Multi-network communication method and device applied to unmanned driving

Abstract

The invention discloses a multi-network communication method and device applied to unmanned driving, and belongs to the technical field of unmanned driving, and the method comprises the steps: obtaining 5G network signal parameter data from a 5G multi-network link channel, and filtering out the 5G multi-network link channel reaching the standard; the NPU neural network is used for conducting quantization processing on the standard 5G multi-network link channel network signal parameters, a comprehensive score of the standard 5G multi-network link channel network signal parameters is obtained, and the sequence of the current optimal link channel and other backup link channels is further obtained; adjusting the optimal link channel in real time according to a comparison result of a preset condition and the real-time 5G network signal parameter data associated with the current optimal link channel; and keeping communication with the cloud server based on the optimal link channel adjusted in real time. The multi-network communication module is composed of at least three independent 5G + C-V2X full-network communication modules and an auxiliary USIM card slot, the network can be optimized in real time along with vehicle operation, and it is ensured that the vehicle always uses the optimal 5G network.

Description

Technical Field

[0001] The present invention relates to a multi-network communication method and device for driverless applications, belonging to the technical field of intelligent networking of driverless vehicles. Background Art

[0002] With the rapid development of 5G + C-V2X intelligent networking technology, 5G communication technology has characteristics such as high bandwidth, low latency, and large connection. Selecting the 5G network can support the connection of a large number of devices, ensuring efficient data exchange between driverless vehicles and other vehicles, infrastructure, and cloud servers in the surrounding environment. It can meet the requirements of driverless vehicles for real-time data transmission and remote control, providing the possibility for remote driverless vehicles. If the 5G cloud server maintains a reliable 5G communication connection with the autonomous driving vehicle, the remote cockpit can directly control the autonomous driving vehicle to achieve remote driverless operation.

[0003] However, the construction density and frequency band planning of existing 5G base stations are inconsistent. Currently, there are still significant problems with the coverage of 5G base station network signals. In some areas covered by 5G base stations, the multiple advantages of 5G such as high speed, low latency, large connection, and wide coverage cannot be fully realized. It is undeniable that there are still some "5G signal blind spots" at present. In addition, in densely populated urban areas, when using a single base station, a large number of users will cause slow network speed and network congestion problems in the transmission resources network. Poor network signal, decreased rate, high packet loss rate, and increased latency will bring serious problems to remote driverless vehicles, resulting in loss of connection and out-of-control, such as sudden stagnation on traffic arteries, causing traffic paralysis problems.

[0004] On the other hand, intelligent networking vehicles that already support driverless generally need to built-in 5G + C-V2X OBU box vehicle-mounted communication units, which only support single 5G card to log in to the network, and few support dual 5G cards and dual-mode access to the 5G network. If only logging in to the 5G frequency bands of 1 or 2 operators, there are obviously still regional limitations in searching for 5G signals to log in and cover, and such a stable 5G network signal login environment cannot be maintained. Therefore, the existing communication modes of driverless vehicles are single, and their promotion and use seriously depend on the high-density 5G base station network coverage rate, and the application scenarios will be limited.

[0005] To solve the problem of insufficient coverage of 5G base station network signals, existing methods mainly rely on increasing the construction density of 5G base stations. However, this method is not only costly but also has a long construction period and is difficult to achieve full coverage in a short time. In addition, the 5G network coverage of a single operator also has certain limitations and cannot guarantee stable network connections in all areas.

[0006] Therefore, in order to achieve stable communication of driverless vehicles in various complex environments, a new method is needed that can make full use of existing network resources, effectively make up for the coverage shortage of a single network, and improve the stability and reliability of network connections. Summary of the Invention

[0007] The object of the present invention is to provide a multi-network communication method and device for driverless applications. By building a 5G + C-V2X full-network communication OBU integrated with the in-vehicle intelligent cockpit that supports at least three networks and three-way communication, it can simultaneously support the 5G networks of the three major operators online. In addition to collecting the 5G network signal parameters of multiple operators in the area where the driverless vehicle is located, this integrated device can also communicate through C-V2X, extend the range by 300 meters with RSU roadside single devices and surrounding vehicles, detect and sense the 5G base station signals and network speeds. These network data are transmitted to the NPU neural network processing unit of the multi-core heterogeneous AP processor in the intelligent cockpit of this integrated device, and this unit will execute the intelligent algorithm module function. The intelligent algorithm module can comprehensively evaluate and score the performance of each network according to multiple indicators such as signal strength (RSRP), signal quality (SINR), mileage residence ratio, duration residence ratio, network speed, and latency. After the indicators are quantified, weights are added to each indicator, and then the weighted summation algorithm is used for calculation. After screening and sorting, the levels of the current best 5G channel and other backup channels are obtained. At the same time, the system will continuously monitor the network status and environmental changes. When it is detected that the comprehensive score of the current best channel is lower than others, it will automatically switch, so that the vehicle and the cloud server always communicate with the best 5G network, and finally the vehicle can maintain a stable 5G network application environment, effectively make up for the blind area of the 5G base station of a single operator, reduce network latency, enhance data transmission capabilities, and improve the safety and response speed of driverless vehicles.

[0008] To solve the above technical problems, the present invention is implemented by the following technical solutions.

[0009] In a first aspect, the present invention provides a 5G multi-network fusion communication method for driverless applications, including: Obtaining 5G network signal parameter data from 5G multi-network network link channels respectively; Obtaining the currently qualified 5G multi-network network link channels according to the comparison results between the set thresholds and the 5G network signal parameter data of each channel; Counting the currently qualified 5G multi-network network link channels to obtain the mileage residence ratio and duration residence ratio of the currently qualified 5G multi-network network link channels; Using an NPU neural network to perform quantization processing on the mileage residence ratio, duration residence ratio, and 5G network signal parameter data of the currently qualified 5G multi-network network link channels to obtain the comprehensive score of the currently qualified 5G multi-network network link channels; Sort the comprehensive scores of the currently qualified 5G multi-network link channels in descending order to obtain the ranking of the current best link channel and other backup link channels; According to the comparison result between the preset conditions and the real-time 5G network signal parameter data associated with the current best link channel, adjust the best link channel in real time; Based on the best link channel adjusted in real time, communicate with the cloud server for driverless remote communication.

[0010] Optionally, obtain 5G network signal parameter data from the 5G multi-network link channels respectively, including: When the vehicle is running, perform C-V2X communication with multiple 5G base stations of different operators within the preset range, the RSU roadside unit of the current road section, and surrounding vehicles running respectively to form a 5G multi-network link channel; among them, the 5G multi-network link channel includes the 5G network link channel at the local position where the vehicle is located and the C-V2X vehicle network transmission link channel detected from the RSU roadside unit and surrounding vehicles running; Detect the network mode and communication frequency band of the 5G network link channel at the local position where the vehicle is located and the C-V2X vehicle network transmission link channel passing through the positions of the RSU roadside unit and surrounding vehicles running in real time. When the network mode of the 5G network link channel supports both 5GNR SA and NSA dual modes, and the 5G network link channel and the C-V2X vehicle network transmission link channel support 5GNR full-network communication frequency band, 5GNR SUL super uplink frequency band and C-V2X frequency band communication, obtain the network signal data of the 5G network link channel at the local position where the vehicle is located, receive the network signal data of the 5G network link channel detected from the RSU roadside unit and surrounding vehicles running, and obtain the vehicle driving mileage data respectively; Analyze the information of the 5G network link channel at the local position where the vehicle is located and the network signal data of the 5G network link channel passing through the C-V2X vehicle network transmission link channel from the positions of the RSU roadside unit and surrounding vehicles running to obtain the 5G network signal parameter data associated with the 5G multi-network link channel, where the 5G network signal parameter data includes the signal strength RSRP, signal quality SINR, network speed, time delay and residence duration of each link channel.

[0011] Optionally, obtain the currently qualified 5G multi-network link channels according to the comparison result between the set threshold and the 5G network signal parameter data of each channel, including: Based on the comparison results of the set signal strength RSRP threshold, signal quality SINR threshold, average uplink and downlink speed thresholds, and minimum uplink and downlink speed thresholds with the signal strength RSRP, signal quality SINR, average uplink and downlink speed, and minimum uplink and downlink speed associated with the 5G multi-network link channel, the initially qualified 5G multi-network link channel is obtained; Based on the heartbeat mechanism, further delay tests are performed on the initially qualified 5G multi-network link channel to obtain a 5G multi-network link channel with a smaller delay, which is used as the currently qualified 5G multi-network link channel.

