Methods, devices, equipment and storage media for switching communication channels

Abstract

This invention provides a method, apparatus, device, and storage medium for switching communication channels, relating to the field of vehicle-to-everything (V2X) communication technology. The method includes: acquiring current vehicle status data of a target vehicle, current network status data of multiple available communication channels, and historical network quality data of each communication channel when other vehicles have historically passed through the target area; based on the acquired data, predicting the predicted network quality of each communication channel within a future preset time period; determining a target communication channel from among the multiple communication channels, and switching to the target communication channel. This invention incorporates the vehicle's current status and the historical network quality of the area into the judgment criteria, transforming channel switching from a reactive response to a predictive selection for a future preset time period. This allows for early channel switching when a vehicle is about to enter a network anomaly area, improving the continuity, timeliness, and overall reliability of communication services.

Description

Technical Field

[0001] This invention relates to the field of vehicle-to-everything (V2X) communication technology, specifically to a method, apparatus, device, and storage medium for switching communication channels. Background Technology

[0002] Currently, vehicle-mounted terminals widely adopt access and handover methods based on network status monitoring to achieve data communication between vehicles and cloud servers. For example, in a typical cellular network environment, the vehicle-mounted terminal monitors the network quality of the current wireless network in real time. When the current network quality meets the preset communication requirements, the vehicle-mounted terminal maintains data transmission on the current communication channel; once the current network quality is detected to be lower than a set threshold, it attempts to switch to other available communication channels.

[0003] However, existing solutions mostly rely on fixed thresholds or preset rules to trigger the switching of communication channels, which is essentially a passive response decision-making mechanism, and this approach has the problem of switching lag. Summary of the Invention

[0004] The purpose of this invention is to provide a method, apparatus, device, and storage medium for switching communication channels, so as to solve the problem of switching lag in the current communication channel switching methods.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] In a first aspect, the present invention provides a method for switching communication channels, applied to an in-vehicle terminal, comprising: acquiring current vehicle status data of a target vehicle, and acquiring current network status data of multiple available communication channels of the target vehicle; determining a target area based on the area where the target vehicle is located, and acquiring historical network quality data of each communication channel when other vehicles have historically passed through the target area; predicting the predicted network quality of each communication channel within a future preset time period based on the current vehicle status data, the current network status data, and the historical network quality data; determining a target communication channel among multiple communication channels whose predicted network quality meets preset network quality requirements, and switching to the target communication channel.

[0007] Furthermore, based on current vehicle status data, current network status data, and historical network quality data, the predicted network quality of each communication channel within a preset future time period is predicted, including: determining the congestion level identifier of each communication channel based on the current network status data, the congestion level identifier being used to indicate the network congestion status of the corresponding communication channel; obtaining historical network status data of each communication channel within a preset historical time period; and predicting the predicted network quality of each communication channel within a preset future time period based on the congestion level identifier and / or historical network status data, as well as the current vehicle status data, current network status data, and historical network quality data.

[0008] Furthermore, among multiple communication channels, a target communication channel whose predicted network quality meets the preset network quality requirements is identified, and the user switches to the target communication channel. This includes: obtaining the service type of the vehicle terminal and determining the service priority corresponding to the service type; obtaining the preset network quality requirements corresponding to the service priority; if the predicted network quality of the current communication channel does not meet the preset network quality requirements, identifying a target communication channel among multiple communication channels that meets the preset network quality requirements; and switching from the current communication channel to the target communication channel.

[0009] Furthermore, switching from the current communication channel to the target communication channel includes: establishing a communication link with the target communication channel; when the quality deviation between the predicted network quality of the target communication channel and the predicted network quality of the current communication channel is greater than the switching threshold, migrating the corresponding transmission link from the current communication channel to the target communication channel, wherein during the migration of the communication channel, the communication link with the current communication channel is not interrupted and the transmission control protocol session and application layer connection of the upper layer application of the vehicle terminal are not interrupted.

[0010] Furthermore, predicting network quality includes: network quality scoring, and determining target communication channels that meet preset network quality requirements among multiple communication channels, including: if the service priority is preset high priority, determining the communication channel with the highest network quality score as the target communication channel among multiple communication channels; if the service priority is preset medium priority, determining at least two communication channels with network quality scores greater than the score threshold as target communication channels among multiple communication channels, wherein at least two target communication channels are used for concurrent transmission of corresponding service data.

[0011] Furthermore, the method for predicting network quality includes a signal identifier, which indicates whether a target vehicle will enter the network anomaly area of ​​the corresponding communication channel within a preset future time period. The method also includes: when the service priority is a preset low priority and the signal identifier indicates that the target vehicle will enter the network anomaly area of ​​the current communication channel within a preset future time period, storing the corresponding service data locally; and when the target vehicle leaves the network anomaly area of ​​the current communication channel, transmitting the corresponding service data based on the current communication channel.

[0012] Furthermore, it also includes: when the signal indicates that the area the target vehicle will enter in the future within a preset time period is a network abnormal area of ​​each communication channel, if the service priority is a preset high priority or a preset medium priority, the corresponding service data will be stored separately; if the service priority is a preset low priority, the corresponding service data will be stored centrally; if the target vehicle leaves the network abnormal area, the service data with higher priority will be transmitted first.

[0013] Furthermore, based on current vehicle status data, current network status data, and historical network quality data, the predicted network quality of each communication channel within a preset future time period is predicted. This includes: preprocessing the current vehicle status data, current network status data, and historical network quality data through the input layer of a pre-trained network prediction model to obtain preprocessed data features; performing prediction processing on the data features through the hidden layer of the network prediction model to obtain predicted data features; and outputting the predicted data features through the output layer of the network prediction model to obtain the predicted network quality of each communication channel.

[0014] Furthermore, the method also includes: receiving a network prediction model sent by a cloud server, the network prediction model being trained based on historical sample data from multiple vehicles; acquiring locally collected training sample data; and adjusting the model parameters of the network prediction model based on the training sample data.

[0015] Secondly, embodiments of the present invention provide a communication channel switching device applied to a vehicle-mounted terminal, comprising:

[0016] The first acquisition module is used to acquire the current vehicle status data of the target vehicle and the current network status data of multiple available communication channels of the target vehicle.

[0017] The second acquisition module is used to determine the target area based on the area where the target vehicle is located, and to acquire historical network quality data of each communication channel when other vehicles pass through the target area in the past.

[0018] The prediction module is used to predict the network quality of each communication channel within a preset time period based on current vehicle status data, current network status data, and historical network quality data.

[0019] The switching module is used to switch to the target communication channel whose predicted network quality meets the preset network quality requirements.

[0020] Thirdly, embodiments of the present invention provide a communication channel switching device, including: a processor, and a memory communicatively connected to the processor;

[0021] The memory stores the instructions that the computer executes;

[0022] The processor executes computer execution instructions stored in memory to implement the communication channel switching method provided in the first aspect above.

[0023] Fourthly, embodiments of the present invention provide a vehicle, including: a vehicle body and a communication channel switching device as provided in the third aspect above.

[0024] Fifthly, embodiments of the present invention provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the communication channel switching method provided in the first aspect above.

[0025] In a sixth aspect, embodiments of the present invention provide a computer program product, comprising: a computer program, wherein when the computer program is executed by a processor, it implements the communication channel switching method provided in the first aspect above.

[0026] The beneficial effects of this invention are:

[0027] This invention provides a method, apparatus, device, and storage medium for switching communication channels. The method includes: acquiring current vehicle status data of a target vehicle and acquiring current network status data of multiple available communication channels of the target vehicle; determining a target area based on the area where the target vehicle is located, and acquiring historical network quality data of each communication channel when other vehicles have historically passed through the target area; predicting the predicted network quality of each communication channel within a preset future time period based on the current vehicle status data, current network status data, and historical network quality data; determining a target communication channel among multiple communication channels whose predicted network quality meets preset network quality requirements, and switching to the target communication channel. This invention no longer relies solely on static network indicators at the current moment, but incorporates the vehicle's current status and the historical network quality of the area into the judgment criteria, transforming channel switching from a reactive response to a predictive selection for a preset future time period. This allows for advance channel adjustment when a vehicle is about to enter a weak network, blind spot, or network fluctuation area, avoiding switching lag and improving the continuity, timeliness, and overall reliability of communication services. Attached Figure Description

[0028] Figure 1 This is a structural diagram illustrating an application scenario provided in an embodiment of the present invention;

[0029] Figure 2 A flowchart illustrating the communication channel switching method provided in an embodiment of the present invention;

[0030] Figure 3 This is a schematic diagram of the structure of the network prediction model provided in an embodiment of the present invention;

[0031] Figure 4 A schematic diagram of the communication channel switching device provided in an embodiment of the present invention;

[0032] Figure 5 This is a schematic diagram of the structure of a communication channel switching device provided in an embodiment of the present invention.

[0033] The accompanying drawings have illustrated specific embodiments of the invention, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the invention in any way, but rather to illustrate the concept of the invention to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0034] The embodiments of the present invention will be described below with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are only for illustrating the present invention and not for limiting the scope of protection of the present invention.

[0035] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0036] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples of apparatuses and methods consistent with some aspects of the invention as detailed in the appended claims.

[0037] The terms "comprising," "including," or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, product, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, product, or apparatus. Without further limitation, the presence of additional identical or equivalent elements in the process, method, product, or apparatus that includes elements is not excluded. For example, the use of terms such as "first," "second," etc., to indicate names does not imply any particular order.

[0038] Vehicle-mounted terminals typically serve as data interaction nodes between vehicles and cloud servers, undertaking communication tasks such as vehicle status reporting, receiving remote control commands, transmitting alarm information, and synchronizing operational data. In practical applications, target vehicles often operate in various complex environments, including urban roads, highways, interchanges, tunnels, underground parking lots, mountain roads, and urban-rural fringe areas, where communication conditions fluctuate rapidly with changes in vehicle location. To ensure continuous network access during vehicle operation, multiple available communication channels can be configured for the vehicle-mounted terminal, allowing switching to a higher-quality communication channel when the current channel's network quality is poor.

[0039] Existing communication channel switching methods typically rely on monitoring current network conditions. The basic idea is to continuously collect network status data from the onboard terminal, such as signal strength, network latency, bit error rate, and connection success rate. When a parameter falls below a preset threshold, or the current connection is interrupted or weakens to a set level, switching to another communication channel is initiated according to preset priority rules. This switching method is relatively straightforward, usually involving periodic detection, threshold comparison, and switching execution. It is characterized by low deployment costs and well-defined rules, and is therefore widely used in existing vehicle communication systems. However, this switching method is essentially a passive response mechanism. Its decision-making is primarily based on the current network condition, rather than the future trend of network quality. In scenarios such as high-speed vehicle movement, or when entering tunnels or underground spaces, network quality often deteriorates rapidly within a short period. Fixed threshold triggering mechanisms only initiate switching after network degradation, inevitably resulting in response lag.

