A security protection method and system based on multi-level automobile Internet of Things

Abstract

The present invention discloses a security protection method and system based on a multi-level automobile Internet of Things, which belongs to the field of automobile security protection, and specifically includes: collecting and encrypting vehicle and environmental data through an encrypted sensor network, and using quantum encryption communication to ensure the safe transmission of data to the local; the local processing unit uses a reinforcement learning AI decryption algorithm to process data, and combines machine learning to identify driving risks; based on a risk assessment model fused with multi-source information, the system assesses risks and optimizes and generates protection measures strategies; when encountering risks, the Internet of Things technology immediately triggers an early warning and executes protection. At the same time, the vehicle data is uploaded to the cloud in real time, a security audit mechanism is established, and comprehensive monitoring and recording are carried out to ensure vehicle driving safety and data privacy protection.

Description

Technical Field

[0001] The present invention relates to the field of automobile safety protection, and specifically to a safety protection method and system based on a multi-level automobile Internet of Things. Background Art

[0002] With the rapid development of the Internet of Things (IoT), cars are no longer isolated means of transportation, but are gradually integrated into a more complex network environment. IoT technology enables cars to communicate with other vehicles, infrastructure, and cloud services in real time, thus achieving more intelligent and interconnected functions. However, this also brings new security challenges, such as cyber attacks and data leakage. Although there are many types of automotive safety protection systems on the market, most of them have single functions and limited protection effects.

[0003] For example, the patent with publication number CN108657073A discloses a car safety protection system based on the Internet of Things, including: installing a controller inside the vehicle, connecting the controller to the vehicle-mounted display, the controller is also connected to a monitoring mobile phone A for sending and receiving monitoring information, the owner of the vehicle holds a monitoring mobile phone B for receiving monitoring information sent by the monitoring mobile phone A, the vehicle-mounted display and the monitoring mobile phone A receive monitoring information sent from the controller and / or the Bluetooth information transmitter, the Bluetooth information transmitter is a Bluetooth connected to a monitoring camera and an infrared detector or a Bluetooth connected to the controller, the monitoring camera and the infrared detector are respectively arranged inside and / or outside the vehicle, the monitoring camera and the infrared detector are also respectively arranged on a drone, the drone remote control is connected to the controller, the vehicle is provided with a drone sending and receiving window, and the vehicle is also provided with vibration, smoke, sensors and sensors and face recognition processors respectively connected to the controller.

[0004] The above existing technologies have the following problems: although they cover the application of vehicle internal monitoring, wireless communication and drones, they are relatively basic in data processing, encrypted communication and intelligent analysis; they rely on traditional monitoring equipment and human judgment to identify potential threats, which makes risk assessment and protection measures lack intelligence; although the connection between the controller and the vehicle display, A monitoring phone and B monitoring phone is mentioned, it does not explicitly mention the integration with the cloud server and the cloud security audit mechanism, which makes it insufficient in real-time data uploading, cloud processing, security monitoring and recording. Summary of the invention

[0005] In view of the shortcomings of the existing technology, the present invention proposes a security protection method and system based on a multi-level automotive Internet of Things. The method collects and encrypts vehicle and environmental data through an encrypted sensor network, and uses quantum encryption communication to ensure the secure transmission of data to the local. The local processing unit uses a reinforcement learning AI decryption algorithm to process data, combined with machine learning to identify driving risks. Based on a risk assessment model fused with multi-source information, the system assesses risks and optimizes and generates protection measures strategies. When encountering risks, the Internet of Things technology immediately triggers an early warning and executes protection. At the same time, the vehicle data is uploaded to the cloud in real time, a security audit mechanism is established, and comprehensive monitoring and recording are carried out to ensure vehicle driving safety and data privacy protection.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A security protection method based on a multi-level automotive Internet of Things, comprising:

[0008] Step S1: using an encrypted sensor network to collect vehicle status information and environmental data, obtaining encrypted data, and using a quantum encryption communication protocol to transmit the collected encrypted data to a local processing unit;

[0009] Step S2: In the local processing unit, the encrypted data is processed and decrypted using an AI decryption algorithm based on reinforcement learning, and based on the decrypted data, driving risks and safety threats are identified through a machine learning model;

[0010] Step S3: Based on the driving risks and safety threats identified in step S2, combined with current driving behavior, vehicle status, road condition information and regional data shared in the cloud, a pre-built risk assessment model is used to conduct a risk assessment, and based on the risk assessment results, an optimization algorithm is used to formulate an optimal protection measure strategy;

[0011] Step S4: When the system detects a risk, based on the risk assessment results and the optimal protective measures strategy developed, the corresponding level of early warning mechanism is automatically triggered through the Internet of Things method, and protective measures are automatically executed according to the triggered early warning level and type;

[0012] Step S5: Upload the vehicle status information and driving behavior data to the cloud server in real time. At the same time, establish a cloud security audit mechanism to monitor and record the uploaded and shared vehicle status information and driving behavior data.