[0012] Optionally, statistics are performed on the currently qualified 5G multi-network link channel to obtain the mileage residence ratio and duration residence ratio of the currently qualified 5G multi-network link channel, including: According to the vehicle driving mileage data and the 5G network signal parameter data of the continuous network access, the mileage of the currently qualified 5G multi-network link channel during continuous network access and the total mileage of the vehicle driving are obtained; The ratio of the mileage of the currently qualified 5G multi-network link channel during continuous network access to the total mileage of the vehicle driving is calculated to obtain the mileage residence ratio of the currently qualified 5G multi-network link channel; The actual application continuous residence duration of the user on the currently qualified 5G multi-network link channel and the total duration of the vehicle driving are statistically obtained from the 5G network signal parameter data; The ratio of the actual application continuous residence duration of the user on the currently qualified 5G multi-network link channel to the total duration of the vehicle driving is calculated to obtain the duration residence ratio of the currently qualified 5G multi-network link channel.

[0013] Optionally, the NPU neural network is used to perform quantization processing on the mileage residence ratio, duration residence ratio, and 5G network signal parameter data of the currently qualified 5G multi-network link channel to obtain the comprehensive score of the currently qualified 5G multi-network link channel, including: Quantization processing is respectively performed on the mileage residence ratio, duration residence ratio, signal strength RSRP, signal quality SINR, network speed, and delay to obtain the real-time parameter index information of the 5G multi-network link channel that eliminates the dimension difference between different parameters; The NPU neural network is used to perform weighted summation calculation on the real-time parameter index information of the 5G multi-network link channel that eliminates the dimension difference between different parameters to obtain the comprehensive score of each link channel.

[0014] Optionally, according to the comparison result of the preset conditions with the real-time 5G network signal parameter data associated with the current best link channel, the best link channel is adjusted in real time, including: Real-time detect the network mode and communication frequency band of the current best link channel and update the 5G network signal parameter data; If it is detected that the current best link channel is in the SUL frequency band, then communicate using the high and low frequency combined frequency band of the 5G network to improve the network uplink coverage rate and uplink network speed rate of the best link channel; If there is no SUL frequency band in the communication frequency band of the current best link channel, then recalculate the comprehensive scores of each link channel and switch the link channel with the highest comprehensive score to the best link channel; If it is detected that the network speed of the current best link channel drops below the preset network speed lower limit value, then select other standby link channels in descending order of the comprehensive score as the new best link channel and continue to receive the signal of the SUL frequency band; If the network speed of all link channels is less than the preset network speed lower limit value, then switch to the 4G network mode for communication with the vehicle network, and the vehicle is switched to intelligent assisted driving.

[0015] Optionally, if it is detected that the current best link channel is in the SUL frequency band, then communicate using the high and low frequency combined frequency band of the 5G network, and it further includes: According to the comparison result between the preset uplink network speed threshold and the uplink network speed rate, continue to communicate using the current best link channel; If the average rate of the uplink network speed is greater than the lower threshold of the average rate, and the lowest rate of the uplink network speed is greater than the lower threshold of the lowest rate, then preferentially jump to the SUL frequency band as the 5G best link channel. The vehicle processor maintains 5G network communication with the cloud server through the best link channel, uploads data to the cloud server and receives the cloud server. The cloud server can monitor driverless driving in real time; If the average rate of the uplink network speed is less than or equal to the lower threshold of the average rate, and the lowest rate of the uplink network speed is less than or equal to the lower threshold of the lowest rate, then recalculate the comprehensive scores of each link channel and switch the link channel with the highest comprehensive score to the best link channel.

[0016] In a second aspect, the present invention provides a 5G multi-network fusion communication device applied to driverless driving, including: a vehicle processor, a multi-network communication module, a positioning module, a vehicle network communication security unit, and a multi-functional vehicle-mounted sensor; The vehicle processor is a multi-core heterogeneous AP processor in the vehicle intelligent cockpit domain, and is composed of an AP processing unit, a GPU processing unit, an NPU neural network processing unit, and a storage unit, and is used to execute the steps of the above method; The multi-network communication module is used for the vehicle to perform 5G full-network communication with more than three 5G network operators respectively, communicate with RSU roadside devices, and perform C-V2X vehicle network communication with surrounding vehicles; The positioning module is connected to the multi-network communication module, and is used for tracking and positioning the vehicle driving path trajectory and sending out PPS second pulse signals to the multi-network communication module; The vehicle networking communication security unit is built with a vehicle networking HSM security encryption chip, and is used for providing security information such as vehicle data encryption, decryption, and user identity authentication when the multi-network communication module conducts vehicle networking communication with surrounding vehicles.

[0017] Optionally, the multi-network communication module consists of at least 3 independent 5G + C-V2X full-network communication modules and an attached USIM card slot; Among them, the 5G + C-V2X communication unit and the USIM communication unit are combined to be used for compatible working modes of 5G network, 4G network, 3G network, and 2G network, and support 5G NR full-network frequency band, 5G NR SUL super uplink frequency band, and C-V2X frequency band; Each 5G + C-V2X communication unit is provided with a 5G network antenna and a C-V2X antenna unit; the 5G network antenna is used for conducting 5G network communication with different operators respectively, and the C-V2X antenna unit is used for conducting C-V2X mode communication with the RSU roadside unit and surrounding running vehicles.

[0018] Optionally, the positioning module consists of a GPS signal receiver, an IMU inertial navigation acceleration sensor, and a gyroscope sensor, and is used for receiving the vehicle's positioning signal, angular velocity, and acceleration; When the vehicle's positioning signal, angular velocity, and acceleration change, the positioning module is started to obtain the vehicle driving trajectory data, and at the same time, the PPS second pulse signal sent out by the positioning module is started, which will activate the automatic synchronization of C-V2X communication in the 5G + C-2X full-network module, and turn on the services of broadcasting and receiving messages of V2I / V2V with the RSU roadside unit and surrounding running vehicles, so as to collect 5G base station signals within the range of 300 meters to 500 meters from the vehicle; it is transmitted to the AP processor in the vehicle intelligent cockpit through the 5G + C-V2X module PCIe2.0 interface for multi-source data analysis and processing to implement the steps of the above method.

[0019] Compared with the prior art, the beneficial effects achieved by the present invention: 1. The at least three independent 5G + C-V2X network-wide communication modules built into the communication device of the present invention can automatically search for the 5G frequency bands of each operator, thereby receiving and collecting the signal conditions and locations of 5G base stations of different operators on the road section where the vehicle is located. At the same time, the positioning module is activated to obtain and track the vehicle's driving trajectory data. The positioning module emits a pulse signal, which will activate the automatic synchronization of C-V2X communication in the 5G + C-V2X network-wide communication module, and start the services of broadcasting and receiving messages with the RSU roadside unit and surrounding moving vehicles for V2I and V2V. The device of the present invention is set to obtain 5G signal messages from the RSU roadside unit and surrounding moving vehicles within 300 meters - 500 meters around the driverless vehicle, that is, collect the signal status information of 5G base stations within 300 meters - 500 meters from the driverless vehicle. The data of the 5G base station signals collected by each operator are transmitted to the AP processor in the in-vehicle intelligent cockpit through the multi-source communication module for multi-source data analysis and processing; and the best network link channel in terms of signal strength, signal quality, continuous signal coverage, network speed, latency, etc. is selected to maintain 5G communication with the server.

[0020] 2. The present invention does not rely on increasing the 5G base station coverage density to ensure the 5G network of the driverless vehicle is online. Through 5G multi-network communication technology, it makes full use of the existing network resources to achieve seamless handover and avoid the impact of signal blind spots on driverless driving.

[0021] 3. The present invention supports 5G multi-network convergence communication of at least three operators, enhances the data transmission ability, and provides higher bandwidth and faster data transmission speed. The driverless vehicle can quickly obtain and process a large amount of environmental information, improving the real-time performance of autonomous driving.

[0022] 4. The present invention has at least three 5G network channels. When one channel fails or the signal is unstable, the other two channels can be used as backups to ensure uninterrupted communication. This redundant design greatly improves the reliability and stability of the system.