[0040] To address the aforementioned issues, this invention first acquires the current vehicle status data of the target vehicle and the current network status data of multiple available communication channels. Then, based on the area where the target vehicle is located, a target area is determined, and historical network quality data of each communication channel is acquired when other vehicles have passed through this target area. Next, based on the current vehicle status data, current network status data, and historical network quality data, the network quality of each communication channel within a preset future time period is predicted. Finally, the system switches to the target communication channel whose predicted network quality meets the preset network quality requirements. This invention no longer relies solely on static network indicators at the current moment but incorporates the vehicle's current status and the historical network quality of its location into the judgment criteria. This transforms channel switching from a reactive response to a predictive selection for a preset future time period, enabling early channel switching when a vehicle is about to enter a weak network, blind spot, or network fluctuation area. This avoids switching lag issues and improves the continuity, timeliness, and overall reliability of communication services.

[0041] The following is a combination of... Figure 1The application scenarios of the embodiments of the present invention are described.

[0042] Figure 1 A structural diagram illustrating an application scenario provided in an embodiment of the present invention. For example... Figure 1 As shown, the application scenarios applicable to embodiments of the present invention include a target vehicle, a network transmission layer, and a cloud server. Specifically, an in-vehicle terminal is deployed in the target vehicle, and the in-vehicle terminal can communicate with the cloud server through the network transmission layer.

[0043] like Figure 1 As shown, the network transport layer includes: cellular network channels, Wi-Fi (Wireless Fidelity) channels, satellite channels, and V2X (Vehicle-to-Everything) channels. Among these, the cellular network channels include communication channels from at least one operator, such as the first operator's 4G (4th Generation Mobile Communication Technology) channel, the first operator's 5G (5th Generation Mobile Communication Technology) channel, the second operator's 4G channel, the second operator's 5G channel, the third operator's 4G channel, and the third operator's 5G channel, etc.

[0044] For example, the target vehicle can be a passenger car or a commercial vehicle, and the vehicle can be in a moving state or a stationary state. In the moving state, the corresponding adapted driving scenario can be an urban road scenario or a highway driving scenario. This embodiment of the invention does not limit the vehicle type or the corresponding adapted driving scenario; it can be determined according to the actual application requirements.

[0045] The vehicle terminal can be based on a vehicle-mounted TBOX (Telematics BOX, vehicle communication gateway) as the core hardware carrier, integrating multi-mode communication modules, such as 5G / 4G cellular communication units, BeiDou / GPS (Global Positioning System) dual-mode satellite communication units, Wi-Fi (Wireless Fidelity) units, etc., and externally connected to the target vehicle's CAN (Controller Area Network Bus), sensor clusters (including but not limited to millimeter-wave radar, cameras, accelerometers, gyroscopes and light sensors), and high-precision positioning modules.

[0046] It should be noted that the cloud server can be a local private cloud device or a public shared cloud device, and this invention does not limit it in this regard.

[0047] This invention can also be applied to other smart terminals, such as smart inspection terminals, smart drones, shipborne communication terminals, etc., without limitation.

[0048] The following is based on the above. Figure 1 The vehicle-mounted terminal shown is the execution subject. The specific implementation of the communication channel switching method provided in the embodiments of the present invention will be described in detail with reference to specific embodiments.

[0049] Figure 2 This is a flowchart illustrating the communication channel switching method provided in an embodiment of the present invention. Figure 2 As shown, a specific implementation of this communication channel switching method may include the following steps:

[0050] S201, Obtain the current vehicle status data of the target vehicle, and obtain the current network status data of multiple available communication channels of the target vehicle.

[0051] Among them, the vehicle terminal is deployed on the target vehicle. The vehicle terminal is the terminal that is executed as the communication channel switching method. The communication channel is the path used to carry business data transmission between the vehicle terminal and the cloud server. Multiple available communication channels represent multiple candidate transmission links that the vehicle terminal can access and use to transmit business data at the current moment.

[0052] In this invention, the vehicle-mounted terminal, as the execution entity, can be deployed in a TBOX (Telematics Box), vehicle gateway, integrated communication controller, or vehicle computing unit with multi-standard wireless access capability, and establish data interaction relationships with the target vehicle's CAN bus, positioning module, inertial measurement unit, and cloud server, respectively.

[0053] The vehicle-mounted terminal can collect current vehicle status data and current network status data in real time at a preset fixed period (such as 500ms).

[0054] In this invention, the current vehicle status data is used to reflect the current operating status of the target vehicle. Specifically, the vehicle terminal obtains the current vehicle status data in real time through the CAN bus and the location positioning module.

[0055] Among them, the vehicle status data of the target vehicle is the data to be transmitted. This vehicle status data is used to reflect the operating status, event status or fault status of the target vehicle at the current moment or within a certain time range.

[0056] Specifically, the vehicle status data of the target vehicle includes at least one of the following: vehicle location, vehicle speed, gear position, driving scenario, and CAN bus data. The vehicle location can be latitude and longitude coordinates collected by a positioning module (such as a GPS / BeiDou positioning module). The vehicle speed is the real-time speed of the target vehicle collected via the CAN bus, in km / h. The gear position includes one of P, D, R, or N. The driving scenario includes: underground parking garage, tunnel, highway driving, urban driving, suburban driving, or stationary driving, etc. CAN bus data includes engine speed, battery voltage, fuel level, brake status, or fault codes, etc.

[0057] In this invention, each communication channel has corresponding current network status data. For example, if the communication channels available to the target vehicle include: a 5G cellular network channel of a first operator, a 4G cellular network channel of a second operator, and a 4G cellular network channel of a third operator, then the vehicle terminal needs to obtain the current network status data a1 of the 5G cellular network channel of the first operator, a2 of the 4G cellular network channel of the second operator, and a3 of the 4G cellular network channel of the third operator.

[0058] Specifically, the current network status data includes at least one of the following: signal strength, signal-to-noise ratio (SNR), transmission delay, jitter, packet loss rate, network standard information, carrier information, and network connection status. Signal strength can be represented by RSRP (Reference Signal Received Power), which corresponds to a preset collection range, such as -30dBm to -150dBm. The unit for signal-to-noise ratio (SINR) can be dm. Network standard information refers to the corresponding communication channel, such as 4G, 5G, V2X, Wi-Fi, or satellite channel. 5G can be in SA (Standalone) mode or NSA (Non-Standalone) mode. Carrier information indicates the carrier to which the 4G and 5G communication channels belong, such as a first carrier, a second carrier, or a third carrier. Network connection status includes normal connection, abnormal connection, or signal interruption.

[0059] S202, based on the area where the target vehicle is located, determine the target area and obtain historical network quality data of each communication channel when other vehicles passed through the target area in the past.

[0060] The target area is the specific spatial region where the target vehicle is currently located, is passing through, or is about to enter, such as a specific road segment. Historical network quality data of each communication channel when other vehicles have passed through the target area is used to characterize the network performance patterns of different communication channels in the historical operation of this area. Historical network quality data is a structured description of the communication effect of communication channels in a certain area and time period.

[0061] In practical implementation, the vehicle-mounted terminal can determine the target area based on the vehicle position, speed, direction of travel, and scene information obtained in S101. One scenario is that the vehicle-mounted terminal pre-stores a pre-built target map, which consists of multiple regions. Each region contains corresponding historical network quality data. This historical network quality data can be obtained by a cloud server collecting historical network quality data of each communication channel when other vehicles pass through each region. The cloud server then pre-sends the historical network quality data of each communication channel corresponding to each region to the vehicle-mounted terminal. After determining the target area, the vehicle-mounted terminal searches locally for the historical network quality data of each communication channel corresponding to that target area. Another scenario is that the vehicle-mounted terminal sends the vehicle position, speed, direction of travel, and scene information to the cloud server. The cloud server determines the target area and the historical network quality data of each communication channel corresponding to the target area, and then sends this historical network quality data to the vehicle-mounted terminal.

[0062] In this invention, the multiple regions in the target map can be constructed using a gridded approach, such as dividing the road space into fixed-size geographic grids; or they can be constructed using a road link segmentation approach, such as dividing the area into multiple regions according to road mileage markers, road topology nodes, tunnel entrances and exits, ramp sections, bridge sections, or parking lot levels; or a composite region model can be used, taking location, road type, elevation changes, and occlusion features as the basis for region division.

[0063] In one possible embodiment, the target vehicle is first located to a map grid cell based on its current position coordinates. Then, based on the vehicle's direction of travel and speed, adjacent grid cells that it will enter within a preset distance are identified. The current grid cell and the forward grid cell are then defined together as the target area. Thus, since the communication channel switching in this invention is for a preset future time period, the target area is not limited to the current location but can also cover the spatial range corresponding to the short-term predicted trajectory of the target vehicle, thereby more closely reflecting the network conditions of the area the target vehicle will enter in the future.

[0064] Historical network quality data includes numerical values ​​used to identify network quality, such as "2" indicating good network quality, "1" indicating poor network quality, and "0" indicating a weak network or a network blind spot. Furthermore, historical network quality data may also include at least one of the following within a specific target area: signal strength variation trajectory, latency variation trajectory, packet loss rate variation trajectory, throughput statistics, network quality mean, variance, fluctuation amplitude, stable duration, disconnection probability, reconnection success rate, and peak-hour congestion characteristics. In this invention, historical network quality data may also include other content, which is not limited thereto.

[0065] Furthermore, the historical network quality data for each communication channel is a statistical result. For example, other vehicles b1 upload historical network status data c1 for each communication channel they collected when passing through the target area, other vehicles b2 upload historical network status data c2 for each communication channel they collected when passing through the target area, and other vehicles b3 upload historical network status data c3 for each communication channel they collected when passing through the target area. The cloud server can process the historical network status data c1, c2, and c3 to obtain historical network quality data. For example, it can calculate the average of the historical network status data or normalize the historical network status data. This invention does not limit the specific processing method.

[0066] In this invention, the introduction of the target area makes network judgment no longer limited to the current instantaneous state, but takes the spatial environment in which the target vehicle is located as the analysis boundary, enabling the vehicle terminal to identify potential weak network areas, blind spots and high fluctuation areas by combining the historical network quality data of the target area, thereby enhancing the accuracy of communication channel selection.

[0067] S203, based on current vehicle status data, current network status data and historical network quality data, predicts the predicted network quality of each communication channel within a preset future time period.

[0068] Specifically, the current vehicle status data can be preprocessed to obtain preprocessed current vehicle status data, and the current network status data can be preprocessed to obtain preprocessed current network status data. Then, based on the preprocessed current vehicle status data, the preprocessed current network status data, and historical network quality data, the predicted network quality of each communication channel within a preset future time period can be predicted.

[0069] Specifically, data preprocessing includes preprocessing operations such as standardization, normalization, and outlier correction.

