A load optimization method and load optimization system based on dynamic classification rendering

Abstract

This application provides a load optimization method and system based on dynamic classification rendering, including: acquiring priority information of each layer in the HMI display list; assigning initial rendering attributes and preset rendering attributes to each layer according to the priority information; rendering each layer according to the initial rendering attributes; monitoring the current load status information of the intelligent cockpit system in real time; determining whether the intelligent cockpit system is overloaded based on the load status information; if the intelligent cockpit system is overloaded, rendering each layer in the HMI display list according to the preset rendering attributes. The technical problem solved by this application is that, in the prior art, although the vehicle system has a protection mechanism based on CPU / GPU load thresholds, it usually only takes passive measures such as frequency reduction or restart after the load exceeds the limit, lacking dynamic optimization of the rendering process.

Description

Technical Field

[0001] This invention relates to the field of intelligent cockpit human-computer interaction technology, and more specifically, to a load optimization method and load optimization system based on dynamic classification rendering. Background Technology

[0002] Current intelligent cockpit human-machine interface (HMI) display systems typically employ a static layered rendering architecture. Each functional module (such as navigation and media) creates its own display layer, which is then uniformly composited and refreshed by a rendering engine (such as Weston) at a fixed frequency.

[0003] However, there is at least one problem with the relevant technologies: In the existing technologies, although the vehicle system has a protection mechanism based on CPU / GPU load thresholds, it usually only takes passive measures such as frequency reduction or restart after the load exceeds the limit, and lacks dynamic optimization of the rendering process. Summary of the Invention

[0004] The technical problem solved by this invention is that, in the prior art, although the vehicle system has a protection mechanism based on CPU / GPU load thresholds, it usually only takes passive measures such as frequency reduction or restart after the load exceeds the limit, and lacks dynamic optimization of the rendering process.

[0005] To address the aforementioned issues, this invention provides a load optimization method based on dynamic classification rendering. This load optimization method is applied to an intelligent cockpit system in a vehicle. The method includes: acquiring priority information of each layer in the HMI display list; assigning initial rendering attributes and preset rendering attributes to each layer based on the priority information; rendering each layer according to the initial rendering attributes; monitoring the current load status information of the intelligent cockpit system in real time; determining whether the intelligent cockpit system is overloaded based on the load status information; if the intelligent cockpit system is overloaded, rendering each layer in the HMI display list according to the preset rendering attributes; monitoring the current load status information of the intelligent cockpit system and determining whether the intelligent cockpit system has returned to normal load; if it has returned to normal load, rendering each layer in the HMI display list according to the initial rendering attributes.

[0006] Compared with existing technologies, the technical effects achieved by this solution are as follows: through dynamic classification rendering, the layer rendering attributes are adjusted in real time according to the load status, realizing the transformation from passive protection to active adjustment of layer refresh rate, effectively reducing the impact of high load and ensuring timely recovery after normal load.

[0007] In one embodiment of the present invention, the priority information includes a first priority, a second priority, and a third priority; the priority information of each layer in the HMI display list is obtained, and initial rendering attributes and preset rendering attributes are assigned to each layer according to the priority information, including: assigning the rendering preset attribute of the layer corresponding to the first priority to perform forced refresh at a first refresh frequency; assigning the preset rendering attribute of the layer corresponding to the second priority to perform conditional refresh; and assigning the preset rendering attribute of the layer corresponding to the third priority to not refresh.

[0008] Compared with existing technologies, the technical effects achieved by adopting this technical solution are as follows: by establishing a three-level priority and configuring forced refresh, conditional refresh, and no refresh strategies, precise on-demand resource allocation is achieved, ensuring the consistency of the core experience.

[0009] In one embodiment of the present invention, the load status information includes system CPU load, GPU load, and HMI CPU load; real-time monitoring of the current load status information of the intelligent cockpit system, and determining whether the intelligent cockpit system is overloaded based on the load status information, including: determining whether the system CPU load is greater than a first load threshold, whether the GPU load is greater than a second load threshold, and whether the HMI CPU load is greater than a third load threshold; if so, the intelligent cockpit system is determined to be overloaded.