[0023] 5. The present invention reduces latency through physical layer technologies such as optimizing the network random access process and quickly activating SUL super uplink frequency band aggregation, and optimizing mobility management processes such as handover. It reduces the delay of video transmission, enabling the driverless vehicle to respond more quickly to various instructions and warnings.

[0024] 6. The present invention can reduce 5G communication latency, enabling the vehicle to respond more quickly to instructions and warnings from other vehicles, infrastructure, and cloud services, improving the safety and response speed of driverless autonomous driving. In case of an emergency, a quick response can effectively avoid accidents.

[0025] 7. The present invention can expand the vehicle's perception range. Through real-time V2I / V2V communication with other vehicles and roadside infrastructure, a driverless vehicle can obtain 5G base station signals and network speed information beyond the perception range of its own 5G communication module. The expanded perception range enables the driverless vehicle to better adapt to complex environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 The figure shows a flowchart of the method for multi-network communication applied to driverless driving according to the present invention; Figure 2 The figure shows a flowchart of the method for filtering non-compliant 5G multi-network link channels according to the present invention; Figure 3 The figure shows a flowchart of the method for real-time adjustment of the optimal link channel according to the present invention; Figure 4 The figure shows a hardware structure diagram of the 5G multi-network fusion communication device applied to driverless driving according to the present invention; Figure 5 The figure shows a network architecture diagram of the 5G multi-network fusion communication device applied to driverless driving according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0027] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and should not be used to limit the protection scope of the present invention.

[0028] The applicant has found that in order to solve the problem of insufficient network signal coverage of 5G base stations in existing driverless vehicle autonomous driving technologies, the construction density of 5G base stations is increased. However, this method is not only costly but also has a long construction period, making it difficult to achieve low-cost promotion of driverless vehicle autonomous driving in a short time. Therefore, an improvement to a vehicle communication device and method is proposed. The device is based on an in-vehicle intelligent connected vehicle built-in 5G+C-V2X full-network communication OBU that supports at least triple-network triple-connection and an in-vehicle intelligent cockpit integrated device. The device includes at least 3 independent 5G+C-V2X full-network communication modules and attached USIM card slots, 5G network antennas and C-V2X antenna units that support the three major operators, a high-precision centimeter-level GPS positioning module, a vehicle networking HSM security chip, a multi-core heterogeneous AP processor for intelligent cockpit applications (built-in GPU and NPU neural network processing units) and a program operation storage core unit, a central control screen and an instrument panel unit. Through this device, it is possible to simultaneously detect the 5G base station network signals and network speed signals of different operators in the area where the driverless vehicle is located, and even establish V2X real-time point-to-point direct communication with other vehicles and roadside infrastructure. It can extend and expand the detection and acquisition of 5G base station signals and network speed information beyond the perception range of the vehicle's 5G communication module (>300 meters). Then, through a weighted sum averaging intelligent algorithm, the intensity and quality of each network signal are monitored in real time. By adaptively and dynamically allocating access to the 5G networks of the three major operators, the best 5G network channel is automatically selected for data transmission, achieving stable online 5G network for the in-vehicle computer.

[0029] The 5G multi-network convergence communication method is an innovative communication solution designed to provide a more efficient and stable 5G network connection for driverless vehicles. The core idea of this method is to achieve the collaborative utilization and dynamic switching of multi-network resources by simultaneously leveraging the 5G network access of at least three different operators. Specifically, when a driverless vehicle starts, the system automatically detects and connects to all available 5G networks. Each 5G communication module is respectively connected to the network of the corresponding operator, and the network connection parameters are initialized. During operation, each 5G communication module real-time monitors the strength and quality of its respective network signal, and transmits this data to the multi-core heterogeneous AP processor NPU neural network processing unit in the intelligent cockpit, which will execute the intelligent algorithm module function. The intelligent algorithm module can comprehensively evaluate the performance of each network based on multiple metrics such as signal strength, network latency, and data transmission rate, and select the network with the best signal strength and the lowest latency as the main working mode. At the same time, the system continuously monitors the network status and environmental changes, and automatically switches when a new and better network is detected to ensure that the best network is always used. In addition, the system will also anticipate the network coverage situation in advance according to the vehicle's position and driving route, and make preparations for network switching. For example, when entering an area with poor signal coverage such as a tunnel or an underground parking lot, the system will switch to a network with stronger signal in advance to ensure the continuity of the network connection. This 5G multi-network convergence communication method not only improves the stability and reliability of the network connection, but also effectively solves problems such as insufficient 5G base station coverage and network congestion, providing a strong communication guarantee for driverless vehicles. The specific embodiments are as follows: Embodiment

[0030] This embodiment provides a multi-network communication method applied to driverless, as Figure 1 shown including: Step 1: Obtain 5G network signal parameter data from the 5G multi-network network link channels respectively; Step 2: Obtain the currently qualified 5G multi-network network link channels according to the comparison results between the set thresholds and the 5G network signal parameter data of each channel; Step 3: Statistically analyze the currently qualified 5G multi-network network link channels to obtain the mileage residence ratio and duration residence ratio of the currently qualified 5G multi-network network link channels; Step 4: Use the NPU neural network to perform quantization processing on the mileage residence ratio, duration residence ratio, and 5G network signal parameter data of the currently qualified 5G multi-network network link channels to obtain the comprehensive score of the currently qualified 5G multi-network network link channels; Step 5: Sort the comprehensive scores of the currently qualified 5G multi-network network link channels in descending order to obtain the sorting of the currently best link channel and other backup link channels; Step 6: Adjust the optimal link channel in real time according to the comparison result between the preset conditions and the real-time 5G network signal parameter data associated with the current optimal link channel. Step 7: Maintain communication with the cloud server based on the optimally adjusted link channel in real time for driverless remote communication.

[0031] Optionally, in step 1, obtaining 5G network signal parameter data from 5G multi-network link channels respectively includes: When the vehicle is running, perform C-V2X communication with multiple 5G base stations of different operators within a preset range, the RSU roadside unit of the current road section, and surrounding moving vehicles in real time to form a 5G multi-network link channel; among them, the 5G multi-network link channel includes the 5G network link channel of the local location where the vehicle is located and the C-V2X vehicle networking transmission link channel detected from the RSU roadside unit and surrounding moving vehicles; Detect the network mode and communication frequency band of the 5G network link channel of the local location where the vehicle is located and the C-V2X vehicle networking transmission link channel passing through the locations of the RSU roadside unit and surrounding moving vehicles in real time. When the network mode of the 5G network link channel supports both 5GNR SA and NSA dual modes, and the 5G network link channel and the C-V2X vehicle networking transmission link channel support 5GNR full-network frequency band, 5GNR SUL super uplink frequency band, and C-V2X frequency band communication, obtain the 5G network link channel network signal data of the local location where the vehicle is located, receive the 5G network link channel network signal data detected from the RSU roadside unit and surrounding moving vehicles, and obtain the vehicle driving mileage data respectively; Analyze the information of the 5G network link channel of the local location where the vehicle is located and the 5G network link channel network signal data transmitted through the C-V2X vehicle networking transmission link channel from the locations of the RSU roadside unit and surrounding moving vehicles to obtain the 5G network signal parameter data associated with the 5G multi-network link channel. Among them, the 5G network signal parameter data includes the signal strength RSRP, signal quality SINR, network speed, time delay, and residence duration of each link channel.

[0032] In this embodiment, 3 independent 5G + C-V2X full-network communication modules built into the integrated device are used to communicate with different operators. Among them, the three 5G communication unit channels respectively receive and collect the 5G base station signal conditions of the three major operators in the road section where the vehicle is located: the location of the cell, PLMN and operator network coding, network working mode, working frequency point, 5G signal RSRP (Reference Signal Receiving Power), RSRQ (Reference Signal Receiving Quality), RSSI (Received Signal Strength Indicator), SINR (Signal-to-Interference-plus-Noise Ratio), etc. to judge the normal parameters of the 5G signal; At the same time, start the high-precision centimeter-level positioning module to obtain the driving trajectory data of the tracked vehicle. The PPS second pulse signal sent from the positioning module will activate the automatic synchronization of C-V2X communication in the 5G+C-2X network-wide communication module, and start the V2I / V2V broadcast and receive message services with the RSU roadside unit and surrounding operating vehicles. The three C-V2X communication unit channels in the 5G+C-V2X network-wide communication module respectively receive and collect 5G signal messages from the RSU roadside unit and surrounding moving vehicles within 300 meters to 500 meters from the driverless vehicle, that is, collect the signal status information parameters of the 5G base station within 300 meters to 500 meters from the driverless vehicle; The above-mentioned various 5G network signal parameter data are transmitted to the AP processor in the in-vehicle intelligent cockpit through the PCIe2.0 interface of the 5G+C-V2X module for multi-source data analysis and processing.

