Network switching method of vehicle terminal and related equipment

Abstract

This invention relates to the field of vehicle communication technology, and more particularly to a network switching method and related equipment for an in-vehicle terminal. The in-vehicle terminal is equipped with a first communication card and a second communication card. The method includes: acquiring current network status information and current vehicle motion status information of the vehicle where the in-vehicle terminal is located; inputting the current network status information and the current vehicle motion status information into a pre-trained first model; the first model outputting a corresponding target switching action based on the current network status information and the current vehicle motion status information; and controlling the in-vehicle terminal to execute the target switching action.

Description

Technical Field

[0001] This invention relates to the field of vehicle communication technology, and in particular to a network switching method and related equipment for an in-vehicle terminal. Background Technology

[0002] With the rapid development of intelligent connected vehicles, in-vehicle terminals (such as TBOXes), as the core unit for vehicle-to-the-cloud and vehicle-to-infrastructure (V2I) communication, directly determine the reliability of key functions such as remote control, real-time navigation, and autonomous driving data transmission through their network connections. Currently, most mainstream TBOXes adopt a single-card single-network mode or a dual-card redundancy switching mode based on signal strength. However, in actual driving, vehicles frequently encounter complex environments such as urban high-rise building obstructions, tunnel crossings, and high-speed movement, resulting in strong fluctuations and sudden attenuation of network signals. Existing switching solutions rely solely on single indicators such as network signal strength for decision-making, failing to fully consider the impact of the vehicle's own motion state (such as speed, acceleration, and deceleration) on network quality changes. This leads to a disconnect between switching decisions and real-world driving scenarios, easily resulting in switching delays, erroneous switching, and even communication interruptions, posing a serious threat to safety-related functions such as autonomous driving. Therefore, how to achieve intelligent and adaptive network switching by incorporating vehicle motion states has become a pressing technical problem to be solved in this field. Summary of the Invention

[0003] In view of this, embodiments of the present invention provide a network switching method and related equipment for an in-vehicle terminal, which can realize adaptive perception and intelligent switching decision-making for driving scenarios, significantly improving the stability and reliability of in-vehicle communication.

[0004] In a first aspect, embodiments of the present invention provide a network switching method for a vehicle-mounted terminal, wherein the vehicle-mounted terminal is equipped with a first communication card and a second communication card, the method comprising: Obtain the current network status information and current vehicle motion status information of the vehicle where the in-vehicle terminal is located; The current network state information and the current vehicle motion state information are input into a pre-trained first model; the first model is used to output a corresponding target switching action based on the current network state information and the current vehicle motion state information. Control the vehicle-mounted terminal to perform the target switching action.

[0005] In one possible implementation, the vehicle motion state information includes vehicle speed and / or driving state; wherein the driving state includes any one of being stationary, moving at a constant speed, accelerating, or decelerating.

[0006] In one possible implementation, inputting the current network state information and the current vehicle motion state information into a pre-trained first model includes: The current network state information is used to extract features to obtain a first feature set; The current vehicle motion state information is used to extract features to obtain a second feature set; The first feature set and the second feature set are fused to obtain a fused multidimensional feature vector; The multidimensional feature vector is input into the first model.

[0007] In one possible implementation, after obtaining the current network status information of the vehicle, the method further includes: The current network state information is optimized using the Kalman filter algorithm to obtain optimized network state information; The step of inputting the current network state information and the current vehicle motion state information into a pre-trained first model includes: The optimized network state information and the current vehicle motion state information are input into the first model.

[0008] In one possible implementation, the state transition matrix of the Kalman filter algorithm is determined based on the current vehicle speed; the higher the current vehicle speed, the more the weights of the state transition matrix are biased towards the predicted value.

[0009] In one possible implementation, the target switching action includes: The first action is to maintain the first communication card as the main link for data transmission. The second action is to switch the second communication card to the primary link; The third action involves simultaneously using both the first and second communication cards to transmit data.

[0010] In one possible implementation, the first model includes a reward function; the reward function includes: A network quality reward item based on changes in network quality after handover; wherein, the network quality reward item is determined by the difference in network quality scores before and after handover, and the network quality score is obtained by weighted summation of at least two parameters in the current network state information; Based on the handover cost penalty incurred in performing the handover action; and Interruption penalty based on whether a communication interruption occurs.

