Redundant power supply control system for intelligent vehicle computing platform

Abstract

The present invention provides a redundant power supply control system for a reconfigurable AI intelligent vehicle-mounted computing power platform; it includes a power supply control strategy and a computing power control strategy; the power supply control strategy includes: S1: real-time monitoring of the health of the main power supply, and if the main power supply fails, proceed to S2; S2: supply power to the vehicle-mounted power system, the computing power sleep instruction shuts down the corresponding computing power module, and the computing power call instruction sends a computing power borrowing request; S3: skip frames to select video data for processing to generate primary vehicle control information; predict computing power demand for those that are not selected and hand them over to other computing power platforms for processing; S4: transmit video data; S5: receive the returned secondary vehicle control information, and perform fusion calculations to form the ultimate vehicle control information; the computing power control strategy includes: C1: review the borrowable computing power value after receiving the borrowing request; C2: call the cache cleaning instruction set, clean up the storage space, call the computing power allocation instruction set, and provide a computing power module; C3: generate secondary vehicle control information, and return it to the computing power platform that sent the borrowing request.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle-mounted computing power platforms, and in particular to a redundant power supply control system for an intelligent vehicle-mounted computing power platform. Background Art

[0002] With the rapid development of the automotive industry, intelligence, electrification and networking have become important trends in the development of modern automobiles. As the core component for vehicles to realize functions such as autonomous driving, intelligent interaction, and data processing, the importance of AI intelligent in-vehicle computing platforms is becoming increasingly prominent.

[0003] In order to ensure the stable operation of the vehicle's power system, a backup battery is generally configured in the prior art to maintain basic vehicle operation in the event of a sudden failure of the main power supply, such as parking at the side of the road or driving to a nearby maintenance point. For example, an electric vehicle with a backup power supply is proposed in Publication (Announcement) No.: CN109466302A; however, in actual use, the backup battery is generally idle for a long time. Apart from regular maintenance and charging and discharging, it is difficult to obtain efficient utilization. Summary of the Invention

[0004] In view of this, the problem to be solved by the present invention is to provide a redundant power supply control system for an intelligent vehicle-mounted computing power platform.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0006] Redundant power supply control system for intelligent vehicle computing platform,

[0007] The vehicle computing platform includes: a redundant power conversion unit powered by a backup power supply, a communication switching unit, and a processing unit;

[0008] The redundant power supply control system is configured with a power supply control strategy, the power supply control strategy comprising the following steps;

[0009] S1: Real-time monitoring of the health of the main power supply for the vehicle power system, sending idle computing power information of the computing power platform and receiving idle computing power information of other computing power platforms. If the main power supply fails, proceed to S2;

[0010] S2: The redundant power conversion unit supplies power to the vehicle power system according to a preset quota ratio, and at the same time obtains a corresponding power supply switching value, and calls a corresponding computing power sleep instruction set and a computing power call instruction set according to the power supply switching value. The computing power sleep instruction set includes a plurality of computing power sleep instructions, and the computing power sleep instructions correspond to the computing power module settings and are used to shut down the corresponding computing power module. The computing power call instruction set includes a plurality of computing power call instructions, and the computing power call instructions are used to send a computing power loan request through the communication exchange unit;

[0011] S3: Frame skipping selects video data collected by the on-board camera and sends it to the processing unit for processing, thereby generating primary vehicle control information. The computing power scheduling and allocation model predicts the computing power demand for the unselected video data and selects other computing power platforms that are compatible with the idle computing power information based on the predicted computing power demand.

[0012] S4: The video data is segmented and each segment is independently encapsulated and transmitted through the real-time streaming protocol;

[0013] S5: Receive the secondary vehicle control information sent back by other computing power platforms, fuse and calculate the primary vehicle control information and the secondary vehicle control information to form the ultimate vehicle control information;

[0014] The redundant power supply control system is configured with a computing power control strategy, which includes the following steps:

[0015] C1: After receiving a loan request from another computing power platform, the borrowable computing power value of the computing power platform is checked. If the borrowable computing power value is greater than the borrowed computing power value, C2 is carried out; otherwise, the loan request is rejected.