[0070] The standardization process can be achieved using the Z-score (standardized value) expression: ,in, This represents the i-th parameter in the current vehicle status data or the current network status data, such as the vehicle position, speed, signal strength, or signal-to-noise ratio mentioned above. Indicates to Standardized parameters obtained after standardization processing. Indicates parameters The corresponding average value, which can be obtained in advance based on the corresponding historical data. express The corresponding variance value can also be obtained in advance based on the corresponding historical data.

[0071] Furthermore, the standardized parameters are normalized to obtain normalized parameters. The expression for normalization is: .in, Indicates to The normalized parameters obtained after normalization. Indicates parameters The corresponding minimum standardized parameter. Indicates parameters The corresponding maximum normalized parameter. Wherein, and All of them were pre-calibrated based on historical data.

[0072] In this invention, the outlier correction process specifically involves the following steps: If a normalized parameter is outside a preset range and its deviation from the preset range is less than or equal to a difference threshold, it can be corrected using linear interpolation. If a normalized parameter is outside a preset range and its deviation from the preset range is greater than a difference threshold, the normalized parameter can be deleted, or a normal normalized parameter from the same historical period can be used as a substitute. For example, a normalized parameter in the range [0, 1] is considered normal. A normalized parameter in the range [-0.09, 0) or (1, 0.01] is considered slightly abnormal and is corrected linearly. A normalized parameter less than -0.09, 0, or greater than 0.01 is considered severely abnormal and is deleted.

[0073] In this step, the future preset time period is used to define the prediction time window. The predicted network quality is an evaluation result of whether each communication channel is suitable for carrying service communication within this time window. The current vehicle status data, current network status data, and historical network quality data are fused together to predict network quality changes that have not yet occurred in the future preset time period, so that the handover action is triggered before the network quality changes, instead of after the network quality changes.

[0074] The preset future time period can range from 3 to 60 seconds, for example, predicting the network quality of each communication channel within the next 5 seconds. The selection of this preset time period is related to vehicle speed, service type, and handover execution latency. When the target vehicle is in a high-speed scenario and carrying low-latency services such as remote control and alarm uploading, the preset future time period can be set to a shorter time window. When the target vehicle is carrying periodic status synchronization or log uploading services, a longer time window can be set to enhance trend prediction capabilities.

[0075] In this invention, setting the future preset time period too short will make the prediction too close to the current measurement and difficult to reflect foresight; setting it too long may introduce too much uncertainty and reduce the accuracy of switching. Therefore, the embodiments of this application can combine the preset service latency tolerance and vehicle motion characteristics to select a reasonable future preset time period, so that the subsequent communication channel switching decision takes into account both timeliness and stability.

[0076] In one possible implementation, a mapping relationship can be preset between current vehicle status data, current network status data, historical network quality data and the predicted network quality of each communication channel, and then the predicted network quality of each communication channel in a future preset time period can be determined based on this mapping relationship.

[0077] In another possible implementation, a pre-trained network prediction model can be used to process current vehicle status data, current network status data, and historical network quality data to obtain the predicted network quality of each communication channel within a preset future time period.

[0078] In this embodiment, after the vehicle terminal powers on, it automatically initializes all its functional modules. Simultaneously, it completes multi-network access initialization, starts the multi-mode communication module, establishes an initial connection with the cloud server, receives the latest network prediction model and network quality parameters from the cloud server, completes system initialization, and enters a standby state. After entering the standby state, the vehicle terminal can continuously input the data required for the network prediction model at fixed intervals of 500ms.

[0079] In one possible implementation, a weighted calculation method can be used to calculate the predicted network quality of each communication channel. The expression for this weighted calculation method can be: .in, This represents the predicted network quality of the j-th communication channel. , and Indicates the corresponding weight. This represents the quality score calculated based on the current vehicle status data from the j-th communication channel. This represents the quality score calculated based on the current network status data of the j-th communication channel. This represents the quality score calculated based on historical network quality data of the j-th communication channel. This invention does not limit the specific calculation process.

[0080] In this invention, the predicted network quality of each communication channel within a preset time period can also be predicted in other ways based on current vehicle status data, current network status data, and historical network quality data, without limitation.

[0081] The predicted network quality includes at least one of the following: network quality score, network stability level, and signal indicator. The network quality score can range from 0 to 100, with 80 to 100 indicating excellent network quality, 60 to 79 indicating good network quality, 40 to 59 indicating average network quality, and 0 to 39 indicating poor network quality. The network stability level can be any one of the following: excellent, good, average, or poor.

[0082] Signal identifiers are used to indicate whether a target vehicle has entered a network anomaly area of ​​the corresponding communication channel within a preset future time period. For example, a signal identifier of 1 for the 5G cellular network channel of the first operator indicates that the target vehicle has entered the network anomaly area of ​​the corresponding 5G cellular network channel of the first operator within the preset future time period. A signal identifier of 0 for the 4G cellular network channel of the second operator indicates that the target vehicle has not entered the network anomaly area of ​​the corresponding 4G cellular network channel of the second operator within the preset future time period. Here, a network anomaly area refers to a weak network area or a network blind spot.

[0083] This step, by fusing current vehicle status data, current network status data for each communication channel, and historical network quality data for the target area where the target vehicle is located, enables the prediction of future trends in network quality for each communication channel. Especially in scenarios involving high-speed movement, terrain occlusion, and significant coverage alternation, network degradation is typically determined first by regional attributes and then by real-time fluctuations, rather than being solely determined by the current instantaneous value. Therefore, using historical network quality data as a regional prior, current network status data as an immediate observation, and current vehicle status data as a motion trend condition, can more realistically reflect the communication environment over a predetermined time period, thus leading to more accurate predictions of network quality.

[0084] In one possible implementation, the current network quality score of each communication channel can be determined based on the current vehicle status data and the current network status data. The service data can then be transmitted using the communication channel corresponding to the current network quality score that is compatible with the service data to be transmitted.

[0085] Based on the above analysis, this step enables the vehicle terminal to complete the identification of the target communication channel before the network quality actually deteriorates, thereby reducing service interruption, reducing handover jitter, and improving the transmission continuity of remote service data.

[0086] S204, determine the target communication channel among multiple communication channels whose predicted network quality meets the preset network quality requirements, and switch to the target communication channel.

[0087] In this step, the preset network quality requirements are used to constrain the quality conditions that the selected communication channels should meet. The target communication channel is selected from multiple available communication channels after prediction and judgment for subsequent service transmission. Based on the above, after obtaining the predicted network quality of each communication channel within a preset future time period, the vehicle terminal needs to complete the selection of the target communication channel according to the preset network quality requirements and complete the communication channel switching when the switching execution conditions are met to ensure the continuous operation of vehicle communication services.

[0088] In the specific implementation process, the vehicle-mounted terminal can first determine the predicted network quality of each communication channel obtained in S103. Preset network quality requirements may include a predicted network quality score higher than a first score threshold, a network stability level higher than a preset level, or not entering a network anomaly area. If multiple communication channels meet the requirements, the vehicle-mounted terminal can determine the target communication channel according to preset rules. These preset rules can be either the highest predicted network quality score or the highest network stability level.

[0089] Furthermore, to avoid frequent switching of communication channels due to short-term fluctuations, the vehicle-mounted terminal can confirm the switching conditions before officially executing the switch. These switching conditions may include the target communication channel meeting preset network quality requirements for several consecutive detection cycles, the target communication channel maintaining an advantage over the current communication channel for a duration exceeding a threshold, the current service being in a switch-allowed state, and successful access authentication for the target channel.

[0090] The vehicle-mounted terminal can first establish a connection with the target communication channel in the background, such as completing network attachment, address acquisition, session establishment, or security authentication in advance, and then switch to the target communication channel when the switching trigger conditions are met. The vehicle-mounted terminal supports concurrent data transmission using multiple communication channels, and can also achieve smooth switching of communication channels by first establishing a communication link with the target communication channel and then disconnecting the communication link with the current communication channel.

[0091] In one possible implementation, if the current communication channel predicts that the network quality still meets the preset network quality requirements and is better than other communication channels, the vehicle terminal can maintain the current connection and continue to periodically execute S201 to S203.

[0092] In this invention, after a successful communication channel switch, the vehicle terminal can also record the execution result of this switch, including the channel identifier before the switch, the channel identifier after the switch, the switch time, the switch reason, the actual network indicators before and after the switch, the service interruption duration, and the switch success marker, and upload the above information to the cloud server for updating subsequent historical network quality data.

[0093] If the communication channel switching fails, such as when the target communication channel fails to connect or when the network quality of the target communication channel drops rapidly after the switch, the vehicle terminal can restore to the original communication channel or select other alternative communication channels according to the fallback mechanism to ensure business continuity.

[0094] Based on the above, this step, by combining predicted network quality with preset network quality requirements, can suppress repeated handovers and communication jitter while ensuring handover accuracy. Compared to the current mechanism of passively handing over only after the network has deteriorated, this step enables the vehicle terminal to establish a connection and complete the handover in advance on the expected target communication channel within a preset future time period. Therefore, it can shorten the duration of service disruption and improve the reliability and timeliness of data transmission.

[0095] In this invention, the current vehicle status data of the target vehicle, the current network status data of multiple communication channels, and the historical network quality data accumulated by other vehicles in the target area are fused and processed. The network quality of each communication channel is predicted within a future preset time window. This enables the vehicle terminal to identify a more suitable target communication channel and complete the switch before the communication quality deteriorates. In this way, the continuity, stability and overall reliability of data transmission of the vehicle terminal in complex road environments and high-speed moving scenarios can be improved.

[0096] Optionally, based on current vehicle status data, current network status data, and historical network quality data, the predicted network quality of each communication channel within a preset future time period is predicted, including: determining the congestion level identifier of each communication channel based on the current network status data, wherein the congestion level identifier is used to indicate the network congestion status of the corresponding communication channel; obtaining historical network status data of each communication channel within a preset historical time period; and predicting the predicted network quality of each communication channel within a preset future time period based on the congestion level identifier and / or historical network status data, as well as the current vehicle status data, current network status data, and historical network quality data.

[0097] The congestion level identifier is used to characterize the degree of congestion of a communication channel at a given moment. It is typically categorized based on channel occupancy, queuing delay, packet loss rate, access failure rate, or service contention intensity, thus mapping different communication channels to different congestion states. For example, a congestion level identifier of "3" indicates a high degree of congestion for the corresponding communication channel at the current moment, "2" indicates a moderate degree of congestion, "1" indicates a low degree of congestion, and "2" indicates no congestion.

[0098] In this invention, historical network status data refers to the evolution of the network status of each communication channel of the target vehicle within a preset historical time period, for example, the preset historical time period is within 2 minutes from the current time. This historical network status data can be sequential data that changes over time. For example, if network status data is collected every 30 seconds, then the historical network status data includes: network status data collected 30 seconds before the current time, network status data collected 1 minute before the current time, network status data collected 1 minute and 30 seconds before the current time, and network status data collected 2 minutes before the current time.