[0011] In one possible implementation, the network quality reward is determined by the difference in network quality scores before and after the handover; the network quality score is obtained by weighted summation of at least two parameters in the current network state information.

[0012] In a second aspect, embodiments of the present invention provide an electronic device, comprising: At least one processor; and At least one memory communicatively connected to the processor, wherein: The memory stores program instructions that can be executed by the processor, and the processor can execute the method described in the first aspect by calling the program instructions.

[0013] Thirdly, embodiments of the present invention provide a computer-readable storage medium storing computer instructions that cause the computer to perform the method described in the first aspect.

[0014] This invention integrates vehicle motion and network states, utilizing a pre-trained machine learning model to achieve network switching for the in-vehicle terminal. This enables the switching to perceive dynamic scenarios such as vehicle speed and driving status, solving the problems of handover lag and high misjudgment rate caused by traditional solutions relying solely on signal strength. It also improves communication stability in complex driving environments. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 A flowchart illustrating a network switching method for an in-vehicle terminal provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of a vehicle-mounted terminal provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0017] To better understand the technical solution of the present invention, the embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0018] It should be understood that the described embodiments are merely some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0019] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. The singular forms “a,” “the,” and “the” as used in the embodiments of this invention and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0020] It should be understood that the term "and / or" used in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0021] It should be understood that although terms such as first, second, third, etc., may be used to describe numbers in embodiments of the present invention, these numbers should not be limited to these terms. These terms are only used to distinguish numbers from each other. For example, without departing from the scope of embodiments of the present invention, a first number may also be referred to as a second number, and similarly, a second number may also be referred to as a first number.

[0022] To address the problem that existing vehicle-mounted terminals (TBOX) cannot adapt to complex and dynamic driving environments when switching between multiple SIM cards, resulting in switching delays, incorrect switching, or even communication interruptions, which in turn affect autonomous driving data transmission and user experience, this invention provides a network switching method for vehicle-mounted terminals. Figure 1 This is a flowchart illustrating a network switching method for a vehicle-mounted terminal provided in an embodiment of the present invention. Figure 1 As shown, the method includes: Step 101: Obtain the current network status information and current vehicle motion status information of the vehicle where the vehicle terminal is located.

[0023] The vehicle-mounted terminal (such as TBOX) contains a first communication card and a second communication card. As the core unit for communication between vehicles and the cloud, vehicles and vehicles (V2V), and vehicles and infrastructure (V2I), the vehicle-mounted terminal needs to monitor the quality of the current communication link and the vehicle's own operating status in real time to provide a data foundation for subsequent intelligent decision-making.

[0024] The current network status information refers to the communication quality of the network currently connected to the first communication card and / or the second communication card. In some embodiments, this network status information may include, but is not limited to, one or more of the following parameters: Reference Signal Receiving Power (RSRP), used to characterize signal strength; Signal to Interference plus Noise Ratio (SINR), used to characterize signal quality; Network latency, used to characterize the round-trip time of data packets from the sender to the receiver; Packet loss rate, used to characterize the proportion of data packets lost during data transmission; Network load, used to characterize the current level of congestion at the base station. The vehicle terminal reads these parameters in real time through its internally integrated communication module (such as a 4G / 5G module). Specifically, the communication module can report the above network parameters through AT commands or the MBIM (Mobile Broadband Interface Model) protocol interface, and the main control unit (such as an MCU) periodically (e.g., every 100 milliseconds) obtains the current network status information through data interaction with the communication module.

[0025] Current vehicle motion state information refers to data reflecting the vehicle's current driving state. In some embodiments, this vehicle motion state information includes vehicle speed and driving state. The driving state can include any one of being stationary, moving at a constant speed, accelerating, or decelerating. The main control unit further determines the vehicle's driving state based on the rate of change of vehicle speed. For example, when the vehicle speed is zero, the driving state is determined to be stationary. When the vehicle speed is not zero and the acceleration is zero, the driving state is determined to be moving at a constant speed. When the vehicle speed is not zero and the acceleration is greater than zero, the driving state is determined to be accelerating. When the vehicle speed is not zero and the acceleration is less than zero, the driving state is determined to be decelerating. By introducing vehicle motion state information, the network switching decision can perceive the vehicle's physical environment, such as whether the vehicle is traveling at high speed on a highway or slowly in a congested urban area, thereby customizing differentiated switching strategies for different scenarios.

