format

[0030] The data packet format used in the proposed framework plays a key role in effectively supporting various infotainment services, particularly in the VNDN environment of autonomous vehicles. Each data packet contains essential information that guides packet processing, delivery decisions, and the provision of Quality of Service (QoS). The following [Table 2] shows the attributes included in the data packet format and their importance.

[0031] attribute explanation Content Name A unique identifier for content transmitted by data packets. Enables content discovery and routing in VNDN. Size Data packet size in bytes. Indicates the amount of data being transmitted. Class Classification of data packets based on content type (e.g., critical information, non-critical information, entertainment). Provides guidance on prioritization and QoS provision. Weight Weight values ​​calculated based on parameters such as centrality and direction from the previous node. They influence forwarding decisions to select the optimal messenger and more efficiently mitigate data broadcast storms.

[0032] In VNDN, the content name is a critical attribute that uniquely identifies the content delivered by a data packet. The size attribute specifies the size of the data packet in bytes, providing information about the amount of data being transmitted and contributing to the efficient allocation of network resources. The class attribute classifies data packets based on content types, such as critical / non-critical information required for autonomous driving or entertainment data, guiding QoS provision and prioritization to ensure critical data is processed preferentially. The weight attribute is calculated by the previous node and can be used to determine whether the next node should forward the packet. This weight influences forwarding decisions to select the optimal forwarder according to the proposed adaptive forwarding strategy. The combination of these attributes facilitates adaptive forwarding, prioritization, and QoS provision within the VNDN environment, thereby enhancing the efficiency and effectiveness of infotainment services for autonomous vehicles.

[0034] 3) Beacon format

[0035] In the proposed EKF-based mobility management scheme, all vehicles, including RSUs and autonomous vehicles, periodically exchange beacons using the WAVE (Wireless Access in Vehicular Environments) communication channel. These beacons contain attributes such as RSUc (currently connected RSU), RSS (received signal strength), location (GPS coordinates), speed, PIT (Pending Interest Table) / FIB (Forwarding Information Base) entries, and timestamps. Through beacon exchange, nodes can monitor network status and adjust operations accordingly. The beacon format is as shown in [Table 3] below.

[0036] attribute explanation RSUc RSU currently connected to the moving vehicle RSS Received signal strength value measured by the moving vehicle from the RSUc Location GPS coordinates of the moving vehicle Speed Instantaneous speed of a moving vehicle PIT / FIB PIT / FIB items of the previous RSU associated with the vehicle in motion Timestamp Time the beacon was generated

[0037] RSS, GPS coordinates, and speed information can be used for vehicle tracking of autonomous vehicles in the proposed EKF-based method. The RSUc attribute indicates the RSU to which the vehicle is currently connected. The PIT / FIB attribute indicates that the PIT / FIB entry information of the previous RSU associated with the moving vehicle must be shared with the related RSU.

[0039] 4) Classification of Infotainment Services

[0040] The proposed framework classifies infotainment services required for autonomous vehicle operation in a VNDN environment into three classes based on their characteristics: Critical Information, Non-critical Information, and Entertainment. This classification is essential for enabling customized data propagation that meets specific QoS requirements. The following [Table 4] shows examples of data types belonging to each service class.

[0041] Critical non-critical Entertainment Emergency Alerts Traffic Alerts Road Alerts Accidents Alerts Hazardous Warnings Construction Alerts Speed ​​Limit Alerts Lane Closure Alerts Disaster Warnings Security Alerts Health Alerts Navigation weather updates, real-time traffic information, gas station information, traffic signal information, speed limit information Music, Video, Radio, Podcast, Video, Phone (Video call), E-book, News, Social Media, Live, Streaming, Virtual Reality

[0042] - Critical Information: Essential to public safety, requiring highly reliable real-time transmission, particularly for the safe operation of autonomous vehicles. This category includes emergency alerts, traffic warnings, and similar information that can contribute to life protection; it must be delivered with the lowest possible latency. It is given the highest priority in terms of delivery and resource allocation to ensure timely delivery. - Non-critical Information: Helps drivers optimize road navigation efficiency. It includes navigation guidance and real-time traffic updates; classified as medium priority, it requires general delivery services.

[0043] - Entertainment Data: Includes media-rich and bandwidth-intensive applications such as video streaming and social media, designed to enhance the in-vehicle experience. It is assigned the lowest priority to ensure the stable delivery of critical and non-critical information.

[0045] 5) Adaptive Forwarding

[0046] In autonomous vehicle communication, data is transmitted and distributed based on content names in a VNDN environment, differentiating it from existing IP address-based location-dependent methods. In the proposed framework, data packets are delivered efficiently and in a timely manner to intended recipients, such as autonomous vehicles or related infrastructure, through multiple stages.

[0048] - Traffic Classification and Class Queues: The framework classifies various data packets into three class queues (critical information queue, non-critical information queue, and entertainment data queue).