[0099] In practical implementation, after the vehicle terminal acquires the current vehicle status data and the current network status data, it uses the aforementioned data preprocessing method to preprocess the current vehicle status data and the current network status data, obtaining preprocessed current vehicle status data and current network status data. Alternatively, the aforementioned data preprocessing method can also be used to preprocess historical network status data, obtaining preprocessed historical network status data.

[0100] In this invention, the vehicle terminal can determine the predicted network quality of each communication channel based on preprocessed current vehicle status data, preprocessed current network status data, preprocessed historical network status data, historical network quality data, and congestion level identifier.

[0101] In one possible implementation, the vehicle terminal retrieves historical network status data for each communication channel from a local cache or a cloud-based historical database, and inputs the congestion level identifier and / or historical network status data together with the current vehicle status data into the network prediction model. The network prediction model estimates the predicted network quality of each communication channel within a preset future time period and outputs the predicted network quality of each communication channel within the preset future time period.

[0102] This invention incorporates the current congestion level of each communication channel with historical network data collected from the target vehicle for each communication channel into the process of predicting network quality. This allows the predicted network quality to not only reflect the current link status but also to infer communication environment trends by considering vehicle movement characteristics. Consequently, the vehicle-mounted terminal can know the network quality of each communication channel in advance for a preset period, providing a basis for subsequent selection of target communication channels.

[0103] By adopting the above method, communication quality can be predicted before the vehicle enters an area with weak coverage, congestion, or frequent handover, reducing handover lag caused by network spikes, improving the accuracy and stability of target communication channel selection, and reducing the probability of repeated handovers, service interruptions, and sudden increases in latency, thereby enhancing the continuity and reliability of the vehicle communication link.

[0104] Optionally, determining a target communication channel whose predicted network quality meets preset network quality requirements among multiple communication channels, and switching to the target communication channel, includes: obtaining the service type of the vehicle terminal and determining the service priority corresponding to the service type; obtaining the preset network quality requirements corresponding to the service priority; determining a target communication channel that meets the preset network quality requirements among multiple communication channels when the predicted network quality of the current communication channel does not meet the preset network quality requirements; and switching from the current communication channel to the target communication channel.

[0105] Among them, business data refers to the data that the vehicle terminal is currently uploading to the cloud server or receiving from the cloud service. Business type is used to characterize the business category corresponding to the data. Business priority is used to reflect the importance of the transmission of the business. Preset network quality requirements are used to limit the available conditions of the communication channel that match the priority of the business.

[0106] After detecting service data to be sent or received, the vehicle-mounted terminal first identifies the service type based on the service identifier, message type field, service subject field, or protocol header information of the service data. Then, it matches the service type with a preset mapping table to determine the corresponding service priority. The mapping table can be pre-distributed by a cloud server and stored locally on the vehicle-mounted terminal, allowing different priorities to be assigned to different services such as remote control, alarm reporting, and status synchronization. Based on this priority, the vehicle-mounted terminal further reads the corresponding preset network quality requirements and uses these requirements as constraints for subsequent channel selection.

[0107] In practical implementation, a mapping relationship between various service priorities and corresponding preset network quality requirements can be pre-established. Then, after determining the service priority of the service type of the currently transmitted data, the corresponding preset network quality requirements can be determined based on this mapping relationship. It can be understood that different service priorities each have their corresponding preset network quality requirements.

[0108] In this invention, the type of service currently being executed by the vehicle terminal is identified, and the priority of the current service is determined according to a preset service priority classification rule. Service priorities can be divided into high-priority services, medium-priority services, and low-priority services.

[0109] For example, vehicle remote control services can be classified as high-priority services, such as remote start, remote unlocking, remote air conditioning control, or vehicle location. High-priority services require transmission latency less than a first latency threshold (e.g., 100 milliseconds), packet loss rate less than a first packet loss rate threshold (e.g., 1%), and reliability greater than or equal to a first reliability threshold (e.g., 99.9%). Vehicle emergency event reporting services can be classified as medium-priority services, such as collision alarms, fault alarms, and abnormal status reporting. Medium-priority services require transmission latency less than a second latency threshold (e.g., 500 milliseconds), packet loss rate less than a second packet loss rate threshold (e.g., 3%), and reliability greater than or equal to a second reliability threshold (e.g., 99.5%). Regular vehicle condition reporting services can be classified as low-priority services. Regular vehicle condition reporting services include the regular reporting of data such as mileage, voltage, fuel level, and location information. Low-priority services require transmission latency to be less than the third latency threshold (e.g., 1000 milliseconds), packet loss rate to be less than the third packet loss rate threshold (e.g., 5%), and reliability to be greater than or equal to the third reliability threshold (e.g., 99%).

[0110] Furthermore, each service priority has corresponding preset network quality requirements, which include preset network quality score thresholds. For example, the preset network quality score threshold for high-priority services is 80 points, for medium-priority services it is 60 points, and for low-priority services it is 40 points.

[0111] Specifically, if the network quality score of the current communication channel is lower than the network quality score threshold corresponding to the current service priority, the system switches to a target communication channel with a network quality score higher than that threshold. For example, if the current service type corresponds to a high-priority service, and the network quality score of the current communication channel is 70, then the corresponding network quality score threshold (80) is not met. Therefore, it is necessary to select a communication channel with a network quality score greater than or equal to 80 from among multiple communication channels as the target communication channel, and then switch to the target communication channel.

[0112] In one alternative implementation, if the target vehicle is about to enter a network anomaly area of ​​the current communication channel, and the network of other communication channels in that area is normal, then the target communication channel with a normal network in that area can be switched to.

[0113] In one alternative implementation, if new service data is to be transmitted, and the service type corresponding to the new service data is different from the current service type, and the predicted network quality of the current communication channel does not match the network quality requirements corresponding to the new service type, then the system can switch to a target communication channel whose predicted network quality matches the network quality requirements corresponding to the new service type.

[0114] This invention shifts communication channel switching from solely relying on a single threshold based on the current network state to incorporating different network quality constraints based on the service priority corresponding to the service type. This allows high-priority services to obtain more stable communication resources first, while low-priority services can complete transmission on communication channels that meet basic requirements. This reduces the risk of increased latency, retransmissions, or interruptions caused by insufficient communication channel quality for critical services, improving the reliability and real-time performance of service data transmission.

[0115] Optionally, switching from the current communication channel to the target communication channel includes: establishing a communication link with the target communication channel; when the quality deviation between the predicted network quality of the target communication channel and the predicted network quality of the current communication channel is greater than a switching threshold, migrating the corresponding transmission link from the current communication channel to the target communication channel, wherein during the migration of the communication channel, the communication link with the current communication channel is not interrupted and the connection between the transmission control protocol session of the upper layer application of the vehicle terminal and the application layer is not interrupted.

[0116] The current communication channel refers to the initial channel that carries business data transmission before the communication channel switchover, the target communication channel refers to the channel to be migrated to receive business data transmission after the switchover, and the communication link is used to complete the pre-connection with the target communication channel before the communication channel switchover.

[0117] In this invention, quality deviation is used to characterize the difference in predicted network quality between two communication channels, and a switching threshold is used to limit the extent to which this difference will trigger migration. For example, the quality deviation is the difference between the network quality score of the target communication channel and the network quality score of the previous communication channel. The switching threshold in this invention is preset, for example, it can be set to a value between 5 and 10, such as 8 points. Setting a switching threshold can avoid frequent switching when the quality deviation is less than the switching threshold.

[0118] Furthermore, the transmission link for business data is used to indicate the path that actually carries the business messages, and the connection between the Transmission Control Protocol (TCP) session and the application layer is used to maintain the continuity of the upper-layer business session of the vehicle terminal.

[0119] In practical implementation, the vehicle terminal can obtain the quality deviation by comparing the predicted network quality of the current communication channel and the predicted network quality of the target communication channel output by the aforementioned network prediction model. When the quality deviation exceeds a preset switching threshold, the vehicle terminal first establishes an available communication link on the target communication channel side, and then smoothly migrates the transmission path of the service data to the target communication channel corresponding to that communication link. This can shorten the migration process time (e.g., less than 100 milliseconds) while preserving the connection status of the current communication channel to ensure normal interaction of retransmission, acknowledgment, and session maintenance information.

[0120] To avoid the upper-layer application of the vehicle terminal not perceiving the interruption, the present invention maintains the continuity of the transmission control protocol session and the application layer connection through session persistence during the migration process, thereby enabling remote control, status reporting and alarm transmission to continue during the handover.

[0121] By employing the above methods, the target communication channel is adopted only when it offers significantly superior quality compared to the current communication channel. This reduces frequent switching caused by minor fluctuations and maintains network connectivity and application session stability during link migration, thereby improving the continuity, reliability, and service transmission integrity of in-vehicle communication.

[0122] Optionally, predicting network quality includes: network quality scoring, and determining target communication channels that meet preset network quality requirements among multiple communication channels, including: if the service priority is preset high priority, determining the communication channel with the highest network quality score as the target communication channel among multiple communication channels; if the service priority is preset medium priority, determining at least two communication channels with network quality scores greater than a score threshold as target communication channels among multiple communication channels, wherein at least two target communication channels are used for concurrent transmission of corresponding service data.

[0123] Among them, the network quality score is used to quantitatively characterize the overall communication capability of the communication channel. The network quality score is determined based on the above method, which will not be elaborated here.

[0124] For high-priority service types, the preset network quality requirement refers to selecting the target communication channel with the highest network quality score from multiple communication channels to reduce switching complexity and link jitter. In this invention, if the communication channel with the highest network quality score among multiple communication channels is the current communication channel, then there is no need to switch communication channels, and the current communication channel can be used for data transmission.

[0125] For medium-priority service types, the preset network quality requirement refers to identifying at least two communication channels with network quality scores greater than a certain threshold as target communication channels. This allows for the simultaneous maintenance of at least two channels meeting the score threshold online, and the transmission of service data in fragments and concurrent transmission, or the synchronous transmission of the same data on different channels to achieve redundant transmission. In this invention, the score threshold is preset, for example, it could be 50 points.

[0126] This invention specifically defines predicted network quality as a network quality score and uses a preset score threshold as the criterion, making comparisons between different communication channels consistent and operable. For high-priority services, the system can quickly identify the communication channel with the highest network quality score, ensuring stable transmission and controllable latency. For medium-priority services, multiple communication channels can be enabled for concurrent transmission, provided the score threshold is met, thereby improving transmission efficiency and reducing the impact of single-channel fluctuations on service continuity.

[0127] By adopting the above method, the vehicle-mounted terminal can accurately select communication channels based on network quality scores, reducing misjudgments caused by inconsistent subjective thresholds, and establishing a clear channel selection method among different priority services. High-priority services can obtain more stable single-link guarantees, while medium-priority services can obtain higher concurrent transmission capabilities and redundancy reliability, thereby improving the accuracy of communication channel switching, the continuity of service transmission, and the overall communication quality in complex vehicle-mounted network environments.