[0016] C2: Calls the cache cleanup instruction set and the computing power allocation instruction set based on the borrowed computing power value;

[0017] The cache cleaning instruction set includes a number of cache cleaning instructions, which are set in correspondence with the data storage module and are used to pre-clean out corresponding storage space for the packaged data to be received;

[0018] The computing power allocation instruction set includes several computing power allocation instructions, corresponding to the computing power module settings, which are used to provide the corresponding computing power module for the received packaged data;

[0019] C3: The processing unit processes the encapsulated data and generates secondary vehicle control information, and transmits the secondary vehicle control information back to the computing power platform that sent the secondment request.

[0020] The redundant power control system is also configured with a vehicle control strategy;

[0021] W1: Real-time monitoring of whether there is power switching. If power switching occurs, proceed to W2; otherwise, maintain the current driving state;

[0022] W2: Obtain the corresponding power supply switching value, and call the corresponding deceleration instruction set and the primary vehicle control instruction set according to the power supply switching value. The deceleration instruction set corresponds to the vehicle power system setting, including a number of deceleration instructions for decelerating the vehicle power system to a preset value;

[0023] The primary vehicle control instruction set corresponds to the non-sleep computing power module setting, including several primary vehicle control instructions, which are used to enable the on-board power system to execute primary vehicle control information. The primary vehicle control instruction set takes precedence over the deceleration instruction set.

[0024] W3: Retrieve the ultimate vehicle control instruction set, which includes several ultimate vehicle control instructions used to enable the vehicle power system to execute ultimate vehicle control information.

[0025] The computing power scheduling and allocation model includes a computing power assessment module and a borrowable computing power calculation module. The computing power assessment module is used to predict the computing power demand for unselected video data, and the borrowable computing power calculation module is used to calculate the idle computing power available from other on-board computing power platforms.

[0026] The evaluation formula of the computing power evaluation module is:

[0027] D i =αR i +βF i +γC i

[0028] R i is the resolution of the i-th video clip, F i is the frame rate of the i-th video clip, C i is the coding complexity of the i-th video segment; α, β, γ are preset weight coefficients;

[0029] The video data collected by the vehicle camera includes MPEG-2, H.263, AVC, HEVC, and AV1 in order of encoding complexity. As the encoding complexity increases, C i The value of and the corresponding preset weight coefficient also increase synchronously.

[0030] The calculation formula of the borrowable computing power calculation module is:

[0031] S=AB·H

[0032] B=(1-H)·A

[0033] S is the idle computing power, H is the health index of the main power supply (ranging from 0 to 1), A is the total available computing power, and B is a linear function of H, which represents the impact of the main power supply health on the idle computing power. When H is 0, it means that the main power supply is completely faulty, in which case B = A. When H is 1, it means that the main power supply is completely normal, in which case B = 0.

[0034] The computing power scheduling and allocation model also includes a stability analysis module, which is used to calculate the stability score of the computing power platform. The computing power scheduling and allocation model selects computing power platforms with stability scores within a threshold from the computing power platforms that meet the predicted computing power requirements;

[0035] The calculation formula of the stability analysis module is:

[0036] G=R+X i ·P i

[0037] G is the stability score, R is the scoring constant, X i is the signal strength score of the i-th vehicle computing platform, P i is the response time score of the i-th vehicle computing platform.

[0038] The navigation data is exchanged with several computing power platforms to be selected, the spatial change position of the vehicle is determined based on the navigation data, and the computing power platform whose moving distance within the preset time does not exceed the preset stable distance is selected.

[0039] The system exchanges driving information with several computing power platforms to be selected, determines the same direction of the two vehicles based on the direction of the vehicle heads, and selects the computing power platform with the same direction of the vehicles.

[0040] Vehicle information is exchanged with several selected computing power platforms. Among the selected computing power platforms, the computing power platform with the same video data encoding type as the vehicle-mounted camera is selected as the highest priority.

[0041] Repeatedly calculate the stability score of the computing power platform according to the preset time.

[0042] The calculation formula for the ultimate vehicle control information is:

[0043] U=f(W P P+W Z Z,diff(P,Z))

[0044] diff(P,Z)=|PZ|

[0045] P represents the priority of primary vehicle control information, Z represents the priority of secondary vehicle control information, WP and WZ represent the preset weight factors of primary and secondary information respectively, U represents the ultimate vehicle control information, and diff(P,Z) is the correlation function used to calculate the degree of difference between primary and secondary information.