[0013] Specifically, the specific steps of step S2 include:

[0014] S2.1: Initialize the smart scheduler based on the received encrypted data Request, combined with the load of the local processing unit, data type priority, and resource availability factors, intelligent scheduling decision is made. The intelligent scheduling decision is expressed as:

[0015]

[0016] Among them, A(r) represents the intelligent scheduling decision, Represents the Nth encrypted vehicle status information, represents the Mth encrypted environmental data, N represents the number of encrypted vehicle status information, M represents the number of encrypted environmental data, P(r) represents the priority evaluation function, r represents the current attribute set of the encrypted data request, q1 represents the amount of allocated resources, g1(r) represents the type of encrypted data, g2(r) represents the urgency, g3(r) represents the size of the encrypted data, α, β, δ represent weight coefficients;

[0017] S2.2: According to the intelligent scheduling decision result, the local processing unit receives the encrypted data from the encrypted sensor network through a communication connection and stores the encrypted data in a storage medium.

[0018] Specifically, the specific steps of step S2 also include:

[0019] S2.3: Define the state space S = {s1, s2, s3} and action space D = {d1, d2} of the decryption algorithm, and build a reinforcement learning training framework based on the defined state space and action space, where s1 represents the current decryption progress, s2 represents the historical decryption efficiency, s3 represents the remaining encrypted data volume, d1 represents the selection process of the decryption method, and d2 represents the adjustment action of the decryption parameters;

[0020] S2.4: Use historical encryption data and corresponding historical decryption results as training data, and input the training data into the reinforcement learning training framework. Through repeated iterative training, the optimal decryption model is obtained.

[0021] Specifically, the specific steps of step S2 also include:

[0022] S2.5: The weighted fusion method of multiple models is used to fuse the optimal decryption model and the quantum decryption algorithm to form a fusion decryption model, and the trained fusion decryption model is deployed to the local processing unit. When new encrypted data arrives, the fusion decryption model is used for real-time decryption processing. The formula is:

[0023] Y(x,θ,ε)=w1×M1(x,θ)+w2×M2(x,ε);

[0024] Among them, Y(x,θ,ε) represents the output result of the fusion decryption model, x represents the input encrypted data, θ represents the optimal decryption model parameter, ε represents the optimal decryption model parameter, M1(x,θ) represents the decryption result of the optimal decryption model, M2(x,ε) represents the decryption result of the quantum decryption algorithm, w1 and w2 represent weights;

[0025] S2.6: Pre-process the decrypted vehicle status information and environmental data, use the machine learning model to identify driving risks and safety threats to the processed data, and generate corresponding alarm information based on the identification results.

[0026] Specifically, the specific steps of S2.6 include:

[0027] S2.61: Receive encrypted data, decrypt the data using a quantum decryption algorithm, restore the original vehicle status information and environmental data, perform preprocessing and feature extraction, and generate decrypted data;

[0028] S2.62: Load a pre-built decision tree machine learning model, train the decision tree machine learning model using the decrypted data, adjust the decision tree machine learning model parameters through iterative optimization, and evaluate the adjusted decision tree machine learning model;

[0029] S2.63: According to the evaluation results, adjust the decision tree machine learning model parameters again to obtain a trained decision tree machine learning model, and deploy the trained decision tree machine learning model to practical applications;

[0030] S2.64: Set the error threshold h1, receive the decrypted data in real time, and use the deployed decision tree machine learning model to make predictions.

[0031] like The current state is safe, where Represents the prediction results of the decision tree machine learning model;

[0032] like The corresponding alarm information is generated.

[0033] Specifically, the specific steps of step S3 include:

[0034] S3.1: Collect real-time driving behavior data and vehicle status data through vehicle sensors, obtain road condition information and regional data from the cloud, and fuse the driving behavior data, vehicle status data, road condition information and regional data using a head-to-tail fusion method to obtain fused data;

[0035] S3.2: Load the pre-built risk assessment model, input the fused data into the risk assessment model for risk assessment, and output the quantitative risk value based on the assessment results;

[0036] S3.3: Introduce reinforcement learning algorithm, take risk assessment results as input, and formulate the optimal protection measures strategy through the pre-trained reinforcement learning model.