[0128] Optionally, the network quality prediction includes a signal identifier, which indicates whether the target vehicle will enter the network anomaly area of ​​the corresponding communication channel within a preset future time period. The method further includes: when the service priority is a preset low priority and the signal identifier indicates that the target vehicle will enter the network anomaly area of ​​the current communication channel within a preset future time period, storing the corresponding service data locally; and when the target vehicle leaves the network anomaly area of ​​the current communication channel, transmitting the corresponding service data based on the current communication channel.

[0129] In this invention, the signal identifier can be obtained in the manner described above, and will not be repeated here. A network anomaly area refers to a situation where the corresponding communication channel experiences signal attenuation, increased latency, increased packet loss, or communication unavailability within the corresponding road area.

[0130] If the predicted network quality indicates that the target vehicle is about to enter a weak network or blind spot area of ​​the current communication channel, the service data to be transmitted can be stored locally. After the target vehicle leaves the weak network or blind spot area, the service data can be automatically uploaded in batches based on the current communication channel.

[0131] Specifically, after acquiring the service data to be transmitted and determining its priority, if the vehicle-mounted terminal determines that the service priority belongs to a preset low priority and the signal indicates that the target vehicle will enter the network anomaly area of ​​the current communication channel within a preset time period, then the service data to be transmitted is temporarily written to the local storage unit. The local storage unit can be the vehicle-mounted terminal's built-in flash memory, memory card, or other erasable and rewritable storage media to avoid the loss of low-priority data during weak network conditions. After the vehicle-mounted terminal re-determines that the target vehicle has left the network anomaly area of ​​the current communication channel, it reads the service data stored locally and completes the transmission through the data link established or restored by the current communication channel.

[0132] In this invention, the vehicle-mounted terminal first determines whether to implement a local caching strategy based on the signal identifier in the predicted network quality and the service priority. This allows low-priority services to be temporarily stored before entering an abnormal area and continue transmission after leaving the abnormal area. This method can avoid packet loss or retransmission of low-priority services in areas with weak coverage or communication interruptions, while reducing communication jitter caused by frequent channel switching, allowing the current communication channel to continue its transmission task after it becomes available again.

[0133] By adopting this implementation method, the vehicle terminal can proactively identify abnormal areas in the future and adopt a storage-before-transmission processing strategy for low-priority services, thereby improving the integrity of service data, reducing the probability of transmission failure in abnormal areas, reducing unnecessary channel switching and bandwidth occupation, and improving the continuity and stability of the vehicle communication system in complex road environments.

[0134] Optionally, it also includes: when the signal indicates that the area the target vehicle enters in a future preset time period is a network abnormal area of ​​each communication channel, if the service priority is a preset high priority or a preset medium priority, the corresponding service data is stored separately; if the service priority is a preset low priority, the corresponding service data is stored centrally; if the target vehicle leaves the network abnormal area, the service data with higher priority is transmitted first.

[0135] The signal identifier is used to indicate whether the target vehicle will enter the network anomaly area corresponding to each communication channel within a preset time period in the future. For example, the target vehicle is about to enter the network anomaly area, and the target vehicle will experience network anomalies in each communication channel in the network anomaly area.

[0136] Among them, the network abnormal area corresponding to each communication channel refers to the area where the communication link quality of each communication channel is abnormal within a preset period of time in the future. This network abnormal area can be a tunnel, underground parking garage, valley road section with obstruction, or area with insufficient base station coverage overlap.

[0137] Separate storage refers to establishing independent cache queues or independent storage spaces for high-priority or medium-priority business data to avoid retrieval and recovery transmission sequence disorder caused by mixing with other business data. Centralized storage refers to writing low-priority business data into a shared cache area or a unified file area to reduce storage management overhead.

[0138] In practical implementation, after obtaining the signal identifier, the vehicle terminal first reads the service priority identifier corresponding to the service data to be transmitted, and writes the high-priority or medium-priority service data into the corresponding independent storage unit. The storage unit can be composed of flash memory or eMMC (Embedded Multi Media Card). Low-priority service data is written to a centralized buffer, and its writing time and data length are managed by a unified index.

[0139] Furthermore, independent storage units and centralized buffers can be configured with different access flags to resume reading according to priority after the target vehicle leaves the network anomaly area. If the signal indicates that multiple communication channels will enter the network anomaly area, the vehicle terminal can establish a recovery transmission sequence table based on the service priority of each service data. When the network recovery detection result meets the conditions for continued transmission, high-priority service data is read first, followed by medium-priority and low-priority service data, and the transmission is completed according to the availability status of the corresponding communication channel.

[0140] This invention enables vehicle-mounted terminals to pre-distinguish storage methods and resume transmission according to priority after leaving the abnormal area, thereby preventing high-priority data from having its cache space occupied or its transmission delayed by low-priority data. Because high-priority and medium-priority data are stored independently, they can be quickly located and uploaded first during recovery. Centralized storage of low-priority data reduces management complexity, thus improving overall the reliability of business data preservation and the orderly nature of recovery transmission in abnormal network environments.

[0141] Optionally, based on current vehicle status data, current network status data, and historical network quality data, the predicted network quality of each communication channel within a preset future time period is predicted, including: preprocessing the current vehicle status data, current network status data, and historical network quality data through the input layer of a pre-trained network prediction model to obtain preprocessed data features; performing prediction processing on the data features through the hidden layer of the network prediction model to obtain predicted data features; and performing output processing on the predicted data features through the output layer of the network prediction model to obtain the predicted network quality of each communication channel.

[0142] In this invention, the pre-trained network prediction model can employ neural networks, time-series prediction models, or machine learning models that integrate multi-source features. Its training samples consist of data collected from different vehicles under different regions and channel conditions, enabling the network prediction model to learn the correspondence between vehicle state data, network state data, historical network quality data, and future network quality. In practical applications, other types of prediction models can also be selected, and this application does not limit this approach.

[0143] Reference Figure 3 The network prediction model consists of an input layer, a hidden layer, and an output layer. The input layer takes into account current vehicle status data, current network status data, and historical network quality data. Additionally, the input layer can accept other data depending on the actual situation, such as congestion level indicators and / or historical network status data. The input layer performs preprocessing operations on the input data, including standardization, normalization, and outlier correction as described above. This preprocessing eliminates interference from different units and numerical ranges on model training, mapping the data to a unified numerical interval, laying the foundation for subsequent time-series feature extraction. Details are described above and will not be repeated here.

[0144] Hidden layers can be based on LSTM (Long Short-Term Memory) networks or Transformers, for example, in Figure 4 In this model, the hidden layers include LSTM layer 1 (128 units), LSTM layer 2 (64 units), a multi-head attention layer, and a regularization layer (Dropout layer). The hidden layers process the data features preprocessed by the input layer to obtain the predicted data features.

[0145] In the specific implementation, the data features output from the input layer are first processed by LSTM layer 1 (an LSTM layer containing 128 units) to capture the short-term and long-term dependencies of the input data features, uncovering the basic patterns of network signal changes over time, and obtaining the first hidden feature. Then, LSTM layer 2 (an LSTM layer containing 64 units) is used to reduce the dimensionality and refine the first hidden feature, filtering redundant information and focusing on core temporal features, while reducing subsequent computational complexity, resulting in the second hidden feature. Next, a regularization layer is used to regularize the second hidden feature, obtaining the regularized feature. Finally, this regularized feature is passed through a multi-head attention layer to focus on key features from different dimensions, automatically identifying the predictive data features that characterize the quality of the predictive network.

[0146] Furthermore, the output layer can perform multi-task parallel prediction on the predicted data features based on a preset fully connected function to obtain the predicted network quality, which includes network quality score, signal identification, and network stability level, etc.

[0147] In this embodiment, after completing multi-source data acquisition, the vehicle-mounted terminal performs unified feature encoding on the current vehicle status data, current network status data, and historical network quality data, and sends them to the network prediction model according to a preset input format. Based on the learned spatiotemporal correlations and channel change patterns, the network prediction model outputs the predicted network quality of each communication channel within a preset future time period. This predicted network quality can reflect the network quality of each communication channel in the area the vehicle is about to enter, thus providing a forward-looking basis for subsequent communication channel switching.

[0148] With this implementation method, the vehicle terminal no longer makes judgments based solely on the current instantaneous network status. Instead, it obtains the channel quality change results over a future period through a pre-trained network prediction model. This allows channel selection to avoid weak network areas and abnormal network areas in advance, reducing delayed and repeated handovers, and improving communication continuity, stability, and the timeliness of service data transmission.

[0149] Optionally, the method further includes: receiving a network prediction model sent by a cloud server, the network prediction model being trained based on historical sample data from multiple vehicles; acquiring locally collected training sample data; and adjusting the model parameters of the network prediction model based on the training sample data.

[0150] The cloud server is used to distribute a general network prediction model to the vehicle terminal. The model parameters of the network prediction model are used to characterize the mapping relationship within the model. Historical sample data from multiple vehicles typically includes vehicle status data, network status data, and historical network quality data for the corresponding area. This data is collected by multiple vehicles under different road conditions and uploaded to the cloud server so that the cloud server can train a network prediction model with cross-regional generalization capabilities.

[0151] In some alternative implementations, the vehicle terminal deploys a network prediction model issued by a cloud server and then directly uses that network prediction model.

[0152] In other alternative implementations, the vehicle-mounted terminal can adjust the parameters of the network prediction model based on the network prediction model it collects.

[0153] The locally collected training sample data is the sample data acquired by the vehicle terminal during operation, including: vehicle status data collected by the target vehicle, network status data, and historical network quality data.

[0154] The training sample data includes sample data collected from the target vehicle's frequently traveled routes. Using this sample data for model training can improve the accuracy of network quality prediction for the target vehicle based on the network prediction model.

[0155] In this invention, the vehicle-mounted terminal can adjust the parameters of the network prediction model at preset intervals, and this invention does not limit this.

[0156] In practical applications, after receiving the network prediction model sent by the cloud server, the vehicle terminal first stores it in the local model library and then fine-tunes the model parameters by combining it with the locally collected training sample data, so that the network prediction model is more suitable for the current target vehicle.

[0157] The model parameter adjustment can be completed locally on the in-vehicle terminal, or assisted by edge computing nodes under the premise of meeting safety policies. During the adjustment process, gradient descent, incremental learning, or online learning methods can be used to further adapt the trained network prediction model to the historical communication characteristics of the target vehicle while maintaining the general rules in the cloud. The network prediction model after parameter adjustment can be used to predict the network quality of each communication channel in the future within a preset time period, thereby improving the consistency between the prediction results and the actual network changes.