[0010] Compared with existing technologies, the technical effects achieved by adopting this technical solution are as follows: by jointly interpreting multi-dimensional load indicators (system CPU, GPU and HMI CPU), the accuracy of overload determination is improved, and strategy fluctuations caused by misjudgment of a single indicator are avoided.

[0011] In one embodiment of the present invention, configuring the preset rendering attribute of the layer corresponding to the second priority to perform conditional refresh includes: traversing the HMI display list and detecting whether there is a layer with priority information of first priority; if so, determining that the preset rendering attribute of the layer corresponding to the second priority is configured not to refresh; if not, traversing the HMI display list and detecting whether there is a layer with priority information of second priority; if there is a layer with priority information of second priority, selecting to refresh according to the importance of the layer.

[0012] Compared with existing technologies, the technical effects achieved by adopting this technical solution are as follows: By introducing layer importance filtering in conditional refresh, the existence of the first priority layer can be dynamically identified and the actions of the second priority layer can be adjusted accordingly, thereby further refining resource scheduling.

[0013] In one embodiment of the present invention, the layers include a security application layer and other application layers; selecting layers for refresh based on their importance includes: traversing layers with priority information of second priority to determine if a security application layer exists; if so, determining that the preset rendering attribute configuration of the security application layer with priority information of second priority is to perform refresh at a second refresh frequency, and assigning the preset rendering attribute configuration of the other application layers with priority information of second priority to not refresh; if not, determining that the preset rendering attribute configuration of the other application layers with priority information of second priority is to perform refresh at a second refresh frequency; wherein, the first refresh frequency is greater than the second refresh frequency.

[0014] Compared with existing technologies, the technical effects achieved by adopting this technical solution are as follows: by subdividing the second priority layer into safe and non-safe applications, it ensures that the display of safety-critical functions such as reversing cameras is prioritized under specific load conditions, thereby enhancing the driving safety of the system.

[0015] In one embodiment of the present invention, the safety application layers include a reversing camera (RVC) layer and a dashcam (DVR) layer; and / or, other application layers include a multimedia layer and a weather layer.

[0016] Compared with existing technologies, the technical effects achieved by adopting this technical solution are as follows: it enumerates specific layers, clarifies the specific scope of security applications and other applications, enhances the certainty of the scope of protection of the claims, and facilitates logical mapping and implementation in actual systems.

[0017] In one embodiment of the present invention, the intelligent cockpit system includes multiple current operating scenarios, and a single layer has different priority information in different current operating scenarios; after assigning initial rendering attributes and preset rendering attributes to each layer according to the priority information, the load optimization method further includes: detecting the current operating scenario of the intelligent cockpit system, detecting whether the current operating scenario has been switched; if so, determining that the priority information of the layer is changed to the priority information associated with the switched current operating scenario.

[0018] Compared with existing technologies, the technical effects achieved by adopting this technical solution are as follows: the introduction of a scene detection mechanism enables dynamic updating of layer priorities (such as in reversing scenarios), allowing the load optimization strategy to be flexibly switched according to the user's real-time driving status, thus giving this method a wider range of adaptability.

[0019] In one embodiment of the present invention, monitoring the current load status information of the intelligent cockpit system and determining whether the intelligent cockpit system has returned to normal load includes: determining whether the system CPU load is less than a fourth load threshold, whether the GPU load is less than a fifth load threshold, and whether the HMI CPU load is less than a sixth load threshold; if so, it is determined that the intelligent cockpit system has returned to normal load.

[0020] Compared with existing technologies, the technical effects achieved by adopting this technical solution are as follows: a clear load recovery threshold is set, and the system is ensured to smoothly switch back to the initial rendering mode through scientific interval determination, effectively preventing the refresh strategy from frequently oscillating at the critical point.