[0033] Optionally, according to the comparison result between the set threshold and the 5G network signal parameter data of each channel, obtain the currently qualified 5G multi-network link channels, such as Figure 2 shown including: According to the comparison results of the set signal strength RSRP threshold, signal quality SINR threshold, average up / downlink speed threshold of the network speed, and minimum up / downlink speed threshold of the network speed with the signal strength RSRP, signal quality SINR, average up / downlink speed of the network speed, and minimum up / downlink speed associated with the 5G multi-network link channels respectively, obtain the initially qualified 5G multi-network link channels; Based on the heartbeat mechanism, perform a delay test on the initially qualified 5G multi-network link channels, obtain the 5G multi-network link channels with smaller delays, and use them as the currently qualified 5G multi-network link channels.

[0034] In this embodiment, when the network mode of the 5G multi-network link channel supports both 5G NRSA and NSA dual modes, and the 5G network link channel and the C-V2X vehicle networking transmission link channel support 5G NR network-wide frequency bands, 5G NRSUL super uplink frequency bands, and C-V2X frequency band communications, according to the set signal strength RSRP threshold of -100 dBM, signal quality SINR threshold of 20 dB, and network speed measurement thresholds are respectively: the average up / downlink speed threshold of the network speed is 100 Mbps, and the minimum up / downlink speed threshold of the network speed is 20 Mbps. When the average up / downlink speed of the network speed > 100 Mbps, and the minimum up / downlink speed of the network speed > 20 Mbps, and whether RSRP is greater than -100 dBM; the 5G network signal parameter data of each channel are respectively compared with the above thresholds, filter out the unqualified 5G network link channels, and generate heartbeat packets for each link channel for testing the network delay. Send the heartbeat packets of each link channel to the V2X server and receive the response message from the V2X server; Based on the comparison result between the timestamp of the response message from the V2X server and the delay threshold, obtain multiple 5G network link channels with smaller delays and use them as qualified 5G network link channels. Select multiple link channels with smaller delays to obtain 5G multi-network data and analyze the signal status of the link channels based on the 5G multi-network data.

[0035] Optionally, in step 3, count the currently qualified 5G multi-network link channels to obtain the mileage residence ratio and duration residence ratio of the currently qualified 5G multi-network link channels, including: Based on the vehicle driving mileage data and the 5G network signal parameter data of the continuous network access, obtain the mileage of the currently qualified 5G multi-network link channels during continuous online access and the total mileage of the vehicle driving. Calculate the ratio of the mileage of the currently qualified 5G multi-network link channels during continuous online access to the total mileage of the vehicle driving to obtain the mileage residence ratio of the currently qualified 5G multi-network link channels. Statistically obtain the actual application continuous residence duration of the user on the currently qualified 5G multi-network link channels and the total driving duration of the vehicle from the 5G network signal parameter data. Calculate the ratio of the actual application continuous residence duration of the user on the currently qualified 5G multi-network link channels to the total driving duration of the vehicle to obtain the duration residence ratio of the currently qualified 5G multi-network link channels.

[0036] Optionally, in step 4, use the NPU neural network to perform quantization processing on the mileage residence ratio, duration residence ratio, and 5G network signal parameter data of the currently qualified 5G multi-network link channels to obtain the comprehensive score of the currently qualified 5G multi-network link channels, including: Perform quantization processing on the mileage residence ratio, duration residence ratio, signal strength RSRP, signal quality SINR, network speed, and delay respectively to obtain the real-time parameter index information of the 5G multi-network link channels that eliminates the dimensional differences between different parameters, and assign weights to each index; when eliminating the dimensional differences between different parameters, use the NPU neural network to standardize the indexes with different dimensions, such as using Min-Max normalization to convert the data to the range of 0-1.

[0037] Then calculate the weighted sum of the weight values of the real-time parameter indexes to obtain the comprehensive score of each link channel.

[0038] In this embodiment, the NPU neural network is used to perform weighted processing on the real-time parameter index information of each link channel to obtain the weight value of each link channel, and then sum the weight values to obtain the comprehensive score. Based on the comprehensive evaluation requirements of the best 5G network, weights are assigned to each indicator. For example: Signal Strength (RSRP): 20%; Signal Quality (SINR): 20%; Duration Residence Ratio: 15%; Mileage Residence Ratio: 15%; Network Speed: 15%; Latency: 15%. Then, the indicators with different dimensions are standardized. For example, using Min-Max normalization, the data is converted to the range of 0 - 1. Then, the weighted summation method is used to calculate the comprehensive score of each network: Comprehensive Score = w1×RSRP + w2×SINR + w3×Duration Residence Ratio + w4×Mileage Residence Ratio + w5×Speed + w6×Latency, where wi is the weight of each indicator. Calculate the comprehensive scores of each channel respectively, and select the one with the highest score as the best channel. Screening and sorting: The best 5G network link channel; The second backup link channel of the 5G network; The third backup link channel of the 5G network.

[0039] List the real-time data of the mileage residence ratio, duration residence ratio, signal strength RSRP, signal quality SINR, network speed, and latency of Channel A and Channel B respectively, as shown in Table 1:

[0040] Table 1 After NPU neural network quantization processing, the evaluation data tables of Channel A and Channel B that eliminate the dimensional differences between different parameters are shown in Table 2:

[0041] Table 2 The formula for calculating the comprehensive score of Channel A is: 0.2×0.9 + 0.2×0.95 + 0.15×0.95 + 0.15×0.98 + 0.15×0.95 + 0.15×0.9 = 0.9320.2×0.9 + 0.2×0.95 + 0.15×0.95 + 0.15×0.98 + 0.15×0.95 + 0.15×0.9 = 0.932 The formula for calculating the comprehensive score of Channel B is: 0.2×0.8 + 0.2×0.9 + 0.15×0.9 + 0.15×0.95 + 0.15×0.9 + 0.15×0.8 = 0.8670.2×0.8 + 0.2×0.9 + 0.15×0.9 + 0.15×0.95 + 0.15×0.9 + 0.15×0.8 = 0.867 Therefore, the comprehensive score of Channel A is higher and should be selected as the best channel.

[0042] Optionally, in step 5, according to the comparison result of the preset conditions and the real-time 5G network signal parameter data associated with the current best link channel, the best link channel is adjusted in real time, such as Figure 3 shown including: Detect the network mode and communication frequency band of the current best link channel in real time and update the 5G network signal parameter data; If it is detected that the current best link channel is the SUL frequency band, then communicate with the high and low frequency combined frequency band of the 5G network to improve the network uplink coverage rate and uplink network speed rate of the best link channel; If there is no SUL frequency band in the communication frequency band of the current best link channel, then recalculate the comprehensive scores of each link channel and switch the link channel with the highest comprehensive score to the best link channel; If it is detected that the network speed of the current best link channel drops to less than the preset network speed lower limit value, then select other standby link channels in descending order of the comprehensive score as the new best link channel and continue to receive the SUL frequency band signal; If the network speed of all link channels is less than the preset network speed lower limit value, then switch to the 4G network mode to communicate with the vehicle networking, and the vehicle is switched to intelligent assisted driving.

[0043] In this embodiment, for the communication frequency band judgment: whether the SUL super uplink frequency bands n81 / n83 / n84 signals are received; after meeting the communication frequency band judgment conditions, start the 5G communication module 5G NR SUL high and low combined frequency band to improve the 5G uplink coverage and uplink rate, and further shorten the delay; In this embodiment, the lower limit value of the network speed drop of the current best link channel is set as: when the network speed of the best 5G network channel drops to <20 Mbps, select other standby link channels in descending order as the new best link channel and continue to receive the SUL frequency band signal; if the network speed of all link channels is <20 Mbps, then switch to the 4G network mode to communicate with the vehicle networking, and the vehicle is switched to intelligent assisted driving.