[0037] Specifically, the scheduler in S2.1 determines the order and rate of receiving the encrypted data based on the encrypted data tag and the load of the current processing unit, and the encrypted data tag includes the urgency and the data type.

[0038] A security protection system based on a multi-level automotive Internet of Things, comprising: an encryption module, a risk identification module, a risk assessment module, and a protection execution module;

[0039] The encryption module is used to collect vehicle status information and environmental data and perform encryption processing;

[0040] The risk identification module is used to decrypt, process and analyze the encrypted data, and identify driving risks and safety threats based on the decrypted data;

[0041] The risk assessment module is used to conduct risk assessment on the current vehicle status and formulate the optimal protection measures strategy;

[0042] The protection execution module is used to automatically trigger the early warning mechanism and execute protection measures when the system detects a risk.

[0043] Specifically, the risk identification module includes: a decryption algorithm unit and a machine learning model unit;

[0044] The decryption algorithm unit uses an AI decryption algorithm based on reinforcement learning to optimize the decryption process and decrypt the encrypted data;

[0045] The machine learning model unit analyzes the decrypted data through a trained machine learning model to identify potential driving risks and safety threats.

[0046] Specifically, the risk assessment module includes: a risk assessment unit and an optimization algorithm unit;

[0047] The risk assessment unit is used to comprehensively assess the current driving risk level based on historical data using a pre-built risk assessment model and combining multiple factors;

[0048] The optimization algorithm unit is used to formulate the best protection measure strategy according to the risk assessment result through the optimization algorithm.

[0049] Compared with the prior art, the present invention has the following beneficial effects:

[0050] 1. The present invention proposes a security protection method based on a multi-level automotive Internet of Things. By integrating encrypted sensor networks, quantum encryption communications, reinforcement learning AI decryption algorithms and machine learning models, it realizes highly intelligent monitoring and analysis of vehicle status, environmental data, potential driving risks and safety threats, which not only ensures the security of data transmission, but also improves the accuracy and timeliness of risk identification.

[0051] 2. The present invention proposes a security protection method based on a multi-level automotive Internet of Things. By combining real-time driving behavior, vehicle status, road condition information and cloud-shared data, the risk assessment model and optimization algorithm are used to formulate the optimal protection measure strategy, and the early warning mechanism is automatically triggered and protective measures are executed when risks are detected. The system effectively reduces the probability of traffic accidents and improves driving safety and reliability. At the same time, vehicle data is uploaded to the cloud in real time and a security audit mechanism is established, which further enhances the traceability and security of the data. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 This is a framework diagram of a security protection method based on a multi-level automotive Internet of Things in the present invention;

[0053] Figure 2 This is a workflow diagram of an algorithm for a security protection method based on a multi-level automotive Internet of Things in the present invention;

[0054] Figure 3 This is an architecture diagram of a security protection system based on a multi-level automotive Internet of Things in the present invention. DETAILED DESCRIPTION

[0055] Example 1

[0056] See also Figure 1-Figure 2 , an embodiment provided by the present invention: a security protection method based on a multi-level automobile Internet of Things, comprising the following steps:

[0057] Step S1: using an encrypted sensor network to collect vehicle status information and environmental data, obtaining encrypted data, and using a quantum encryption communication protocol to transmit the collected encrypted data to a local processing unit;

[0058] Among them, vehicle status information includes speed, acceleration, and steering angle, and environmental data includes surrounding vehicle locations, road conditions, and weather conditions.

[0059] Furthermore, the specific steps of step S1 include:

[0060] (1) Deployment of encrypted sensor networks:

[0061] Install multiple encrypted sensors on the vehicle. The encrypted sensors have data collection and encryption functions and are used to monitor the vehicle status and perceive the environment. Axle temperature sensors, speed sensors, and acceleration sensors are used to monitor the vehicle status, and radars, cameras, and temperature sensors are used to perceive the environment.

[0062] (2) Data collection and encryption:

[0063] The sensor collects vehicle status information and environmental data in real time. While collecting data, an encryption module inside the sensor or in the sensor network encrypts the data. The present invention selects a quantum encryption algorithm, wherein the process of the quantum encryption algorithm includes:

[0064] 1) Quantum Key Distribution:

[0065] Information is encoded using the polarization or phase properties of quantum states, such as photons. When a quantum state is observed, its state changes. This property is called the "uncertainty principle", which ensures the security of key transmission. At the same time, the communicating parties compare their respective observed quantum state results, eliminate the inconsistent parts caused by interference or errors, and finally generate a set of secure keys;

[0066] Among them, any third party who attempts to eavesdrop on the quantum channel will be discovered due to the uncertainty principle of quantum mechanics, because the third party's measurement will interfere with the quantum state and leave traces. At the same time, the sender and receiver communicate through optical fiber to confirm the correctness of the key and discard the interfered part;

[0067] 2) Quantum encryption algorithm:

[0068] 1) Generate a pair of entangled quantum states, held by the sender and receiver respectively;

[0069] 2) The sender signs the message using the quantum state it holds, and then sends the signature and the message together to the receiver;

[0070] 3) The recipient uses the entangled quantum state it holds to verify the authenticity of the signature.

