Performance evaluation, performance optimization method, device, electronic device and storage medium

Abstract

The present disclosure provides a performance evaluation, performance optimization method, device, electronic device, and storage medium, which relate to the field of image processing, and particularly to the technical fields of autonomous driving, high-precision maps, intelligent transportation, and computer image processing. The specific implementation solution is as follows: determining the transmission bandwidth information of the autonomous driving system; determining the latency information of the autonomous driving system; determining the power consumption information of the autonomous driving system; and determining the performance evaluation information of the autonomous driving system according to the transmission bandwidth information, latency information, and power consumption information.

Description

Technical Field

[0001] The present disclosure relates to the field of image processing, and in particular, to the fields of autonomous driving, high-precision maps, intelligent transportation, and computer image processing technologies. Specifically, it relates to a performance evaluation, performance optimization method, device, electronic device, and storage medium. Background Art

[0002] With the continuous development of autonomous driving technology, the demand for the data processing ability of autonomous driving systems is getting higher and higher. For example, for an autonomous driving system, it is necessary to evaluate the performance of the hardware system of the autonomous driving system. Summary of the Invention

[0003] The present disclosure provides a performance evaluation, performance optimization method, device, electronic device, and storage medium.

[0004] According to one aspect of the present disclosure, there is provided a performance evaluation method for an autonomous driving system, including: determining the transmission bandwidth information of the autonomous driving system, where the transmission bandwidth information includes at least one of the following: sensor input bandwidth, processor storage bandwidth, and processor interconnection bandwidth, the processor storage bandwidth represents the bandwidth between the processor and the memory, and the processor interconnection bandwidth represents the interconnection bandwidth between multiple processors; determining the latency information of the autonomous driving system, where the latency information includes at least one of the following: input latency, processing latency, and output latency, the input latency represents the data path input latency, the processing latency represents the latency of the processor executing a processing operation, and the output latency represents the latency between the processor and the actuator; determining the power consumption information of the autonomous driving system, where the power consumption information includes at least one of the following: non-dynamic load power consumption and dynamic load power consumption, the non-dynamic load power consumption represents the static no-load power consumption; and, determining the performance evaluation information of the autonomous driving system according to the transmission bandwidth information, the latency information, and the power consumption information.

[0005] According to another aspect of the present disclosure, there is provided a performance optimization method for an autonomous driving system, including: using the performance evaluation information of the autonomous driving system determined according to the performance evaluation method of the autonomous driving system; determining a performance optimization strategy according to the performance evaluation information; and, optimizing the autonomous driving system according to the performance optimization strategy.

[0006] According to another aspect of the present disclosure, there is provided a performance evaluation device for an autonomous driving system, including: a transmission bandwidth determination module configured to determine the transmission bandwidth information of the autonomous driving system, where the transmission bandwidth information includes at least one of the following: sensor input bandwidth, processor storage bandwidth, and processor interconnection bandwidth, the processor storage bandwidth represents the bandwidth between the processor and the memory, and the processor interconnection bandwidth represents the interconnection bandwidth between multiple processors; a latency determination module configured to determine the latency information of the autonomous driving system, where the latency information includes at least one of the following: input latency, processing latency, and output latency, the input latency represents the input latency of the data path, the processing latency represents the latency of the processor to execute a processing operation, and the output latency represents the latency between the processor and the actuator; a power consumption determination module configured to determine the power consumption information of the autonomous driving system, where the power consumption information includes at least one of the following: non-dynamic load power consumption and dynamic load power consumption, the non-dynamic load power consumption represents the static no-load power consumption; and a performance evaluation module configured to determine the performance evaluation information of the autonomous driving system based on the transmission bandwidth information, the latency information, and the power consumption information.

[0007] According to another aspect of the present disclosure, there is provided a performance optimization device for an autonomous driving system, including: a performance evaluation information determination module configured to determine the performance evaluation information of the autonomous driving system by using the performance evaluation device for the autonomous driving system; a performance optimization strategy determination module configured to determine a performance optimization strategy based on the performance evaluation information; and an optimization module configured to optimize the autonomous driving system according to the performance optimization strategy.

[0008] According to another aspect of the present disclosure, there is provided an electronic device, including: at least one processor; and a memory communicatively connected to the at least one processor; where the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method as described above in the present disclosure.

[0009] According to another aspect of the present disclosure, there is provided a non-transitory computer-readable storage medium storing computer instructions, where the computer instructions are used to cause the computer to execute the method as described above in the present disclosure.

[0010] According to another aspect of the present disclosure, there is provided a computer program product including a computer program, where the computer program, when executed by a processor, implements the method as described above in the present disclosure.

[0011] According to another aspect of the present disclosure, there is provided an autonomous driving vehicle including the electronic device as described above.

[0012] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it used to limit the scope of the present disclosure. Other features of the present disclosure will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] The drawings are used to better understand the solution and do not constitute a limitation to the present disclosure. Among them:

[0014] Figure 1 Schematically shows an exemplary system architecture to which a performance evaluation method, a performance optimization method, and a device of an autonomous driving system according to an embodiment of the present disclosure can be applied;

[0015] Figure 2 Schematically shows a flowchart of a performance evaluation method of an autonomous driving system according to an embodiment of the present disclosure;

[0016] Figure 3 Schematically shows a flowchart of determining performance evaluation information of an autonomous driving system according to transmission bandwidth information, delay information, and power consumption information according to an embodiment of the present disclosure;

[0017] Figure 4 Schematically shows a schematic diagram of the principle of a performance evaluation method of an autonomous driving system according to an embodiment of the present disclosure;

[0018] Figure 5A Schematically shows a flowchart of determining the non-dynamic load power consumption of an autonomous driving system according to an embodiment of the present disclosure;

[0019] Figure 5B Schematically shows a schematic diagram of the principle of determining the non-dynamic load power consumption of an autonomous driving system according to an embodiment of the present disclosure;

[0020] Figure 5C Schematically shows a schematic diagram of the principle of determining the visual sensor input bandwidth of an autonomous driving system according to an embodiment of the present disclosure;

[0021] Figure 6 Schematically shows a block diagram of a hardware system of an autonomous driving system according to an embodiment of the present disclosure;

[0022] Figure 7 Schematically shows a flowchart of a performance optimization method of an autonomous driving system according to an embodiment of the present disclosure.

[0023] Figure 8 Schematically shows a block diagram of a performance evaluation device of an autonomous driving system according to an embodiment of the present disclosure;

[0024] Figure 9 Schematically shows a block diagram of a performance optimization device of an autonomous driving system according to an embodiment of the present disclosure; and

[0025] Figure 10 A block diagram of an electronic device suitable for implementing a performance evaluation method and a performance optimization method of an autonomous driving system according to an embodiment of the present disclosure is schematically shown. Detailed implementation manners

[0026] The following describes exemplary embodiments of the present disclosure with reference to the accompanying drawings. Various details of the embodiments of the present disclosure are included to assist understanding, and they should be considered merely exemplary. Therefore, those of ordinary skill in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, descriptions of well-known functions and structures are omitted below for clarity and conciseness.

[0027] The hardware system of the autonomous driving system is a hardware platform for sensing the surrounding environment information of the autonomous driving vehicle and responding to the surrounding environment. For example, various sensors of the hardware system are used to collect and interact with the surrounding environment information to implement operations such as sensing, positioning, and predicting various obstacles.

[0028] During the performance evaluation process of the autonomous driving system, due to the complex processing link of the autonomous driving system, for example, involving multiple processors, and the inaccuracy of existing performance data, for example, involving theoretical performance data provided by suppliers or test performance data obtained based on existing application instances, it is difficult to accurately evaluate or optimize the performance of the autonomous driving system.

[0029] To this end, an embodiment of the present disclosure proposes a performance evaluation scheme for an autonomous driving system. For example, determining the transmission bandwidth information of the autonomous driving system. The transmission bandwidth information includes at least one of the following: sensor input bandwidth, processor storage bandwidth, and processor interconnection bandwidth. The processor storage bandwidth represents the bandwidth between the processor and the memory, and the processor interconnection bandwidth represents the interconnection bandwidth between multiple processors. Determining the latency information of the autonomous driving system. The latency information includes at least one of the following: input latency, processing latency, and output latency. The input latency represents the input latency of the data path, the processing latency represents the latency of the processor executing the processing operation, and the output latency represents the latency between the processor and the actuator. Determining the power consumption information of the autonomous driving system. The power consumption information includes at least one of the following: non-dynamic load power consumption and dynamic load power consumption. The non-dynamic load power consumption represents the static no-load power consumption. Determining the performance evaluation information of the autonomous driving system according to the transmission bandwidth information, latency information, and power consumption information.

[0030] According to an embodiment of the present disclosure, since the performance evaluation information of the autonomous driving system is determined based on the transmission bandwidth information, delay information, and power consumption information of the autonomous driving system, and the transmission bandwidth information, delay information, and power consumption information can comprehensively cover the hardware requirements of modules such as the operating system, algorithms, and applications carried by the hardware system of the autonomous driving system, therefore, the embodiments of the present disclosure can improve the accuracy of the performance evaluation information, thereby facilitating the rational utilization of system resources and reducing overhead.

[0031] Figure 1 Schematically shows an exemplary system architecture to which the performance evaluation method, performance optimization method, and device of the autonomous driving system according to the embodiments of the present disclosure can be applied.

[0032] It should be noted that, Figure 1 The illustration is only an example of the system architecture to which the embodiments of the present disclosure can be applied, to help those skilled in the art understand the technical content of the present disclosure, but it does not mean that the embodiments of the present disclosure cannot be used in other devices, systems, environments, or scenarios.

[0033] As Figure 1 shown, the system architecture 100 according to this embodiment may include a vehicle 101, a radar sensor 102, a vision sensor 103, a network 104, and an electronic device 105. The network 104 is used to provide a medium for the communication link between the vehicle 101 and the electronic device 104. The network 104 may include various connection types. For example, at least one of wired and wireless communication links, etc.

[0034] The vehicle 101 may include an automobile that uses an unconventional vehicle fuel as a power source (or uses a conventional vehicle fuel and adopts a new vehicle power device), integrating advanced technologies in the aspects of vehicle power control and drive, forming an automobile with advanced technical principles, new technologies, and new structures. For example, pure electric vehicles, range-extended electric vehicles, hybrid vehicles, fuel cell electric vehicles, hydrogen engine vehicles, etc.

[0035] For example, the vehicle 101 may be a vehicle configured with an autonomous driving system. The vehicle 101 may be an autonomous driving vehicle. The vehicle 101 may be equipped with a collection device for collecting surrounding environment information.

[0036] The radar sensor 102 may be used to collect point cloud data of the site for map building. The radar sensor 102 may include a laser scanner, at least one laser source, and at least one detector. The radar sensor 102 may be configured on the vehicle 101.

[0037] The vision sensor 103 is used to obtain images of the vehicle 101 during the autonomous driving process. The vision sensor 103 may be installed on the vehicle 101. For example, it may be installed on the outer top of the vehicle 101, or installed inside the vehicle 101.

[0038] The vision sensor 103 can be used to acquire images of objects during the driving of the vehicle 101. The vision sensor 103 can be various models of cameras.

[0039] In addition, the vision sensor 103 can be integrated with the vehicle 101 or can be discrete from the vehicle 101, which is not limited herein.