[0044] Optionally, if it is detected that the current best link channel is the SUL frequency band, then communicating with the high and low frequency combined frequency band of the 5G network further includes: According to the comparison result of the preset uplink network speed threshold and the uplink network speed rate, continue to communicate with the current best link channel; If the average rate of the uplink network speed is greater than the lower threshold of the average rate, and the lowest rate of the uplink network speed is greater than the lower threshold of the lowest rate, then preferentially jump to the SUL frequency band as the 5G best link channel, and the vehicle processor maintains 5G network communication with the cloud server through the best link channel, uploads data to the cloud server and receives the cloud server, and the cloud server can monitor the driverless in real time; If the average rate of the uplink network speed is less than or equal to the lower threshold of the average rate, and the lowest rate of the uplink network speed is less than or equal to the lower threshold of the lowest rate, then recalculate the comprehensive scores of each link channel and switch the link channel with the highest comprehensive score to the best link channel.

[0045] In this embodiment, the average uplink network speed in the SUL frequency band is set to 150 Mbps, and the minimum uplink speed is 80 Mbps. The current best link channel uses whether the SUL super uplink frequency band signal is detected as the criterion. If a stable SUL frequency band signal is received, the 5G communication module activates the SUL high-low frequency combined frequency band, such as the n41 + n83 frequency band combination. By means of the cooperation and complementarity between high frequency and low frequency, and the aggregation of time domain and frequency domain, it makes full use of the spectrum resources of the low frequency band and the good wireless propagation characteristics to supplement the uplink capacity of the 5G TDD mid-frequency band, thereby improving the 5G uplink coverage and capacity and shortening the network delay. The best 5G network link channel uses whether the average network speed of the SUL high-low frequency combined frequency band exceeds 150 Mbps as the criterion. If so, the driverless vehicle preferably uses this channel for video data transmission. If the network speed of the current best 5G network link channel drops to below the lowest network speed threshold of 20 Mbps during the process, the AP processor in the intelligent cockpit automatically switches to the second or third backup 5G network link channel to maintain a stable 5G network. If the signals of the 3 5G network link channels are poor and it drops to the LTE network mode, the in-vehicle computer automatically switches to the intelligent connected vehicle automatic assisted driving mode, that is, through the fusion of radar and multiple camera sensors, combined with intelligent connection, integrating vehicle-road-cloud integration, and through intelligent driving algorithms, to control the driving.

[0046] The AP processor system in the intelligent cockpit continuously monitors the network status and environmental changes, and adjusts the network configuration in real time as needed. If a new and better network is detected, the system will automatically switch to ensure that the best network is always used. At the same time, the system will also predict the network coverage in advance according to the vehicle's location and driving route, and make preparations for network switching.

[0047] In terms of specific implementation details, the system will set up a real-time monitoring module responsible for monitoring the network status and environmental changes. The real-time monitoring module will collect network status information regularly (such as every 500 milliseconds), including signal strength, network latency, and data transmission rate, etc. When a change in the network status is detected, such as a decrease in signal strength or an increase in network latency, the real-time monitoring module will send a notification to the intelligent algorithm module. The intelligent algorithm module will re-evaluate the performance of each network based on the new network status information and decide whether to perform a network switch. In addition, the system will also predict the network coverage in advance according to the vehicle's geographical location and driving route, and make preparations for network switching. For example, when the vehicle is about to enter an area with poor signal coverage such as a tunnel or an underground parking lot, the system will switch to a network with stronger signal in advance to ensure the continuity of the network connection. The system will use high-precision maps and vehicle positioning technology to identify the upcoming signal blind spots or areas with weak network coverage in advance and make preparations for network switching in advance.

[0048] Furthermore, to further improve network stability, a network pre-switching mechanism is introduced. When it is detected that the signal quality of the current working network drops to a preset threshold, the system will automatically initiate the pre-switching process, searching and evaluating backup network channels in advance to ensure a smooth switch before the primary network fails and avoid network interruption. The specific advantages are as follows: 1. Real-time monitoring of the signal quality of the current working network. 2. Initiating the pre-switching process when the signal quality drops to the preset threshold. 3. Searching and evaluating backup network channels, selecting the best backup channel for switching to ensure uninterrupted network. Application examples are as follows: Application 1. Urban complex road environment On a busy road in a certain city, driverless vehicles are equipped with the three-network three-connection 5G communication device of this patent method. During driving, the vehicle receives network signals from different operators in real time through three independent 5G communication modules. The intelligent algorithm module continuously monitors the strength and quality of each network signal. When the vehicle enters an area with weak signal coverage, the system automatically selects the network channel with the strongest signal for data transmission. For example, when the vehicle passes through a certain tunnel, the original signal may become weak, but the system quickly switches to the alternative network with a stronger signal to ensure that the vehicle can continuously receive information about the surrounding environment, such as the positions and speeds of other vehicles, and thus make timely responses and decisions to ensure driving safety.

[0049] Application 2. Remote area testing In remote areas, due to insufficient 5G base station coverage, a single network may not be able to provide a stable connection. Driverless vehicles equipped with this patent method can still achieve a stable 5G network connection in this environment. Through the three-network three-connection technology, the vehicle can make full use of the network resources of different operators. When the network signal of one operator is weak, the system will automatically switch to the network of other operators to ensure the continuity and stability of the network connection. For example, in a certain test, when the vehicle passed through a mountainous area with weak signal coverage, the system maintained a stable 5G connection through dynamic network switching and successfully completed the test task.

[0050] Through the implementation and verification of the above specific examples, it is proved that the patent method is effective and reliable in different environments and scenarios, can provide an efficient and stable 5G network connection for driverless vehicles, and improve the safety and efficiency of autonomous driving.

[0051] Application 3. Remote transportation testing A certain logistics company uses driverless vehicles equipped with the patented method for cargo transportation. During a long-distance transportation mission, the vehicle needs to pass through various complex environments, including urban roads, highways, and remote mountainous areas. On urban roads, the vehicle maintains a stable network connection through the three-network and three-connection 5G communication device, receiving real-time traffic information and surrounding environment data to ensure driving safety. On highways, the vehicle utilizes the low-latency feature of the 5G network to quickly respond to the dynamic changes of the vehicle ahead and avoid potential dangers. In remote mountainous areas, although the 5G signal coverage is weak, the system ensures that the vehicle always maintains a stable network connection through intelligent network switching and successfully completes the entire transportation mission. This not only improves transportation efficiency but also reduces the risks that may be brought by human driving.

[0052] Through the implementation and verification of the above specific examples, it is proved that the patented method is effective and reliable in different environments and scenarios, and can provide an efficient and stable 5G network connection for driverless vehicles, improving the safety and efficiency of autonomous driving.

[0053] In summary, the 5G multi-network fusion communication method proposed in this patent realizes the collaborative utilization and dynamic switching of multi-network resources by simultaneously using at least three 5G networks of different operators. Specifically, when the driverless vehicle starts, the system will automatically detect and connect to all available 5G networks. Each 5G communication module is respectively connected to the network of the corresponding operator, and the network connection parameters are initialized. During operation, each 5G communication module real-time monitors the strength and quality of its own network signal and transmits this data to the intelligent algorithm module. The intelligent algorithm module running on the multi-core heterogeneous AP processor and NPU neural network processing unit of the intelligent cockpit in the integrated device of the present invention can comprehensively evaluate the performance of each network according to multiple indicators such as signal strength, network latency, and data transmission rate, and select the network with the best signal strength and the lowest latency as the main working mode. At the same time, the system will continuously monitor the network status and environmental changes, and automatically switch when a new and better network is detected to ensure that the best network is always used. In addition, the system will also predict the network coverage situation in advance according to the vehicle's position and driving route and make preparations for network switching. For example, when entering an area with poor signal coverage such as a tunnel or an underground parking lot, the system will switch to a network with stronger signal in advance to ensure the continuity of the network connection. This 5G multi-network fusion communication method not only improves the stability and reliability of the network connection but also effectively solves problems such as insufficient 5G base station coverage and network congestion, providing a strong communication guarantee for driverless vehicles. Embodiment

[0054] This embodiment provides a multi-network communication device applied to driverless, including: a vehicle processor, a multi-network communication module, a positioning module, a vehicle networking communication security unit, and a multi-functional vehicle-mounted sensor; The vehicle processor is a multi-core heterogeneous AP processor in the vehicle intelligent cockpit domain, which is composed of an AP processing unit, a GPU processing unit, an NPU neural network processing unit, and a storage unit, and is used to execute the steps of the above method; The multi-network communication module is used for the vehicle to perform 5G full-network communication with more than three operators respectively, communicate with the RSU roadside device, and perform C-V2X vehicle networking communication with surrounding vehicles; The positioning module is connected to the multi-network communication module and is used to send out PPS second pulse signals issued by the multi-network communication module; The vehicle networking communication security unit is built-in with a vehicle networking HSM security encryption chip, and is used to provide security information for vehicle data encryption, decryption, and user identity authentication when the multi-network communication module performs vehicle networking communication with surrounding vehicles.