[0046] The advantages and positive effects of the present invention are:

[0047] (1) The on-board computing power platform is powered by a backup power supply. When the main power supply fails, the redundant power conversion unit supplies power to the on-board power system. At the same time, by shutting down some CPUs, the power consumption of the backup power supply is reduced. On the one hand, it can ensure that the on-board power system is powered according to the preset quota ratio and keep the on-board power system running. On the other hand, shutting down some computing power modules with large power consumption can also reduce the probability of overload of the backup power supply and avoid causing the whole vehicle to lose control. The tasks of this computing power platform can be handled by borrowing other computing power platforms with idle computing power, which not only realizes the efficient use of the backup power supply, but also enables the whole vehicle to operate normally after the backup power supply takes over the main power supply.

[0048] (2) Each computing power platform in the networked state can be used as an object to borrow computing power or be borrowed computing power, so that the computing power of each vehicle-mounted computing power platform in the entire regional network can be efficiently allocated and utilized. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0050] In the attached figure:

[0051] Figure 1 It is a flowchart of the power control strategy in the redundant power control system of the intelligent vehicle-mounted computing power platform of the present invention;

[0052] Figure 2 It is a flowchart of the computing power control strategy in the redundant power supply control system of the intelligent vehicle-mounted computing power platform of the present invention. DETAILED DESCRIPTION

[0053] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0054] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention pertains. The terms used in this specification of the present invention are for the purpose of describing specific embodiments only and are not intended to limit the present invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0055] The present invention provides a redundant power supply control system for an intelligent vehicle-mounted computing platform;

[0056] The on-board computing power platform includes: a redundant power conversion unit, a communication exchange unit, and a processing unit powered by a backup power supply. The backup power supply uses several small power supplies connected in parallel. This way, even if one of the small power supplies is damaged, there will be no power outage, thereby ensuring the long-term operation of the on-board computing power platform and realizing the utilization of the backup power supply.

[0057] The redundant power supply control system is configured with a power supply control strategy, the power supply control strategy comprising the following steps;

[0058] S1: Real-time monitoring of the health of the main power supply for the vehicle power system, sending idle computing power information of the computing power platform and receiving idle computing power information of other computing power platforms. If the main power supply fails, proceed to S2;

[0059] Each computing power platform can interact with each other through a star network, chain network, etc., and obtain real-time information on the computing power of each platform and the health of the main power supply.

[0060] The health of the main power source is expressed as a percentage, indicating the remaining proportion of the battery capacity relative to a new battery. For example, a power source health of 80% means that the current power provided by the battery is 80% of its original capacity when it was brand new and unused. The higher the battery health value, the closer the battery performance is to its new state. Conversely, the lower the value, the greater the battery loss and the more obvious the performance degradation. The main reason for the decline in power source health is the aging of the electrode material, which causes changes in its structure, such as changes in the crystal structure, particle growth or cracking, which in turn affects the battery performance. This aging process is irreversible, so as the health decreases, the risk of damage also increases accordingly, and the overall performance of the battery will also decrease. When the power source health is 0, it means that the main power source is completely unusable. The power source health can be monitored and obtained by the on-board battery management system.

[0061] Idle computing power refers to the remaining computing power of the computing power platform's CPU when it is not processing full-load tasks. The main reason for its generation is the performance difference of the CPU. For example, if the computing power platform is equipped with a high-performance CPU, its computing speed and processing power far exceed the requirements of the application. Even when running some more complex programs, these performances may not be fully utilized, resulting in some computing power being idle. Conversely, if the computing power platform is equipped with a low-performance CPU, then when running some more complex programs, there may be insufficient computing power, making it impossible to complete them on time.