[0071] (3) Encrypted data transmission:

[0072] The encrypted data is transmitted to a local processing unit through a quantum encryption communication protocol to prevent the data from being stolen, tampered with or forged during transmission. The local processing unit can be an on-board computer or an edge computing device.

[0073] (4) Local processing unit receives and decrypts

[0074] After receiving the encrypted data, the local processing unit uses the corresponding quantum key to perform decryption operations to restore the original vehicle status information and environmental data.

[0075] Step S2: In the local processing unit, the encrypted data is processed and decrypted using an AI decryption algorithm based on reinforcement learning, and based on the decrypted data, driving risks and safety threats are identified through a machine learning model;

[0076] Step S3: Based on the driving risks and safety threats identified in step S2, combined with the current driving behavior, vehicle status, road condition information and vehicle and regional data shared in the cloud, a pre-built risk assessment model is used to conduct a risk assessment, and based on the risk assessment results, an optimization algorithm is used to formulate the optimal protection measures strategy;

[0077] Step S4: When the system detects a risk, based on the risk assessment results and the optimal protective measures strategy developed, the corresponding level of early warning mechanism is automatically triggered through the Internet of Things method, and protective measures are automatically executed according to the triggered early warning level and type;

[0078] Among them, the protection measures include: automatic deceleration, lane change, automatic emergency braking, forward collision warning, lane departure warning, and fatigue driving detection.

[0079] Furthermore, the specific steps of step S4 include:

[0080] (1) According to the results output by the risk assessment model in step S3, it is determined whether there is a risk and the level of the risk. According to the risk assessment results, the system searches for the corresponding warning level from a preset warning level mapping table, where the warning level is divided into four levels: low, medium, high, and emergency;

[0081] (2) Use IoT methods to send warning signals to IoT devices to trigger the corresponding level of warning mechanism, where IoT devices include vehicle display screens, mobile phone apps, and remote monitoring centers. At the same time, the warning signal contains risk type, risk level, location, and recommended measures;

[0082] (3) According to the optimal protection measure strategy formulated in step S3, the system sends control instructions to the vehicle control system through the Internet of Things interface. The control instructions include adjusting the vehicle speed, changing the driving route, initiating emergency braking, and shutting down unnecessary electrical equipment;

[0083] (4) The system monitors the implementation of protective measures and collects feedback information after implementation, including vehicle speed changes, route adjustment results, and braking effects. Based on the feedback information, the system further adjusts the protective measures strategy.

[0084] Step S5: Upload the vehicle status information and driving behavior data to the cloud server in real time. At the same time, establish a cloud security audit mechanism to monitor and record the uploaded and shared vehicle status information and driving behavior data.

[0085] Furthermore, the specific steps of step S5 include:

[0086] (1) Using quantum encryption communication protocols and communication links, the encrypted data is uploaded to the cloud server in real time;

[0087] (2) The cloud server receives and distributes the uploaded vehicle status information and driving behavior data;

[0088] (3) Design and deploy a cloud-based security audit system, which includes security policy formulation, audit rule configuration, and audit log recording modules, and formulate data access permission control policies to ensure that only authorized users can access and operate vehicle data;

[0089] (4) Configure audit rules. The cloud security audit system monitors the uploaded and shared vehicle status information and driving behavior data in real time according to the configured audit rules, and records all operation logs related to vehicle data, including operation time, operation type, operation result, and information of the operating user;

[0090] (5) Regularly conduct security analysis on audit logs to identify potential security threats and non-compliant behaviors, and take security response measures based on the analysis results, including blocking illegal access and performing security reinforcement.