[0055] Optionally, the multi-source communication module is composed of at least 3 independent 5G + C-V2X full-network communication modules and affiliated USIM card slots; Among them, the 5G + C-V2X communication unit and the USIM communication unit are combined to be compatible with the working modes of 5G network, 4G network, 3G network, and 2G network, and support 5GNR full-network frequency bands, 5GNR SUL super uplink frequency bands, and C-V2X frequency bands; Each 5G + C-V2X communication unit is provided with a 5G network antenna and a C-V2X antenna unit; the 5G network antenna is used to perform 5G network communication with different operators respectively, and the C-V2X antenna unit is used to perform C-V2X mode communication with the RSU roadside unit and surrounding running vehicles.

[0056] Optionally, the positioning module is composed of a GPS signal receiver, an IMU inertial navigation acceleration sensor, and a gyro sensor, and is used to receive the positioning signal, angular velocity, and acceleration of the vehicle; When the positioning signal, angular velocity, and acceleration of the vehicle change, the positioning module is started to obtain the vehicle driving trajectory data. At the same time, the PPS second pulse signal issued by the positioning module is started, which will activate the automatic synchronization of C-V2X communication in the 5G + C-2X full-network module, and start the V2I / V2V broadcast and receive message services with the RSU roadside unit and surrounding running vehicles, so as to collect 5G base station signals within a range of 300 meters to 500 meters from the vehicle; it is transmitted to the AP processor in the in-vehicle intelligent cockpit through the 5G + C-V2X module PCIe2.0 interface for multi-source data analysis and processing to implement the steps of the method in Embodiment 1.

[0057] This embodiment will be described with an intelligent connected vehicle integrated with a 5G+C-V2X all-network OBU and a vehicle infotainment system that supports three-network three-connection built-in. The embodiment does not limit the present invention to only the 5G+C-V2X all-network OBU that supports three-network three-connection, but can cover the 5G+C-V2X all-network OBU with multiple-network multiple-connection.

[0058] As Figure 4 and Figure 5 shown, a 5G+C-V2X all-network OBU that supports at least three-network three-connection and a vehicle infotainment system are installed in an intelligent connected driverless vehicle. This device is the core equipment for realizing the 5G three-network fusion communication of the present invention and provides a stable network connection for the driverless vehicle. The integrated device includes a multi-source communication module including at least 3 independent 5G+C-V2X all-network communication modules and an attached USIM card slot, 5G network antennas for supporting the three major operators and a C-V2X antenna unit; a positioning module includes a high-precision centimeter-level GPS positioning module, a vehicle networking communication security unit includes a vehicle networking security chip, a vehicle processor - a multi-core heterogeneous AP processor for intelligent cockpit applications (built-in GPU and NPU neural network processing units) and a program running storage core unit, a central control screen and an instrument panel unit, and multi-functional vehicle-mounted sensors - vehicle-mounted cameras and radars.

[0059] The multi-source communication module includes 3 independent 5G+C-V2X all-network communication modules and an attached USIM card slot, which are modules compatible with 5G / 4G / 3G / 2G communication; support for 5G NR SA / NSA dual mode; support for 5G NR all-network frequency bands: n1, n28, n41, n78, n79; support for 5G NR SUL super uplink frequency bands: n41+n83, n78+n81, n78+n84; support for C-V2X frequency band: B47(1T / 2R). Based on covering the current existing commercial working frequency bands of operators to the greatest extent and reflecting the different 5G working frequency bands of different operators, at least 3 operators can be selected to access the 5G network. For example, these 3 5G+C-V2X all-network communication modules can be used for accessing the 5G networks of Operator A, Operator B, and Operator C respectively, covering the n83 frequency band (700MHz frequency band), n41 frequency band (2.6GHz frequency band) 2515-2675MHz, and n79 frequency band (4.9GHz frequency band) 4800-4900MHz of Operator A's 5G network; the n78 frequency band (3.5GHz frequency band) 3400MHz-3500MHz of Operator C's 5G network; the n78 frequency band (3.5GHz frequency band) 3500MHz-3600MHz of Operator B's 5G network, thus breaking through the current single 5G network access.

[0060] Three independent 5G + C-V2X all-net communication modules, in addition to respectively supporting different operators' 5G network bands, also support the C-V2X vehicle networking B47 band, support the V2X vehicle networking communication protocol, and perform NR-V2X receive / transmit communication with the RSU roadside unit and surrounding operating vehicles. Each communication module has independent signal reception and data transceiver transmission processing capabilities, and can simultaneously receive and process 5G communication data and V2X communication data from different operators. These data will perform data transmission communication with the multi-core heterogeneous AP processor of the intelligent cockpit application through the USB2.0 or PCIE2.0 interface.

[0061] One of the modules can communicate with a high-precision centimeter-level positioning module through the UART serial port, be used to run the RTK differential positioning algorithm, and provide a high-precision centimeter-level positioning signal.

[0062] Support three operators' 5G network antennas and C-V2X antenna units, which are respectively used for receiving and transmitting communication with the 5G networks of the three major operators, and performing NR-V2X receive / transmit communication with the RSU roadside unit and surrounding operating vehicles. The 5G network antenna is integrated with the 5G + C-V2X all-net communication module, and can be used for real-time collection of 5G network signals / speed signals of the vehicle body position, uploading video data of driverless vehicles, and sending down 5G + V2X cloud control commands, etc.; the C-V2X antenna unit is integrated with the C-V2X communication unit in the 5G + C-V2X all-net communication module, and through V2X point-to-point direct communication, it can be used to collect 5G network signals, speed signals of the RSU roadside unit and surrounding operating vehicles within a range of 300 - 500 meters outside the vehicle body, and receive vehicle video data, warning signals, etc. of the intelligent networked application scenario.

[0063] The high-precision centimeter-level GPS positioning module supports built-in L1 / L5 multi-frequency Beidou / GPS receiving and positioning chips, is connected to the 5G + C-V2X all-net communication module through the UART serial port, can obtain raw observation data, and provides a PPS second pulse for C-V2X communication synchronization; through the built-in acceleration + gyro sensor for inertial navigation, the entire module is used for high-precision centimeter-level positioning of driverless vehicles under the support of the RTK differential positioning algorithm.

[0064] The vehicle networking security chip supports national cryptographic algorithms. By providing security functions such as encryption, decryption, and identity authentication, it protects the vehicle from hacker attacks, ensures the confidentiality and integrity of vehicle data, and the security of communication between the vehicle and external devices. The chip is connected to the 5G + C-V2X all-net communication module through the SPI bus.

[0065] The multi-core heterogeneous AP processor (built-in GPU and NPU neural network processing unit) and the program running storage core unit for intelligent cockpit applications. It integrates multiple processing units to support various complex functions in the intelligent cockpit, can provide powerful computing capabilities and efficient energy management for the intelligent cockpit, and at the same time support complex AI functions and rich multimedia experiences (support for H.264 / H.265 high-definition video encoding and decoding). It has rich external peripheral interfaces: supports 3 PCIe2.0 interfaces for high-speed transmission with 3 5G+C-2X full-network communication modules respectively, and 2 lanes of PCIe2.0 can support a transmission rate of 1GB / s; 3 to 4 USB2.0 / USB3.0 interfaces; MIPI DSI display output interface; there are SDIO / SPI / I2C / UART / CAN bus, among which the CAN bus interface, by connecting to a high-speed CAN bus transceiver, is used to access the in-vehicle central gateway to achieve communication with other ECU domain controls in the vehicle; there is also an RGMII interface that supports connection to an Ethernet PHY chip for expanding in-vehicle Ethernet and communicating with the in-vehicle central gateway.