[0158] This invention combines cloud-based pre-training with local fine-tuning, enabling the vehicle-mounted terminal to inherit general network change patterns extracted from historical samples of multiple vehicles while simultaneously correcting local biases using locally collected training sample data. Therefore, it maintains high prediction accuracy even in various complex scenarios. Consequently, the vehicle-mounted terminal can identify network quality change trends earlier when selecting communication channels, reducing handover lag and the probability of misjudgment, improving the continuity and stability of service data transmission, and enhancing the applicability of the communication channel switching method across different vehicle models, regions, and operating network conditions.

[0159] In summary, the following example illustrates the points:

[0160] Example 1:

[0161] The application scenario is as follows: A user parks their vehicle in an underground garage and prepares to remotely turn on the vehicle's air conditioning via a mobile application. However, network coverage in underground garages is often weak, and signal coverage varies between different mobile operators.

[0162] The execution process is as follows: The vehicle-mounted terminal collects vehicle status data and network status data every 500ms. The target vehicle is currently stationary (gear in P, speed 0km / h) at level B2 of the underground parking garage. The collected network status data is as follows: 5G cellular network channel of the first operator: RSRP=-115dBm, latency=180ms, packet loss rate=8%; 4G cellular network channel of the second operator: RSRP=-95dBm, latency=60ms, packet loss rate=2%; 4G cellular network channel of the third operator: RSRP=-105dBm, latency=90ms, packet loss rate=4%. The collected data is then input into the network prediction model. Based on the input data, the network prediction model predicts the historical network quality data for each communication channel within the next 5 seconds: 5G cellular network channel of the first operator: 35 points (poor, predicted to continue to decline); 4G cellular network channel of the second operator: 75 points (good and stable); 4G cellular network channel of the third operator: 55 points (average, slightly fluctuating). Furthermore, since this service is a remote initiation request, it is a high-priority service, requiring a network quality score ≥ 80 points, latency < 100ms, and packet loss rate < 1%. The current communication channel is the 5G cellular network channel of the first operator. The network quality score (35) of this communication channel is lower than the high-priority service threshold of 80 points. The network quality score (75) of the 4G cellular network channel of the second operator is higher than the current communication channel's 35 points and is the highest among all communication channels. Therefore, the handover conditions are met, and the handover is triggered. During the handover process, the vehicle terminal establishes a pre-connection with the 4G cellular network channel of the second operator in advance, completes identity authentication and parameter configuration, and while maintaining the 5G cellular network channel link of the first operator, migrates the service data to the 4G cellular network channel of the second operator, and keeps the TCP session uninterrupted. The upper-layer application is unaware of this. In addition, a handover threshold of 8 points can be set to avoid ping-pong handover.

[0163] In this exemplary case, the remote activation command is transmitted via the 4G cellular network channel of a second operator, and the vehicle-mounted terminal provides real-time feedback on the execution status. If the network quality of the communication channel deteriorates, the reception of the command is automatically retried to ensure successful execution. The number of receptions should be less than or equal to a preset threshold, such as 3 times. Therefore, in this example, the remote activation command can be efficiently sent to the vehicle-mounted terminal.

[0164] Example 2:

[0165] The application scenario is as follows: The target vehicle is traveling at a relatively high speed (e.g., 100 km / h) on a highway, and the tunnel entrance is 3 kilometers ahead. Based on the obtained historical network quality data, the 5G cellular network channel of the first operator has weak signal coverage inside the tunnel, while the 4G cellular network channel of the second operator has relatively stable signal coverage.

[0166] The execution process is as follows: The vehicle terminal collects real-time vehicle status data, including vehicle speed (100km / h), gear position (D), and GPS location showing the vehicle is approximately 130 meters from the tunnel entrance. Current network status data includes: the first operator's 5G cellular network channel: RSRP = -75dBm, latency = 30ms; the second operator's 4G cellular network channel signal: RSRP = -85dBm, latency = 50ms. Then, the network prediction model uses the current vehicle status data, current network status data, and historical network quality data to predict that the target vehicle will enter the tunnel in approximately 5 seconds. Inside the tunnel, the network quality score of the first operator's 5G cellular network channel is below 20 points, and the network quality score of the second operator's 4G cellular network channel signal is 65 points. Although the current network quality of the first operator's 5G cellular network channel signal is excellent, the network prediction model predicts that the target vehicle is about to enter the tunnel, meeting the handover condition of "approaching a weak network or blind spot," thus triggering handover preparation in advance. Before the vehicle enters the tunnel, the onboard terminal establishes a pre-connection with the second operator's 4G cellular network channel. Before entering the tunnel, the link migration is completed, making the second operator's 4G cellular network channel the primary network, while the first operator's 5G cellular network channel remains connected as a backup network. Once inside the tunnel, the network quality score of the second operator's 4G cellular network channel is maintained between 60 and 70 points, meeting the needs of low-to-medium priority services. In case of an emergency, a multi-channel concurrent mechanism is activated.

[0167] As can be seen in this example, it can be ensured that the target vehicle switches the communication channel in advance before entering the network abnormal area, avoiding communication channel switching delay. In addition, it can also ensure the communication of the vehicle terminal in the network abnormal area.

[0168] Example 3:

[0169] The application scenario is as follows: when a target vehicle is driving in a tunnel and a collision occurs, the airbags deploy, the vehicle terminal detects the collision signal, and needs to immediately report the collision alarm to the cloud server.

[0170] The execution process is as follows: The vehicle terminal detects the airbag deployment signal and collision sensor signal via the CAN bus, identifies it as a collision emergency, and classifies it as a medium-priority service. Current network status data is as follows: First operator's 4G cellular network channel: RSRP = -108dBm, latency = 150ms; Second operator's 4G cellular network channel: RSRP = -102dBm, latency = 120ms; Third operator's 4G cellular network channel: RSRP = -110dBm, latency = 160ms. The network prediction model can also predict the network quality score of each communication channel. For example, the network quality score of the first operator's 4G cellular network channel is 55 points, the second operator's 4G cellular network channel is 62 points, and the third operator's 4G cellular network channel is 50 points. Since the collision alarm is an emergency event with a reliability requirement of ≥99.5%, a single network cannot guarantee delivery; therefore, the vehicle terminal decides to activate a multi-channel concurrent transmission mechanism. Specifically, collision alarm information can be sent simultaneously through two networks: the 4G cellular network channel of the first operator and the 4G cellular network channel of the second operator. The alarm information includes: collision time, vehicle position, vehicle speed, airbag status, and seatbelt status. Each alarm message is uniquely identified for deduplication by the cloud server, and is also stored locally in a cache (non-volatile storage) as a backup. Furthermore, the in-vehicle terminal awaits confirmation from the cloud server. If no confirmation is received within a preset time, the system can automatically retry sending the alarm. During retrying, a third operator's 4G cellular network channel is added, creating a three-way concurrent transmission to increase the probability of successful alarm information transmission.

[0171] As can be seen in this example, when a collision occurs, the alarm information can be transmitted efficiently and successfully by using multi-channel concurrent data.

[0172] Example 4:

[0173] The application scenario is as follows: the target vehicle is in an underground parking garage with a weak network signal, and routine vehicle status reporting is required at regular intervals.

[0174] The execution process is as follows: The vehicle terminal collects current vehicle status data: mileage 15230km, battery voltage 12.6V, fuel level 45%, location B3 level of underground parking garage, and engine status is off. Current network status data is as follows: First operator's 4G cellular network channel: RSRP = -125dBm, latency = 500ms; Second operator's 4G cellular network channel: signal interrupted; Third operator's 4G cellular network channel: RSRP = -120dBm, latency = 400ms. A network quality score of 30 is predicted for both the first and third operator's 4G cellular network channels using a network prediction model. Since both network quality scores are below the low-priority service threshold of 40, they do not meet the reporting conditions, and the vehicle terminal can store the data to be transmitted in its local buffer. The cache uses non-volatile storage media, such as an 8GB non-volatile storage medium. If a single vehicle status data entry is approximately 2KB, then 8GB can store about 4 million data entries, supporting approximately 2222 days of data storage based on a 3-minute reporting cycle. Furthermore, the cached data is sorted by timestamp for fast retrieval. When the target vehicle leaves the underground parking garage and the network returns to normal (network quality score ≥ 40), the vehicle terminal can automatically trigger cached data transmission. The transmission method can be: prioritize transmitting the latest data, then transmit historical data. If the network deteriorates again during transmission, transmission can be paused and the transmission progress saved. After successful data transmission, the cached data is marked as "uploaded" and automatically cleared after a preset cache duration (e.g., 7 days).

[0175] As can be seen in this example, when the network quality of each communication channel is poor, low-priority data can be stored first and then transmitted after the network is restored, thus ensuring the successful transmission of low-priority data.

[0176] Example 5:

[0177] The application scenario is as follows: the target vehicle is driving on urban roads and has multiple business needs, such as: the vehicle terminal performing routine vehicle condition data reporting, users initiating remote vehicle search, and the target vehicle needing to report abnormal tire pressure.

[0178] The execution process is as follows: The vehicle terminal simultaneously detects three types of services: regular vehicle condition data reporting (low priority), two-dimensional remote vehicle location requests (high priority), and tire pressure anomaly alarms (medium priority). The currently available communication channel resources are assessed as follows: First operator's 5G cellular network channel, with a network quality score of 85, latency of 25ms, and sufficient bandwidth. Second operator's 4G cellular network channel: network quality score of 72, latency of 55ms, and medium bandwidth. Third operator's 4G cellular network channel: network quality score of 68, latency of 65ms, and medium bandwidth. Therefore, for the high-priority remote vehicle location, the first operator's 5G cellular network channel can be allocated to ensure low-latency response. For the medium-priority tire pressure alarm, the second operator's 4G cellular network channel can be allocated, while the first operator's 5G cellular network channel is used as a backup channel. For the low-priority vehicle condition reporting, the third operator's 4G cellular network channel can be allocated, and buffering can be implemented if the network deteriorates. Furthermore, the network quality and service status can be reassessed every 500ms. If the quality of the first operator's 5G cellular network channel drops below 80 points, the remote vehicle location will switch to the second operator's 4G cellular network channel. If the tire pressure alarm is successfully sent, the backup channel resources will be released. The timing of vehicle condition reporting will be dynamically adjusted based on network conditions. After the service is completed, the occupied network resources will be released, and the default network configuration will be restored.

[0179] In summary, this example demonstrates how efficient transmission of multi-service data can be achieved when multiple service requirements exist.

[0180] Example 6:

[0181] The application scenario is as follows: During daily operation, the vehicle-mounted terminal continuously collects network status data and vehicle status data, uses a network prediction model to make predictions, and then uploads the collected data and model prediction results to a cloud server. The cloud server uses the large amount of data collected from various vehicles to continuously iterate and optimize the network prediction model, and then distributes the optimized network prediction model to the vehicle-mounted terminal.