[0091] The specific steps of step S2 include:

[0092] S2.1: Initialize the smart scheduler based on the received encrypted data Request, combined with the load of the local processing unit, data type priority, and resource availability factors, intelligent scheduling decision is made. The intelligent scheduling decision is expressed as:

[0093]

[0094] Among them, A(r) represents the intelligent scheduling decision, Represents the Nth encrypted vehicle status information, represents the Mth encrypted environmental data, N represents the number of encrypted vehicle status information, M represents the number of encrypted environmental data, P(r) represents the priority evaluation function, r represents the current attribute set of the encrypted data request, q1 represents the amount of allocated resources, g1(r) represents the type of encrypted data, g2(r) represents the urgency, g3(r) represents the size of the encrypted data, α, β, δ represent weight coefficients, and the setting method of α, β, δ is determined by those skilled in the art through a large number of experiments;

[0095] Among them, the scheduler in S2.1 determines the order and rate of receiving encrypted data based on the encrypted data tag and the load of the current processing unit, and the encrypted data tag includes the urgency and data type.

[0096] Furthermore, the specific steps of S2.1 include:

[0097] (1) Initialize the smart scheduler

[0098] 1) Configure the scheduling strategy: In the present invention, a priority queue algorithm is selected;

[0099] 2) Set priority rules: Set different priorities for different data requests based on data type, urgency, and data size;

[0100] 3) Configure resource allocation strategy: determine the unit and method of resource allocation, such as CPU time slice, memory space, network bandwidth, and design resource allocation algorithm to ensure that resources can be allocated on demand and avoid resource waste and conflict;

[0101] (2) Intelligent scheduling decision

[0102] 1) Receive encrypted data requests: monitor encrypted data requests from encrypted sensor networks, including data type, size, and priority information;

[0103] 2) Evaluate the local processing unit load: monitor the CPU usage, memory usage, and network bandwidth indicators of the local processing unit in real time, and analyze the current load to determine whether there are sufficient resources to process new data requests;

[0104] 3) Analyze data type priority: Evaluate the importance and urgency of the data request based on the data type and priority information in the received data request;

[0105] 4) Make decisions based on resource availability: Comprehensively consider the load of the local processing unit, data type priority and resource availability factors, use an intelligent scheduling algorithm to make decisions, and determine the processing order of data requests and the resource allocation plan. In the present invention, the intelligent scheduling algorithm selects a heuristic algorithm, and the specific implementation process of the heuristic algorithm is the prior art content in this field, and is not the creative solution of this application, so it will not be repeated here.

[0106] S2.2: According to the intelligent scheduling decision result, the local processing unit receives the encrypted data from the encrypted sensor network through a communication connection, and stores the encrypted data in a storage medium;

[0107] Furthermore, the specific steps of S2.2 include:

[0108] (1) Based on the results of the intelligent scheduling decision, the local processing unit establishes a communication connection with the encrypted sensor network through a quantum encryption communication protocol and ensures the security of the communication connection;

[0109] (2) Once the communication connection is established, the local processing unit begins to receive encrypted data from the encrypted sensor network. This data may be real-time data captured by the sensor and sent to the local processing unit after encryption processing;

[0110] (3) Before the received encrypted data is stored, the integrity is verified by checking the checksum or signature of the data to ensure that the data has not been tampered with during transmission;

[0111] (4) After verification, the local processing unit stores the encrypted data in a predetermined storage medium. In the present invention, the storage medium refers to a local hard disk. During the storage process, data backup and redundant storage technology are required to ensure the security and reliability of the data.

[0112] (5) Manage and index the stored encrypted data, including establishing a database, data table or index file.

[0113] S2.3: Define the state space S = {s1, s2, s3} and action space D = {d1, d2} of the decryption algorithm, and build a reinforcement learning training framework based on the defined state space and action space, where s1 represents the current decryption progress, s2 represents the historical decryption efficiency, s3 represents the remaining encrypted data volume, d1 represents the selection process of the decryption method, and d2 represents the adjustment action of the decryption parameters;

[0114] S2.4: Use historical encryption data and corresponding historical decryption results as training data, and input the training data into the reinforcement learning training framework to obtain the optimal decryption model through repeated iterative training;

[0115] Furthermore, the specific steps of S2.4 include:

[0116] (1) Collect historical encrypted data and their corresponding successful decryption results, and perform preprocessing to ensure that the data format is suitable for input into the reinforcement learning model, and divide the historical encrypted data and their corresponding successful decryption results into training set, verification set and test set;

[0117] (2) Obtain the state space, action space and Policy Gradient training framework in step S2.3, and design the policy network;

[0118] (3) Initialize the weights of the policy network and the value function network, randomly sample a batch of encrypted data from the training set as input, and for each encrypted data, use the policy network to generate an action, i.e., the decryption strategy;

[0119] (4) Execute the decryption action and record the decryption result, reward obtained, and next state;

[0120] The obtained reward and state information are used to calculate the policy gradient, and at the same time, the parameters of the policy network are updated according to the policy gradient, wherein the specific calculation formula of the policy gradient update of the REINFORCE algorithm is the prior art content in the art, which is not the inventive solution of the present application and is not described in detail here;

[0121] (5) Repeat the above process until the preset training rounds are reached, and use the validation set to evaluate the performance of the trained model. Adjust the model parameters or training strategy based on the evaluation results.