[0026] Step 102: Input the current network state information and the current vehicle motion state information into the pre-trained first model. The first model is used to output the corresponding target switching action based on the current network state information and the current vehicle motion state information. The first model can be a machine learning-based reinforcement learning model, such as the SARSA model, DQN model, DDPG model, or Q-learning model.

[0027] In this process, after acquiring the current network state and vehicle motion state, the onboard terminal uses this multi-dimensional data as input to a pre-trained first model. Taking the first model as an example, the Q-learning model is a reinforcement learning-based decision-making model. Its core is a Q-table, which stores the expected rewards (i.e., Q-values) obtained by performing different actions under different states. In this embodiment of the invention, the Q-learning model learns the mapping relationship between the current network quality and the current vehicle motion state—which switching action is optimal—through a large amount of offline training data and / or data during online driving.

[0028] The Q-learning model receives the vehicle's current state as input. In this invention, this state is composed of both network state information and vehicle motion state information. Before inputting the current network state and vehicle motion state into the first model, feature extraction is required to avoid interference from irrelevant information. Specifically, feature extraction is performed on the current network state information to obtain a first feature set. Feature extraction is also performed on the current vehicle motion state information to obtain a second feature set. For example, feature extraction of the current network state information yields key information such as RSRP and SINR, which constitute the first feature set. Feature extraction of the current vehicle motion state information yields key information such as vehicle speed and driving state, which constitute the second feature set.

[0029] The first and second feature sets are then fused to obtain a fused multidimensional feature vector. Finally, this multidimensional feature vector is used as input to the Q-learning model.

[0030] To eliminate the influence of different dimensions between features, the current network state information can be standardized. For example, for RSRP, which is known to typically range from -120dBm to -50dBm, Min-Max normalization can be used to map it to the [0,1] interval. Then, the standardized network state parameters (such as RSRP, SINR, network latency, packet loss rate, network load, etc.) are combined with vehicle motion state parameters (such as vehicle speed, position, and driving state, etc.) to form a multidimensional feature vector. This multidimensional feature vector represents the current state. The Q-learning model will use this state vector to look up its internal Q-table, find the action with the highest expected reward in this state, and output it as the target switching action.

[0031] In some embodiments, to accelerate model convergence and decision-making speed, the aforementioned network state information and vehicle state information can be discretized. For example, the standardized network quality can be weighted and summed to obtain a comprehensive network quality score, which is then divided into three levels: excellent, medium, and poor. Simultaneously, vehicle speed is divided into three intervals: low, medium, and high, and driving states (stationary, constant speed, accelerating, and decelerating) are categorized into four types. This discretized state space reduces model complexity, making the model more suitable for real-time operation on resource-constrained in-vehicle MCUs. Thus, the state space S is composed of network quality levels (excellent, medium, and poor), vehicle speed levels (low, medium, and high), and driving states (stationary, constant speed, accelerating, and decelerating), totaling 3×3×4=36 discrete states. This state space design ensures coverage of major vehicle driving scenarios while maintaining model simplicity, facilitating real-time decision-making on resource-constrained in-vehicle terminals.

[0032] In a specific example, to further optimize the input data quality of the Q-learning model and ensure that decisions are based on a smoother and more realistic network state, this invention preprocesses the network state information after step 101 and before step 102. Specifically, after obtaining the vehicle's current network state information, a Kalman filter algorithm is used to optimize the current network state information, resulting in optimized network state information. Correspondingly, the current network state information and the current vehicle motion state information are input into the pre-trained Q-learning model, specifically by inputting the optimized network state information and the current vehicle motion state information into the Q-learning model.

[0033] Kalman filtering is a highly efficient recursive filter capable of estimating the state of a dynamic system from a series of incomplete and noisy measurements. In vehicular environments, due to factors such as rapid vehicle movement, multipath effects, and shadow fading, the raw network signals (such as RSRP and SINR) often contain a large amount of noise and fluctuate wildly. Directly using this noisy data for decision-making will lead to frequent erroneous handovers.

[0034] Specifically, the Kalman filtering process includes: First, defining the network state vector, denoted as x(k). Here, x(k) = [RSRP(k), SINR(k), Delay(k)], where k is the discrete time step. RSRP(k) is the received power of the reference signal at time k. SINR(k) is the signal-to-interference-plus-noise ratio at time k. Delay(k) is the network delay at time k.