[0066] In the multi-core heterogeneous AP processor of the intelligent cockpit, the collaborative work of the GPU and NPU provides powerful computing capabilities for data processing and analysis. The GPU is responsible for high-performance graphics processing and parallel computing tasks, while the NPU focuses on deep learning and neural network computing, achieving efficient data processing and multi-source data fusion. The specific division of labor is as follows: The GPU is responsible for high-performance graphics processing and parallel computing tasks to improve data processing efficiency; the NPU focuses on deep learning and neural network computing to achieve efficient data processing and multi-source data fusion; the GPU and NPU work together to provide powerful computing support for driverless vehicles.

[0067] The NPU neural network processing unit built into the AP processor in the core unit is used for multi-data fusion processing of the collected 5G base station signals / network speed signals and sensor data, performing intelligent algorithm data analysis, training, and processing; the memory in the core unit contains LPDDR4 or LPDDR5 memory that supports high-speed program operation and EMMC Flash for storing the collected data.

[0068] The central control screen and instrument panel unit are directly driven by the multi-core heterogeneous AP processor for intelligent cockpit applications through the MIPI DSI display interface to display the external central control screen and digital instrument panel; among them, the central control screen is used to display navigation maps, multimedia entertainment, and control the ECU domain control units of the vehicle, and the instrument panel is used to display the mileage, power, or fuel consumption status of driverless vehicles.

[0069] Such as Figure 4The intelligent cockpit domain controller, radar / multi-channel camera unit, and inertial navigation IMU sensor shown will be used to obtain and collect perception data information such as the driving status and vehicle speed of the driverless vehicle. These data information are connected to the vehicle central gateway through the CAN BUS and in-vehicle Ethernet interfaces of the intelligent driving domain controller to achieve communication connection with the intelligent cockpit domain controller. After data packet processing, the data is transmitted to the 5G+C-V2X communication module, and the 5G cloud server can receive video data in real time.

[0070] The above-mentioned hardware devices together constitute the physical basis of the 5G multi-network fusion communication device for driverless vehicles.

[0071] In this integrated device, based on the 5G frequency band ranges supported by different 5G network operators, it is possible to support inserting at least 3 5G vehicle networking cards of different 5G network operators. For example, the 5G vehicle networking cards corresponding to Operator A, Operator C, and Operator B can support logging in to the 5G frequency band networks of each operator respectively, realizing three-way communication among at least 3 5G networks. Once any one of the cards is registered and logged in to the network, the vehicle head unit can perform Uu communication with the cellular communication base station. The vehicle head unit can read the ICCID number of each SIM card. The CCID (Integrated Circuit Card Identification Number) is the unique identification number of the SIM card, consisting of 20 digits in total. It is equivalent to the ID card of the SIM card and is used to identify and distinguish different SIM cards. The coding format of the CCID is usually as follows: The first six digits are the operator code. For example, for Operator A, it is 898600; for Operator B, it is 898601; for Operator C, it is 898603. The following digits represent different information, such as the code of the province, autonomous region, or municipality directly under the Central Government where this SIM card is issued, and the last two digits of the year number, etc. The last few digits are the user identification code and the check bit, from which it is possible to determine which operator's network. Different operator 5G networks have different coverage ranges and signal strengths. By simultaneously accessing three networks, it is possible to standby and automatically search for mobile cellular communication base stations at the same time, which can maximize the stability of the network connection.

[0072] When the driverless vehicle is driving, the 3 independent 5G+C-V2X full-network communication modules built into this integrated device adopt the method of automatically setting and searching for the 5G frequency bands of each operator to lock the frequency, and respectively receive and collect the signal parameter status of the 5G base station networks of the three major operators in the road section area where the vehicle is located.

[0073] Start the high-precision centimeter-level positioning module to obtain the driving trajectory data of the tracked vehicle. Synchronously activate C-V2X communication to establish point-to-point direct communication within a range of 300 meters with the RSU roadside unit and surrounding running vehicles, and perform V2I / V2V broadcast and receive message services. It can receive 5G base station network signal parameters collected from the RSU roadside unit and surrounding running vehicles, achieving the purpose of collecting 5G base station-related network signal data through multiple channels. By real-time monitoring and collecting network signal data, it provides an important basis for subsequent network selection and handover.

[0074] In summary, the communication device of the present invention is built-in with at least 3 independent 5G + C-V2X full-network communication modules, which can automatically search for 5G frequency bands of each operator, so as to receive and collect the signal status and location of 5G base stations of different operators in the section where the vehicle is located. At the same time, the positioning module is activated to obtain and track the vehicle driving trajectory data. The positioning module emits a pulse signal, which will activate the automatic synchronization of C-V2X communication in the 5G + C-V2X full-network communication module, and start the services of broadcasting and receiving messages with the RSU roadside unit and surrounding running vehicles for V2I and V2V. The device of the present invention is set to obtain 5G signal messages from the RSU roadside unit and surrounding running vehicles within 300 meters to 500 meters around the driverless vehicle, that is, collect the signal status information of 5G base stations within 300 meters to 500 meters from the driverless vehicle. The data of 5G base station signals collected by each operator are transmitted to the AP processor in the in-vehicle intelligent cockpit through the multi-source communication module for multi-source data analysis and processing; and the best network link channel in terms of signal strength, signal quality, signal continuous coverage, network speed, latency, etc. is selected to maintain 5G communication with the server. The present invention does not rely on increasing the 5G base station coverage density to ensure that the driverless 5G network is online. Through 5G multi-network communication technology, the existing network resources are fully utilized to achieve seamless handover and avoid the impact of signal blind spots on driverless driving. The present invention supports 5G multi-network fusion communication of at least three operators, enhances the data transmission ability, and provides higher bandwidth and faster data transmission speed. The driverless vehicle can quickly obtain and process a large amount of environmental information, improving the real-time performance of autonomous driving. The present invention has at least three 5G network channels. When one channel fails or the signal is unstable, the other two channels can be used as backups to ensure that the communication is not interrupted. This redundant design greatly improves the reliability and stability of the system. The present invention reduces latency through physical layer technologies such as optimizing the network random access process and quickly activating SUL super uplink frequency band aggregation, and optimizes mobility management processes such as handover. Reduce the latency of video transmission, enabling the driverless vehicle to respond to various instructions and warnings more quickly. The present invention can reduce 5G communication latency, enabling the vehicle to respond more quickly to instructions and warnings from other vehicles, infrastructure, and cloud services, improving the safety and response speed of driverless autonomous driving. In case of an emergency, a quick response can effectively avoid accidents. The present invention can expand the vehicle's perception range. Through V2I / V2V real-time communication with other vehicles and roadside infrastructure, the driverless vehicle can obtain 5G base station signals and network speed information beyond the perception range of its own 5G communication module. The expanded perception range enables the driverless vehicle to better adapt to complex environments.

[0075] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) that contain computer-usable program code.

[0076] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be realized by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for realizing the functions specified in one Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0077] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing devices to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including instruction means, and the instruction means realizes the functions specified in one Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0078] These computer program instructions can also be loaded onto a computer or other programmable data processing devices, so that a series of operation steps are executed on the computer or other programmable devices to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable devices provide steps for realizing the functions specified in one Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0079] The embodiments of the present invention have been described above in conjunction with the accompanying drawings. However, the present invention is not limited to the above specific embodiments. The above specific embodiments are merely illustrative and not restrictive. Those of ordinary skill in the art, under the inspiration of the present invention and without departing from the spirit and scope protected by the present invention's purpose and claims, can still make many forms, and all of these fall within the protection scope of the present invention.

Claims

1. A multi-network communication method for unmanned driving, characterized in that: include: Obtain 5G network signal parameter data from 5G multi-network link channels respectively; According to the comparison result between the set threshold and the 5G network signal parameter data of each channel, the 5G multi-network link channel that currently meets the standard is obtained; Statistics are collected on the currently qualified 5G multi-network link channels to obtain the mileage residence ratio and duration residence ratio of the currently qualified 5G multi-network link channels; Use the NPU neural network to quantify the mileage residence ratio, duration residence ratio and 5G network signal parameter data of the currently qualified 5G multi-network link channels to obtain a comprehensive score for the currently qualified 5G multi-network link channels; Sort the comprehensive scores of the currently qualified 5G multi-network link channels from high to low to obtain the ranking of the current best link channel and other backup link channels; Adjust the best link channel in real time according to the comparison result of the preset conditions and the real-time 5G network signal parameter data associated with the current best link channel; The optimal link channel based on real-time adjustment maintains communication with the cloud server for unmanned driving remote communication.