[0021] On the other hand, the present invention also provides a load optimization system based on dynamic classification rendering. The load optimization system includes: an acquisition module, used to acquire priority information of each layer in the HMI display list, and assign initial rendering attributes and preset rendering attributes to each layer according to the priority information; a monitoring module, used to monitor the current load status information of the intelligent cockpit system in real time; and a rendering control module, used to determine whether the intelligent cockpit system is overloaded according to the load status information, and when overloaded, to render each layer in the HMI display list according to the preset rendering attributes; wherein, the rendering control module is also used to render each layer according to the initial rendering attributes when the intelligent cockpit system returns to normal load.

[0022] Compared with existing technologies, the technical effects achieved by adopting this technical solution are as follows: it can achieve the technical effects corresponding to any of the above examples, which will not be elaborated here.

[0023] In one embodiment of the present invention, the rendering control module is integrated into the Weston rendering engine; the rendering control module dynamically controls the Weston rendering engine to perform or skip compositing operations for specific layers based on the layer attributes in the preset rendering attributes.

[0024] Compared with existing technologies, the technical effects achieved by adopting this technical solution are: integrating control logic into the Weston rendering engine, reducing overhead by skipping specific operations directly in the compositing layer, and having the advantages of high execution efficiency and minimal changes to the existing HMI architecture.

[0025] By adopting the technical solution of the present invention, the following technical effects can be achieved: (1) By dynamically classifying and rendering, the layer rendering attributes are adjusted in real time according to the load status, realizing the transformation from passive protection to active adjustment of layer refresh rate, effectively reducing the impact of high load and ensuring timely recovery after normal load. (2) By setting up three priority levels and configuring forced refresh, conditional refresh and no refresh strategies, precise on-demand resource allocation was achieved, ensuring the consistency of the core experience; (3) By combining multi-dimensional load indicators (system CPU, GPU and HMI CPU) for joint judgment, the accuracy of overload judgment is improved and the strategy fluctuation caused by misjudgment of a single indicator is avoided. Attached Figure Description

[0026] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings to be used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Figure 1 This is a flowchart illustrating a load optimization method based on dynamic classification rendering, provided in an embodiment of the present invention. Detailed Implementation

[0027] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0028] See Figure 1 This is a flowchart illustrating a load optimization method based on dynamic classification rendering provided by an embodiment of the present invention. The load optimization method is used in an intelligent cockpit system in a vehicle, and includes: Step S100: Obtain the priority information of each layer in the HMI display list, assign initial rendering attributes and preset rendering attributes to each layer according to the priority information, and render each layer according to the initial rendering attributes; Specifically, during system startup, rendering attributes are injected into each functional module layer in the HMI display list by loading a preset configuration file (such as a JSON or XML script). The initial rendering attributes are usually configured to high-performance mode (such as a 60Hz full refresh rate) to ensure smooth user interaction.

[0029] Step S200: Monitor the current load status information of the intelligent cockpit system in real time, and determine whether the intelligent cockpit system is overloaded based on the load status information; Step S300: If the intelligent cockpit system is overloaded, render each layer in the HMI display list according to the preset rendering attributes; At this point, the system switches from "static full-speed refresh" to "dynamic on-demand refresh". Preset rendering properties actively reduce the frequency of the rendering engine's requests to underlying hardware resources by lengthening the rendering cycle or directly freezing unnecessary layers.

[0030] Step S400: Monitor the current load status information of the intelligent cockpit system and determine whether the intelligent cockpit system has returned to normal load; Step S500: If normal load is restored, render each layer in the HMI display list according to the initial rendering attributes.

[0031] By dynamically classifying and rendering layers, the rendering properties of the layers are adjusted in real time according to the load status, realizing the transformation from passive protection to active adjustment of the layer refresh rate, effectively reducing the impact of high load and ensuring timely recovery after normal load.