[0049] - Enhanced Weighted Process Queuing (EWFQ): The proposed framework introduces an EWFQ policy for handling packet scheduling in a VNDN environment (see Fig. 2). EWFQ allows for dynamic scheduling and assigns variable weights based on various parameters. For example, upon receiving a data packet, temporal requirements are verified by considering the arrival time, and packets that arrive first have priority in processing and transmission. For spatial validity verification, signal strength is used to determine the geographical location and spatial proximity of adjacent autonomous vehicles / RSUs, with higher signal strength indicating greater proximity. Packets that pass the temporal and spatial verification obtain queuing weights through a fuzzy logic system. Packets with higher weights are processed and transmitted preferentially (see Fig. 2), and the packet delivery order is readjusted according to the calculated weights. That is, the packet with the highest weight W1 is delivered first, followed by W2 and W3. The fuzzy logic system calculates appropriate weights by considering various parameters such as priority, link quality, network density, and arrival time. Link quality controls data flow and reduces packet loss by accurately measuring signal strength, bandwidth availability, and reliability, while network density identifies network load from data flow. Fuzzy logic consists of a fuzzifier (converting input variables into a fuzzy set), an inference engine (using fuzzy rules), and a defuzzifier (converting the fuzzy set into an output). EWFQ is an efficient scheduler that uses a fuzzy logic system to assign more dynamic and accurate weights to service flows compared to existing WFQ and FIFO policies.

[0050] - Advanced Push-based Data Forwarding (APDF): In VNDN, push-based data forwarding refers to the distribution of critical and non-critical information to all vehicles within the network, including autonomous vehicles. It enables information transmission without explicit requests, thereby eliminating the need for PIT items. The proposed framework introduces APDF (Advanced Push-based Data Forwarding) for the distribution of critical and non-critical information to mitigate the broadcast storm effects of information data essential for autonomous driving in the VNDN environment. The APDF approach provides more optimal forwarding by selecting a forwarder using the Current Node's Weight (CNW) and the Previous Node's Weight (PNW), enabling more robust and efficient information distribution to all vehicles in the VNDN environment, including autonomous vehicles. The CNW field represents the weight of the current node, while the PNW field represents the weight of the previous node that facilitated packet delivery. To be selected as the optimal forwarder, each node maintains CNW and PNW attribute tables. Weight W at vehicle n for information i in The value is calculated using the following equation (1).

[0051]

[0052] RSSI here in represents the Received Signal Strength Indication of vehicle n for packet i, and D in represents the directionality of vehicle n with respect to packet i, and V in represents the speed of vehicle n for packet i. HC in represents the number of hops packet i passed through vehicle n, and NC inrepresents the node centrality of vehicle n for information packet i. All parameters are normalized by dividing by the maximum value, and the α coefficients determine the optimal carrier by reflecting the importance of each term.

[0053] When critical or non-critical information packets are generated by producer vehicles, such as autonomous vehicles or RSUs that generate information, the process begins by initializing Time-To-Live (TTL) parameters to limit the data propagation range. The producer vehicle broadcasts the data packet with an initial weight calculated based on parameters such as RSSI, speed, and direction. Each receiving autonomous vehicle checks if it is within the RSU coverage area; if so, the RSU acts as the primary carrier. Otherwise, the vehicle calculates the CNW using the same parameters and compares it to the PNW. Data is delivered only if the CNW of the critical or non-critical information is higher than the PNW; otherwise, the packet is dropped. This selective delivery continues as each carrier decrements the TTL, and until the TTL expires, data reaches vehicles, including all autonomous vehicles, via a more robust and efficient path, effectively mitigating the broadcast storm effect. The basic process of APDF is illustrated in Fig. 3. Critical and non-critical information data is generated and transmitted only by the optimal carriers within the network using the APDF method. By selectively delivering through nodes with higher weights, APDF achieves more efficient push-based data distribution for all autonomous vehicles.

[0054] - Enhanced Pull-based Data Forwarding (EPDF): The proposed framework also presents Enhanced Pull-based Data Forwarding (EPDF) for entertainment data distribution, which is an improved version of the NameCent method. The EPDF method enables the selection of a more optimal forwarder to facilitate better data delivery. While NameCent uses two parameters—the number of packets of interest and the RSSI value—the EPDF method introduces three additional parameters—Speed, Direction, and Hop Count—to select a more robust optimal forwarder. Weight W at vehicle n for entertainment data packet i. i The value can be calculated using the same equation (1) as the APDF described above. However, in the case of EPDF, the NC of equation (1) in The term represents Name Centrality, which indicates the number of Interest packets received at vehicle n for entertainment data packet i. The basic process of the proposed EPDF is demonstrated in Fig. 3. If the CNW of the entertainment content is higher than the PNW, EPDF selects the current node as the optimal carrier. Therefore, in the figure, the vehicle nodes for data / vid01 are maintained as optimal carriers. In EPDF, the TTL value is used to indicate the maximum number of hops a data packet can pass before being dropped.

[0056] 6) Mobility management support

[0057] Mobility management is an essential element for ensuring smooth handovers and reliable QoS, particularly in VNDN environments, due to the dynamic vehicle movements of autonomous vehicles. The proposed framework utilizes an Extended Kalman Filter (EKF) to provide a robust solution for smoother mobility coordination of autonomous vehicles. The EKF-based mobility management strategy can be applied to both consumer and producer autonomous vehicles, and uses signal strength, speed, direction, and GPS input to estimate real-time location and manage handovers based on the movement of the autonomous vehicle (see Fig. 4).