[0040] The electronic device 105 can include at least one of a terminal device and a server. The electronic device can include a Central Processing Unit (CPU) and a Graphics Processing Unit (GPU). The CPU and the GPU can be communicatively connected through a PCIE (Peripheral Component Interconnect Express, a high-speed serial computer expansion bus standard) transmission link. The CPU and the vision sensor 102 can be communicatively connected through a PCIE transmission link. The terminal device can be various electronic devices having a display screen and supporting web browsing. For example, the terminal device can include at least one of a smart phone, a tablet computer, a laptop computer, a desktop computer, etc.

[0041] The server can be a server providing various services. For example, the server can be a cloud server, also known as a cloud computing server or a cloud host, which is a host product in the cloud computing service system, and solves the defects of difficult management and weak business scalability existing in traditional physical hosts and VPS services (Virtual Private Server).

[0042] For example, the electronic device 105 can be used to execute the performance evaluation method of the automatic driving system according to the embodiments of the present disclosure. The electronic device 105 can receive the transmission bandwidth information, latency information, and power consumption information of the automatic driving system sent by the vehicle 101. The electronic device 105 can determine the performance evaluation information of the automatic driving system of the vehicle 101 according to the acquired transmission bandwidth information, latency information, and power consumption information.

[0043] For example, the electronic device 105 can also be used to execute the performance optimization method of the automatic driving system according to the embodiments of the present disclosure. The electronic device 105 can acquire the performance evaluation information of the automatic driving system, determine the performance optimization strategy of the automatic driving system, and optimize the automatic driving system according to the above performance optimization strategy.

[0044] For example, the electronic device 105 may also receive point cloud data from the radar sensor 102 and receive autonomous driving images from the vision sensor 103, so that the electronic device 105 can execute the performance evaluation method of the above-mentioned autonomous driving system and the performance optimization method of the above-mentioned autonomous driving system.

[0045] It should be noted that the performance evaluation method of the autonomous driving system and the performance optimization method of the autonomous driving system provided by the embodiments of the present disclosure may also be executed by the vehicle 101. Correspondingly, the performance evaluation device of the autonomous driving system and the performance optimization device of the autonomous driving system provided by the embodiments of the present disclosure may also be provided in the vehicle 101.

[0046] It should be understood that Figure 1 the numbers of vehicles, radar sensors, vision sensors, networks, and servers in

[0047] are merely illustrative. According to the implementation requirements, there may be any number of vehicles, vision sensors, networks, and servers.

[0048] Figure 2 Schematically shows a flowchart of a performance evaluation method of an autonomous driving system according to an embodiment of the present disclosure.

[0049] As Figure 2 shown, the method 200 includes operations S210 to S240.

[0050] In operation S210, determine the transmission bandwidth information of the autonomous driving system.

[0051] In operation S220, determine the latency information of the autonomous driving system.

[0052] In operation S230, determine the power consumption information of the autonomous driving system.

[0053] In operation S240, determine the performance evaluation information of the autonomous driving system according to the transmission bandwidth information, latency information, and power consumption information.

[0054] According to an embodiment of the present disclosure, the autonomous driving system may include multiple types of sensors, multiple memories, multiple processors, and other multiple hardware devices. During autonomous driving, the autonomous driving system may perform data interaction by controlling multiple hardware devices, thereby realizing autonomous driving data processing and controlling the safe operation of the autonomous driving vehicle.

[0055] According to an embodiment of the present disclosure, performance indicators such as transmission bandwidth, latency, and power consumption will affect the controllability of the autonomous driving system over the autonomous driving vehicle.

[0056] According to an embodiment of the present disclosure, the transmission bandwidth information of the autonomous driving system may include at least one of the following: sensor input bandwidth, processor storage bandwidth, and processor interconnection bandwidth. The processor storage bandwidth may characterize the bandwidth between the processor and the memory. The processor interconnection bandwidth may characterize the interconnection bandwidth between multiple processors.

[0057] According to an embodiment of the present disclosure, the latency information of the autonomous driving system may include at least one of the following: input latency, processing latency, and output latency. The input latency may characterize the input latency of the data path. The processing latency may characterize the latency of the processor executing a processing operation. The output latency may characterize the latency between the processor and the actuator.

[0058] According to an embodiment of the present disclosure, the sensor may be used to collect environmental data around the autonomous driving vehicle and input the collected environmental data into the autonomous driving system. The sensor may include a vision sensor for collecting image information, for example, a camera. Alternatively, the sensor may include a radar for collecting information such as the position and speed of obstacles during autonomous driving. For example, a lidar, a millimeter-wave radar, etc. Alternatively, the sensor may include an infrared sensor for collecting various types of information. Alternatively, the transmitter may include an inertial measurement unit for measuring the position information of the autonomous driving vehicle itself.

[0059] According to an embodiment of the present disclosure, the sensor input bandwidth may characterize the bandwidth for the sensor to input the collected data into the autonomous driving system. For example, the sensor may transmit various types of information collected through multiple data paths to the processor or the memory. For example, the vision sensor may transmit the collected image data through at least one data path. The vision sensor and the lidar may also transmit the collected image data or point cloud data through at least one data path.

[0060] According to an embodiment of the present disclosure, the input latency may characterize the input latency of the data path. Since there may be multiple data paths in the autonomous driving system, different input latencies may exist for the multiple data paths during the process of transmitting the environmental data collected by the sensor through the data path to the processor or the memory.

[0061] For example, data path A is used to transmit image data of 200M pixels at a transmission speed of 1 frame per second, data path B is used to transmit image data of 300M pixels at a transmission speed of 2 frames per second, and data path C is used to transmit point cloud data. At the same moment, the data of the data paths undergoing data transmission is different, which may result in different input latencies of the data paths.

[0062] According to embodiments of the present disclosure, the processor storage bandwidth may characterize the bandwidth between the processor and the memory. The processor storage bandwidth may limit the amount of data or the data transfer speed for data interaction between the processor and the memory.

[0063] According to embodiments of the present disclosure, an autonomous driving system may process data input by sensors through a processor and transmit the processing result to the memory for storing the processing result. Alternatively, the processor may also obtain parameter data to be processed from the memory for subsequent processing of the parameter data.

[0064] According to embodiments of the present disclosure, the processor interconnection bandwidth may characterize the interconnection bandwidth between multiple processors. During autonomous driving, the autonomous driving system may control the operation of an autonomous vehicle through one or more processors. For example, functions such as obstacle positioning and prediction.

[0065] According to embodiments of the present disclosure, during the process of controlling the normal operation of an autonomous vehicle through multiple processors, one or more data transmissions may occur between the multiple processors. The processor interconnection bandwidth will affect the data transmission between the multiple processors, thereby affecting the performance of the autonomous driving system.

[0066] For example, for the obstacle positioning function, the autonomous driving system may process multiple steps of positioning an obstacle through multiple processors. During the above process, the processor interconnection bandwidth between the multiple processors will affect the processing performance of the positioning function.

[0067] According to embodiments of the present disclosure, the processing latency may characterize the latency of the processor executing a processing operation. When the autonomous driving system processes the data collected during autonomous driving through one or more processors, due to factors such as the chip processing ability and data reading and writing ability of the processor, different processing latencies will be generated by the processor. The processing latency of the processor will also affect the performance of the autonomous driving system.

[0068] According to embodiments of the present disclosure, the output latency may characterize the latency between the processor and the actuator. After the data processing is completed by the processor, the processing result needs to be transmitted to the actuator so that the actuator can execute the processing result to implement the corresponding autonomous driving function. The processing result may be a processing instruction generated by the processor, or a processing parameter or an execution result.

[0069] For example, the processor may transmit the processing result to one or more actuators through one or more instruction execution paths. For example, a graphics processor may transmit an image processing result to actuator 1 through instruction execution path 1. A central processing unit may transmit a driving processing result to actuator 2 through instruction execution path 2.

[0070] According to an embodiment of the present disclosure, during the process of the processor transmitting the processing result to the actuator, the output delay caused by the processor outputting the processing result will also affect the performance of the autonomous driving system.

[0071] According to an embodiment of the present disclosure, autonomous driving vehicles in different states correspond to different load states of the autonomous driving system. For example, when the autonomous driving system only controls the start of the autonomous driving vehicle and does not process the collected data such as image data and point cloud data, the load state of the autonomous driving system is static no-load, and the power consumption information generated by the autonomous driving system includes non-dynamic load power consumption.

[0072] According to an embodiment of the present disclosure, when the autonomous driving system controls the normal operation of the autonomous driving vehicle, the load state of the autonomous driving system is dynamic load, and the power consumption information generated by the autonomous driving system includes dynamic load power consumption. Alternatively, when the autonomous driving system controls the normal operation of the autonomous driving vehicle and the autonomous driving system processes the collected data such as image data and point cloud data, the load state of the autonomous driving system is dynamic load, and the power consumption information generated by the autonomous driving system includes dynamic load power consumption. Different dynamic loads correspond to different dynamic load power consumptions.

[0073] According to an embodiment of the present disclosure, the autonomous driving system can determine the transmission bandwidth information based on the chip parameter information provided by the supplier or standard data such as the interface protocol used in the data transmission process.

[0074] According to an embodiment of the present disclosure, the autonomous driving system can determine the delay information of the autonomous driving system by determining the time difference between the transmitted data and the received data.

[0075] According to an embodiment of the present disclosure, the autonomous driving system can collect power consumption metrics during the autonomous driving process, such as voltage or current, through a controller, and determine the power consumption information of the autonomous driving system based on the power consumption metrics.

[0076] According to an embodiment of the present disclosure, the controller can be fixedly installed inside the autonomous driving vehicle, such as a chip, to collect power consumption metrics during the autonomous driving process. Alternatively, the controller can also include an external detection device, which can be flexibly installed on the autonomous driving vehicle when collecting the power consumption metrics of the autonomous driving vehicle.

[0077] According to an embodiment of the present disclosure, after determining the transmission bandwidth information, delay information, and power consumption information of the autonomous driving system, the performance evaluation information of the autonomous driving system can be jointly determined by integrating the transmission bandwidth information, delay information, and power consumption information.

[0078] For example, according to information such as the balance of resources within the autonomous driving system and the core business direction, one or more metrics are selected from the transmission bandwidth information, latency information, and power consumption information, and the performance evaluation information of the autonomous driving system is comprehensively determined.

[0079] According to an embodiment of the present disclosure, one or more metrics are selected from the transmission bandwidth information, latency information, and power consumption information according to the degree of influence of the transmission bandwidth information, latency information, and power consumption information on performance, and the performance evaluation information of the autonomous driving system is comprehensively determined. Alternatively, according to the degree of influence of the transmission bandwidth information, latency information, and power consumption information on performance, according to a predetermined ratio or predetermined weight, the performance evaluation information is determined according to the transmission bandwidth information, latency information, and power consumption information.

[0080] The above are only exemplary embodiments, but are not limited thereto. It may also include methods for determining performance metrics known in the art, as long as the performance evaluation method can be determined by comprehensively considering multiple performance metrics.

[0081] According to an embodiment of the present disclosure, since the performance evaluation information of the autonomous driving system is determined according to the transmission bandwidth information, latency information, and power consumption information of the autonomous driving system, and the transmission bandwidth information, latency information, and power consumption information can comprehensively cover the hardware requirements of modules such as the operating system, algorithms, and applications carried by the hardware system of the autonomous driving system, therefore, the embodiments of the present disclosure can improve the accuracy of the performance evaluation information, and thus are beneficial to the reasonable utilization of system resources and the reduction of overhead.

[0082] The following refers to Figure 3 、 Figure 4 、 Figure 5A 、 Figure 5B 、 Figure 5C and Figure 6 , and further illustrate the performance evaluation method of the autonomous driving system according to the embodiments of the present disclosure in combination with specific embodiments.