[0122] S2.5: The weighted fusion method of multiple models is used to fuse the optimal decryption model and the quantum decryption algorithm to form a fusion decryption model, and the trained fusion decryption model is deployed to the local processing unit. When new encrypted data arrives, the fusion decryption model is used for real-time decryption processing. The formula is:

[0123] Y(x,θ,ε)=w1×M1(x,θ)+w2×M2(x,ε);

[0124] Wherein, Y(x,θ,ε) represents the output result of the fusion decryption model, x represents the input encrypted data, θ represents the optimal decryption model parameter, ε represents the optimal decryption model parameter, M1(x,θ) represents the decryption result of the optimal decryption model, M2(x,ε) represents the decryption result of the quantum decryption algorithm, w1 and w2 represent weights, and the setting method of w1 and w2 is determined by those skilled in the art through a large number of experiments;

[0125] Furthermore, the quantum decryption algorithm refers to the quantum encryption algorithm in step S1, and its specific calculation formula is the prior art content in this field, which is not the inventive solution of the present application and will not be elaborated here.

[0126] S2.6: Pre-process the decrypted vehicle status information and environmental data, use the machine learning model to identify driving risks and safety threats to the processed data, and generate corresponding alarm information based on the identification results.

[0127] The specific steps of S2.6 include:

[0128] S2.61: Receive encrypted data, decrypt the data using a quantum decryption algorithm, restore the original vehicle status information and environmental data, clean the decrypted data, remove noise, errors and outliers, perform feature extraction, and generate decrypted data;

[0129] S2.62: Load a pre-built decision tree machine learning model, train the decision tree machine learning model using the decrypted data, adjust the decision tree machine learning model parameters through iterative optimization to minimize the prediction error, and evaluate the adjusted decision tree machine learning model, wherein, in the present invention, the iterative optimization algorithm uses the gradient descent method, and the gradient descent and decision tree machine learning model construction process are prior art contents in the art, and are not the inventive solution of the present application, and are not described in detail here;

[0130] S2.63: According to the evaluation results, adjust the decision tree machine learning model parameters again to obtain a trained decision tree machine learning model, and deploy the trained decision tree machine learning model to practical applications;

[0131] S2.64: Set the error threshold h1, receive the decrypted data in real time, and use the deployed decision tree machine learning model to make predictions.

[0132] like The current state is safe, where Represents the prediction results of the decision tree machine learning model;

[0133] like The corresponding alarm information is generated.

[0134] The specific steps of step S3 include:

[0135] S3.1: Collect real-time driving behavior data and vehicle status data through vehicle sensors, obtain road condition information and regional data from the cloud, and fuse the driving behavior data, vehicle status data, road condition information and regional data using a head-to-tail fusion method to obtain fused data;

[0136] S3.2: Load a pre-built risk assessment model, input the fused data into the risk assessment model for risk assessment, and output a quantitative risk value based on the assessment result. The construction process of the risk assessment model is the prior art content in this field, and is not an inventive solution of the present application, and will not be described in detail here;

[0137] S3.3: Introduce a reinforcement learning algorithm, take the risk assessment results as input, and formulate the optimal protection measures strategy through the pre-trained reinforcement learning model. The specific implementation process of the reinforcement learning algorithm is the existing technology content in this field, which is not the inventive solution of this application and will not be elaborated here. At the same time, the training process of the reinforcement learning model is the existing technology content in this field, which is not the inventive solution of this application and will not be elaborated here.

[0138] Example 2

[0139] See also Figure 3Another embodiment provided by the present invention is a security protection system based on a multi-level automotive Internet of Things, comprising:

[0140] Encryption module, risk identification module, risk assessment module, protection execution module;

[0141] An encryption module is used to collect vehicle status information and environmental data and perform encryption processing;

[0142] A risk identification module, which is used to decrypt, process and analyze encrypted data, and identify driving risks and safety threats based on the decrypted data;

[0143] Risk assessment module, used to conduct risk assessment on the current vehicle status and formulate the optimal protection measures strategy;

[0144] The protection execution module is used to automatically trigger the early warning mechanism and execute protection measures when the system detects risks.