[0058] Compared to existing research, the proposed EKF is a real-time positioning technology that dynamically estimates the location of autonomous vehicles with greater accuracy by utilizing a Kalman model. By providing accurate real-time location information, EKF enables more reliable data retrieval during autonomous vehicle movement in VNDN environments. The proposed EKF technology recursively processes measurements collected during movement to accurately track the location of consumer / producer autonomous vehicles and predict the next target RSU. This provides better, seamless mobility support while minimizing overhead. In contrast to TNS algorithms, which rely on geographic guidance and are limited by the length of Tabu lists, EKF technology provides more precise location estimates to track the movement of autonomous vehicles and more accurately predict future RSU connections. This enables preemptive handovers to minimize service interruptions during the movement of consumer / producer autonomous vehicles.

[0060] - Position Estimation Using EKF: According to Figure 4, the process begins by initializing the EKF algorithm with the autonomous vehicle's current position. When new values, such as GPS coordinates, speed, direction, and signal strength, are received, they are compared with the predicted values. Based on this comparison, the Kalman gain is calculated, which determines the amount of adjustment required for the prediction according to the new values. The Kalman gain is a key parameter of the EKF algorithm that determines how much the filter trusts the measured value versus the predicted value. The adjusted prediction is combined with the newly received values ​​to obtain an updated estimate of the autonomous vehicle's position. This updated estimate serves as a new starting point for the next prediction. By repeatedly performing these steps, the system can accurately track the position of the autonomous vehicle over time.

[0061] The EKF algorithm includes three main steps—Prediction, Measurement Update, and State Estimate—to accurately estimate the position of the autonomous vehicle and predict the next target RSU.

[0062] a) Step 1 (Prediction): In this step, the predicted state estimate and the predicted state covariance matrix from the previous time step are propagated to the current time step using a system dynamics model.

[0063]

[0064] Predicted state estimate here is the previous state estimate of the non-linear state transition function f(·). and control input It is obtained by applying to. Predicted state covariance matrix ε is the Jacobian matrix F representing the partial derivatives of f(·) with respect to the state variables. and previous state covariance matrix and process noise covariance matrix It is calculated using .

[0065] b) Step 2 (Measure Update): In this step, the predicted state estimates and the predicted state covariance matrix are the current measurements It is updated using .

[0066]

[0067] Only the knife is advantageous here. is the predicted state covariance matrix , Jacobian matrix representing the partial derivatives of the measurement function h(·) with respect to the state variables , measurement noise covariance matrix and is calculated using the measurement update equation. Updated state estimate is the predicted state estimate, Kalman gain, and measurement It is obtained by combining the difference between the predicted state estimate and the evaluated measurement function value. Updated state covariance matrix It is obtained by adjusting the predicted state covariance matrix using Kalman gain and Jacobi matrix.

[0068] c) Step 3 (State Estimation): The state estimate at time step k is the updated state estimate after the measurement update step. It is obtained as.

[0069]

[0070] The EKF algorithm estimates the position of the autonomous vehicle by repeating the three steps above and predicts the next RSU based on the measurements. This algorithm incorporates the system's nonlinearity through state transition and measurement functions, and accounts for uncertainties in measurement and system dynamics using process noise and measurement noise covariance matrices. It should be noted that specific implementation details (state vector selection, derivation of state transition and measurement functions, noise covariance matrix tuning, etc.) may vary depending on the specific requirements and characteristics of the VNDN environment under consideration.

[0072] - Handover Management: The EKF algorithm is used to estimate the location of autonomous vehicles and facilitate the handover management process in the VNDN environment. The decision to initiate a handover is made based on predefined criteria, such as the Signal-to-Interference-plus-Noise Ratio (SINR) threshold. If these criteria indicate that a handover is necessary to maintain a seamless connection with the new RSU, the process is initiated preemptively based on the location estimation provided by EKF.

[0073] The EKF-based mobility management strategy ensures reliable data delivery during the movement of autonomous vehicles. For example, when an RSU receives a packet of interest from a consumer (autonomous vehicle), it uses an EKF model to track the consumer's location. As the consumer moves, the EKF continuously updates its estimated location. This tracking process also applies when the producer (autonomous vehicle or information source) changes its location. In the reverse path of data delivery, if the consumer (autonomous vehicle) is within the initial RSU's range, the data packet is delivered directly. However, if the consumer (autonomous vehicle) moves out of range, the PIT entry associated with the consumer is shared with the next expected RSU based on the predicted vehicle location. This ensures that the data packet is delivered correctly using the EKF location input, even if the consumer (autonomous vehicle) moves to a new RSU. Similarly, if a producer (autonomous vehicle or information source) moves to a new RSU, that RSU updates the Forwarding Information Base (FIB) entry related to that producer (autonomous vehicle or information source).

[0074] PIT items are shared with the next target RSU via beacons. Each RSU generates a Bloom filter that conveys a summary of the PIT items. This Bloom filter is embedded in the beacon marked as the PIT and broadcast to adjacent RSUs. The target RSU receives the beacon, extracts the Bloom filter, and updates the PIT items associated with that consumer. The update process involves comparing the target RSU's local PIT items with the new PIT items using existing data structures, such as hash tables. By utilizing the Bloom filter as a probabilistic data structure, potential matches can be efficiently identified and updated in the local PIT items. This approach facilitates efficient data delivery based on the predicted locations of consumers and producers (autonomous vehicles or related nodes) while preserving NDN principles.