[0083] Figure 3 Schematically shows a flowchart for determining the performance evaluation information of the autonomous driving system according to the transmission bandwidth information, latency information, and power consumption information according to an embodiment of the present disclosure.

[0084] As Figure 3 shown, the method 300 further defines the operation S240 in Figure 2 , and the method 300 includes operations S341 to S342.

[0085] In operation S341, a weighted statistical value is determined according to the transmission bandwidth information, latency information, and power consumption information.

[0086] In operation S342, a performance evaluation value of the autonomous driving system is determined according to the weighted statistical value.

[0087] According to an embodiment of the present disclosure, the weighted statistical value may include a transmission bandwidth weighted statistical value, a delay weighted statistical value, and a power consumption weighted statistical value. The transmission bandwidth weighted statistical value may be determined according to a transmission bandwidth weight and transmission bandwidth information. The delay weighted statistical value may be determined according to a delay weight and delay information. The power consumption weighted statistical value may be determined according to a power consumption weight and power consumption information.

[0088] According to an embodiment of the present disclosure, the transmission bandwidth weight may include at least one of a sensor input bandwidth weight corresponding to a sensor input bandwidth, a processor storage bandwidth weight corresponding to a processor storage bandwidth, and a processor interconnection bandwidth weight corresponding to a processor interconnection bandwidth. The transmission bandwidth weighted statistical value may be determined according to the transmission bandwidth weight and the transmission bandwidth information, and may include: the sensing bandwidth weighted statistical value may be determined according to at least one of the sensor input bandwidth weighted statistical value, the processor storage bandwidth weighted statistical value, and the processor interconnection bandwidth weighted statistical value. The sensor input bandwidth weighted statistical value may be determined according to the sensor input bandwidth weight and the sensor input bandwidth. The processor storage bandwidth weighted statistical value may be determined according to the processor storage bandwidth weight and the processor storage bandwidth. The processor interconnection bandwidth weighted statistical value may be determined according to the processor interconnection bandwidth weight and the processor interconnection bandwidth.

[0089] According to an embodiment of the present disclosure, the delay bandwidth weight may include at least one of an input delay weight corresponding to an input delay, a processing delay weight corresponding to a processing delay, and an output delay weight corresponding to an output delay. The delay weighted statistical value may be determined according to the delay weight and the delay information, and may include: the delay weighted statistical value may be determined according to at least one of the input delay weighted statistical value, the processing delay weighted statistical value, and the output delay weighted statistical value. The input delay weighted statistical value may be determined according to the input delay weight and the input delay. The processing delay weighted statistical value may be determined according to the processing delay weight and the processing delay. The output delay weighted statistical value may be determined according to the output delay weight and the output delay.

[0090] According to an embodiment of the present disclosure, the power consumption weight may include at least one of a non-dynamic load power consumption weight corresponding to a non-dynamic load power consumption and a dynamic load power consumption weight corresponding to a dynamic load power consumption. The power consumption weighted statistical value may be determined according to the power consumption weight and the power consumption information, and may include: the power consumption weighted statistical value may be determined according to at least one of the non-dynamic load power consumption weighted statistical value and the dynamic load power consumption weighted statistical value. The non-dynamic load power consumption weighted statistical value may be determined according to the non-dynamic load power consumption weight and the non-dynamic load power consumption. The dynamic load power consumption weighted statistical value may be determined according to the dynamic load power consumption weight and the dynamic load power consumption.

[0091] According to embodiments of the present disclosure, the transmission bandwidth weight, the latency weight, and the power consumption weight can be configured according to actual service requirements, which are not limited herein. For example, an artificial intelligence method can be used to determine the transmission bandwidth weight, the latency weight, and the power consumption weight.

[0092] For example, by learning the historical information of the transmission bandwidth information, the latency information, and the power consumption information in the standard database through an index model, the importance levels of the transmission bandwidth information, the latency information, and the power consumption information in the performance evaluation information are determined, so as to determine the transmission bandwidth weight, the latency weight, and the power consumption weight.

[0093] For example, for an autonomous driving vehicle, the importance levels of the latency information and the power consumption information are higher than that of the transmission bandwidth, and the importance level of the latency information is higher than that of the power consumption information. Therefore, the latency weight is higher than the power consumption weight, and the power consumption weight is higher than the transmission bandwidth weight. For example, the latency weight can be greater than 0 and less than or equal to 1, that is, the input latency weight can be greater than 0 and less than or equal to 1. The processing latency weight can be greater than 0 and less than or equal to 1. The output latency weight can be greater than 0 and less than or equal to 1. The power consumption weight can be greater than 0 and less than or equal to 1, that is, the non-dynamic load power consumption weight can be greater than 0 and less than or equal to 1. The dynamic load power consumption weight can be greater than 0 and less than or equal to 1.

[0094] According to embodiments of the present disclosure, the statistical values corresponding to the transmission bandwidth weighted statistical value, the latency weighted statistical value, and the power consumption weighted statistical value can be determined. The statistical values corresponding to the transmission bandwidth weighted statistical value, the latency weighted statistical value, and the power consumption weighted statistical value are determined as the weighted statistical value. The statistical values corresponding to the transmission bandwidth weighted statistical value, the latency weighted statistical value, and the power consumption weighted statistical value can include the average value. The average value can include the geometric average value.

[0095] According to embodiments of the present disclosure, after the weighted statistical value, the weighted statistical value is determined as the performance evaluation value of the autonomous driving system. The larger the transmission bandwidth, the higher the performance evaluation value of the autonomous driving system, which can indicate that the performance of the autonomous driving system is better. The smaller the latency, the higher the performance evaluation value of the autonomous driving system, which can indicate that the performance of the autonomous driving system is better. The smaller the power consumption, the higher the performance evaluation value of the autonomous driving system, which can indicate that the performance of the autonomous driving system is better.

[0096] According to an embodiment of the present disclosure, from the perspective of the actual application of the autonomous driving system, by determining the weight values of the transmission bandwidth information, latency information, and power consumption information respectively, and determining the weighted statistical values of the transmission bandwidth information, latency information, and power consumption information, in the actual application scenario of autonomous driving, a comprehensive coverage of the performance indicators of the autonomous driving system is achieved. Since the performance evaluation value can reflect the specific transmission bandwidth information, latency information, and power consumption information, as well as the influence degree of the transmission bandwidth information, latency information, and power consumption information on the performance evaluation, the accuracy of the performance evaluation information is improved.

[0097] Figure 4 Schematically shows a schematic diagram of the principle of a method for evaluating the performance of an autonomous driving system according to an embodiment of the present disclosure.

[0098] As Figure 4 shown, in 400, the transmission bandwidth information 401 of the autonomous driving system is determined. The transmission bandwidth information 401 may include at least one of the sensor input bandwidth 4011, the processor storage bandwidth 4012, and the processor interconnection bandwidth 4013.

[0099] The latency information 402 of the autonomous driving system is determined. The latency information 402 may include at least one of the input latency 4021, the processing latency 4022, and the output latency 4023.

[0100] The power consumption information 403 of the autonomous driving system is determined. The power consumption information 403 may include at least one of the non-dynamic load power consumption 4031 and the dynamic load power consumption 4032.

[0101] According to the transmission bandwidth information 401, the latency information 402, and the power consumption information 403, the performance evaluation information 404 is determined. In the case where a certain performance indicator does not exist in the determined transmission bandwidth information 401, latency information 402, and power consumption information 403, the performance indicator is set to 0.

[0102] Figure 5A Schematically shows a flowchart for determining the non-dynamic load power consumption of an autonomous driving system according to an embodiment of the present disclosure.

[0103] As Figure 5A shown, 500A includes operations S501 to S505, and can be used as an exemplary embodiment of operation S230.

[0104] In operation S531, does the power consumption information include non-dynamic load power consumption?; if so, perform operations S532 to S534; if not, perform operation S535.

[0105] In operation S532, determine at least one first power consumption value in the first predetermined time period.

[0106] In operation S533, a first statistical value corresponding to at least one first power consumption value is determined.

[0107] In operation S534, based on the first statistical value, the non-dynamic load power consumption of the autonomous driving system is determined.

[0108] In operation S535, non-dynamic load power consumption is output.

[0109] According to an embodiment of the present disclosure, it is determined whether the power consumption information includes non-dynamic load power consumption. In the case where it is determined that the power consumption information does not include non-dynamic load power consumption, operation S535 is executed to output non-dynamic load power consumption. For example, "non-dynamic load power consumption" is output. Alternatively, "the non-dynamic load power consumption is 0" is output. In the case where it is determined that the power consumption information includes non-dynamic load power consumption, operation S532 is executed, and operations S532 to S534 are sequentially executed to determine the non-dynamic load power consumption of the autonomous driving system.

[0110] According to an embodiment of the present disclosure, the first predetermined time period may be a certain predetermined time period when the autonomous driving system controls the start of the autonomous driving vehicle. For example, the first predetermined time period may be within 1 min - 10 min after the autonomous driving vehicle starts.

[0111] According to an embodiment of the present disclosure, during the process of determining the non-dynamic load power consumption of the autonomous driving system, at least one first power consumption value within the first predetermined time period may be randomly selected, or at least one first power consumption value within a certain sub-time period within the first predetermined time period may be randomly selected.

[0112] According to an embodiment of the present disclosure, after determining at least one first power consumption value, at least one first power consumption value may be sorted in descending or ascending order to obtain a first power consumption value sequence.

[0113] According to an embodiment of the present disclosure, the largest first power consumption value may be selected from the first power consumption value sequence, and the above largest first power consumption value may be used as the first statistical value. Alternatively, the smallest first power consumption value may be selected from the first power consumption value sequence, and the above smallest first power consumption value may be used as the first statistical value. Alternatively, the median or mode may be selected from the first power consumption value sequence, and the above median or mode may be used as the first statistical value.

[0114] According to an embodiment of the present disclosure, the average value of all first power consumption values within the first power consumption value sequence may also be determined, and this average value may be used as the first statistical value.

[0115] According to an embodiment of the present disclosure, before determining the first statistical value, at least one first power consumption value may also be screened to delete abnormal data.

[0116] According to an embodiment of the present disclosure, after determining the first statistical value, the first statistical value is determined as the non-dynamic load power consumption of the autonomous driving system.

[0117] According to an embodiment of the present disclosure, by determining at least one first power consumption value in a first predetermined period, determining a first statistical value corresponding to the at least one first power consumption value, and then determining the non-dynamic load power consumption of the autonomous driving system, the influence of abnormal data on the non-dynamic load power consumption can be excluded, the accuracy of the non-dynamic load power consumption can be improved, and the accuracy of the performance evaluation information can be improved.

[0118] According to an embodiment of the present disclosure, the first power consumption value may be determined according to at least one of a first current value and a first voltage value. The first current value and the first voltage value may be obtained by detecting the power supply of the autonomous driving system via a controller.

[0119] According to an embodiment of the present disclosure, the controller may have a voltage detection function and a current detection function. The controller determines the first power consumption value by directly collecting at least one of the first current value and the first voltage value of the power supply of the autonomous driving vehicle.

[0120] According to an embodiment of the present disclosure, the controller may output at least one of the collected first current value and the first voltage value, and the processor determines the first power consumption value according to at least one of the first current value and the first voltage value.

[0121] According to an embodiment of the present disclosure, the controller may also collect at least one of the first current value and the first voltage value, determine the first power consumption value according to at least one of the first current value and the first voltage value, and directly transmit the first power consumption value to the processor.