[0035] Then, state equations and observation equations are constructed based on the network state vector. The state equation is denoted as X(k). The observation equation is denoted as Z(k). Where X(k) = A·X(k-1) + B·u(k-1) + w(k-1). Z(k) = H·X(k) + v(k). Where A is the state transition matrix, B is the control matrix, u(k) is the control input, w(k) is the process noise (following a Gaussian distribution with variance Q), H is the observation matrix, and v(k) is the observation noise (following a Gaussian distribution with variance R). Filtering iterations are then performed.

[0036] The filtering iteration is divided into a prediction step and an update step. The prediction step calculates the prior state and the prior covariance. The prior state is calculated using formula (1). The prior covariance is calculated using formula (2).

[0037] Formula (1): X^(k|k-1) = A·X^(k-1|k-1) + B·u(k-1).

[0038] Formula (2): P(k|k-1) = A·P(k-1|k-1)·A + Q.

[0039] The update step then calculates the Kalman gain (denoted as K(k)), the posterior state, and the updated covariance (denoted as P(k|k)). Here, K(k) = P(k|k-1)·H·[H·P(k|k-1)·H + R]. P(k|k) = [I - K(k)·H]·P(k|k-1). Here, I is the identity matrix. The posterior state is calculated using formula (3).

[0040] Formula (3): X^(k|k) = X^(k|k-1) + K(k)·[Z(k) - H·X^(k|k-1)].

[0041] The optimal network state parameters are output through iteration. The Kalman filter can estimate the true, smooth network state X^(k|k) from noisy observations Z(k), which can then be used as input to the subsequent Q-learning model.

[0042] In some embodiments, to further improve the adaptability of Kalman filtering in vehicular environments and enable it to adjust the filtering strategy according to the dynamic changes of the vehicle, the state transition matrix of the Kalman filtering algorithm is determined based on the current vehicle speed. In Kalman filtering, the state estimate is a weighted average of the predicted and observed values. The weights are determined by the Kalman gain. The larger the Kalman gain, the more the Kalman filter trusts the current observations. The smaller the Kalman gain, the more the Kalman filter trusts the predicted values. The higher the current vehicle speed, the more drastic the state changes, and the worse the timeliness of the observations. In this case, the Kalman gain is reduced, thus making the Kalman filter trust the predicted values ​​more.

[0043] The state transition matrix A describes the change of the system state over time. When the vehicle is moving at high speed, the changes in network state (such as signal strength) are much more drastic than at low speed. In this case, relying solely on measurements from the previous moment to predict the current moment is insufficient, and the confidence of the prediction model should be reduced. Therefore, this embodiment of the invention dynamically adjusts the weights of the state transition matrix A according to the vehicle speed. For example, a function relationship A = f(vehicle speed) can be pre-defined. As the vehicle speed increases, the elements in matrix A are adjusted so that the weight of the prediction step is relatively larger than the weight of the observation step. This allows the filter to place more trust in the prediction value based on the physical motion model (such as a uniform motion model) rather than the observation value, which may have a significant delay. Conversely, when the vehicle speed is low, the filter places more trust in the real-time observation value. This method of dynamically adjusting the Kalman filter parameters according to vehicle speed makes the signal optimization process itself scene-adaptable, providing more accurate data support for subsequent intelligent decision-making.

[0044] Correspondingly, the steps to fuse the filtered network parameters with the vehicle state parameters to construct a multi-dimensional feature vector may include: First, selecting multiple core features (such as filtered RSRP (F_RSRP), filtered SINR (F_SINR), network delay (Delay), packet loss rate (Loss), vehicle speed (V), and driving state (S, 0=stationary, 1=uniform speed, 2=acceleration, 3=deceleration, etc.).

[0045] Next, the constructed multi-dimensional feature vectors are standardized. Specifically, Min-Max normalization is used to map each feature to the [0,1] interval to eliminate the difference in units, such as F_RSRP_norm = (F_RSRP - RSRP_min) / (RSRP_max - RSRP_min), where RSRP_min=-120dBm and RSRP_max=-50dBm.

[0046] In some embodiments, the Q-learning model includes a reward function, which is crucial for guiding the model to learn the optimal handover action. The design of the reward function directly affects whether the behavior ultimately learned by the model meets expectations. The reward function designed in this embodiment is a composite function, including multiple reward items to comprehensively evaluate the merits of a handover action. Specifically, the reward items in the reward function include: a network quality reward item based on the change in network quality after the handover, a handover cost penalty item based on the handover action, and an interruption penalty item based on whether a communication interruption occurs.