[0076] FIG. 5 is a flowchart for providing comprehensive infotainment services in the VNDN of the present invention.

[0077] Figure 5 shows the workflow of the proposed framework, which focuses on adaptive forwarding and mobility management in a VNDN environment. The workflow consists of two main segments: (a) adaptive forwarding and (b) mobility management support.

[0078] The proposed framework uses an adaptive forwarding strategy that considers the priority and type of each data packet in an autonomous driving environment. When a new packet arrives (S110), it is classified as important information, non-important information, or entertainment data and placed into the important information queue, non-important information queue, or entertainment data packet queue, respectively (S120). Subsequently, an appropriate weight is assigned to each packet by the EWFQ scheduler (S130), and the output order is determined based on the weight. Important information receives the highest priority. Each data packet waits in the queue in the order of the weight assigned by the EWFQ scheduler.

[0079] For important and non-important information related to autonomous driving safety, the APDF method is invoked to determine the optimal deliverer based on direction, speed, node centrality, etc. (S150, S160). For entertainment content, data delivery and content search are processed using the EPDF method (S150, S170). Subsequently, data packets determined to be forwarded are forwarded (S180).

[0081] The proposed framework addresses the management of vehicle mobility for autonomous vehicles. When a consumer autonomous vehicle acquires RSS values, GPS coordinates, and speed values ​​(S210) and moves away from the current RSU to near a new RSU, an Extended Kalman Filter (EKF)-based position estimation technique is invoked (S220). This technique utilizes the consumer autonomous vehicle's previous location to estimate the expected proximity between the current location and the new RSU (S230, S240, S250). This information triggers a handover request to the new RSU (S260). During the handover process, the new expected RSU updates the PIT items associated with the consumer autonomous vehicle (S270). This update is facilitated by a Bloom filter procedure that summarizes the PIT items. A Bloom filter embedded in the beacon is used to ensure that the new RSU appropriately updates the PIT items. Similarly, when a producer (autonomous vehicle or information source) moves to the new RSU, the EKF estimates its location. Based on this estimated location, the new RSU updates the FIB entry related to the corresponding producer (autonomous vehicle or information source) (S270)

[0082] .

[0083] The proposed framework makes final decisions regarding data delivery using adaptive forwarding and EKF-based mobility management. This parallel procedure enables the selection of a more optimal deliverer based on the forwarding mechanism to reflect mobility support for autonomous vehicles, thereby more effectively ensuring the mitigation of redundant infotainment data in VNDN environments.

[0085] Below, the computational complexity and network overhead of the framework of the present invention described above are analyzed in a dense autonomous vehicle environment.

[0087] 1) Computational Complexity Analysis

[0088] This framework provides a comprehensive computational complexity analysis for each algorithm used, namely EWFQ, Adaptive Forwarding (APDF / EPDF), and EKF.

[0089] The computational complexity of the proposed EWFQ mechanism consists of two main stages: dynamic weight calculation using fuzzy logic and the scheduling process. The dynamic calculation of weights depends on the number of input variables and fuzzy rules. If the number of input variables is m and the number of fuzzy rules is r, the computational complexity of the fuzzy rules is O(mr). The computational complexity of the scheduling process is O(log f), where f represents the number of flows. Since m and r are constant values, the computational complexity of EWFQ can be simplified to O(log f) per packet, indicating that it increases logarithmically with respect to the number of flows f.

[0090] The computational complexity of the proposed adaptive forwarding strategy is analyzed for both APDF and EPDF methods. Although the APDF method handles push-based data forwarding and the EPDF method is performed during pull-based data forwarding, both methods have similar computational complexity characteristics. To calculate the weight for each packet at the node / autonomous vehicle, various parameters such as RSS, speed, direction, hop count, and node centrality / name centrality are considered. The complexity of calculating the weight value is O(v) depending on the number of autonomous vehicles (v). For optimal decision-making, the adaptive forwarding strategy considers the weight of the previous node for each packet. Therefore, the computational complexity of the adaptive forwarding strategy is O(vp), where p is the number of packets. Thus, the overall adaptive forwarding complexity for both APDF and EPDF methods is O(v+vp), which increases linearly with the number of autonomous vehicles v and the number of packets p.

[0091] To analyze the computational complexity of the proposed EKF algorithm, we consider the three stages of prediction, measurement update, and state estimation.

[0092] The prediction step of the EKF involves computing the predicted state estimates and the predicted state covariance matrix. Assuming the dimension of the state vector is n and the complexity of the state transition function is constant, the computational complexity of the predicted state estimates is O(n). The computational complexity of the predicted state covariance matrix is ​​O(n) due to matrix multiplication operations. 2 ) is. Therefore, the total computational complexity of the prediction step is O(n 2 )am.