[0032] Furthermore, the priority information includes first priority, second priority, and third priority; step S100 includes: Step S110: Assign the rendering preset properties of the corresponding first priority layer to the first refresh frequency and perform a forced refresh; Specifically, the first priority corresponds to Layer A (first priority layer), which covers key interactive functions such as dynamic navigation paths and ADAS alarm prompts. The first refresh frequency can be configured to perform composition once every 3 refresh cycles (e.g., 48ms) to ensure real-time feedback of key information.

[0033] Step S120: Assign the preset rendering properties of the corresponding second priority layer as conditional refresh; Specifically, the second priority corresponds to Layer B (second priority layer), such as media playback and weather information. This type of layer only participates in compositing under specific conditions, reducing unnecessary overhead.

[0034] Step S130: Assign the preset rendering properties of the corresponding third priority layer to not refresh.

[0035] Specifically, the third priority corresponds to Layer C (third priority layer), such as the static background bottom bar. Dynamic freezing is performed under high load, meaning the compositing stage is completely skipped.

[0036] By establishing a three-tier priority system and configuring forced refresh, conditional refresh, and no refresh strategies, precise on-demand resource allocation was achieved, ensuring the consistency of the core experience.

[0037] Furthermore, the load status information includes system CPU load, GPU load, and HMI CPU load; step S200 includes: Step S210: Determine whether the system CPU load is greater than the first load threshold, whether the GPU load is greater than the second load threshold, and whether the HMI CPU load is greater than the third load threshold; In this embodiment, the first and second load thresholds can be set to 90%, and the third load threshold can be set to 3500 dmips based on the chip performance. This three-dimensional judgment logic enables a more accurate prediction of load trends.

[0038] Step S220: If yes, then the intelligent cockpit system is determined to be overloaded.

[0039] By combining multi-dimensional load metrics (system CPU, GPU, and HMI CPU) for joint analysis, the accuracy of overload determination is improved, avoiding strategy fluctuations caused by misjudgment of a single metric.

[0040] Further, step S120 includes: Step S121: Traverse the HMI display list and check if there is a layer with priority information of first priority; Step S122: If yes, then determine that the preset rendering attribute configuration of the corresponding second priority layer is not refreshed; Step S123: If not, traverse the HMI display list and check if there is a layer with priority information of second priority; Step S124: If there is a layer with priority information of second priority, then refresh according to the importance of the layer.

[0041] By introducing layer importance filtering in conditional refresh, the existence of the first priority layer can be dynamically identified and the actions of the second priority layer can be adjusted accordingly, further refining resource scheduling.

[0042] Furthermore, the layers include a security application layer and other application layers; step S124 includes: Step S1241: Iterate through the layers with priority information of second priority to see if there is a security application layer; Step S1242: If yes, then determine that the preset rendering attribute configuration of the security application layer in the second priority is to be refreshed at the second refresh frequency, and the preset rendering attribute configuration of the other application layers in the second priority is not refreshed. The second refresh rate can be configured to refresh once every 10 refresh cycles (e.g., 160ms), thereby balancing display smoothness and system power consumption.

[0043] Step S1243: If not, determine that the preset rendering properties of other application layers in the second priority are configured to perform refresh at the second refresh frequency. The first refresh frequency is greater than the second refresh frequency.

[0044] The second priority layer is further divided into safe and non-safe applications to ensure that the display of safety-critical functions such as the reversing camera is prioritized under specific load conditions, thereby enhancing the driving safety of the system.

[0045] Furthermore, security application layers include a reversing camera (RVC) layer and a dashcam (DVR) layer; and / or, other application layers include a multimedia layer and a weather layer.

[0046] Listing specific layers clarifies the specific scope of security applications and other applications, enhances the certainty of the scope of protection of the claims, and facilitates logical mapping and implementation in actual systems.