2. The multi-network communication method for unmanned driving according to claim 1 is characterized in that: Obtain 5G network signal parameter data from 5G multi-network link channels respectively, including: When the vehicle is running, it conducts C-V2X communication with multiple 5G base stations of different operators within a preset range, the RSU roadside unit of the current road section, and surrounding vehicles in real time, forming a 5G multi-network link channel; the 5G multi-network link channel includes the 5G network link channel at the local location of the vehicle and the C-V2X vehicle network transmission link channel detected by the RSU roadside unit and surrounding vehicles; Real-time detection of the 5G network link channel at the local location of the vehicle and the network mode and communication frequency band of the C-V2X vehicle network transmission link channel from the RSU roadside unit and the surrounding vehicles. When the network mode of the 5G network link channel supports 5GNR SA and NSA dual-mode, and the 5G network link channel and the C-V2X vehicle network transmission link channel support 5GNR full network frequency band, 5GNR SUL super uplink frequency band and C-V2X frequency band communication, obtain the 5G network link channel network signal data at the local location of the vehicle, receive the 5G network link channel network signal data from the RSU roadside unit and the surrounding vehicles, and obtain the vehicle mileage data; The 5G network link channel at the local location of the vehicle and the network signal data of the 5G network link channel transmitted from the RSU roadside unit and the surrounding vehicles through the C-V2X vehicle network transmission link channel are analyzed to obtain the 5G network signal parameter data associated with the 5G multi-network network link channel, where the 5G network signal parameter data includes the signal strength RSRP, signal quality SINR, network speed, latency and dwell time of each link channel.

3. The multi-network communication method for unmanned driving according to claim 2 is characterized in that: According to the comparison results between the set threshold and the 5G network signal parameter data of each channel, the currently qualified 5G multi-network link channel is obtained, including: According to the comparison results of the set signal strength RSRP threshold, signal quality SINR threshold, average uplink and downlink speed threshold and minimum uplink and downlink speed threshold respectively with the signal strength RSRP, signal quality SINR, average uplink and downlink speed, and minimum uplink and downlink speed associated with the 5G multi-network network link channel, a preliminary qualified 5G multi-network network link channel is obtained; Based on the heartbeat mechanism, a latency test is performed on the preliminarily qualified 5G multi-network link channel to obtain a 5G multi-network link channel with smaller latency, which is used as the currently qualified 5G multi-network link channel.

4. The multi-network communication method for unmanned driving according to claim 3 is characterized in that: Statistics are collected on the currently qualified 5G multi-network link channels to obtain the mileage retention ratio and duration retention ratio of the currently qualified 5G multi-network link channels, including: According to the vehicle mileage data and the 5G network signal parameter data of the network access, the mileage of the current 5G multi-network link channel network access and the total mileage of the vehicle are obtained; The ratio of the continuous online mileage of the currently qualified 5G multi-network link channel to the total mileage of the vehicle is calculated to obtain the mileage retention ratio of the currently qualified 5G multi-network link channel; The actual application duration of users in the currently qualified 5G multi-network link channel and the total driving time of vehicles are calculated from the 5G network signal parameter data; The ratio of the actual continuous residence time of users in the currently qualified 5G multi-network network link channel and the total driving time of the vehicle is calculated to obtain the current qualified 5G multi-network network link channel residence ratio.

5. The multi-network communication method for unmanned driving according to claim 4 is characterized in that: The NPU neural network is used to quantify the mileage residence ratio, duration residence ratio and 5G network signal parameter data of the currently qualified 5G multi-network link channels to obtain a comprehensive score of the currently qualified 5G multi-network link channels, including: The mileage dwell ratio, duration dwell ratio, signal strength RSRP, signal quality SINR, network speed and latency are quantified to obtain real-time parameter indicator information of 5G multi-network link channels that eliminates the dimensional differences between different parameters. The NPU neural network is used to perform weighted summation calculation on the real-time parameter index information of 5G multi-network link channels that eliminates the dimensional differences between different parameters, and obtain a comprehensive score for each link channel.

6. The multi-network communication method for unmanned driving according to claim 5 is characterized in that: According to the comparison result of the preset conditions and the real-time 5G network signal parameter data associated with the current best link channel, the best link channel is adjusted in real time, including: Real-time detection of the network mode and communication frequency band of the current best link channel and update of 5G network signal parameter data; If it is detected that the current best link channel is the SUL frequency band, the 5G network will communicate with the high and low frequency combined frequency bands to improve the network uplink coverage and uplink speed of the best link channel; If there is no SUL band in the communication frequency band of the current best link channel, the comprehensive scores of each link channel are recalculated and the link channel with the highest comprehensive score is switched to the best link channel; If it is detected that the network speed of the current best link channel drops below the preset network speed lower limit, other backup link channels are selected from high to low according to the comprehensive score as the new best link channel and continue to receive the signal of the SUL frequency band; If the network speed of all link channels is less than the preset lower limit of network speed, the communication with the Internet of Vehicles will be reduced to 4G network mode, and the car will be switched to intelligent assisted driving.

7. The multi-network communication method for unmanned driving according to claim 6 is characterized in that: If it is detected that the current best link channel is the SUL frequency band, communication is performed using the high and low frequency combined frequency bands of the 5G network, which also includes: According to the comparison result between the preset uplink speed threshold and the uplink speed rate, the communication is continued with the current best link channel; If the average rate of the uplink network speed is greater than the lower threshold of the average rate, and the minimum rate of the uplink network speed is greater than the lower threshold of the minimum rate, the SUL frequency band is preferentially jumped to the 5G best link channel. The vehicle processor maintains 5G network communication with the cloud server through the best link channel, uploads data to the cloud server and receives data from the cloud server. The cloud server can monitor the unmanned driving in real time; If the average rate of the uplink network speed is less than or equal to the lower limit threshold of the average rate, and the minimum rate of the uplink network speed is less than or equal to the lower limit threshold of the minimum rate, the comprehensive scores of each link channel are recalculated and the link channel with the highest comprehensive score is switched to the best link channel.

8. A multi-network communication device for unmanned driving, characterized in that: include: Vehicle processor, multi-network communication module, positioning module and Internet of Vehicles communication security unit; The vehicle processor is a multi-core heterogeneous AP processor in the vehicle intelligent cockpit domain, which is composed of an AP processing unit, a GPU processing unit, an NPU neural network processing unit and a storage unit, and is used to execute the steps of the method according to any one of claims 1 to 7; The multi-network communication module is used for the vehicle to perform 5G full network communication with more than three 5G network operators, communicate with RSU roadside devices, and perform C-V2X vehicle networking communication with surrounding vehicles; The positioning module is connected to the multi-network communication module, and is used for tracking and positioning the vehicle's driving path, and sends a PPS second pulse signal to the multi-network communication module; The Internet of Vehicles communication security unit has a built-in Internet of Vehicles HSM security encryption chip, which is used to provide security information for vehicle data encryption, decryption and user identity authentication when the multi-network communication module communicates with surrounding vehicles in the Internet of Vehicles.

9. The multi-network communication device for unmanned driving according to claim 8, characterized in that: The multi-network communication module is composed of at least three independent 5G+C-V2X full network communication modules and an attached USIM card slot; Among them, the 5G+C-V2X communication unit and USIM communication unit are combined to be compatible with 5G network, 4G network, 3G network and 2G network working modes, and support 5G NR full network frequency band, 5G NRSUL super uplink frequency band and C-V2X frequency band; Each 5G+C-V2X communication unit is equipped with a 5G network antenna and a C-V2X antenna unit; the 5G network antenna is used for 5G network communications with different operators, and the C-V2X antenna unit is used for C-V2X mode communications with the RSU roadside unit and surrounding vehicles.

10. The multi-network communication device for unmanned driving according to claim 8, characterized in that: The positioning module is composed of a GPS signal receiver, an IMU inertial navigation acceleration sensor and a gyroscope sensor, and is used to receive the vehicle's positioning signal, angular velocity and acceleration; When the vehicle's positioning signal, angular velocity and acceleration change, the positioning module is started to obtain the vehicle's driving trajectory data. At the same time, the PPS second pulse signal emitted by the positioning module is started to activate the C-V2X communication automatic synchronization in the 5G+C-2X full network module, and start the V2I / V2V broadcast and message receiving service with the RSU roadside unit and surrounding vehicles, thereby collecting 5G base station signals within 300 meters to 500 meters from the vehicle.

Images