[0093] The measurement update step of the EKF calculates three values: the Kalman gain, the updated state estimate, and the updated state covariance matrix. The calculation of the Kalman gain involves matrix inversion and multiplication, and when the dimension of the state vector is n, the complexity is O(n²). 3The complexity of the updated state estimate and the updated state covariance matrix is ​​O(n) and O(n), respectively. 2 ) is. Therefore, the computational complexity of the measurement update step is O(n²) which is dominated by the Kalman gain calculation. 3 )am.

[0094] The state estimation step involves a simple assignment operation, and its time complexity is constant O(1). Therefore, the total computational complexity of the EKF algorithm is O(n) in each iteration. 3 It can be simplified to ), where n is the dimension of the state vector.

[0095] Overall, the computational complexity of this framework is O(log f + v + vp + n 3 It can be expressed as ), and the dominant factor is O(n 3 )am.

[0097] 2) Network Overhead Analysis

[0098] In VNDN, network overhead refers to additional data transmitted within the network in addition to the actual desired data. This overhead includes redundant data packets and beacon packets, which are essential to VNDN framework functions but consume network resources. Mathematically, network overhead NO can be expressed as follows.

[0099]

[0100] Here, D a is the actual number of data packets, D r is the number of duplicate data packets, D b represents the number of beacon packets. While actual data packets carry the desired information content, duplicate data packets are copies of the actual data packets. Beacon packets are small control packets that carry essential information for network management and coordination, such as RSS, GPS coordinates, autonomous vehicle speed, and PIT / FIB entries.

[0101] The actual data packet size is Sa , the size of the beacon packet is S b In that case, network overhead can be expressed as follows.

[0102]

[0103] Packet size plays a crucial role in determining communication overhead and network overhead. As packet sizes increase, overhead increases because more network resources are required for transmission and processing. This framework aims to minimize network overhead by reducing the amount of redundant data transmitted within the network. Adaptive Forwarding (APDF and EPDF) methods select the optimal carrier and redundant data packets (D r Reduces overall network overhead by minimizing the distribution of ). Considering the minimized number of redundant data packets, the network overhead of this framework can be expressed as follows.

[0104]

[0105] Here, D' r represents the number of reduced redundant data packets, and D' r ≪D r am.

[0106] Generally, since beacon packets transmit only important information, their size is very small compared to actual data packets (S a >>>> S b ) can be ignored. Therefore, the total network overhead of this framework can be approximated as follows.

[0107]

[0108] By minimizing the number of redundant data packets transmitted, this framework effectively reduces network overhead.

[0110] 3) Comparison with existing representative methods

[0111] When comparing the proposed framework with existing representative approaches used in VNDN, the computational complexity of the proposed framework is O(n 3 ) is O(n) of the conventionally presented method 2 It is higher than the complexity. However, it should be noted that the increased computational complexity is a trade-off for the enhanced functionality and performance provided by the framework. By integrating advanced technologies such as EWFQ, APDF / EPDF, and EKF into queuing, data forwarding, and mobility management, it provides more efficient and reliable data distribution in autonomous driving VNDN environments.

[0112] On the other hand, the proposed framework addresses network overhead By reducing the conventionally presented network overhead It shows a significant reduction compared to. This framework significantly reduces communication and network overhead by ultimately minimizing the distribution of redundant data packets within the network through the selection of the optimal carrier using input parameters such as signal strength, autonomous vehicle speed, direction, hop count, and node centrality / name centrality. Figures 9a and 10a illustrate the reduction of redundant data packet copies using EPDF and APDF schemes, respectively. The minimization of redundant data packets leads to a significant reduction in network overhead, which is illustrated in Figure 11a by an example that supports the mobility of autonomous vehicles while mitigating unnecessary data transmissions using an EKF-based mobility management scheme. While the network overhead reduction effect by this framework increases computational complexity, it brings improved performance and scalability in an autonomous vehicle VNDN environment.

[0114] FIG. 6 is a diagram illustrating a scenario as an embodiment used to simulate the framework of the present invention in a dense VNDN environment.

[0115] The simulation was performed in the ndnSIM simulator. This simulator implements NDN protocols and networks over real-world topologies. This simulation was conducted to evaluate the performance of the proposed framework in a VNDN environment and to compare it with existing methods. A mesh topology, a 5,000-meter bidirectional highway, and nine RSUs equipped with IEEE 802.11p radio technology (WAVE) were considered. The simulation utilized the Simulation of Urban Mobility (SUMO) traffic simulator within the simulation environment to generate realistic autonomous vehicle movement patterns by randomly distributing the locations of the autonomous vehicles. The simulation scenarios more accurately reproduce actual traffic situations, and Figure 6, which captures the dynamic and heterogeneous movements of autonomous vehicles in an urban environment, conceptually illustrates the arrangement and relationships between the various components used in the simulation.