[0047] Furthermore, the intelligent cockpit system includes multiple current operating scenarios, and a single layer has different priority information in different current operating scenarios; after step S100, the load optimization method also includes: Step S140: Detect the current operating scenario of the intelligent cockpit system and detect whether the current operating scenario has changed; Step S150: If yes, then determine that the priority information of the layer is changed to the priority information associated with the current running scene after the switch.

[0048] For example, when a vehicle switches from a normal driving scenario to a reversing scenario, the reversing image RVC layer is upgraded from Layer B (second priority) to Layer A (first priority); if the navigation interface switches from the foreground to the background, it is downgraded from Layer A to Layer B.

[0049] By introducing a scene detection mechanism, the layer priority is updated dynamically (e.g., in reversing scenarios), enabling the load optimization strategy to switch flexibly according to the user's real-time driving status, thus giving this method a wider range of adaptability.

[0050] Further, step S400 includes: Step S410: Determine whether the system CPU load is less than the fourth load threshold, the GPU load is less than the fifth load threshold, and the HMI CPU load is less than the sixth load threshold; Specifically, the fourth and fifth load thresholds can be set to 70%, and the sixth load threshold can be set to 2500 dmips. By establishing this "safety level" fallback strategy, a buffer is reserved.

[0051] Step S420: If yes, then determine that the intelligent cockpit system has returned to normal load.

[0052] A clear load recovery threshold was set, and scientific interval determination was used to ensure that the system smoothly switched back to the initial rendering mode, effectively preventing the refresh strategy from oscillating frequently at the critical point.

[0053] On the other hand, the present invention also provides a load optimization system based on dynamic classification rendering. The load optimization system includes: an acquisition module, used to acquire priority information of each layer in the HMI display list, and assign initial rendering attributes and preset rendering attributes to each layer according to the priority information; a monitoring module, used to monitor the current load status information of the intelligent cockpit system in real time; and a rendering control module, used to determine whether the intelligent cockpit system is overloaded according to the load status information, and when overloaded, to render each layer in the HMI display list according to the preset rendering attributes; wherein, the rendering control module is also used to render each layer according to the initial rendering attributes when the intelligent cockpit system returns to normal load.

[0054] The technical effects that can be achieved in any of the above examples will not be elaborated here.

[0055] Furthermore, the rendering control module is integrated into the Weston rendering engine; the rendering control module dynamically controls the Weston rendering engine to execute or skip compositing operations for specific layers based on the layer attributes in the preset rendering properties.

[0056] For example, the system modifies Weston's underlying compositing logic to read the preset rendering attribute flags of each layer before compositing each frame. If the determination result is "skip compositing," the engine will no longer perform pixel blending and memory copying operations for that specific layer, thereby reducing hardware overhead from a physical perspective.

[0057] By integrating control logic into the Weston rendering engine, overhead can be reduced directly at the compositing layer by skipping specific operations, which has the advantages of high execution efficiency and minimal changes to the existing HMI architecture.

[0058] While the present invention has been disclosed above, it is not limited thereto. Any person skilled in the art can make various modifications and alterations without departing from the spirit and scope of the invention; therefore, the scope of protection of the present invention should be determined by the scope defined in the claims.

Claims

1. A load optimization method based on dynamic classification rendering, the load optimization method being used in an intelligent cockpit system on a vehicle, characterized in that, The load optimization method includes: Obtain priority information for each layer in the HMI display list, assign initial rendering attributes and preset rendering attributes to each layer based on the priority information, and render each layer based on the initial rendering attributes; Real-time monitoring of the current load status information of the intelligent cockpit system, and determination of whether the intelligent cockpit system is overloaded based on the load status information; If the intelligent cockpit system is overloaded, then each layer in the HMI display list is rendered according to the preset rendering attributes; Monitor the current load status information of the intelligent cockpit system to determine whether the intelligent cockpit system has returned to normal load; If normal load is restored, each layer in the HMI display list is rendered according to the initial rendering attributes.