[0145] The encryption module includes: an encryption sensor unit and a quantum encryption unit;

[0146] The encrypted sensor unit is composed of multiple encrypted sensors and is used to collect vehicle status and environmental data in real time. The sensors have built-in encryption functions to ensure that the collected data is encrypted before transmission;

[0147] The quantum encryption unit uses advanced quantum encryption communication protocols to ensure the security of encrypted data during transmission and prevent data from being eavesdropped or tampered with.

[0148] The risk identification module includes: decryption algorithm unit and machine learning model unit;

[0149] The decryption algorithm unit uses an AI decryption algorithm based on reinforcement learning to optimize the decryption process and decrypt encrypted data while maintaining high security and efficiency;

[0150] The machine learning model unit analyzes the decrypted data through the trained machine learning model to identify potential driving risks and safety threats, such as sudden braking, speeding, and collision risks.

[0151] The risk assessment module includes: risk assessment unit and optimization algorithm unit;

[0152] The risk assessment unit is used to comprehensively assess the current driving risk level based on historical data using a pre-built risk assessment model and combining multiple factors;

[0153] The optimization algorithm unit is used to formulate the best protection measures strategy according to the risk assessment results through the optimization algorithm.

[0154] The protection execution module includes: early warning unit, protection measures execution unit, cloud server unit, and security audit unit;

[0155] The early warning unit is used to automatically trigger the early warning mechanism of the corresponding level according to the risk assessment results and the formulated protective measures strategy to remind the driver to pay attention;

[0156] A protective measure execution unit is used to automatically execute corresponding protective measures according to the triggered warning level and type to reduce or avoid potential risks;

[0157] The cloud server unit is responsible for receiving and storing uploaded vehicle status information and driving behavior data, and providing data sharing and query services;

[0158] The security audit unit is used to monitor and record uploaded and shared data in real time to ensure the integrity and security of the data. At the same time, it provides early warning and response to abnormal access and data leakage.

[0159] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0160] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0161] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in the field may also change, modify, replace and modify the above-mentioned embodiments without departing from the purpose and scope of protection of the present invention, and all of these are within the protection of the present invention.

Claims

1. A security protection method based on multi-level automotive Internet of Things, characterized in that: include: Step S1: using an encrypted sensor network to collect vehicle status information and environmental data, obtaining encrypted data, and using a quantum encryption communication protocol to transmit the collected encrypted data to a local processing unit; Step S2: In the local processing unit, the encrypted data is processed and decrypted using an AI decryption algorithm based on reinforcement learning, and based on the decrypted data, driving risks and safety threats are identified through a machine learning model; Step S3: Based on the driving risks and safety threats identified in step S2, combined with current driving behavior, vehicle status, road condition information and regional data shared in the cloud, a pre-built risk assessment model is used to conduct a risk assessment, and based on the risk assessment results, an optimization algorithm is used to formulate an optimal protection measure strategy; Step S4: When the system detects a risk, based on the risk assessment results and the optimal protective measures strategy developed, the corresponding level of early warning mechanism is automatically triggered through the Internet of Things method, and protective measures are automatically executed according to the triggered early warning level and type; Step S5: Upload the vehicle status information and driving behavior data to the cloud server in real time. At the same time, establish a cloud security audit mechanism to monitor and record the uploaded and shared vehicle status information and driving behavior data.

2. A security protection method based on a multi-level automotive Internet of Things as claimed in claim 1, characterized in that: The specific steps of step S2 include: S2.1: Initialize the smart scheduler based on the received encrypted data Request, combined with the load of the local processing unit, data type priority, and resource availability factors, intelligent scheduling decision is made. The intelligent scheduling decision is expressed as: Among them, A(r) represents the intelligent scheduling decision, Represents the Nth encrypted vehicle status information, represents the Mth encrypted environmental data, N represents the number of encrypted vehicle status information, M represents the number of encrypted environmental data, P(r) represents the priority evaluation function, r represents the current attribute set of the encrypted data request, q1 represents the amount of allocated resources, g1(r) represents the type of encrypted data, g2(r) represents the urgency, g3(r) represents the size of the encrypted data, α, β, δ represent weight coefficients; S2.2: According to the intelligent scheduling decision result, the local processing unit receives the encrypted data from the encrypted sensor network through a communication connection and stores the encrypted data in a storage medium.

3. A security protection method based on multi-level automotive Internet of Things as claimed in claim 2, characterized in that: The specific steps of step S2 also include: S2.3: Define the state space S = {s1, s2, s3} and action space D = {d1, d2} of the decryption algorithm, and build a reinforcement learning training framework based on the defined state space and action space, where s1 represents the current decryption progress, s2 represents the historical decryption efficiency, s3 represents the remaining encrypted data volume, d1 represents the selection process of the decryption method, and d2 represents the adjustment action of the decryption parameters; S2.4: Use historical encryption data and corresponding historical decryption results as training data, and input the training data into the reinforcement learning training framework. Through repeated iterative training, the optimal decryption model is obtained.