[0116] The density of autonomous vehicles varied from 10 to 200, consisting of consumer autonomous vehicles (25%), intermediate autonomous vehicles (50%), and content-producing autonomous vehicles (25%). The autonomous vehicles moved according to a random waypoint movement model, and speeds varied from 1 m / s to 100 m / s to simulate a dynamic city scenario. Each RSU periodically exchanged beacons with nearby connected autonomous vehicles and other RSUs to share useful information. In each iteration, each consumer autonomous vehicle generated 0 to 4 interest packets and propagated them across the network to retrieve data packets containing necessary content. Each RSU was equipped with network caching capabilities to store temporary copies of returned data packets in its local RSU cache. The buffer size was set to 100 packets, and the maximum, minimum, and optimal thresholds were set to 80, 20, and 50 packets, respectively. Network load varied between 0.3 and 0.9, the simulation duration was set to 500 seconds, and a 50-second warm-up period was included. Monte Carlo simulations were performed by repeating random seed values ​​1,000 times for 1,000 iterations, and the results were verified within a 95% confidence interval to ensure statistical reliability. The performance of the proposed framework was evaluated using various metrics (interest fulfillment rate, mean end-to-end latency, network overhead, and data transfer rate). This was conducted under various autonomous vehicle densities, interest rates, and movement patterns, and the results were compared with existing segmentation schemes to demonstrate that the proposed framework is effective for improving overall network performance and efficient mobility management in dense autonomous vehicle VNDN environments. The simulation parameters and their values ​​are presented in Table 5.

[0117] parameters value simulator nbnSIM Traffic simulator SUMO Topology Mesh Coverage area 5000 m RSU number 9 Radio Technology (WAVE) IEEE 802.11p RSU range 300m Number of vehicles 10~200 Vehicle speed 1~100 m / s Vehicle location Random Mobile model Random Waypoint Interest rate 0~4 interest / consumer Buffer size 100 packets maximum threshold 80 packets minimum threshold 20 packets Appropriate threshold 50 packets Network load 0.3~0.9 Simulation time 500 seconds Warm-up time 50 seconds confidence interval 95% Number of repetitions 1000

[0118] Figure 7 is a diagram showing the relationship between network load and queuing delay for various schedulers and priorities.

[0119] Figure 7 shows the average wait latency of EWFQ, WFQ, and FIFO schedulers under increasing network load. In the EWFQ scheme, three priority classes—critical, non-critical, and entertainment—are considered. The proposed EWFQ achieves the lowest latency at all loads compared to WFQ and FIFO for all data types in an autonomous driving environment. The main conclusion is that EWFQ optimizes the wait latency of critical, non-critical, and entertainment data through priority scheduling.

[0121] FIGS. 8a to 8d are drawings showing a performance comparison between the framework of the present invention with adaptive forwarding including extended Kalman filter (EKF)-based mobility management, the framework of the present invention with only adaptive forwarding without EKF-based mobility management, and a simple VNDN scheme.

[0122] Figures 8a through 8d attempt to verify the potential of the framework in a VNDN environment. Figure 8a shows the average CDPP according to the number of autonomous vehicles, categorized by critical, non-critical, and entertainment data types. The CDPP (Copies of Data Packet Processed) metric measures the number of duplicate copies of data packets processed by autonomous vehicles while they are transmitted over a network. For all data types, adaptive delivery combined with EKF-based mobility management significantly improves CDPP, gradually increasing from 22 to 1,736 as the density of autonomous vehicles increases from 10 to 200. In contrast, when adaptive delivery is used without EKF-based mobility management, critical data increases rapidly from 42 to 3,255, non-critical data from 45 to 3,472, and entertainment data from 50 to 3,906.

[0123] Figure 8b shows that the average CDPP increases with the number of packets of interest. For critical data, the CDPP increases from 7 to 183 when only adaptive forwarding is applied, but decreases for all data types when adaptive forwarding with an EKF-based mobility management strategy is applied, as it optimizes forwarding through EKF tracing. For example, with 5 packets of interest, the CDPP for critical data is 37 without EKF, but decreases to 13 when EKF support is used.

[0124] Figure 8c shows the average latency according to the number of autonomous vehicles. Average latency refers to the time required for a data packet to be transmitted from a producer (autonomous vehicle or information source) to a consumer (autonomous vehicle) in a VNDN environment. When only adaptive forwarding for critical data is used, the latency increases from 2ms to 240ms. On the other hand, adaptive forwarding with an EKF-based mobility management strategy reduces the latency from 1.6ms to 111ms. Optimizing forwarding through EKF tracking significantly reduces latency regardless of the data type. For example, with 100 autonomous vehicles, the latency for critical data is 79ms without EKF support, but it decreases to 45ms with EKF optimization. The main reason is that EKF enables accurate tracking of autonomous vehicles. Adaptive forwarding with EKF optimizes data broadcast storm mitigation by transmitting data packets through a more optimized path. On the other hand, simple VNDN expansion allows only pool-based forwarding without mobility awareness, causing data broadcast storms and resulting in higher latency.

[0125] Figure 8d shows the results of improving latency for all data types with increasing distance through an adaptive delivery strategy combined with EKF mobility management. For example, at 2,000 meters, the latency for critical data is 104ms without EKF support, but decreases to 39ms with EKF support. In contrast, the naive VNDN method exhibits a high latency of 256ms at a distance of 2,000 meters.

[0126] As can be seen from the results in Figures 8a to 8d, integrating adaptive forwarding and EKF-based mobility management is essential for more efficient and reliable data distribution in a dynamic autonomous driving VNDN environment.

[0128] Figures 9a to 9d show a comparison of performance between EPDF and a conventional fragment method in a vehicle speed range of 70 to 100 km / h.