[0122] According to an embodiment of the present disclosure, by collecting at least one of the first voltage value and the first current value of the power supply of the autonomous driving system by the controller, an accurate first power consumption value can be determined from the power source, and the influence of other driving modules or functional processing modules can be avoided, resulting in inaccurate measurement of the first power consumption value of the autonomous driving system, thereby improving the accuracy of the non-dynamic load power consumption.

[0123] Figure 5B Schematically shows a schematic diagram of the principle of determining the non-dynamic load power consumption of the autonomous driving system according to an embodiment of the present disclosure.

[0124] As Figure 5BAs shown, in 500B, a load state set 501 may correspond to a first predetermined time period. The load state set 501 may have N load states. For example, load state 501_1,......, load state 501_n,......, load state 501_N. A first current value 501_n1 and a first voltage value 501_n2 correspond to the load state 501_n. N may be an integer greater than or equal to 1. n ∈ {1, 2,......, N - 1, N}.

[0125] The first power consumption value 502 may include M first power consumption values. For example, the first power consumption value 502_1_1 to the first power consumption value 502_1_I corresponding to the load state 501_1. The first power consumption value 502_n_J to the first power consumption value 502_n_U corresponding to the load state 501_n. The first power consumption value 502_N_V to the first power consumption value 502_N_M corresponding to the load state 501_N. M may be an integer greater than or equal to 1. N is less than or equal to M. I, J, U, and V may be integers greater than or equal to 1, and I, J, U, and V increase in sequence, and V is less than or equal to M.

[0126] At least one first voltage value and at least one first current value can be collected in each load state. For example, in the case of the load state 501_1, i first voltage values and i first current values can be collected to determine the first power consumption value 502_1_1 to the first power consumption value 502_1_I. In the case of the load state 501_n, (U - J + 1) first current values 501_n1 and (U - J + 1) first current values 501_n2 are collected to determine the first power consumption value 502_n_J to the first power consumption value 502_n_U. In the case of the load state 501_N, (M - V + 1) first voltage values and (M - V + 1) first current values can be collected to determine the first power consumption value 502_N_V to the first power consumption value 502_N_M.

[0127] After determining the M first power consumption values, the average value of the M first power consumption values can be determined as the first statistical value 503, and the first statistical value 503 can be determined as the non - dynamic load 504.

[0128] According to an embodiment of the present disclosure, in the case where the power consumption information includes dynamic load power consumption, operation S230 may include the following operations.

[0129] Determine at least one second power consumption value for a second predetermined time period. Determine a second statistical value corresponding to the at least one second power consumption value. Determine the dynamic load power consumption of the autonomous driving system according to the second statistical value.

[0130] According to an embodiment of the present disclosure, first determine whether the power consumption information includes dynamic load power consumption. When it is determined that the power consumption information does not include dynamic load power consumption, output no dynamic load power consumption. For example, output "no dynamic load power consumption". Alternatively, output "the dynamic load power consumption is 0". When it is determined that the power consumption information includes dynamic load power consumption, perform the above operation of determining the dynamic load power consumption.

[0131] As an embodiment, the second predetermined time period may be a certain predetermined time period during the normal operation of the autonomous driving system controlling the autonomous vehicle. For example, the second predetermined time period may be 15 min to 30 min after the autonomous vehicle starts.

[0132] According to an embodiment of the present disclosure, similar to the method of determining the non-dynamic load power consumption, a second statistical value corresponding to at least one second power consumption value may be determined. The dynamic load power consumption of the autonomous driving system is determined according to the second statistical value, which will not be elaborated here.

[0133] According to an embodiment of the present disclosure, by determining at least one second power consumption value of the second predetermined time period, determining a second statistical value corresponding to at least one second power consumption value, and then determining the dynamic load power consumption of the autonomous driving system, the influence of abnormal data on the dynamic load power consumption can be excluded, the accuracy of the dynamic load power consumption can be improved, so as to improve the accuracy of the performance evaluation information.

[0134] According to an embodiment of the present disclosure, the second power consumption value may be determined according to at least one of a second current value and a second voltage value. The second current value and the second voltage value may be obtained by detecting the power supply of the autonomous driving system via a controller. At least two of the at least one second power consumption value may be obtained under different load conditions.

[0135] According to an embodiment of the present disclosure, at least one of the second current value and the second voltage value of the power supply of the autonomous vehicle may be collected by a controller having a voltage detection function and a current detection function to determine the second power consumption value.

[0136] According to an embodiment of the present disclosure, the controller may output at least one of the collected second current value and second voltage value, and the processor determines the second power consumption value according to at least one of the second current value and the second voltage value.

[0137] According to an embodiment of the present disclosure, the controller may also collect at least one of the second current value and the second voltage value, determine the second power consumption value according to at least one of the second current value and the second voltage value, and directly transmit the second power consumption value to the processor.

[0138] According to an embodiment of the present disclosure, during the process of determining the dynamic load power consumption, at least two different load states are selected within a second predetermined period, and at least one second power consumption value can be determined in one load state.

[0139] For example, for each load state, at least one second power consumption value can be collected by a fine-grained multi-channel method. For example, for the load state where the autonomous driving system implements obstacle prediction, the power consumption value of this load state is divided into L + 1 power consumption levels by L predetermined power consumption values. At least one second power consumption value is collected at each power consumption level. The L predetermined power consumption values have the same granularity unit, for example, 1 watt.

[0140] According to an embodiment of the present disclosure, by collecting at least one of the second voltage value and the second current value of the power supply of the autonomous driving system through a controller, an accurate second power consumption value can be determined from the power source, avoiding inaccurate measurement of the second power consumption value of the autonomous driving system caused by the influence of other driving modules or functional processing modules, and improving the accuracy of the dynamic load power consumption. In addition, under different load conditions, obtaining the power consumption under the defined granularity unit can fully verify the system power consumption under different loads, and improve the accuracy of the dynamic load power consumption.

[0141] According to an embodiment of the present disclosure, the sensor input bandwidth may include at least one of the following: a visual sensor input bandwidth and a radar input bandwidth. The radar input bandwidth may include at least one of the following: a lidar input bandwidth and a millimeter-wave radar input bandwidth.

[0142] According to an embodiment of the present disclosure, when the transmission bandwidth information includes the sensor input bandwidth,

[0143] When the sensor input bandwidth includes the visual sensor input bandwidth, operation S210 may include the following operations.

[0144] Determine the first input bandwidth corresponding to at least one first combination information in a third predetermined period to obtain at least one first input bandwidth. According to the at least one first input bandwidth, determine the visual sensor input bandwidth of the autonomous driving system.

[0145] According to an embodiment of the present disclosure, the first combination information may include a first transmission frame rate and the resolution of video data. The at least one first combination information may be different from each other.

[0146] According to an embodiment of the present disclosure, when the sensor input bandwidth includes the radar input bandwidth, operation S210 may include the following operations.

[0147] Determine a second input bandwidth corresponding to at least one second combination information in a fourth predetermined period to obtain at least one second input bandwidth. Determine the radar input bandwidth of the autonomous driving system according to the at least one second input bandwidth.

[0148] According to an embodiment of the present disclosure, the second combination information may include a second transmission frame rate and the resolution of point cloud data. At least one second combination information may be different from each other.

[0149] According to an embodiment of the present disclosure, when the transmission bandwidth information includes the sensor input bandwidth, it may be determined whether the sensor input bandwidth includes the visual sensor input bandwidth, and it may be determined whether the sensor input bandwidth includes the radar input bandwidth.

[0150] According to an embodiment of the present disclosure, it may be first determined whether the sensor input bandwidth includes the visual sensor input bandwidth, and then it may be determined whether the sensor input bandwidth includes the radar input bandwidth. Alternatively, it may also be first determined whether the sensor input bandwidth includes the radar input bandwidth, and then it may be determined whether the sensor input bandwidth includes the visual sensor input bandwidth. Alternatively, it may also be determined whether the sensor input bandwidth includes the visual sensor input bandwidth and whether the sensor input bandwidth includes the radar input bandwidth simultaneously.

[0151] According to an embodiment of the present disclosure, in the process of determining the visual sensor input bandwidth, the third predetermined period may be a predetermined period during the autonomous driving process, or may be a predetermined period during the startup process or the stop process of the autonomous driving vehicle.

[0152] In the process of determining the radar input bandwidth, the fourth predetermined period may be a predetermined period during the autonomous driving process, or may be a predetermined period during the startup process or the stop process of the autonomous driving vehicle.

[0153] Figure 5C Schematically shows a schematic diagram of the principle of determining the visual sensor input bandwidth of an autonomous driving system according to an embodiment of the present disclosure.

[0154] As Figure 5C shown, in 500C, the first combination information 505 may include a first transmission frame rate 5051 and the resolution 5052 of video data. The first transmission frame rate 5051 may include P first transmission frame rates. For example, the first transmission frame rate 5051_1,......, the first transmission frame rate 5051_p,......, the first transmission frame rate 5051_P. P may be an integer greater than or equal to 1. The resolution 5052 of video data may include Q resolutions. For example, the resolution 5052_1,......, the resolution 5052_q,......, the resolution 5052_Q. Q may be an integer greater than or equal to 1.

[0155] According to an embodiment of the present disclosure, the first combined information 505 may include combinations of P first transmission frame rates and Q resolutions of video data, that is, it may include P*Q pieces of first combined information.

[0156] For example, the pixels of the video data that can be collected by the autonomous driving system are divided into P first transmission frame rates, and the same granularity unit exists between the P first transmission frame rates. For example, 1 frame / second, etc. The resolutions of the video data that can be collected by the vision sensor are divided into Q resolutions, and the same granularity unit exists between the Q resolutions. For example, 1 pixel, 10 pixels, or 20 pixels.

[0157] According to an embodiment of the present disclosure, for each piece of first combined information, a first input bandwidth corresponding to the first combined information can be determined, and P*Q first input bandwidths are obtained. The first input bandwidth 506 includes first input bandwidth 506_1,......, first input bandwidth 506_P*Q.

[0158] For example, the Q resolutions can be arranged in ascending order, and the smallest resolution is 200M pixels. The first first transmission frame rate is 1 frame / second, and starting from the first resolution of 200M pixels, the resolution is gradually increased according to the granularity unit of the resolution. For each piece of first combined information, the first input bandwidth is obtained through each data path until Q first input bandwidths corresponding to Q resolutions are collected. Then, the first transmission frame rate is gradually increased according to the granularity unit of the first transmission frame rate until P*Q first input bandwidths corresponding to P first transmission frame rates and Q resolutions are collected.

[0159] After determining the P*Q first input bandwidths, the vision sensor input bandwidth 507 can be determined according to the P*Q first input bandwidths. For example, the smallest first input bandwidth among the P*Q first input bandwidths can be used as the vision sensor input bandwidth of the autonomous driving system. Alternatively, the largest first input bandwidth among the P*Q first input bandwidths can also be used as the vision sensor input bandwidth of the autonomous driving system.

[0160] According to an embodiment of the present disclosure, for the radar input bandwidth, the second combined information includes a second transmission frame rate and the resolution of the point cloud data. The point cloud data can be collected by a lidar or a millimeter-wave radar.

[0161] According to an embodiment of the present disclosure, similar to the first combined information, the second combined information may include combinations of R second transmission frame rates and S resolutions of the point cloud data, that is, it may include R*S pieces of second combined information. R and S can be integers greater than or equal to 1.

[0162] For example, the point cloud data that can be collected by the autonomous driving system is divided into R second transmission frame rates, and the R second transmission frame rates have the same granularity unit, for example, 100K points per second, etc. The resolution of the point cloud data that can be collected by lidar or millimeter-wave radar is divided into S resolutions, and the S resolutions have the same granularity unit, for example, 1 pixel, 10 pixels, or 20 pixels, etc.