4. A security protection method based on multi-level automotive Internet of Things as claimed in claim 3, characterized in that: The specific steps of step S2 also include: S2.5: The weighted fusion method of multiple models is used to fuse the optimal decryption model and the quantum decryption algorithm to form a fusion decryption model, and the trained fusion decryption model is deployed to the local processing unit. When new encrypted data arrives, the fusion decryption model is used for real-time decryption processing. The formula is: Y(x,θ,ε)=w1×M1(x,θ)+w2×M2(x,ε); Among them, Y(x,θ,ε) represents the output result of the fusion decryption model, x represents the input encrypted data, θ represents the optimal decryption model parameter, ε represents the optimal decryption model parameter, M1(x,θ) represents the decryption result of the optimal decryption model, M2(x,ε) represents the decryption result of the quantum decryption algorithm, w1 and w2 represent weights; S2.6: Pre-process the decrypted vehicle status information and environmental data, use the machine learning model to identify driving risks and safety threats to the processed data, and generate corresponding alarm information based on the identification results.

5. A security protection method based on multi-level automotive Internet of Things as claimed in claim 4, characterized in that: The specific steps of S2.6 include: S2.61: Receive encrypted data, decrypt the data using a quantum decryption algorithm, restore the original vehicle status information and environmental data, perform preprocessing and feature extraction, and generate decrypted data; S2.62: Load a pre-built decision tree machine learning model, train the decision tree machine learning model using the decrypted data, adjust the decision tree machine learning model parameters through iterative optimization, and evaluate the adjusted decision tree machine learning model; S2.63: According to the evaluation results, adjust the decision tree machine learning model parameters again to obtain a trained decision tree machine learning model, and deploy the trained decision tree machine learning model to practical applications; S2.64: Set the error threshold h1, receive the decrypted data in real time, and use the deployed decision tree machine learning model to make predictions. like The current state is safe, where Represents the prediction results of the decision tree machine learning model; like A corresponding alarm message is generated.

6. A security protection method based on multi-level automotive Internet of Things as claimed in claim 5, characterized in that: The specific steps of step S3 include: S3.1: Collect real-time driving behavior data and vehicle status data through vehicle sensors, obtain road condition information and regional data from the cloud, and fuse the driving behavior data, vehicle status data, road condition information and regional data using a head-to-tail fusion method to obtain fused data; S3.2: Load the pre-built risk assessment model, input the fused data into the risk assessment model for risk assessment, and output the quantitative risk value based on the assessment results; S3.3: Introduce reinforcement learning algorithm, take risk assessment results as input, and formulate the optimal protection measures strategy through the pre-trained reinforcement learning model.

7. A security protection method based on multi-level automotive Internet of Things as claimed in claim 6, characterized in that: The scheduler in S2.1 determines the order and rate of receiving the encrypted data based on the encrypted data tag and the load of the current processing unit, and the encrypted data tag includes the urgency and the data type.

8. A security protection system based on a multi-level automobile Internet of Things, which is used to implement a security protection method based on a multi-level automobile Internet of Things as described in any one of claims 1 to 7, characterized in that: include: Encryption module, risk identification module, risk assessment module, protection execution module; The encryption module is used to collect vehicle status information and environmental data and perform encryption processing; The risk identification module is used to decrypt, process and analyze the encrypted data, and identify driving risks and safety threats based on the decrypted data; The risk assessment module is used to conduct risk assessment on the current vehicle status and formulate the optimal protection measures strategy; The protection execution module is used to automatically trigger the early warning mechanism and execute protection measures when the system detects a risk.

9. A security protection system based on multi-level automotive Internet of Things as claimed in claim 8, characterized in that: The risk identification module includes: a decryption algorithm unit and a machine learning model unit; The decryption algorithm unit uses an AI decryption algorithm based on reinforcement learning to optimize the decryption process and decrypt the encrypted data; The machine learning model unit analyzes the decrypted data through a trained machine learning model to identify potential driving risks and safety threats.

10. A security protection system based on multi-level automotive Internet of Things as claimed in claim 9, characterized in that: The risk assessment module includes: a risk assessment unit and an optimization algorithm unit; The risk assessment unit is used to comprehensively assess the current driving risk level based on historical data using a pre-built risk assessment model and combining multiple factors; The optimization algorithm unit is used to formulate the best protection measure strategy according to the risk assessment result through the optimization algorithm.

Images