[0129] Figures 9a to 9d each show the key performance indicators of the EPDF method of the present invention compared with the existing methods NameCent, CODIE, and simple VNDN.

[0130] Figure 9a shows CDPP values ​​as network density increases. The EPDF method of the present invention consistently shows lower CDPP values ​​compared to other methods, which means more efficient data transmission.

[0131] Figure 9b shows the Interest Satisfaction Rate (ISR), which represents the ratio of successfully transmitted data packets among data transmission requests. EPDF consistently achieves a higher ISR, demonstrating a superior data transmission success rate.

[0132] Figure 9c shows the data transmission rate (DDR). This represents the ratio of transmitted data packets. EPDF consistently delivers superior performance compared to other methods by better mitigating data broadcast congestion through the selection of the optimal path and the selection of a more robust optimal carrier.

[0133] Figure 9d shows the Data Packet Latency (DPL). This represents the average time taken to transmit a data packet. EPDF consistently shows a lower latency (Data Packet Latency) and represents the average time taken to transmit a data packet. EPDF consistently shows a lower latency, ensuring timely delivery of data.

[0134] Overall, Figures 9a to 9d highlight the efficiency and effectiveness of our EPDF method in a VNDN environment by consistently demonstrating superior performance compared to other methods in terms of CDPP, ISR, DDR, and DPL.

[0136] FIGS. 10a and FIGS. 10b are diagrams showing a comparison of data forwarding performance between the APDF of the present invention and a conventional method at an average speed of 80 km / h.

[0137] Figure 10a shows the results of comparing the CDPP values ​​of three data transmission methods (proposed APDF, fuzzy logic, and simple VNDN). The results are displayed as network density increases, with an average autonomous vehicle speed of 80 km / h. The proposed APDF method achieves significantly lower CDPP values ​​compared to fuzzy logic and simple VNDN methods. This is because the proposed APDF optimizes the selection of the next carrier.

[0138] On the other hand, the fuzzy logic method uses a cluster head, which is prone to failure due to the movement of autonomous vehicles. It achieved a significantly lower CDPP value compared to the VNDN method. This is because the proposed APDF selects the optimal next messenger, and only the optimal messenger broadcasts information data. Conversely, the fuzzy logic method uses a cluster head, which has a high probability of failure due to the movement of autonomous vehicles. By selecting the optimal messenger, APDF ensures that the network is not overloaded with unnecessary broadcast messages, thereby realizing more efficient and reliable data distribution.

[0139] Figure 10b compares the efficiency of the proposed APDF, fuzzy logic, and simple VNDN schemes under increasing network densities. Efficiency is defined as the proportion of vehicles that successfully receive transmitted information data. The results show that the proposed APDF achieves higher efficiency than the other two schemes at all network densities. The improved efficiency of APDF stems from the optimal selection of the next forwarder and adaptation to changes in network density. This provides higher durability in a highly dynamic VNDN environment.

[0140] Overall, Figures 10a and 10b show that the proposed APDF method is excellent at minimizing redundant transmission and achieving higher efficiency.

[0142] Figures 11a and 11b are drawings showing a performance comparison between the EKF and the conventional method at an average speed of 80 km / h.

[0143] The results shown in Figures 11a and 11b were obtained by considering mobility management for pull-based traffic in VNDN. The proposed framework utilized EKF-based mobility management in conjunction with EPDF, while TNS and MMV methods used their respective mobility management methods and data delivery mechanisms. The simple VNDN method used a basic data delivery method without a mobility management mechanism.

[0144] Figure 11a illustrates network overhead that varies with network density. The average speed of the autonomous vehicle was assumed to be 80 km / h. Network overhead is defined as the total number of packets generated and transmitted during the network expansion process. The results show that the proposed EKF method generates lower network overhead during autonomous vehicle movement compared to other methods. The performance improvement of EKF is attributed to improved accuracy in autonomous vehicle location prediction and increased handover speeds between RSUs. This minimizes the number of data packets due to improved mobility management. In contrast, the high network overhead of TNS and MMV indicates frequent handovers and connection loss. This reduction is attributed to the utilization of a modified data packet format that includes attributes such as class and weight. Periodic beacon exchange eliminates unnecessary data packet transmissions during autonomous vehicle movement, enabling more efficient data delivery decisions.

[0145] Figure 11b compares the absorption of the proposed EKF with existing mobility management methods such as TNS, MMV, and simple VNDN. The results show that the proposed EKF method records the lowest absorption as network density increases. By accurately tracking the location of autonomous vehicles, the proposed EKF minimizes absorption by sharing PITs or FIBs with the next RSU. In contrast, the inaccurate location prediction and slow handover of existing methods result in unoptimized routing and longer paths. Overall, the proposed EKF minimizes absorption by sharing PITs or FIBs with the next RSU.

[0146] In contrast, the inaccurate location prediction and slow handover of the existing method result in unoptimized and longer paths. Overall, Figures 11a and 11b show that the proposed EKF minimizes data packet redundancy through accurate prediction of the autonomous vehicle's location and seamless handover between RSUs. This reduces network overhead and the number of hops.