[0047] A network quality reward is used to encourage the model to switch the network to a better state. In some embodiments, the network quality reward is determined by the difference in network quality scores before and after the switch. First, a network quality score is defined, which is obtained by a weighted sum of at least two parameters from the current network state information. For example, the same network quality score formula as when discretizing the state space can be used. The formula is: Q_score = α·RSRP_norm + β·SINR_norm + γ·(1-Delay_norm) + δ·(1-Loss_norm), where α, β, γ, and δ are weight coefficients (which can be determined through offline training, such as α=0.3, β=0.3, γ=0.2, δ=0.2). RSRP_norm is the normalized reference signal received power. SINR_norm is the normalized signal-to-interference-plus-noise ratio, Delay_norm is the normalized network delay, and Loss_norm is the normalized packet loss rate. After performing an action (such as a switch), a new network quality score, denoted as Q_score_new, is calculated and subtracted from the network quality score before the action (denoted as Q_score_old). The difference is denoted as ΔQ. ΔQ can serve as the basis for network quality rewards. Specifically, ΔQ = Q_score_new - Q_score_old. If ΔQ is positive, it indicates that the action improved network quality, and a positive reward is given. If ΔQ is negative, it indicates that the action caused a decrease in network quality, and a negative reward (i.e., a penalty) is given.

[0048] A handover cost penalty term is used to prevent the model from switching too frequently. Each handover (especially the second action, primary / backup switch) introduces additional signaling overhead and the risk of brief data interruption. If the model switches frequently for minor network quality improvements, it is detrimental to the overall communication stability. Therefore, this embodiment of the invention introduces a handover cost penalty term. For example, if the second action (primary / backup switch) is performed, a fixed negative reward (e.g., -0.1) is given. If the third action (bandwidth aggregation) is performed, since it involves dual-card collaboration and its complexity is slightly lower than a complete handover, a smaller negative reward (e.g., -0.05) can be given. If the first action (no handover) is performed, this reward is 0.

[0049] The interruption penalty is used to severely punish actions that cause communication interruptions. Communication interruptions are the most unacceptable situation in automotive scenarios (especially those involving autonomous driving). If a communication interruption occurs after performing an action (e.g., a failed handover resulting in a disconnection), a large negative reward (e.g., -0.5) is given. Conversely, if communication remains normal after the action is performed, a small positive reward (e.g., +0.2) is given to encourage stable connections.

[0050] The final immediate reward R can be expressed as the sum of these three terms, for example: R = (Q_score_new - Q_score_old) + switching cost penalty + interruption penalty. The model continuously updates the Q-table through temporal difference learning, aiming to maximize the long-term cumulative reward, thereby learning a robust policy that pursues high-quality networks while avoiding frequent switching and interruptions.

[0051] Specifically, the Q-learning model uses temporal difference learning to update the Q-table, with the update formula: Q(S_t,A_t) = Q(S_t, A_t) + η·[R_t + γ·maxQ(S_{t+1}, A) - Q(S_t, A_t)]. Here, S_t is the current state, A_t is the currently executed action, R_t is the immediate reward obtained after executing the action, S_{t+1} is the new state after executing the action, maxQ(S_{t+1}, A) represents the maximum Q-value of all possible actions in the new state, η is the learning rate (ranging from 0.1 to 0.3, decreasing with iterations), and γ is the discount factor (ranging from 0.8 to 0.9, used to weigh the importance of future rewards). Through continuous iterative updates, the Q-table gradually converges to the optimal policy.

[0052] Step 103: Control the vehicle terminal to perform the target switching action.

[0053] When the first model outputs a target switching action, the vehicle terminal executes the action accordingly. The target switching action is the optimal decision made by the first model based on the current state, thereby improving or maintaining the current communication quality.

[0054] In some embodiments, the target switching action includes three core actions: a first action, a second action, and a third action. The first action refers to maintaining the first communication card as the primary link for data transmission. This means that the network quality of the current primary link (e.g., the first communication card) is acceptable, and no switching is necessary; the vehicle terminal continues to use the first communication card for data transmission and reception. The second action refers to switching the second communication card to the new primary link. This means that the model determines that the current primary link is about to deteriorate or has already deteriorated, while the quality of the backup link (the second communication card) is better, thus requiring a primary / backup switch. The third action refers to simultaneously utilizing both the first and second communication cards for data transmission. This means that the model determines that the quality of both links is good, or that there is a large amount of data to be transmitted (e.g., map downloads, OTA upgrades). In this case, bandwidth aggregation is performed, using both links simultaneously for data transmission, which can significantly improve throughput.