[0163] For example, the S resolutions can be arranged in ascending order, and the smallest resolution is 200M pixels. The first second transmission frame rate is 300K points per second, and the first resolution starts from 200M pixels, and the resolution is gradually increased according to the granularity unit of the resolution. For each second combination information, the second input bandwidth is obtained through each data path until the R second input bandwidths corresponding to the S resolutions are collected. Then, the second transmission frame rate is gradually increased according to the granularity unit of the second transmission frame rate until the R*S second input bandwidths corresponding to the R second transmission frame rates and the S resolutions are collected.

[0164] After determining the R*S second input bandwidths, the radar input bandwidth can be determined based on the R*S second input bandwidths. For example, the smallest second input bandwidth among the R*S second input bandwidths can be used as the radar input bandwidth of the autonomous driving system. Alternatively, the largest second input bandwidth among the R*S second input bandwidths can also be used as the radar input bandwidth of the autonomous driving system.

[0165] According to the embodiments of the present disclosure, by obtaining the input bandwidth under the defined granularity unit under different first combination information and second combination information, the input bandwidth under different data combinations can be fully verified, improving the accuracy of the sensor input bandwidth.

[0166] According to the embodiments of the present disclosure, when the transmission bandwidth information includes the processor storage bandwidth, operation S210 may include the following operations.

[0167] Adjust the expected processor storage bandwidth according to the storage bandwidth adjustment information to obtain the processor storage bandwidth of the autonomous driving system.

[0168] According to the embodiments of the present disclosure, the storage bandwidth adjustment information may include at least one of the following: read / write operation ratio information, read / write operation type information, and storage bandwidth allocation information of at least one functional unit of the processor in the case of a shared memory. The read / write operation type information may include at least one of the following: continuous read operation, continuous write operation, discontinuous read operation, and discontinuous write operation.

[0169] According to an embodiment of the present disclosure, the expected processor memory bandwidth is the bandwidth between the processor and the memory provided by the processor or memory chip vendor. It should be noted that since the expected processor memory bandwidth provided by the vendor is the theoretical bandwidth data obtained under specific test conditions, the processor memory bandwidth is not the expected processor memory bandwidth in any case.

[0170] According to an embodiment of the present disclosure, the expected processor memory bandwidth is adjusted according to the storage bandwidth adjustment information of the autonomous driving system. For example, in the case where the read / write operation ratio is 30% and 80%, the bandwidth between the processor and the memory is different. For example, the bandwidth with a read / write operation ratio of 80% is less than the bandwidth with a read / write operation ratio of 30%. For example, in the case of continuous read / write and discontinuous read / write, the bandwidth between the processor and the memory is different. For example, the bandwidth of continuous read operations and continuous write operations is less than the bandwidth of discontinuous read operations and discontinuous write operations. For example, in the case where there are multiple functional units inside the processor, such as the central processing unit and the image processing unit, when directly sharing the storage unit, the storage bandwidth allocation ratio of the above functional units in the processor during the actual business process is used to adjust the expected processor memory bandwidth.

[0171] According to an embodiment of the present disclosure, by using the information including the read / write operation ratio information, the read / write operation type information, and the storage bandwidth allocation information, the theoretical bandwidth data provided by the vendor can be adjusted to improve the accuracy of the processor memory bandwidth.

[0172] According to an embodiment of the present disclosure, when the transmission bandwidth information includes the processor interconnection bandwidth, operation S210 may include the following operations.

[0173] In the case where the interconnection relationship between two processors among multiple processors is a direct interconnection relationship, the maximum rate corresponding to the interconnection interface protocol is determined as the processor interconnection bandwidth corresponding to the two processors. In the case where the interconnection relationship between two processors among multiple processors is an indirect interconnection relationship, the processor interconnection bandwidth corresponding to the two processors is determined according to the interconnection configuration status.

[0174] According to an embodiment of the present disclosure, the direct interconnection relationship indicates that there is a direct data transmission interaction between two processors. Since during the data transmission process, the transmitted data needs to meet the requirements of both the data sender and the data receiver at the same time. Thus, in the case where it is determined that the interconnection relationship between two processors among multiple processors is a direct interconnection relationship, the maximum rate supported by the interconnection interface protocol provided by the vendor is determined as the processor interconnection bandwidth corresponding to the two processors.

[0175] According to an embodiment of the present disclosure, an indirect interconnection relationship indicates that there is no direct data transmission interaction between two processors, and data transmission can be indirectly achieved through other hardware devices such as a memory. For example, processor 1 stores target data in the memory, and processor 2 obtains the above target data from the memory.

[0176] According to an embodiment of the present disclosure, when there is an indirect interconnection relationship between two processors, the indirect data interaction between the two processors will be affected by the data path or the actual service type. Therefore, according to the interconnection configuration status, the processor interconnection bandwidth corresponding to two processors with an indirect interconnection relationship can be determined.

[0177] According to an embodiment of the present disclosure, by determining the interconnection relationship between two processors among multiple processors, and then determining the bandwidth for data transmission between the processors, the application scenarios for data processing among multiple processors can be improved, and the accuracy of determining the processor interconnection bandwidth can be increased.

[0178] According to an embodiment of the present disclosure, determining the processor interconnection bandwidth corresponding to two processors according to the interconnection configuration status may include the following operations.

[0179] In response to detecting that the interconnection configuration status is a static configuration status, determine the minimum interconnection bandwidth between the two processors as the processor interconnection bandwidth corresponding to the two processors.

[0180] According to an embodiment of the present disclosure, the static configuration status may indicate that there is a fixed transmission form between two processors. For example, a processor and a processor only achieve indirect interconnection through one or more fixed hardware devices. For example, processor A and processor B only achieve indirect interconnection through one or more data paths of memory C.

[0181] According to an embodiment of the present disclosure, when the interconnection relationship between two processors is an indirect interconnection relationship and the interconnection configuration status is a static configuration status, the indirect data transmission between the two processors preferentially uses the data path with the minimum bandwidth for transmission. Therefore, when it is determined that the interconnection relationship between two processors is an indirect interconnection relationship and it is detected that the interconnection configuration status is a static configuration status, determine the minimum interconnection bandwidth between the two processors as the processor interconnection bandwidth corresponding to the two processors.

[0182] According to an embodiment of the present disclosure, when it is determined that the interconnection relationship between two processors is an indirect interconnection relationship and it is detected that the interconnection configuration status is a static configuration status, by determining the minimum interconnection bandwidth between the two processors as the processor interconnection bandwidth corresponding to the two processors, in line with the actual application scenario of the autonomous driving system, the measurement operations and the amount of measurement data can be reduced.

[0183] According to an embodiment of the present disclosure, determining the processor interconnection bandwidth corresponding to two processors according to the interconnection configuration state may further include the following operations.

[0184] In response to detecting that the interconnection configuration state is a dynamic configuration state, determine at least one interconnection bandwidth between the two processors during a fifth predetermined period. Determine a third statistical value corresponding to the at least one interconnection bandwidth. According to the third statistical value, determine the processor interconnection bandwidth corresponding to the two processors.

[0185] According to an embodiment of the present disclosure, the dynamic configuration state may indicate that there is no fixed transmission form between the two processors, and the data transmission form needs to be adjusted according to the actual service situation. For example, adjust the data path.

[0186] According to an embodiment of the present disclosure, the fifth predetermined period may be a period when the autonomous driving system executes a predetermined function or a predetermined algorithm. For example, a period when executing an obstacle positioning function or an obstacle prediction function.

[0187] For example, according to the service function being processed by the autonomous driving system, determine the data path between the two processes and the corresponding processor interconnection bandwidth. For example, when the autonomous driving system is executing the obstacle positioning function, the data path between processor Y and processor Z is XXX-1, and correspondingly, the processor interconnection bandwidth corresponds to the obstacle positioning function. After executing the obstacle positioning function and when the obstacle prediction function is to be executed, the data path between processor Y and processor Z is XXX-2, and correspondingly, the processor interconnection bandwidth corresponds to the obstacle prediction function.

[0188] According to an embodiment of the present disclosure, when it is detected that the interconnection configuration state is a dynamic configuration state, determine at least one interconnection bandwidth between the two processors during the fifth predetermined period, and then the corresponding third statistical value can be obtained by determining the first predetermined percentile value of the at least one interconnection bandwidth. The first predetermined percentile can be configured according to actual service requirements and is not limited herein. For example, the first predetermined percentile may be 99. Alternatively, the maximum interconnection bandwidth can also be selected from the at least one interconnection bandwidth and determined as the third statistical value. After determining the third statistical value, the third statistical value can be directly determined as the processor interconnection bandwidth corresponding to the two processors.

[0189] According to an embodiment of the present disclosure, when it is determined that the interconnection relationship between the two processors is an indirect interconnection relationship and it is detected that the interconnection configuration state is a dynamic configuration state, by statistically analyzing at least one interconnection bandwidth during the fifth predetermined period, the application scenario covering the indirect interconnection relationship between the processors can be improved, and the accuracy of determining the processor interconnection bandwidth can be improved.

[0190] According to an embodiment of the present disclosure, when the latency information includes the input latency, operation S220 may include the following operations.

[0191] Determine at least one first time difference in a sixth predetermined time period. Determine the transceiver latency of the transceiver according to a fourth statistical value corresponding to the at least one first time difference. Determine the transmission line latency of the printed circuit board in the sixth predetermined time period. Determine the input latency of the autonomous driving system according to the transceiver latency and the transmission line latency.

[0192] According to an embodiment of the present disclosure, the first time difference may represent the time difference between a first transmission time and a first reception time. The first transmission time may represent the time when the transceiver of the data path transmits the first transmission data. The first reception time may represent the time when the transceiver receives the received data. The received data may represent the response data corresponding to the first transmission data.

[0193] According to an embodiment of the present disclosure, the sixth predetermined time period may be a time period when the autonomous driving system executes a predetermined function or a predetermined algorithm. For example, a time period for executing an obstacle positioning function or an obstacle prediction function.

[0194] According to an embodiment of the present disclosure, the autonomous driving system may collect environmental data around the autonomous driving vehicle through sensors. After the sensors collect the environmental data, the environmental data may be transmitted to the autonomous driving system through the data path. For example, the autonomous driving system may implement data transmission through the transceiver of the data path.

[0195] For example, by sending the first transmission data to the sensor or other hardware devices, the transceiver of the data path can receive the received data corresponding to the first transmission data from the sensor or other hardware devices to complete the data transmission. The transceiver may be a transceiver chip having functions of receiving data, receiving instructions, initiating data, and initiating instructions.

[0196] According to an embodiment of the present disclosure, the transceiver may determine a first time difference for the data path to transmit data according to the first transmission time of the first transmission data and the first reception time of the received data. Within the sixth predetermined time period, for the same data, the transceiver may determine at least one first time difference according to at least one first transmission time and at least one reception time corresponding to the first transmission time.

[0197] According to an embodiment of the present disclosure, after determining the at least one first time difference, a fourth statistical value corresponding to the at least one first time difference may be determined according to the average value of the at least one first time difference, and the fourth statistical value may be determined as the transceiver latency of the above transceiver. Alternatively, a fourth statistical value corresponding to the at least one first time difference may be determined according to the median of the at least one first time difference, and the fourth statistical value may be determined as the transceiver latency of the above transceiver.

[0198] For example, for the same data, determine ten first time differences for five data paths transmitting ten times. And use the average value of the ten first time differences as the transceiver delay of the transceiver.