[0148] This invention presents the first comprehensive framework that enables holistic information entertainment services in a highly dynamic autonomous driving VNDN environment. Class attributes are inserted into data packet formats to classify data into critical, non-critical, and entertainment categories. An enhanced weighted process queuing (EWFQ) scheduler is proposed to meet various QoS requirements. EWFQ scheduling improves the latency and reliability of critical information data. An adaptive push / pull distribution strategy for selective delivery based on content type is proposed for the first time. Active mobility prediction utilizing an extended Kalman filter (EKF) is proposed to track the locations of consumer and producer autonomous vehicles to provide better fault-free handover. This maintains seamless connectivity while the consumer and producer autonomous vehicles are in motion. Simulation results showed that the proposed framework offers advantages over existing fragmented VNDN methods. Compared to existing partitioned VNDN methods, it demonstrated superior performance in terms of reduced redundant data, reduced overhead, improved data delivery, and reduced latency.

[0150] FIG. 12 is a diagram showing the configuration of a comprehensive infotainment service data forwarding system (100) in a VNDN environment including an autonomous vehicle.

[0151] A comprehensive infotainment service data forwarding system (100) in a VNDN environment including an autonomous vehicle comprises a processor (110), a non-volatile storage unit (120) for storing programs and data, a volatile memory (130) for storing programs currently running, a communication unit (140) for communicating with an external device (300), and a bus which is an internal communication channel between these devices. Programs currently running may include device drivers, operating systems, and various applications. Although not illustrated, the comprehensive infotainment service data forwarding system (100) in a VNDN environment including an autonomous vehicle may include a power supply unit such as a battery.

[0152] The VNDN environment infotainment service data forwarding application (210) is a program installed and operated in a comprehensive infotainment service data forwarding system (100) in a VNDN environment including an autonomous vehicle, and performs data forwarding for a comprehensive infotainment service in a VNDN environment including an autonomous vehicle as described above with reference to FIGS. 1 to 11. Explanation of the symbols

[0153] 100: Comprehensive infotainment service data forwarding system in a VNDN environment including autonomous vehicles

Claims

Claim 1 A method for an autonomous vehicle or a Road Side Unit (RSU) to perform adaptive data forwarding for comprehensive infotainment services in a vehicular named data networking (VNDN) environment including an autonomous vehicle, comprising: (a) the autonomous vehicle or the Road Side Unit (RSU) generating or receiving a data packet; (b) classifying the data packet according to a class attribute included in the data packet; (c) calculating a current node weight (CNW) for the data packet from parameters for calculating weights for the data packet including the class attribute, and including the corresponding weight value in the weight attribute of the data packet; (d) determining the forwarding order of the data packet according to the calculated current node weight; and, (e) a step of forwarding the data packets according to the determined order, wherein in step (c), the current node weight (CNW) is calculated by weighted sum of normalized values ​​of a combination of parameters including a centrality (NC) parameter whose definition is converted to node centrality or name centrality according to the class attribute according to the classification of step (b), and the weight of the centrality term is set as a coefficient such that the sum of the weights of other parameters is 1. In a vehicular named data networking (VNDN) environment including an autonomous vehicle, the autonomous vehicle or the Road Side Unit (RSU) performs adaptive data forwarding for comprehensive infotainment services. Claim 2 delete Claim 3 A method for an autonomous vehicle or a Road Side Unit (RSU) to perform adaptive data forwarding for comprehensive infotainment services in a vehicular named data networking (VNDN) environment including an autonomous vehicle, wherein, in claim 1, the class attributes are classified into critical, non-critical, and entertainment. Claim 4 In claim 3, in step (c), the current node weight (CNW) is the weight (W) of vehicle n for packet i. in As, mathematical formula Calculated by, where RSSI in represents the Received Signal Strength Indication of vehicle n for packet i, and D in represents the directionality of vehicle n with respect to packet i, and V in represents the speed of vehicle n for packet i, and HC in is the number of hops packet i passed through vehicle n, α1, α2, α3, and α4 are coefficients that adjust the importance of each parameter, and NC in ...indicates the Node Centrality of Vehicle n for information packet i when the above class attribute is important or non-important, and the Name Centrality when the above class attribute is entertainment, which represents the number of interest packets received by Vehicle n for entertainment data packet i, and RSSI max , D max , V max , HC max , NC max is, each of the above parameters RSSI in , D in , V in , HC in , NC in A method for an autonomous vehicle or a Road Side Unit (RSU) to perform adaptive data forwarding for comprehensive infotainment services in a vehicular named data networking (VNDN) environment including an autonomous vehicle, characterized by each representing a maximum value for normalization. Claim 5 An adaptive data forwarding method for comprehensive infotainment services in a VNDN environment including an autonomous vehicle, characterized in that, in claim 1, it further comprises: (f1) a step of obtaining a signal strength (RSS), GPS coordinates of the current node autonomous vehicle, and a speed value; (f2) a step of calculating the estimated position of the current node autonomous vehicle using an extended Kalman filter (EKF); (f3) a step of obtaining a signal strength value from a new RSU if the estimated position is not in the same category as the existing RSU (Road Side Unit); and (f4) a step of handover the vehicle information and PIT / FIB table to the new RSU if the signal strength value of the new RSU is greater than the signal strength value of the previous RSU. Claim 6 delete Claim 7 delete

Images