[0062] S2: Because the main power supply is completely unusable, the redundant power conversion unit supplies power to the on-board power system according to a preset quota ratio, thereby maintaining the operation of the on-board power system, and at the same time obtains the corresponding power supply switching value, and calls the corresponding computing power sleep instruction set and computing power call instruction set according to the power supply switching value. The computing power sleep instruction set includes several computing power sleep instructions, and the computing power sleep instructions correspond to the computing power module settings, which are used to shut down the corresponding computing power module. The computing power module is the CPU (central processing unit) with higher power consumption in the entire on-board computing power platform, that is, the computing power sleep instruction in this application is an instruction for controlling the CPU to stop running. By shutting down some CPUs to reduce the power consumption of the backup power supply, on the one hand, it can ensure that the on-board power system is powered according to the preset quota ratio and maintain the operation of the power system. On the other hand, since the backup power supply adopts a mode of supplying power in parallel with several small power supply bodies, shutting down some CPUs with higher power consumption can also reduce the probability of overload of the backup power supply and avoid loss of control of the entire vehicle;

[0063] The computing power call instruction set includes a plurality of computing power call instructions, and the computing power call instructions are used to send a computing power loan request through the communication exchange unit;

[0064] Since the computing power of this computing power platform has decreased due to the shutdown of some computing power modules, other computing power platforms with idle computing power are used to handle the tasks of this computing power platform.

[0065] S3: Skip frames and select the video data collected by the vehicle camera and hand it over to the processing unit for processing, thereby generating primary vehicle control information;

[0066] Frame skipping selects video data collected by on-board cameras. When processing the video captured by the on-board cameras, instead of using every frame, some frames are skipped according to the preset frame width, and only some frames are selected for processing and analysis, thereby reducing the amount of data to be processed and reducing the transmission pressure. For example, if a video originally has 30 frames per second, the data processing amount can be reduced by half by selecting every other frame through frame skipping.

[0067] Since the computing power of this computing platform has decreased due to the shutdown of some computing modules, we first perform frame skipping on the collected video data, that is, sampling a portion of the data. Due to the reduction in processing volume, the computing modules that are not shut down can also be processed within the time limit, thereby promptly discovering abnormal situations and forming primary vehicle control information to respond. Vehicle control information refers to the data and instruction set for managing and regulating various systems and components of the vehicle to achieve safe and efficient operation of the vehicle. Primary vehicle control information is the first vehicle control information issued and handed over to the on-board power system for execution, so that the vehicle can prepare for possible actions to be performed in advance;

[0068] For example, when there is a warning of vehicle deceleration ahead, if the image information determines that the vehicle ahead is decelerating, the system will apply the brakes accordingly to avoid a collision. If the image information determines that there are speed bumps, ditches, potholes and other obstacles on the road, the system will calculate the avoidance route in advance based on the location of the obstacle and the vehicle's speed.

[0069] For video data that is not selected by frame skipping, the computing power scheduling and allocation model predicts the computing power demand for the unselected video data, and selects other computing power platforms that are adapted to the idle computing power information based on the predicted computing power demand for processing.

[0070] S4: The video data is segmented and each segment is independently encapsulated and transmitted through the real-time streaming protocol;

[0071] Pre-established rules for sending video data: if the file size is less than 30M, it is considered a small data packet and can be transmitted directly; if the file size is larger than 30M, it is considered a large data packet. The sender can divide the video data into multiple smaller segments for transmission and reassemble them at the receiving end. In this way, even if a segment is lost, only the segment needs to be retransmitted instead of the entire data packet, thereby improving transmission efficiency and reliability.

[0072] S5: Receive secondary vehicle control information sent back by other computing power platforms. Secondary vehicle control information is a collection of data and instructions for managing and regulating various vehicle systems and components, generated based on the video data sent to other computing power platforms. The primary vehicle control information and secondary vehicle control information are then integrated and calculated to form the ultimate vehicle control information. The ultimate vehicle control information is the vehicle control information that is ultimately executed by the onboard power system after the secondary vehicle control information.

[0073] If the ultimate vehicle control information requires a parking action, and the vehicle has already performed a braking action according to the primary vehicle control information, the previous braking action will be switched to a braking stop action. If the ultimate vehicle control information requires an acceleration action, the previous braking action will be switched to an accelerator pressing action.

[0074] The redundant power supply control system is configured with a computing power control strategy, which includes the following steps:

[0075] C1: After receiving a loan request from another computing power platform, the platform will check the available computing power of the computing power platform against its own. This means recalculating the current computing power usage of the computing power platform to determine whether there is sufficient idle computing power available for loan. If the available computing power is greater than the computing power value of the loan request, C2 will be carried out. Otherwise, the loan request will be rejected. Maintaining the normal operation of the computing power platform is the top priority.