[0199] According to an embodiment of the present disclosure, when the transceiver vendor provides the standard transceiver delay, the standard transceiver delay marked on the transceiver can also be directly used as the transceiver delay of the transceiver.

[0200] According to an embodiment of the present disclosure, the input delay may include the transceiver delay of the transceiver and may also include the transmission line delay caused by a Printed Circuit Board (PCB). Therefore, within the sixth predetermined time period, it is also necessary to determine the transmission line delay of the printed circuit board.

[0201] According to an embodiment of the present disclosure, the PCB of the autonomous driving vehicle can be simulated, and the simulated transmission line delay corresponding to the PCB wiring method can be obtained in the simulation environment, and the simulated transmission line delay can be used as the transmission line delay of the printed circuit board.

[0202] According to an embodiment of the present disclosure, the sum of the transceiver delay and the transmission line delay is determined as the input delay of the autonomous driving system.

[0203] According to an embodiment of the present disclosure, by covering the transceiver delay caused by the transceiver and the transmission line delay caused by the PCB, the accuracy of determining the input delay of the autonomous driving system is improved.

[0204] According to an embodiment of the present disclosure, when the delay information includes the processing delay, operation S220 may include the following operations.

[0205] Determine at least one second time difference in the seventh predetermined time period. Determine a fifth statistical value corresponding to the at least one second time difference. Determine the processing delay of the autonomous driving system according to the fifth statistical value.

[0206] According to an embodiment of the present disclosure, the second time difference may be the time difference between the start time and the end time of the processing operation.

[0207] According to an embodiment of the present disclosure, during the process of the processor executing the processing operation, due to factors such as the processor chip, parallel processing operations, memory, and reading frequency, the processor will generate a processing delay. The seventh predetermined time period may be a time period for the processor to process one processing operation, or may also be a time period for the processor to process multiple processing operations corresponding to a certain service function or algorithm.

[0208] According to an embodiment of the present disclosure, in the case where the processor executes the same processing operation, the same service function, or the processing operation of the same algorithm, at least one second time difference can be determined according to the time difference between the start time and the end time when the processor executes the above processing operation. Then, according to at least one second time difference, a second predetermined percentile value corresponding to the same processing operation, the same service function, or the same algorithm is determined, and the second predetermined percentile value is determined as the fifth statistical value corresponding to at least one second time difference. The second predetermined percentile can be configured according to actual service requirements and is not limited herein. For example, the second predetermined percentile can be 99. In the case of determining the fifth statistical value, the fifth statistical value can be determined as the processing delay of the autonomous driving system.

[0209] According to an embodiment of the present disclosure, the start time and the end time can be obtained from a high-precision event source provided by the autonomous driving system to ensure the calculation accuracy of the processing delay.

[0210] According to an embodiment of the present disclosure, by determining at least one second time difference and the fifth statistical value corresponding to the second time difference, and then determining the processing delay of the autonomous driving system, the accuracy of determining the processing delay of the processor can be improved.

[0211] According to an embodiment of the present disclosure, in the case where the delay information includes the output delay, operation S220 may include the following operations.

[0212] Determine at least one third time difference of the eighth predetermined time period. Determine the sixth statistical value corresponding to at least one third time difference. Determine the output delay of the autonomous driving system according to the sixth statistical value.

[0213] According to an embodiment of the present disclosure, the third time difference may represent the time difference between the second transmission time and the second reception time. The second transmission time may represent the time when the processor transmits the second transmission data. The second reception time may represent the time when the actuator receives the second transmission data.

[0214] According to an embodiment of the present disclosure, after the processor executes the processing operation, the processing result is sent to the actuator so that the actuator executes the above processing result.

[0215] According to an embodiment of the present disclosure, the eighth predetermined time period may be a time period when the autonomous driving system executes a predetermined function or a predetermined algorithm. For example, a time period for executing an obstacle positioning function or an obstacle prediction function.

[0216] According to an embodiment of the present disclosure, the processor sends second transmission data to the actuator, and the actuator receives the second transmission data from the processor. Based on the second transmission time when the processor sends the second transmission data and the second reception time when the actuator receives the second transmission data, at least one third time difference for transmitting the second transmission data can be calculated.

[0217] According to an embodiment of the present disclosure, an average value of at least one third time difference can be determined, and this average value can be used as a sixth statistical value corresponding to the at least one third time difference. Then, this sixth statistical value can be determined as the output delay of the autonomous driving system.

[0218] According to an embodiment of the present disclosure, the execution mode of the actuator can include a remote loopback mode and a direct loopback mode.

[0219] When it is determined that the execution mode of the actuator is the direct loopback mode, by determining at least one third time difference, a sixth statistical value corresponding to the at least one third time difference can be determined, and based on the sixth statistical value, the output delay of the autonomous driving system can be determined.

[0220] When it is determined that the execution mode of the actuator is the remote loopback mode, by determining at least one third time difference, a sixth statistical value corresponding to the at least one third time difference can be determined, and based on the sixth statistical value, the output delay of the autonomous driving system can be determined.

[0221] Alternatively, the standard output delay in the remote loopback mode provided by the supplier can also be directly obtained and determined as the output delay of the autonomous driving system.

[0222] According to an embodiment of the present disclosure, by determining at least one third time difference, based on the sixth statistical value corresponding to the third time difference, and further determining the output delay of the autonomous driving system, the accuracy of determining the output delay of the processor can be improved.

[0223] Figure 6 A block diagram of the hardware system of the autonomous driving system according to an embodiment of the present disclosure is schematically shown.

[0224] As Figure 6As shown, in 600, there are E sensor data paths, for example, sensor data path 601_1,......, sensor data path 601_E. There are also F memories, for example, memory 602_1,......, memory 602_F. There are also G processors, for example, processor 603_1,......, processor 603_G. There are also H instruction execution paths, for example, instruction execution path 604_1,......, instruction execution path 604_H. Each processor may have at least one memory corresponding to the processor. Multiple processors may have shared memories. E, F, G, and H may be integers greater than or equal to 1.

[0225] According to an embodiment of the present disclosure, the sensor data path is used to transmit the environmental data collected by the sensor to the autonomous driving system, and there will be sensor input bandwidth and input delay. The input bandwidth or input delay between sensor data paths 601_1,......, sensor data path 601_E may be the same or different.

[0226] The processor is used to perform processing operations, and the memory is used to store the execution results of the actuator or the data transmitted by the sensor data path. There will be processing delay and processor interconnection bandwidth in the processor. The processing delays of processors 603_1,......, processor 603_G may be the same or different. The interconnection bandwidth between processors 603_1,......, processor 603_G may be the same or different.

[0227] There will be processor storage bandwidth between the processor and the memory. The processor storage bandwidth between memories 602_1,......, memory 602_F and processors 1 603_1,......, processor 603_G may be the same or different.

[0228] The instruction execution path is used to send data to the actuator, including execution instructions or execution data. During the process of transmitting data through the instruction execution path, there will be output delay. The output delays between instruction execution paths 604_1,......, instruction execution path 604_H may be the same or different.

[0229] According to an embodiment of the present disclosure, there will be power consumption in the sensor data path, memory, actuator, and instruction execution path.

[0230] It should be noted that for the first predetermined time period to the eighth predetermined time period appearing in the present disclosure, it may be the same predetermined time period during the autonomous driving process, or different predetermined time periods during the autonomous driving process, or the same or different predetermined time periods during the startup or stop process of the autonomous driving vehicle.

[0231] According to an embodiment of the present disclosure, the performance evaluation method of the present disclosure is designed from the perspective of autonomous driving applications on the one hand. By integrating transmission bandwidth information, latency information, and power consumption information into the actual application scenario, the performance of the autonomous driving system is evaluated. On the other hand, from the perspective of the system, it is different from directly adding the performance evaluation information of individual processors to evaluate the performance of the autonomous driving system, and has the characteristic of accuracy.

[0232] Figure 7 FIG. schematically shows a flowchart of a method for optimizing the performance of an autonomous driving system according to an embodiment of the present disclosure.

[0233] As Figure 7 shown, the method 700 includes operations S710 to S730.

[0234] In operation S710, the performance evaluation information of the autonomous driving system is determined by using the above-mentioned performance evaluation method of the autonomous driving system.

[0235] In operation S720, according to the performance evaluation information, a performance optimization strategy is determined.

[0236] In operation S730, the autonomous driving system is optimized according to the performance optimization strategy.

[0237] According to an embodiment of the present disclosure, the performance evaluation information of the autonomous driving system may include the performance evaluation value of the autonomous driving system, and may also include the evaluation values corresponding to the transmission bandwidth information, latency information, and power consumption information respectively, so as to determine the performance optimization strategy according to the specific evaluation values.

[0238] After determining the performance evaluation information of the autonomous driving system, a performance optimization strategy can be determined according to the performance evaluation information.

[0239] For example, in the case where it is determined that the performance evaluation value is lower than a predetermined threshold, according to the importance of the transmission bandwidth information, latency information, and power consumption information, the optimization order is determined. For example, latency information - power consumption information - transmission bandwidth information.

[0240] For example, after determining the optimization order, after optimizing the first performance index, the optimized performance evaluation value is compared with the first threshold. In the case where the optimized performance evaluation value reaches the first threshold, the second performance index is optimized. In the case where the optimized performance evaluation value does not reach the first threshold, the performance index is continuously optimized until the optimized performance evaluation value reaches the first threshold.

[0241] After optimizing the second performance metric, compare the optimized performance evaluation value with the second threshold. When the optimized performance evaluation value reaches the second threshold, optimize the third performance metric. When the optimized performance evaluation value does not reach the second threshold, continue to optimize this performance metric until the optimized performance evaluation value reaches the second threshold.

[0242] After optimizing the third performance metric, compare the optimized performance evaluation value with the third threshold. When the optimized performance evaluation value reaches the third threshold, end the optimization. When the optimized performance evaluation value does not reach the third threshold, continue to optimize this performance metric until the optimized performance evaluation value reaches the third threshold.

[0243] According to an embodiment of the present disclosure, the performance optimization strategy may also be to compare the evaluation values corresponding to the transmission bandwidth information, latency information, and power consumption information with the fourth threshold, fifth threshold, and sixth threshold respectively. When the transmission bandwidth information does not reach the fourth threshold, optimize the transmission bandwidth information. When the latency information does not reach the fifth threshold, optimize the latency information. When the power consumption information does not reach the sixth threshold, optimize the power consumption information.

[0244] According to an embodiment of the present disclosure, by obtaining the performance evaluation information of the above-mentioned autonomous driving system and determining a more accurate performance optimization strategy based on the above-mentioned performance evaluation information, it is beneficial to reasonably utilize system resources and reduce overhead.

[0245] In the technical solution of the present disclosure, the processing of collection, storage, use, processing, transmission, provision, and disclosure of the user's personal information involved all comply with the provisions of relevant laws and regulations and do not violate public order and good customs.

[0246] Figure 8 Schematically shows a block diagram of a performance evaluation device of an autonomous driving system according to an embodiment of the present disclosure.

[0247] As Figure 8 shown, the performance evaluation device 800 of the autonomous driving system may include a transmission bandwidth determination module 810, a latency determination module 820, a power consumption determination module 830, and a performance evaluation module 840.

[0248] The transmission bandwidth determination module 810 is used to determine the transmission bandwidth information of the autonomous driving system. The transmission bandwidth information includes at least one of the following: sensor input bandwidth, processor storage bandwidth, and processor interconnection bandwidth. The processor storage bandwidth characterizes the bandwidth between the processor and the memory. The processor interconnection bandwidth characterizes the interconnection bandwidth between multiple processors.