[0055] During bandwidth aggregation, the main control unit monitors the load of the links corresponding to the first and second communication cards in real time and dynamically adjusts the data allocation ratio. In one specific embodiment, when the load rate of the main link (e.g., the link corresponding to the current main communication card) does not exceed 70%, the main link undertakes 70% of the data transmission task, and the auxiliary link undertakes 30%. When the load of the main link exceeds 70%, the data is evenly distributed to the two links to achieve load balancing and maximize bandwidth utilization. This dynamic allocation strategy can be flexibly adjusted according to the actual network conditions and service requirements.

[0056] Through the above steps, this embodiment of the invention incorporates the vehicle's own motion state into the decision-making basis for network switching, so that the switching decision no longer depends solely on signal strength. For example, when a vehicle is speeding on a highway, even if the current signal strength is still acceptable, the model may predict that it is about to enter a signal blind spot due to the high speed, thereby triggering the second action (switching the second communication card to the main link) or the first action (not switching but increasing the monitoring frequency) in advance to avoid communication interruption.

[0057] Corresponding to the network switching method of the vehicle terminal described above, this embodiment of the invention provides a vehicle terminal. Figure 2 This is a schematic diagram of the structure of a vehicle-mounted terminal provided in an embodiment of the present invention. Figure 2 As shown, the vehicle terminal includes: an acquisition module 201, a processing module 202, and a control module 203.

[0058] The acquisition module 201 is used to acquire the current network status information and the current vehicle motion status information of the vehicle.

[0059] Processing module 202 is used to input the current network state information and the current vehicle motion state information into a pre-trained first model. The first model is used to output the corresponding target switching action based on the current network state information and the current vehicle motion state information.

[0060] The control module 203 is used to control the vehicle terminal to perform target switching actions.

[0061] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention, such as... Figure 3 As shown, the aforementioned electronic device may include at least one processor and at least one memory communicatively connected to the processor, wherein the memory stores program instructions executable by the processor, and the processor can execute this specification by calling the program instructions. Figure 1 The embodiment shown provides a network switching method for an in-vehicle terminal.

[0062] like Figure 3 As shown, the electronic device is represented in the form of a general-purpose computing device. The components of the electronic device may include, but are not limited to: one or more processors 310, communication interface 320 and memory 330, and a communication bus 340 connecting different system components (including memory 330, communication interface 320 and processor 310).

[0063] Communication bus 340 represents one or more of several bus architectures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus using any of the various bus architectures. For example, these architectures include, but are not limited to, Industry Standard Architecture (ISA) buses, Micro Channel Architecture (MAC) buses, Enhanced ISA buses, Video Electronics Standards Association (VESA) local buses, and Peripheral Component Interconnect (PCI) buses.

[0064] Electronic devices typically include a variety of computer-readable media. These media can be any available media that can be accessed by the electronic device, including volatile and non-volatile media, and removable and non-removable media.

[0065] Memory 330 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) and / or cache memory. The electronic device may further include other removable / non-removable, volatile / non-volatile computer system storage media. Memory 330 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of the embodiments described herein.

[0066] A program / utility having a set (at least one) of program modules may be stored in memory 330. Such program modules include—but are not limited to—an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment. The program modules typically perform the functions and / or methods described in the embodiments of this specification.

[0067] Processor 310 executes various functional applications and data processing by running programs stored in memory 330, such as implementing the functions described in this specification. Figure 1 The embodiment shown provides a network switching method for an in-vehicle terminal.

[0068] This specification provides a computer program product, which includes a computer program that, when executed by a processor, performs the functions described in this specification. Figure 1 The embodiment shown provides a network switching method for an in-vehicle terminal.

[0069] This specification provides a computer-readable storage medium storing computer instructions that cause a computer to execute this specification. Figure 1 The embodiment shown provides a network switching method for an in-vehicle terminal.