[0076] C2: Calls the cache cleanup instruction set and the computing power allocation instruction set based on the borrowed computing power value;

[0077] The cache cleaning instruction set includes several cache cleaning instructions, which are set in the corresponding data storage module and are used to pre-clean out the corresponding storage space for the packaged video data to be received. For example, it cleans up temporary files that have completed their mission. Temporary files are files that are generated during the program running and occupy a certain amount of storage space. By cleaning up temporary files, it is ensured that there is enough storage space to receive the packaged video data.

[0078] The computing power allocation instruction set includes several computing power allocation instructions, which are commands for managing and allocating computing power modules. The corresponding computing power module settings are used to provide the corresponding computing power module for the received encapsulated video data, and then unpack and calculate the encapsulated video data;

[0079] C3: Generates secondary vehicle control information after processing the encapsulated video data, and transmits the secondary vehicle control information back to the computing power platform that sent the loan request;

[0080] Since encapsulated video data has a certain lag after transmission, in this application, the encapsulated video data can also be pre-processed first. According to the format of the encapsulated video data, such as MP4, AVI, etc., the corresponding decoding library or tool is used to convert it into a sequence of raw video frames suitable for processing. For example, OpenCV can be used to easily open and read video files, convert the encapsulated video data into frame-by-frame image data, and then perform image enhancement operations on the decapsulated video frames to improve the quality and clarity of the image and highlight the characteristics of the target object, including contrast adjustment, brightness enhancement, color correction, etc. At the same time, filtering algorithms such as Gaussian filtering and median filtering are used to remove noise from the image to reduce the interference of noise on subsequent target detection and recognition. Finally, according to the input size required by the processing unit, the video frame is resized and normalized to ensure that the image size meets the input requirements of the image processing model and the pixel values ​​are normalized to a specific range so that it can be processed accurately.

[0081] The redundant power control system is also configured with a vehicle control strategy;

[0082] W1: Determine whether power switching is required by real-time monitoring of the voltage and current of the main power supply and backup power supply. When the main power supply fails or is interrupted, the voltage and current will change significantly, such as the main power supply voltage drops or becomes zero. At the same time, the voltage and current of the backup battery will change accordingly to provide power support. If power switching occurs, proceed to W2; otherwise, maintain the current driving state;

[0083] W2: Obtaining the corresponding power supply switching value, and calling the corresponding deceleration instruction set and primary vehicle control instruction set based on the power supply switching value. The deceleration instruction set corresponds to the vehicle power system settings. The deceleration instruction set is a collection of deceleration instructions. Different deceleration instructions correspond to different deceleration actions of the vehicle. The deceleration instructions of the present invention are directed to the electric power module, that is, reducing the power output of the vehicle to perform a deceleration action. On the one hand, it avoids excessive power consumption, and on the other hand, it is convenient to make corrections to the vehicle's driving action at low speeds;

[0084] The primary vehicle control instruction set corresponds to the non-sleep computing power module setting and is a collection of primary vehicle control instructions. Different primary vehicle control instructions correspond to different vehicle driving speed and direction adjustments. The primary vehicle control instructions of the present invention are targeted at the electric power module, thereby adjusting the vehicle driving speed and direction; the primary vehicle control instruction set takes precedence over the deceleration instruction set.

[0085] W3: Retrieve the ultimate vehicle control instruction set. The ultimate vehicle control instruction set is a collection of several ultimate vehicle control instructions. Different ultimate vehicle control instructions correspond to different vehicle speed and direction adjustments. The ultimate vehicle control instructions of the present invention are targeted at the electric power module, thereby adjusting the vehicle speed and direction; the ultimate vehicle control instruction set takes precedence over the deceleration instruction set.

[0086] The calculation formula for the ultimate vehicle control information is:

[0087] U=f(W P P+W Z Z,diff(P,Z))

[0088] diff(P,Z)=|PZ|

[0089] P represents the priority of the primary vehicle control information, Z represents the priority of the secondary vehicle control information, the priority of the primary vehicle control information is higher than the priority of the secondary vehicle control information, WP and WZ represent the preset weight factors of the primary and secondary information respectively, U represents the ultimate vehicle control information, and diff(P, Z) is a correlation function used to calculate the degree of difference between the primary and secondary information;

[0090] Specifically, a threshold R can be set. If diff(P, Z) exceeds the threshold, it indicates that there is a large difference between the two pieces of information and a more conservative calculation is required, as follows:

[0091]

[0092] V represents another threshold used to determine whether the information difference is large enough to require the safest action.