[0249] The latency determination module 820 is configured to determine the latency information of the autonomous driving system. The latency information includes at least one of the following: input latency, processing latency, and output latency. The input latency represents the input latency of the data path. The processing latency represents the latency of the processor executing the processing operation. The output latency represents the latency between the processor and the actuator.

[0250] The power consumption determination module 830 is configured to determine the power consumption information of the autonomous driving system. The power consumption information includes at least one of the following: non-dynamic load power consumption and dynamic load power consumption. The non-dynamic load power consumption represents the static no-load power consumption.

[0251] The performance evaluation module 840 is configured to determine the performance evaluation information of the autonomous driving system based on the transmission bandwidth information, latency information, and power consumption information.

[0252] According to an embodiment of the present disclosure, the performance evaluation module 840 may include a weight determination unit and a performance evaluation unit

[0253] The weight determination unit is configured to determine a weighted statistical value based on the transmission bandwidth information, latency information, and power consumption information.

[0254] The performance evaluation unit is configured to determine the performance evaluation value of the autonomous driving system based on the weighted statistical value.

[0255] According to an embodiment of the present disclosure, when the power consumption information includes non-dynamic load power consumption, the power consumption determination module 830 may include a first power consumption determination unit, a first statistical value determination unit, and a non-dynamic load power consumption determination unit.

[0256] The first power consumption value determination unit is configured to determine at least one first power consumption value in a first predetermined time period.

[0257] The first statistical value determination unit is configured to determine a first statistical value corresponding to at least one first power consumption value.

[0258] The non-dynamic load power consumption determination unit is configured to determine the non-dynamic load power consumption of the autonomous driving system based on the first statistical value.

[0259] According to an embodiment of the present disclosure, the first power consumption value is determined based on at least one of a first current value and a first voltage value. The first current value and the first voltage value are obtained by detecting the power supply of the autonomous driving system via a controller.

[0260] According to an embodiment of the present disclosure, when the power consumption information includes dynamic load power consumption, the power consumption determination module 830 may include a second power consumption value determination unit, a second statistical value determination unit, and a dynamic load power consumption determination unit.

[0261] The second power consumption value determination unit is configured to determine at least one second power consumption value in a second predetermined time period.

[0262] A second statistical value determination unit, configured to determine a second statistical value corresponding to at least one second power consumption value.

[0263] A dynamic load power consumption determination unit, configured to determine the dynamic load power consumption of the autonomous driving system according to the second statistical value.

[0264] According to an embodiment of the present disclosure, the second power consumption value is determined according to at least one of a second current value and a second voltage value. The second current value and the second voltage value are obtained by detecting the power supply of the autonomous driving system via a controller. At least two of the at least one second power consumption value are obtained under different load conditions.

[0265] According to an embodiment of the present disclosure, the sensor input bandwidth includes at least one of the following: a visual sensor input bandwidth and a radar input bandwidth. The radar input bandwidth includes at least one of the following: a lidar input bandwidth and a millimeter-wave radar input bandwidth.

[0266] According to an embodiment of the present disclosure, when the transmission bandwidth information includes the sensor input bandwidth,

[0267] When the sensor input bandwidth includes the visual sensor input bandwidth, the transmission bandwidth determination module 810 may include a first input bandwidth determination unit and a visual sensor input bandwidth determination unit.

[0268] The first input bandwidth determination unit is configured to determine a first input bandwidth corresponding to at least one first combination information in a third predetermined period to obtain at least one first input bandwidth. The first combination information includes a first transmission frame rate and the resolution of video data, and at least one first combination information is different from each other.

[0269] The visual sensor input bandwidth determination unit is configured to determine the visual sensor input bandwidth of the autonomous driving system according to at least one first input bandwidth.

[0270] When the sensor input bandwidth includes the radar input bandwidth, the transmission bandwidth determination module 810 may include a second input bandwidth determination unit and a radar input bandwidth determination unit.

[0271] The second input bandwidth determination unit is configured to determine a second input bandwidth corresponding to at least one second combination information in a fourth predetermined period to obtain at least one second input bandwidth. The second combination information includes a second transmission frame rate and the resolution of point cloud data, and at least one second combination information is different from each other.

[0272] The radar input bandwidth determination unit determines the radar input bandwidth of the autonomous driving system according to at least one second input bandwidth.

[0273] According to an embodiment of the present disclosure, when the transmission bandwidth information includes the processor storage bandwidth, the transmission bandwidth determination module 810 may include a processor storage bandwidth determination unit.

[0274] The processor storage bandwidth determination unit is configured to adjust the expected processor storage bandwidth according to the storage bandwidth adjustment information to obtain the processor storage bandwidth of the autonomous driving system.

[0275] According to an embodiment of the present disclosure, the storage bandwidth adjustment information includes at least one of the following: read-write operation ratio information, read-write operation type information, and storage bandwidth allocation information of at least one functional unit of the processor in the case of a shared memory. The read-write operation type information includes at least one of the following: continuous read operation, continuous write operation, discontinuous read operation, and discontinuous write operation.

[0276] According to an embodiment of the present disclosure, when the transmission bandwidth information includes the processor interconnection bandwidth, the transmission bandwidth determination module 810 may include a first processor interconnection bandwidth determination unit and a second processor interconnection bandwidth determination unit.

[0277] The first processor interconnection bandwidth determination unit is configured to, when there is a direct interconnection relationship between two processors among multiple processors, determine the maximum rate corresponding to the interconnection interface protocol as the processor interconnection bandwidth corresponding to the two processors.

[0278] The second processor interconnection bandwidth determination unit is configured to, when there is an indirect interconnection relationship between two processors among multiple processors, determine the processor interconnection bandwidth corresponding to the two processors according to the interconnection configuration status.

[0279] According to an embodiment of the present disclosure, the second processor interconnection bandwidth determination unit may include a first processor interconnection bandwidth determination subunit.

[0280] The first processor interconnection bandwidth determination subunit is configured to, in response to detecting that the interconnection configuration status is a static configuration status, determine the minimum interconnection bandwidth between the two processors as the processor interconnection bandwidth corresponding to the two processors.

[0281] According to an embodiment of the present disclosure, the second processor interconnection bandwidth determination unit may further include a second processor interconnection bandwidth determination subunit, an interconnection bandwidth statistical value determination subunit, and a third processor interconnection bandwidth determination subunit.

[0282] In response to detecting that the interconnection configuration status is a dynamic configuration status,

[0283] The second processor interconnection bandwidth determination subunit is configured to determine at least one interconnection bandwidth between the two processors in a fifth predetermined period.

[0284] An interconnection bandwidth statistical value determination subunit, configured to determine a third statistical value corresponding to at least one interconnection bandwidth.

[0285] A third processor interconnection bandwidth determination subunit, configured to determine a processor interconnection bandwidth corresponding to two processors according to the third statistical value.

[0286] According to an embodiment of the present disclosure, when the delay information includes an input delay, the delay determination module 820 may include a first time difference determination unit, a transceiver delay determination unit, a transmission line delay determination unit, and an input delay determination unit.

[0287] The first time difference determination unit is configured to determine at least one first time difference in a sixth predetermined time period. The first time difference represents the time difference between a first transmission time and a first reception time. The first transmission time represents the time when a transceiver of a data path transmits first transmission data. The first reception time represents the time when the transceiver receives received data. The received data represents response data corresponding to the first transmission data.

[0288] The transceiver delay determination unit is configured to determine the transceiver delay of the transceiver according to a fourth statistical value corresponding to at least one first time difference.

[0289] The transmission line delay determination unit is configured to determine the transmission line delay of a printed circuit board in a sixth predetermined time period.

[0290] The input delay determination unit is configured to determine the input delay of the autonomous driving system according to the transceiver delay and the transmission line delay.

[0291] According to an embodiment of the present disclosure, when the delay information includes a processing delay, the delay determination module 820 may include a second time difference determination unit, a processing delay statistical value determination unit, and a processing delay determination unit.

[0292] The second time difference determination unit is configured to determine at least one second time difference in a seventh predetermined time period. The second time difference is the time difference between the start time and the end time of a processing operation.

[0293] The processing delay statistical value determination unit determines a fifth statistical value corresponding to at least one second time difference.

[0294] The processing delay determination unit is configured to determine the processing delay of the autonomous driving system according to the fifth statistical value.

[0295] According to an embodiment of the present disclosure, when the delay information includes an output delay, the delay determination module 820 may include a second time difference determination unit, an output delay statistical value determination unit, and an output delay determination unit.

[0296] A second time difference determination unit is configured to determine at least one third time difference in an eighth predetermined time period. The third time difference represents the time difference between a second transmission time and a second reception time. The second transmission time represents the time when the processor transmits second transmission data. The second reception time represents the time when the actuator receives the second transmission data.

[0297] An output delay statistical value determination unit determines a sixth statistical value corresponding to at least one third time difference.

[0298] An output delay determination unit is configured to determine the output delay of the autonomous driving system according to the sixth statistical value.

[0299] Figure 9 FIG. schematically shows a block diagram of a performance optimization device for an autonomous driving system according to an embodiment of the present disclosure.

[0300] As Figure 9 shown, the performance optimization device 900 of the autonomous driving system may include a performance evaluation information determination module 910, a performance optimization strategy determination module 920, and an optimization module 930.

[0301] The performance evaluation information determination module 910 is configured to determine the performance evaluation information of the autonomous driving system by using the performance evaluation device of the autonomous driving system according to the embodiment of the present disclosure.

[0302] The performance optimization strategy determination module 820 is configured to determine a performance optimization strategy according to the performance evaluation information.

[0303] The optimization module 830 is configured to optimize the autonomous driving system according to the performance optimization strategy.

[0304] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0305] According to an embodiment of the present disclosure, an electronic device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method as described above.

[0306] According to an embodiment of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to execute the method as described above.

[0307] According to an embodiment of the present disclosure, a computer program product includes a computer program, and the computer program implements the method as described above when executed by a processor.

[0308] Figure 10A block diagram of an electronic device suitable for implementing a performance evaluation method and a performance optimization method of an autonomous driving system according to an embodiment of the present disclosure is schematically shown. The electronic device is intended to represent various forms of digital computers, such as, a laptop computer, a desktop computer, a workbench, a personal digital assistant, a server, a blade server, a mainframe computer, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as, a personal digital processor, a cellular phone, a smart phone, a wearable device, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0309] As Figure 10 shown, the electronic device 1000 includes a computing unit 1001, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 1002 or a computer program loaded from a storage unit 1008 into a random access memory (RAM) 1003. In the RAM 1003, various programs and data required for the operation of the electronic device 1000 can also be stored. The computing unit 1001, the ROM 1002, and the RAM 1003 are connected to each other through a bus 1004. An input / output (I / O) interface 1005 is also connected to the bus 1004.

[0310] A plurality of components in the electronic device 1000 are connected to the I / O interface 1005, including: an input unit 1006, such as a keyboard, a mouse, etc.; an output unit 1007, such as various types of displays, speakers, etc.; a storage unit 1008, such as a magnetic disk, an optical disk, etc.; and a communication unit 1009, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 1009 allows the electronic device 1000 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0311] The computing unit 1001 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 1001 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 1001 executes the various methods and processes described above, for example, the performance evaluation method and the performance optimization method of the autonomous driving system. For example, in some embodiments, the performance evaluation method and the performance optimization method of the autonomous driving system can be implemented as a computer software program, which is tangibly included in a machine-readable medium, such as the storage unit 1008. In some embodiments, part or all of the computer program can be loaded and / or installed onto the electronic device 1000 via the ROM 1002 and / or the communication unit 1009. When the computer program is loaded into the RAM 1003 and executed by the computing unit 1001, one or more steps of the performance evaluation method and the performance optimization method of the autonomous driving system described above can be executed. Alternatively, in other embodiments, the computing unit 1001 can be configured to execute the performance evaluation method and the performance optimization method of the autonomous driving system by any other suitable means (such as by means of firmware).