[0182] The execution process is as follows: The vehicle-mounted terminal encapsulates local data at preset intervals (e.g., minutes), including: network status data (such as RSRP, SINR, latency, packet loss rate, etc. for the past hour), vehicle status data (location, speed, driving scenario, etc.), other data required for input to the network prediction model, the prediction results of the network prediction model, and communication channel switching records (communication channels before and after switching). This data is compressed and encrypted before being uploaded to the cloud server via a low-priority service channel. Upload is scheduled when network quality is good (quality score ≥ 70) to avoid impacting high-priority services. Further, the cloud server parses, verifies, stores, removes abnormal data, fills in missing data, standardizes data format, and labels the prediction accuracy based on the difference between the actual network quality and the prediction results. Then, it extracts effective features, including time period features, location features, road segment features, and operator features. Finally, the cloud server performs batch training of the network prediction model based on the accumulated data.

[0183] As can be seen, in this example, the training data for the network prediction model comes from data uploaded by a large number of vehicle-mounted terminals. The training objective can be to minimize the prediction error to improve the accuracy of the network quality score. A reserved validation set can be used to evaluate the trained model, ensuring generalization ability, and to adjust model hyperparameters such as the number of LSTM layers, hidden units, and learning rate based on the evaluation results. Further, after the optimized model is tested and validated, it is packaged into a model update package. The model update package can be delivered to the vehicle-mounted terminal via OTA (Over-The-Air). The vehicle-mounted terminal can perform model updates during idle periods (such as after the vehicle is powered off). The update process can use a differential update method, updating only the changed model parameters to reduce data traffic. After the update is completed, the vehicle-mounted terminal automatically verifies the model integrity and rolls back to the old version (if verification fails). After the model update, the vehicle-mounted terminal records the prediction performance of the new model and uploads the comparison data of the prediction accuracy of the new model and the old model to the cloud server. The cloud server evaluates the model update effect based on the feedback data, guiding the subsequent optimization direction.

[0184] In summary, this example demonstrates how the network prediction model can be continuously updated and optimized to improve its performance.

[0185] Example 7:

[0186] The application scenario is as follows: the network environment in urban-rural fringe areas is complex, characterized by uneven base station coverage, mixed network standards, and severe signal interference. When target vehicles travel in this area, network quality fluctuates frequently, and traditional handover solutions are prone to problems such as ping-pong handover and service interruptions.

[0187] The execution process was as follows: The vehicle traveled from the urban area to the suburbs, initially using the first operator's 5G cellular network channel, with an RSRP of -70dBm, latency of 20ms, and a network quality score of 92. The second operator's 4G cellular network channel had an RSRP of -88dBm, latency of 45ms, and a network quality score of 78. The third operator's 4G cellular network channel had an RSRP of -92dBm, latency of 55ms, and a network quality score of 72. As the vehicle reached the suburban area, the signal of the first operator's 5G cellular network channel began to fluctuate, and the onboard terminal detected a rapid signal attenuation: the RSRP dropped from -70dBm to -105dBm. The signals of the second and third operator's 4G cellular network channels remained relatively stable, with only a slight decrease in signal strength.

[0188] Furthermore, the network prediction model predicts that the quality of the first operator's 5G cellular network channel will continue to decline, with the quality score expected to drop below 40 points within the next 5 seconds. Currently, the routine vehicle condition reporting service (low priority) is being executed. The network quality score of the first operator's 5G cellular network channel has dropped to 45 points, below the low-priority threshold of 40 points. The network quality scores of the second operator's 4G cellular network channel are 75 points and those of the third operator's 4G cellular network channel are 70 points, both higher than the current communication channel. Therefore, a handover can be triggered, selecting the second operator's 4G cellular network channel as the target communication channel. Additionally, a handover threshold of 8 points can be set, meaning that the network quality score of the second operator's 4G cellular network channel must remain at least 8 points higher than that of the first operator's 5G cellular network channel for at least 3 seconds before a handover is executed, to avoid frequent handovers caused by rapid signal fluctuations in suburban areas.

[0189] After the handover is complete, the network quality score of the second operator's 4G cellular network channel can be continuously monitored. If the network quality score of the second operator's 4G cellular network channel remains stable between 70 and 78 points, then the low-priority service requirements can be met. If a high-priority service (such as remote control) subsequently becomes available, and the second operator's 4G cellular network channel cannot meet the requirements, then a reassessment should be made to switch to the third operator's 4G cellular network channel or wait for the first operator's 5G cellular network channel to recover.

[0190] In summary, this example demonstrates how to ensure stable data transmission in areas with complex network environments.

[0191] Example 8:

[0192] The application scenario is as follows: The target vehicle is traveling on a main urban road, where V2X roadside units have been deployed, supporting V2X communication. The target vehicle needs to intelligently switch between the V2X channel and the cellular network (4G / 5G) channel to support different service requirements. The V2X channel is characterized by: low latency (e.g., end-to-end latency can be less than 10ms), short-range communication (e.g., effective communication distance of approximately 300-500 meters), and high reliability (e.g., reliability can reach 99.99% within the effective coverage area). Applicable scenarios for the V2X channel include vehicle-to-vehicle (V2V) communication, vehicle-to-infrastructure (V2I) communication, and safety-related services.

[0193] The execution process is as follows: The target vehicle enters the V2X roadside unit coverage area. The onboard terminal detects the RSU broadcast signal and initiates a registration request to the RSU, completing identity authentication and parameter configuration. Collected V2X network status data includes: RSRP = -65dBm, latency 8ms, and a corresponding network quality score of 95. Corresponding service types include: safety services (such as forward collision warning, requiring latency less than 20ms and reliability ≥99.99%), traffic efficiency services (such as traffic light information push, requiring latency less than 50ms), and information services (such as routine vehicle condition reporting, requiring latency less than 1000ms). Furthermore, for forward collision warning and traffic light information push, prioritizing the use of the V2X channel can achieve low latency and high reliability. For routine vehicle condition reporting, cellular networks (4G / 5G) can be used to avoid consuming V2X safety service resources.

[0194] If the target vehicle gradually moves away from the V2X roadside unit coverage center, the V2X signal strength begins to decrease: RSRP drops from -65dBm to -85dBm. If the network prediction model predicts that the target will leave the V2X coverage area in 8 seconds, then security services need to be switched to the cellular network channel in advance.

[0195] Further, the quality of each cellular network channel was evaluated. For example, the RSRP of the first operator's 5G cellular network channel was -78dBm, with a latency of 25ms and a network quality score of 85. The RSRP of the second operator's 4G cellular network channel was -82dBm, with a latency of 35ms and a network quality score of 80. Then, the forward collision warning service was switched from V2X to the first operator's 5G cellular network channel (highest quality score, lowest latency). For traffic light information push: the V2X channel was continued until the vehicle left the coverage area, after which it was cached. For routine vehicle condition reporting, the cellular network was continued. Then, the vehicle terminal established a front-end connection with the first operator's 5G cellular network channel, completed security authentication, and migrated the forward collision warning service data stream from V2X to the first operator's 5G cellular network channel to maintain service continuity and ensure that the collision warning was not interrupted. At the same time, the V2X network connection was maintained until the vehicle completely left the coverage area and then disconnected.

[0196] After the handover is completed, the first operator's 5G cellular network channel carries both security-related and regular services. The latency increases from 8ms to 25ms, but still meets the requirement of forward collision warning of less than 20m.

[0197] In this example, different service types can use different communication channels for data transmission. After leaving the coverage area of ​​one of the communication channels, the system can switch to another communication channel with better network quality to ensure the stability of the corresponding service data transmission.

[0198] In summary, this invention can predict the network quality change trend of each communication channel within a preset time period through a network prediction model, and proactively predict weak network areas and signal blind spots, thus achieving proactive prediction and communication channel switching. Compared with the passive threshold switching method of existing technologies, which only triggers switching when the network quality drops below a threshold, the switching response time of this invention is advanced by a certain amount of time, effectively avoiding service interruption problems caused by switching lag. Based on the different requirements of services for network latency, packet loss rate, and reliability, this invention divides the services of the vehicle terminal into three levels: high priority (remote control), medium priority (emergency event reporting), and low priority (routine vehicle condition reporting), and formulates differentiated network selection strategies and reliability assurance mechanisms for different priorities. Specifically, high-priority services are forced to use the network with the best quality score and can preempt network resources of low-priority services. Medium-priority services adopt a multi-channel concurrent transmission mechanism to ensure that emergency information is delivered. Low-priority services are automatically cached when the network quality is poor and retransmitted after the network recovers.

[0199] This invention supports unified access and intelligent scheduling of various network standards, including 2G, 3G, 4G, 5G, and 6G cellular networks from multiple operators, V2X vehicle-to-everything (V2X) networks, in-vehicle Wi-Fi, and satellite communication. Through technologies such as alternative network pre-connection, seamless link migration, and TCP session persistence, it ensures uninterrupted service and no data loss during handover. Simultaneously, this invention employs a handover hysteresis mechanism, setting reasonable handover thresholds and duration windows (e.g., 3 seconds) based on network quality score differences to effectively suppress ping-pong handover issues caused by minor fluctuations in network quality. Addressing data transmission issues in weak network and blind spot environments, this invention provides a robust reliability guarantee mechanism. For low-priority routine vehicle condition reporting, a strategy of weak network caching and automatic retransmission after network recovery is adopted. For medium-priority emergency event reporting, a strategy of concurrent transmission through multiple network channels is used to ensure no emergency information is missed. The buffer uses non-volatile storage media to ensure data is not lost after vehicle power failure. During retransmission, the latest data is transmitted first, followed by historical cached data, ensuring data timeliness.

[0200] Furthermore, this invention is adaptable to various vehicle models and in-vehicle terminal hardware, including passenger cars, commercial vehicles, and new energy vehicles, and encompasses different hardware forms such as pre-installed and aftermarket in-vehicle terminals. This invention employs an architecture combining local decision-making and cloud collaboration. The local network prediction model and decision-making unit are responsible for real-time network quality prediction and handover decisions, resulting in a short response time and meeting low-latency service requirements. The cloud server is responsible for model training, iteration, and updates, continuously improving model prediction accuracy by leveraging its powerful computing resources and massive amounts of data. Finally, this invention effectively reduces system overhead caused by frequent handovers through a ping-pong handover suppression mechanism; and avoids unnecessary network searches and connection attempts through network quality prediction and service hierarchical scheduling, reducing the power consumption of the in-vehicle terminal.

[0201] The following are embodiments of the apparatus of the present invention, which can be used to execute embodiments of the method of the present invention. For details not disclosed in the embodiments of the apparatus of the present invention, please refer to the embodiments of the method of the present invention.

[0202] Figure 4 A schematic diagram of the communication channel switching device provided in this embodiment of the invention. (See attached diagram.) Figure 4 As shown, the communication channel switching device 40 includes a first acquisition module 401, a second acquisition module 402, a prediction module 403, and a switching module 404.

[0203] The first acquisition module 401 is used to acquire the current vehicle status data of the target vehicle and the current network status data of multiple available communication channels of the target vehicle.