[0093] Specifically, the computing power scheduling and allocation model includes a computing power evaluation module and a borrowable computing power calculation module. The computing power evaluation module is used to predict the computing power demand of unselected video data and calculate the required computing power. The borrowable computing power calculation module is used to calculate the idle computing power available from other on-board computing power platforms.

[0094] The evaluation formula of the computing power evaluation module is:

[0095] D i =αR i +βF i +γC i

[0096] R i is the resolution of the i-th video clip, F i is the frame rate of the i-th video clip, C i is the coding complexity of the i-th video segment; α, β, γ are preset weight coefficients;

[0097] The video data collected by the vehicle camera includes MPEG-2, H.263, AVC, HEVC, and AV1 in order of encoding complexity. As the encoding complexity increases, C i The value of and the corresponding preset weight coefficient also increase synchronously.

[0098] For the calculation of the borrowable computing power value of the computing power platform, this application introduces the parameter of the health index of the main power supply, because the main reason for the decline in the health of the power supply is the aging of the electrode material, which causes changes in its structure, such as changes in the crystal structure, enlargement or rupture of particles, etc., which in turn affects the performance of the battery; this aging process is irreversible, so as the health index decreases, the risk of main battery failure also increases accordingly. Therefore, for the computing power platform in this situation, some idle computing power must be saved and not borrowed out, because when the main power supply suddenly fails, if some computing power modules are still in During the borrowing period, the computing power module corresponding to the stored idle computing power can be used to handle the instantaneous complex calculations caused by the sudden failure of the main power supply. The higher the health index of the main power supply, the lower the probability of sudden failure of the main power supply, and the less idle computing power needs to be retained. The lower the health index of the main power supply, the higher the probability of sudden failure of the main power supply, and the more idle computing power needs to be retained. That is, the idle computing power retained to deal with sudden failure of the main power supply is subtracted from the current total available computing power to obtain the borrowable idle computing power. The calculation formula is:

[0099] S=AB·H

[0100] B=(1-H)·A

[0101] S is the available idle computing power, H is the health index of the main power supply (ranging from 0 to 1), A is the total available computing power, and B is a linear function of H, which represents the impact of the main power supply health on the idle computing power. When H is 0, it means that the main power supply is completely faulty, in which case B = A. When H is 1, it means that the main power supply is completely normal, in which case B = 0.

[0102] Furthermore, considering that the transmission stability of the computing power platform will easily affect the transmission timeliness of the secondary vehicle control information during the secondment of computing power, this system also designs a stability analysis module to calculate the stability score of each computing power platform in the computing power platform that meets the predicted computing power requirements, and screen the computing power platforms with stability scores within the threshold;

[0103] The calculation formula of the stability analysis module is:

[0104] G=R+X i ·P i

[0105] G is the stability score, R is the scoring constant, X i is the signal strength score of the i-th vehicle computing platform, P i is the response time score of the i-th vehicle computing platform.

[0106] The signal strength score is positively correlated with the signal strength. The higher the signal strength, the higher the corresponding signal strength score. The response time score is negatively correlated with the response time. The longer the response time, the lower the corresponding response time score.

[0107] Furthermore, considering that the signal strength and response time of the two vehicles vary with the relative distance between them, if the relative distance between the vehicles changes significantly during the computing power transfer time, it will cause a sudden change in the stability score, thereby affecting the transmission stability of the computing power platform.

[0108] Therefore, in one embodiment, the computing power platform also exchanges navigation data with several computing power platforms to be selected, determines the spatial change position of the vehicle based on the navigation data, and selects the computing power platform whose moving distance within the preset time does not exceed the preset stability distance from the computing power platforms whose computing power stability scores meet the preset requirements;

[0109] To prevent drivers' privacy concerns from being compromised during navigation data exchange, a distributed computing architecture can be leveraged to distribute navigation data processing tasks across multiple nodes. Each node processes only a portion of the data and remains unaware of the data being processed by other nodes. This prevents a single node from possessing complete vehicle route information, reducing the risk of privacy breaches. Alternatively, a data access rights management system can be established, ensuring that only authorized parties can access and process navigation data, ensuring that information such as the vehicle's route is not disclosed to unauthorized entities, including the owners of other vehicles.