2. The load optimization method according to claim 1, characterized in that, The priority information includes a first priority, a second priority, and a third priority; The step of obtaining priority information for each layer in the HMI display list and assigning initial rendering attributes and preset rendering attributes to each layer based on the priority information includes: The rendering preset properties of the layer corresponding to the first priority are configured to the first refresh frequency to perform a forced refresh; Configure the preset rendering properties of the corresponding second priority layer to perform conditional refresh; The preset rendering properties of the layer corresponding to the third priority are configured to not refresh.

3. The load optimization method according to claim 2, characterized in that, The load status information includes system CPU load, GPU load, and HMI CPU load; The real-time monitoring of the current load status information of the intelligent cockpit system, and the determination of whether the intelligent cockpit system is overloaded based on the load status information, includes: Determine whether the system CPU load is greater than a first load threshold, whether the GPU load is greater than a second load threshold, and whether the HMI CPU load is greater than a third load threshold; If so, the intelligent cockpit system is determined to be overloaded.

4. The load optimization method according to claim 2, characterized in that, The step of configuring the preset rendering attributes of the layer corresponding to the second priority as conditional refresh includes: Traverse the HMI display list and check if there is a layer with priority information of the first priority; If so, then the preset rendering properties of the corresponding second priority layer are configured to not refresh; If not, then traverse the HMI display list to check if there is a layer with the priority information being the second priority; If a layer with priority information of the second priority exists, then refresh is performed according to the importance of that layer.

5. The load optimization method according to claim 4, characterized in that, The layers include security application layers and other application layers; The step of selecting and refreshing based on the importance of the layers includes: The priority information is iterated to determine whether a security application layer exists in the second priority layer. If so, then the preset rendering attribute configuration of the security application layer in the second priority is set to refresh at the second refresh frequency, and the preset rendering attribute configuration of the other application layers in the second priority is set to not refresh. If not, then the preset rendering attribute configuration of other application layers in the second priority is determined to be refreshed at the second refresh frequency. The first refresh frequency is greater than the second refresh frequency.

6. The load optimization method according to claim 4, characterized in that, The security application layers include a reversing camera (RVC) layer and a dashcam (DVR) layer; And / or, the other application layers include multimedia layers and weather layers.

7. The load optimization method according to claim 2, characterized in that, The intelligent cockpit system includes multiple current operating scenarios, and each layer has different priority information in different current operating scenarios; After assigning initial rendering attributes and preset rendering attributes to each layer according to the priority information, the load optimization method further includes: Detect the current operating scenario of the intelligent cockpit system, and detect whether the current operating scenario has switched. If so, the priority information of the layer is determined to be changed to the priority information associated with the current running scene after the switch.

8. The load optimization method according to claim 3, characterized in that, The monitoring of the current load status information of the intelligent cockpit system and the determination of whether the intelligent cockpit system has returned to normal load include: Determine whether the system CPU load is less than the fourth load threshold, whether the GPU load is less than the fifth load threshold, and whether the HMI CPU load is less than the sixth load threshold; If so, the intelligent cockpit system is determined to have returned to normal load.

9. A load optimization system based on dynamic classification rendering, characterized in that, The load optimization system includes: The acquisition module is used to acquire the priority information of each layer in the HMI display list, and assign initial rendering attributes and preset rendering attributes to each layer according to the priority information; The monitoring module is used to monitor the current load status information of the intelligent cockpit system in real time; The rendering control module is used to determine whether the intelligent cockpit system is overloaded based on the load status information, and when overloaded, to render each layer in the HMI display list according to the preset rendering attributes. The rendering control module is also used to render each layer according to the initial rendering attributes when the intelligent cockpit system resumes normal load.

10. The load optimization system according to claim 9, characterized in that, The rendering control module is integrated into the Weston rendering engine; The rendering control module dynamically controls the Weston rendering engine to perform or skip compositing operations for a specific layer based on the layer attributes in the preset rendering attributes.

Images