[0204] The second acquisition module 402 is used to determine the target area based on the area where the target vehicle is located, and to acquire historical network quality data of each communication channel when other vehicles pass through the target area in the past.

[0205] The prediction module 403 is used to predict the network quality of each communication channel within a preset time period based on the current vehicle status data, the current network status data, and historical network quality data.

[0206] The switching module 404 is used to switch to the target communication channel whose predicted network quality meets the preset network quality requirements.

[0207] Furthermore, the prediction module 403 is specifically used for: determining the congestion level identifier of each communication channel based on the current network status data, wherein the congestion level identifier is used to indicate the network congestion status of the corresponding communication channel; obtaining historical network status data of each communication channel within a preset historical time period; and predicting the predicted network quality of each communication channel within a preset time period based on the congestion level identifier and / or historical network status data, as well as the current vehicle status data, current network status data, and historical network quality data.

[0208] Furthermore, the switching module 404 is specifically used to obtain the service type of the vehicle terminal and determine the service priority corresponding to the service type; obtain the preset network quality requirements corresponding to the service priority; if the predicted network quality of the current communication channel does not meet the preset network quality requirements, determine the target communication channel that meets the preset network quality requirements among multiple communication channels; and switch from the current communication channel to the target communication channel.

[0209] Furthermore, when switching from the current communication channel to the target communication channel, the switching module 404 is specifically used to establish a communication link with the target communication channel; when the quality deviation between the predicted network quality of the target communication channel and the predicted network quality of the current communication channel is determined to be greater than the switching threshold, the corresponding transmission link is migrated from the current communication channel to the target communication channel. During the migration of the communication channel, the communication link with the current communication channel is not interrupted, nor is the transmission control protocol session of the upper layer application of the vehicle terminal connected to the application layer.

[0210] Furthermore, the network quality prediction includes: network quality scoring. When the switching module 404 determines the target communication channel that meets the preset network quality requirements among multiple communication channels, it is specifically used for: if the service priority is a preset high priority, determining the communication channel with the highest network quality score among multiple communication channels as the target communication channel; if the service priority is a preset medium priority, determining at least two communication channels with network quality scores greater than the score threshold among multiple communication channels as target communication channels, wherein the at least two target communication channels are used to concurrently transmit the corresponding service data.

[0211] Furthermore, the network quality prediction includes signal identifiers, which indicate whether a target vehicle will enter a network anomaly area of ​​the corresponding communication channel within a preset future time period. The device also includes:

[0212] The storage module (not shown in the figure) is used to locally store the corresponding business data when the business priority is a preset low priority and the signal indicates that the target vehicle will enter the network abnormal area of ​​the current communication channel within a preset time period in the future.

[0213] The transmission module (not shown in the figure) is used to transmit the corresponding service data based on the current communication channel when the target vehicle leaves the network abnormal area of ​​the current communication channel.

[0214] Furthermore, the storage module is also used to store the corresponding service data separately if the service priority is a preset high priority or a preset medium priority, and to centrally store the corresponding service data if the service priority is a preset low priority, when the signal indicates that the area to which the target vehicle will enter in the future is a network abnormal area of ​​each communication channel.

[0215] The transmission module is also used to prioritize the transmission of higher-priority service data if the target vehicle leaves the network abnormal area.

[0216] Furthermore, the prediction module 403 is specifically used to: preprocess the current vehicle state data, current network state data, and historical network quality data through the input layer of the pre-trained network prediction model to obtain preprocessed data features; perform prediction processing on the data features through the hidden layer of the network prediction model to obtain predicted data features; and perform output processing on the predicted data features through the output layer of the network prediction model to obtain the predicted network quality of each communication channel.

[0217] Furthermore, the aforementioned device also includes a training module (not shown in the figure), used to receive a network prediction model sent by a cloud server, the network prediction model being trained based on historical sample data from multiple vehicles; acquire locally collected training sample data; and adjust the model parameters of the network prediction model based on the training sample data.

[0218] The communication channel switching device provided in this embodiment of the invention can be used to execute the method steps of the above method embodiment. The specific implementation and technical effects are similar, and will not be repeated here.

[0219] Figure 5 This is a schematic diagram of a communication channel switching device provided in an embodiment of the present invention. The communication channel switching device 50 includes at least one processor 501 and a memory 502.

[0220] Optionally, the communication channel switching device 50 further includes a communication component 503. The processor 501, memory 502, and communication component 503 are connected via a bus 504.

[0221] In a specific implementation, at least one processor 501 executes computer execution instructions stored in memory 502, causing at least one processor 501 to perform the above-described method.

[0222] The specific implementation process of processor 501 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0223] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0224] The memory may include random access memory (RAM) and non-volatile memory (NVM), such as at least one disk storage device.

[0225] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0226] This invention also provides a vehicle, including: a vehicle body and a communication channel switching device as provided in the above embodiments.

[0227] This invention also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0228] This invention also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the aforementioned communication channel switching method.

[0229] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as ferromagnetic random access memory (FRAM), static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic surface memory, flash memory, magnetic disk, optical disk, or compact disc read-only memory (CD-ROM), etc. The readable storage medium can be any available medium accessible to general-purpose or special-purpose computers.

[0230] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0231] The division of units is merely a logical functional division; 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 indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0232] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0233] In addition, the functional units in the various embodiments of the present invention 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.

[0234] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. 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.

[0235] Those skilled in the art will understand that the present invention may take the form of a hardware embodiment, a software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0236] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus, and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0237] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0238] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0239] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0240] Finally, it should be noted that other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A method for switching communication channels, characterized in that, Applied to vehicle-mounted terminals, the method includes: Obtain the current vehicle status data of the target vehicle, and obtain the current network status data of multiple available communication channels of the target vehicle; Based on the area where the target vehicle is located, determine the target area and obtain historical network quality data of each of the communication channels when other vehicles passed through the target area in the past; Based on the current vehicle status data, the current network status data, and the historical network quality data, predict the predicted network quality of each of the communication channels within a future preset time period; Among the plurality of communication channels, a target communication channel whose predicted network quality meets the preset network quality requirements is determined, and the user switches to the target communication channel.

2. The method according to claim 1, characterized in that, The step of predicting the predicted network quality of each communication channel within a preset future time period based on the current vehicle status data, the current network status data, and the historical network quality data includes: Based on the current network status data, the congestion level identifier of each communication channel is determined, and the congestion level identifier is used to indicate the network congestion status of the corresponding communication channel; Obtain historical network status data for each of the communication channels within a preset historical time period; Based on the congestion level identifier and / or historical network status data, as well as the current vehicle status data, the current network status data, and the historical network quality data, the predicted network quality of each of the communication channels within the future preset time period is predicted.

3. The method according to claim 1 or 2, characterized in that, The step of determining, among the plurality of communication channels, the target communication channel whose predicted network quality meets the preset network quality requirements, and switching to the target communication channel, includes: Obtain the service type of the vehicle terminal and determine the service priority corresponding to the service type; Obtain the preset network quality requirements corresponding to the service priority; If the predicted network quality of the current communication channel does not meet the preset network quality requirements, a target communication channel that meets the preset network quality requirements is determined from the plurality of communication channels. Switch from the current communication channel to the target communication channel.

4. The method according to claim 3, characterized in that, The switching from the current communication channel to the target communication channel includes: Establish a communication link with the target communication channel; When the predicted network quality of the target communication channel deviates from the predicted network quality of the current communication channel by a quality threshold greater than a switching threshold, the corresponding transmission link is migrated from the current communication channel to the target communication channel. During the migration of the communication channel, the communication link with the current communication channel is not interrupted, nor is the transmission control protocol session and application layer connection of the upper layer application of the vehicle terminal interrupted.

5. The method according to claim 3, characterized in that, The predicted network quality includes: a network quality score; the step of determining the target communication channel that meets the preset network quality requirements among the plurality of communication channels includes: If the service priority is a preset high priority, the communication channel with the highest network quality score among the multiple communication channels is determined as the target communication channel; If the service priority is a preset medium priority, at least two communication channels with network quality scores greater than the score threshold are determined as the target communication channels among the plurality of communication channels, wherein the at least two target communication channels are used to concurrently transmit the corresponding service data.

6. The method according to claim 5, characterized in that, The predicted network quality includes a signal identifier, which indicates whether the target vehicle enters a network anomaly area of ​​the corresponding communication channel within the preset future time period. The method further includes: When the service priority is a preset low priority and the signal indicates that the target vehicle has entered the network abnormal area of ​​the current communication channel within the preset future time period, the corresponding service data is stored locally. If the target vehicle leaves the network anomaly area of ​​the current communication channel, the corresponding service data is transmitted based on the current communication channel.

7. The method according to claim 6, characterized in that, Also includes: When the signal indicates that the area the target vehicle enters during the future preset time period is a network abnormal area of ​​each of the communication channels, if the service priority is a preset high priority or a preset medium priority, the corresponding service data is stored separately; if the service priority is a preset low priority, the corresponding service data is stored centrally. If the target vehicle leaves the network anomaly area, higher priority service data will be transmitted first.

8. The method according to claim 1 or 2, characterized in that, The step of predicting the predicted network quality of each communication channel within a preset future time period based on the current vehicle status data, the current network status data, and the historical network quality data includes: The current vehicle state data, the current network state data, and the historical network quality data are preprocessed using the input layer of a pre-trained network prediction model to obtain preprocessed data features. The predicted data features are obtained by performing prediction processing on the data features through the hidden layer of the network prediction model; The predicted data features are processed by the output layer of the network prediction model to obtain the predicted network quality of each communication channel.

9. The method according to claim 8, characterized in that, The method further includes: Receive a network prediction model sent by a cloud server, the network prediction model being trained based on historical sample data from multiple vehicles; Acquire locally collected training sample data, and adjust the model parameters of the network prediction model based on the training sample data.

10. A communication channel switching device, characterized in that, The device, applied to an in-vehicle terminal, includes: The first acquisition module is used to acquire the current vehicle status data of the target vehicle and the current network status data of multiple available communication channels of the target vehicle. The second acquisition module is used to determine the target area based on the area where the target vehicle is located, and to acquire historical network quality data of each of the communication channels when other vehicles pass through the target area in the past. The prediction module is used to predict the network quality of each of the communication channels within a preset future time period based on the current vehicle status data, the current network status data, and the historical network quality data. The switching module is used to switch to the target communication channel whose predicted network quality meets the preset network quality requirements.

11. A communication channel switching device, characterized in that, include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory to implement the communication channel switching method as described in any one of claims 1 to 9.

12. A vehicle, characterized in that, include: The vehicle body and the communication channel switching device as described in claim 11.

13. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the communication channel switching method as described in any one of claims 1 to 9.

14. A computer program product, characterized in that, include: A computer program, when executed by a processor, implements the method for switching communication channels as described in any one of claims 1 to 9.

Images