[0110] In one embodiment, the computing power platform also exchanges driving information with several computing power platforms to be selected, determines the same direction of the bodies of the two vehicles based on the direction of the front of the vehicle, and selects the computing power platform with the same direction of the body from the computing power platforms whose computing power stability scores meet the preset requirements.

[0111] Furthermore, considering that different models of computing platforms are equipped with different on-board cameras, the encoding types of the generated video data are also different. This requires the computing platform to transcode video data of encoding types not supported by this platform before processing it. Therefore, this computing platform also exchanges vehicle information with several selected computing platforms. Among the computing platforms whose moving distance does not exceed the preset stable distance or the computing platforms in the same direction as the vehicle body, the computing platform with the same video data encoding type of the on-board camera is selected as the highest priority. In this way, the transcoding process can be omitted, thereby further improving the parsing and processing speed of the video data.

[0112] Specifically, the stability score of the computing power platform is repeatedly calculated according to the preset time length to ensure the stability of the transmission between the computing power lending platform and the borrowed computing power platform.

[0113] The embodiments of the present invention are described in detail above, but the contents described are only preferred embodiments of the present invention and should not be considered to limit the scope of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of this patent.

Claims

1. The redundant power supply control system of the intelligent vehicle computing platform is characterized by ； The vehicle computing platform includes: a redundant power conversion unit powered by a backup power supply, a communication switching unit, and a processing unit; The redundant power supply control system is configured with a power supply control strategy, the power supply control strategy comprising the following steps; S1: Real-time monitoring of the health of the main power supply for the vehicle power system, sending idle computing power information of the computing power platform and receiving idle computing power information of other computing power platforms. If the main power supply fails, proceed to S2; S2: The redundant power conversion unit supplies power to the vehicle power system according to a preset quota ratio, and at the same time obtains a corresponding power supply switching value, and calls a corresponding computing power dormancy instruction set and a computing power call instruction set according to the power supply switching value. The computing power dormancy instruction set includes a plurality of computing power dormancy instructions, and the computing power dormancy instructions correspond to computing power module settings and are used to shut down the corresponding computing power modules. The computing power call instruction set includes a plurality of computing power call instructions, and the computing power call instructions are used to send a computing power loan request through the communication exchange unit; S3: Frame skipping selects video data collected by the on-board camera and sends it to the processing unit for processing, thereby generating primary vehicle control information. The computing power scheduling and allocation model predicts the computing power demand for the unselected video data and selects other computing power platforms that are compatible with the idle computing power information based on the predicted computing power demand. S4: The video data is segmented and each segment is independently encapsulated and transmitted through the real-time streaming protocol; S5: Receive the secondary vehicle control information sent back by other computing power platforms, fuse and calculate the primary vehicle control information and the secondary vehicle control information to form the ultimate vehicle control information; The redundant power supply control system is configured with a computing power control strategy, which includes the following steps: C1: After receiving a loan request from another computing power platform, the borrowable computing power value of the computing power platform is checked. If the borrowable computing power value is greater than the borrowed computing power value, C2 is carried out; otherwise, the loan request is rejected. C2: Calls the cache cleanup instruction set and the computing power allocation instruction set based on the borrowed computing power value; The cache cleaning instruction set includes a number of cache cleaning instructions, which are set in correspondence with the data storage module and are used to pre-clean out corresponding storage space for the packaged data to be received; The computing power allocation instruction set includes several computing power allocation instructions, corresponding to the computing power module settings, which are used to provide the corresponding computing power module for the received packaged data; C3: The processing unit processes the encapsulated data and generates secondary vehicle control information, and transmits the secondary vehicle control information back to the computing power platform that sent the secondment request.