[0070] The aforementioned computer-readable storage medium may be any combination of one or more computer-readable media. A computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium may be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), or flash memory, optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium may be any tangible medium that contains or stores a program that may be used by or in connection with an instruction execution system, apparatus, or device.

[0071] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0072] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this specification. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0073] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this specification, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0074] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this specification includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which the embodiments of this specification pertain.

[0075] Depending on the context, the word "if" as used here can be interpreted as "when," "when," "in response to determination," or "in response to detection." Similarly, depending on the context, the phrase "if determination" or "if detection (of the stated condition or event)" can be interpreted as "when determination," "in response to determination," "when detection (of the stated condition or event)," or "in response to detection (of the stated condition or event)."

[0076] It should be noted that the devices involved in the embodiments of this specification may include, but are not limited to, personal computers (PCs), personal digital assistants (PDAs), wireless handheld devices, tablet computers, mobile phones, MP3 displays, MP4 displays, etc.

[0077] In the several embodiments provided in this specification, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0078] Furthermore, the functional units in the various embodiments of this specification can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in a combination of hardware and software functional units.

[0079] The integrated units implemented as software functional units described above can be stored in a computer-readable storage medium. These software functional units, stored in a storage medium, include several instructions to cause a computer device (which may be a personal computer, a connector, or a network device, etc.) or a processor to execute some steps of the methods described in the various embodiments of this specification. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0080] The above description is merely a preferred embodiment of this specification and is not intended to limit this specification. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this specification should be included within the scope of protection of this specification.

[0081] The same or similar parts between the various embodiments in this specification can be referred to mutually. In particular, the device embodiments and terminal embodiments are basically similar to the method embodiments, so the description is relatively simple, and the relevant parts can be referred to the description in the method embodiments.

Claims

1. A network switching method of a vehicle terminal, characterized by, The vehicle-mounted terminal is equipped with a first communication card and a second communication card, and the method includes: Obtain the current network status information and current vehicle motion status information of the vehicle where the in-vehicle terminal is located; The current network state information and the current vehicle motion state information are input into a pre-trained first model; the first model is used to output a corresponding target switching action based on the current network state information and the current vehicle motion state information. Control the vehicle-mounted terminal to perform the target switching action.

2. The method of claim 1, wherein, The vehicle motion status information includes vehicle speed and / or driving status; wherein, the driving status includes any one of being stationary, moving at a constant speed, accelerating, or decelerating.

3. The method of claim 1, wherein, The step of inputting the current network state information and the current vehicle motion state information into a pre-trained first model includes: The current network state information is used to extract features to obtain a first feature set; The current vehicle motion state information is used to extract features to obtain a second feature set; The first feature set and the second feature set are fused to obtain a fused multidimensional feature vector; The multidimensional feature vector is input into the first model.

4. The method of claim 1, wherein, After obtaining the current network status information of the vehicle, the method further includes: The current network state information is optimized using the Kalman filter algorithm to obtain optimized network state information; The step of inputting the current network state information and the current vehicle motion state information into a pre-trained first model includes: The optimized network state information and the current vehicle motion state information are input into the first model.

5. The method of claim 4, wherein, The state transition matrix of the Kalman filter algorithm is determined based on the current vehicle speed; the higher the current vehicle speed, the more the weights of the state transition matrix are biased towards the predicted value.

6. The method according to claim 1, characterized in that, The target switching action includes: The first action is to maintain the first communication card as the main link for data transmission. The second action is to switch the second communication card to the primary link; The third action involves simultaneously using both the first and second communication cards to transmit data.

7. The method according to claim 1, characterized in that, The first model includes a reward function; the reward function includes: A network quality reward item based on changes in network quality after handover; wherein, the network quality reward item is determined by the difference in network quality scores before and after handover, and the network quality score is obtained by weighted summation of at least two parameters in the current network state information; Based on the handover cost penalty incurred in performing the handover action; and Interruption penalty based on whether a communication interruption occurs.

8. The method according to claim 7, characterized in that, The network quality reward is determined by the difference in network quality scores before and after the handover; the network quality score is obtained by weighted summation of at least two parameters in the current network status information.

9. An electronic device, characterized in that, include: At least one processor; as well as At least one memory communicatively connected to the processor, wherein: The memory stores program instructions that can be executed by the processor, and the processor can execute the method according to any one of claims 1 to 7 by calling the program instructions.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause the computer to perform the method according to any one of claims 1 to 7.

Images