2. The redundant power supply control system of the intelligent vehicle-mounted computing platform according to claim 1 is characterized in that: The redundant power control system is also configured with a vehicle control strategy; W1: Real-time monitoring of whether there is power switching. If power switching occurs, proceed to W2; otherwise, maintain the current driving state; W2: Obtain the corresponding power supply switching value, and call the corresponding deceleration instruction set and the primary vehicle control instruction set according to the power supply switching value. The deceleration instruction set corresponds to the vehicle power system setting, including a number of deceleration instructions for decelerating the vehicle power system to a preset value; The primary vehicle control instruction set corresponds to the non-sleep computing module settings, including several primary vehicle control instructions, which are used to enable the vehicle power system to execute the primary vehicle control information. The primary vehicle control instruction set takes precedence over the deceleration instruction set; W3: Retrieve the ultimate vehicle control instruction set, which includes several ultimate vehicle control instructions used to enable the vehicle power system to execute ultimate vehicle control information.

3. The redundant power supply control system of the intelligent vehicle-mounted computing platform according to claim 1, characterized in that: The computing power scheduling and allocation model includes a computing power assessment module and a borrowable computing power calculation module. The computing power assessment module is used to predict the computing power demand for unselected video data, and the borrowable computing power calculation module is used to calculate the idle computing power available from other on-board computing power platforms. The evaluation formula of the computing power evaluation module is: D i =αR i +βF i +γC i R i is the resolution of the i-th video clip, F i is the frame rate of the i-th video clip, C i is the coding complexity of the i-th video segment; α, β, and γ are preset weight coefficients; The video data collected by the vehicle-mounted camera includes, in order of encoding complexity, MPEG-2, H.263, AVC, HEVC, and AV1; As the coding complexity increases, C i The value of and the corresponding preset weight coefficient also increase synchronously.

4. The redundant power supply control system of the intelligent vehicle-mounted computing platform according to claim 3 is characterized in that: The calculation formula of the borrowable computing power calculation module is: S=AB·H B=(1-H)·A S is the idle computing power, H is the health index of the main power supply, and the value range is 0-1. A is the total amount of currently available computing power. B is a linear function of H, which represents the impact of the health of the main power supply on the idle computing power. When H is 0, it means that the main power supply is completely faulty, and B = A. When H is 1, it means that the main power supply is completely normal, and B = 0.

5. The redundant power supply control system of the intelligent vehicle-mounted computing platform according to claim 4 is characterized in that: The computing power scheduling and allocation model also includes a stability analysis module, which is used to calculate the stability score of the computing power platform. The computing power scheduling and allocation model selects computing power platforms with stability scores within a threshold from the computing power platforms that meet the predicted computing power requirements; The calculation formula of the stability analysis module is: G=R+X i ·P i G is the stability score, R is the scoring constant, X i is the signal strength score of the i-th vehicle computing platform, P i is the response time score of the i-th vehicle computing platform.

6. The redundant power supply control system of the intelligent vehicle-mounted computing platform according to claim 5, characterized in that: The navigation data is exchanged with several computing power platforms to be selected, the spatial change position of the vehicle is determined based on the navigation data, and the computing power platform whose moving distance within the preset time does not exceed the preset stable distance is selected.

7. The redundant power supply control system of the intelligent vehicle-mounted computing platform according to claim 5, characterized in that: The system exchanges driving information with several computing power platforms to be selected, determines the same direction of the two vehicles based on the direction of the vehicle heads, and selects the computing power platform with the same direction of the vehicles.

8. The redundant power supply control system of the intelligent vehicle-mounted computing platform according to claim 6 or 7, characterized in that: Vehicle information is exchanged with several selected computing power platforms. Among the selected computing power platforms, the computing power platform with the same video data encoding type as the vehicle-mounted camera is selected as the highest priority.

9. The redundant power supply control system of the intelligent vehicle-mounted computing platform according to claim 5, characterized in that: Repeatedly calculate the stability score of the computing power platform according to the preset time.

10. The redundant power supply control system of the intelligent vehicle-mounted computing platform according to claim 1, characterized in that: The calculation formula for the ultimate vehicle control information is: U=f(W P ·P+W Z ·Z,diff(P,Z)) diff(P,Z)=|PZ| P represents the priority of primary vehicle control information, Z represents the priority of secondary vehicle control information, WP and WZ represent the preset weight factors of primary and secondary information respectively, U represents the ultimate vehicle control information, and diff(P,Z) is the correlation function used to calculate the degree of difference between primary and secondary information.

Images