[0312] Based on the foregoing electronic device, the present disclosure further provides an autonomous driving vehicle, which may include the electronic device, and may further include a communication component, a display screen for implementing a human-machine interface, an information acquisition device for acquiring surrounding environment information, etc. The communication component, the display screen, and the information acquisition device are communicatively connected to the electronic device. The electronic device included in the autonomous driving vehicle can implement the object motion trajectory information processing method of the embodiments of the present disclosure.

[0313] Among them, the electronic device can be integrally integrated with the communication component, the display screen, and the information acquisition device, or can be separately provided from the communication component, the display screen, and the information acquisition device.

[0314] The various embodiments of the systems and techniques described above in this specification can be implemented in digital electronic circuitry, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems-on-chip (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include: being implemented in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be a special-purpose or general-purpose programmable processor that receives data and instructions from, and transmits data and instructions to, a storage system, at least one input device, and at least one output device.

[0315] The program code for implementing the methods of the present disclosure can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the program codes, when executed by the processor or controller, cause the functions / operations specified in the flowchart and / or block diagram to be implemented. The program code can be executed entirely on the machine, partly on the machine, as a stand-alone software package partly on the machine and partly on a remote machine or entirely on the remote machine or server.

[0316] In the context of the present disclosure, a machine-readable medium can be a tangible medium that can contain or store a program for use by or in connection with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0317] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and a pointing device (e.g., a mouse or a trackball) through which the user can provide input to the computer. Other kinds of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, speech input, or tactile input).

[0318] The systems and techniques described herein can be implemented in a computing system including backend components (e.g., as a data server), or a computing system including middleware components (e.g., an application server), or a computing system including frontend components (e.g., a user computer having a graphical user interface or a web browser through which the user can interact with an implementation of the systems and techniques described herein), or a computing system including any combination of such backend components, middleware components, or frontend components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local area network (LAN), wide area network (WAN), and the Internet.

[0319] A computer system can include a client and a server. The client and the server are generally remote from each other and typically interact through a communication network. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, a server of a distributed system, or a server incorporating a blockchain.

[0320] It should be understood that various forms of the processes shown above can be used, with steps reordered, added, or deleted. For example, the steps recited in this disclosure can be executed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and no limitation is imposed herein.

[0321] The above specific embodiments do not constitute a limitation on the protection scope of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure shall be included within the protection scope of this disclosure.

Claims

1. A performance evaluation method for an autonomous driving system, comprising: Determining the transmission bandwidth information of the autonomous driving system, where the transmission bandwidth information includes at least one of the following: sensor input bandwidth, processor storage bandwidth, and processor interconnection bandwidth, the processor storage bandwidth characterizing the bandwidth between the processor and the memory, and the processor interconnection bandwidth characterizing the interconnection bandwidth between multiple processors; Determining the latency information of the autonomous driving system, where the latency information includes at least one of the following: input latency, processing latency, and output latency, the input latency characterizing the input latency of the data path, the processing latency characterizing the latency of the processor executing a processing operation, and the output latency characterizing the latency between the processor and the actuator; Determining the power consumption information of the autonomous driving system, where the power consumption information includes at least one of the following: non-dynamic load power consumption and dynamic load power consumption, the non-dynamic load power consumption characterizing the static no-load power consumption; and Determining the performance evaluation information of the autonomous driving system based on the transmission bandwidth information, the latency information, and the power consumption information.

2. The method according to claim 1, wherein, The determining the performance evaluation information of the autonomous driving system based on the transmission bandwidth information, the latency information, and the power consumption information includes: Determining a weighted statistical value based on the transmission bandwidth information, the latency information, and the power consumption information; and Determining the performance evaluation value of the autonomous driving system based on the weighted statistical value.

3. The method according to claim 1, wherein, When the power consumption information includes the non-dynamic load power consumption, the determining the power consumption information of the autonomous driving system includes: Determining at least one first power consumption value in a first predetermined time period; Determining a first statistical value corresponding to the at least one first power consumption value; and Determining the non-dynamic load power consumption of the autonomous driving system based on the first statistical value.

4. The method according to claim 3, wherein, The first power consumption value is determined based on at least one of a first current value and a first voltage value, and the first current value and the first voltage value are obtained by the controller detecting the power supply of the autonomous driving system.

5. The method according to claim 1, wherein, When the power consumption information includes the dynamic load power consumption, the determining the power consumption information of the autonomous driving system includes: Determining at least one second power consumption value in a second predetermined time period; Determining a second statistical value corresponding to the at least one second power consumption value; and Determining the dynamic load power consumption of the autonomous driving system based on the second statistical value.

6. The method according to claim 4, wherein When the power consumption information includes the dynamic load power consumption, the determining the power consumption information of the autonomous driving system includes: Determining at least one second power consumption value in a second predetermined time period; Determining a second statistical value corresponding to the at least one second power consumption value; and Determining the dynamic load power consumption of the autonomous driving system based on the second statistical value; wherein the second power consumption value is determined based on at least one of a second current value and a second voltage value, and the second current value and the second voltage value are obtained by the controller detecting the power supply of the autonomous driving system; wherein at least two of the at least one second power consumption values are obtained under different load conditions.

7. The method according to claim 1, wherein, The sensor input bandwidth includes at least one of the following: a vision sensor input bandwidth and a radar input bandwidth, and the radar input bandwidth includes at least one of the following: a lidar input bandwidth and a millimeter-wave radar input bandwidth; Wherein, when the transmission bandwidth information includes the sensor input bandwidth, the determining the transmission bandwidth information of the autonomous driving system includes: When the sensor input bandwidth includes the vision sensor input bandwidth, Determining a first input bandwidth corresponding to at least one first combination information in a third predetermined period to obtain at least one first input bandwidth, where the first combination information includes a first transmission frame rate and a resolution of video data, and the at least one first combination information is different from each other; and Determining the vision sensor input bandwidth of the autonomous driving system according to the at least one first input bandwidth; and When the sensor input bandwidth includes the radar input bandwidth, Determining a second input bandwidth corresponding to at least one second combination information in a fourth predetermined period to obtain at least one second input bandwidth, where the second combination information includes a second transmission frame rate and a resolution of point cloud data, and the at least one second combination information is different from each other; and Determining the radar input bandwidth of the autonomous driving system according to the at least one second input bandwidth.

8. The method according to claim 1, wherein When the transmission bandwidth information includes the processor storage bandwidth, the determining the transmission bandwidth information of the autonomous driving system includes: Adjusting an expected processor storage bandwidth according to storage bandwidth adjustment information to obtain the processor storage bandwidth of the autonomous driving system; Wherein, the storage bandwidth adjustment information includes at least one of the following: a read-write operation ratio information, a read-write operation type information, and a storage bandwidth allocation information of at least one functional unit of the processor when sharing the memory, and the read-write operation type information includes at least one of the following: a continuous read operation, a continuous write operation, a discontinuous read operation, and a discontinuous write operation.

9. The method according to claim 1, wherein When the transmission bandwidth information includes the processor interconnection bandwidth, the determining the transmission bandwidth information of the autonomous driving system includes: When there is a direct interconnection relationship between two of the multiple processors, determining the maximum rate corresponding to the interconnection interface protocol as the processor interconnection bandwidth corresponding to the two processors; When there is an indirect interconnection relationship between two of the multiple processors, determining the processor interconnection bandwidth corresponding to the two processors according to the interconnection configuration status.

10. The method according to claim 9, wherein The determining the processor interconnection bandwidth corresponding to the two processors according to the interconnection configuration status includes: In response to detecting that the interconnection configuration status is a static configuration status, determining the minimum interconnection bandwidth between the two processors as the processor interconnection bandwidth corresponding to the two processors.

11. The method according to claim 10, wherein, The determining the processor interconnection bandwidth corresponding to the two processors according to the interconnection configuration status further includes: In response to detecting that the interconnection configuration status is a dynamic configuration status, Determine at least one interconnection bandwidth between the two processors in the fifth predetermined time period; Determine a third statistical value corresponding to the at least one interconnection bandwidth; and Determine the processor interconnection bandwidth corresponding to the two processors according to the third statistical value.

12. The method according to any one of claims 1 to 11, wherein When the latency information includes the input latency, the determining the latency information of the autonomous driving system includes: Determine at least one first time difference in a sixth predetermined time period, where the first time difference represents the time difference between a first transmission time and a first reception time, the first transmission time represents the time when a transceiver of a data path transmits first transmission data, the first reception time represents the time when the transceiver receives received data, and the received data represents response data corresponding to the first transmission data; Determine the transceiver latency of the transceiver according to a fourth statistical value corresponding to the at least one first time difference; Determine the transmission line latency of a printed circuit board in the sixth predetermined time period; and Determine the input latency of the autonomous driving system according to the transceiver latency and the transmission line latency.

13. The method according to any one of claims 1 to 11, wherein, When the latency information includes the processing latency, the determining the latency information of the autonomous driving system includes: Determine at least one second time difference in a seventh predetermined time period, where the second time difference is the time difference between the start time and the end time of the processing operation; Determine a fifth statistical value corresponding to the at least one second time difference; and Determine the processing latency of the autonomous driving system according to the fifth statistical value.

14. The method according to any one of claims 1 to 11, wherein, When the latency information includes the output latency, the determining the latency information of the autonomous driving system includes: Determine at least one third time difference in an eighth predetermined time period, where the third time difference represents the time difference between a second transmission time and a second reception time, the second transmission time represents the time when the processor transmits second transmission data, and the second reception time represents the time when the actuator receives the second transmission data; Determine a sixth statistical value corresponding to the at least one third time difference; and Determine the output latency of the autonomous driving system according to the sixth statistical value.

15. A method for optimizing the performance of an autonomous driving system, comprising: Determine the performance evaluation information of the autonomous driving system by using the method according to any one of claims 1 to 14; Determine a performance optimization strategy according to the performance evaluation information; And Optimize the autonomous driving system according to the performance optimization strategy.

16. A performance evaluation device for an autonomous driving system, comprising: A transmission bandwidth determination module, configured to determine the transmission bandwidth information of the autonomous driving system, where the transmission bandwidth information includes at least one of the following: sensor input bandwidth, processor storage bandwidth, and processor interconnection bandwidth, the processor storage bandwidth represents the bandwidth between the processor and the memory, and the processor interconnection bandwidth represents the interconnection bandwidth between multiple processors; A latency determination module, configured to determine the latency information of the autonomous driving system, where the latency information includes at least one of the following: input latency, processing latency, and output latency. The input latency represents the input latency of the data path, the processing latency represents the latency of the processor executing a processing operation, and the output latency represents the latency between the processor and the actuator; A power consumption determination module, configured to determine the power consumption information of the autonomous driving system, where the power consumption information includes at least one of the following: non-dynamic load power consumption and dynamic load power consumption. The non-dynamic load power consumption represents the static no-load power consumption; and A performance evaluation module, configured to determine the performance evaluation information of the autonomous driving system according to the transmission bandwidth information, the latency information, and the power consumption information.

17. A performance optimization device for an autonomous driving system, comprising: A performance evaluation information determination module, configured to determine the performance evaluation information of the autonomous driving system by using the device according to claim 16; A performance optimization strategy determination module, configured to determine a performance optimization strategy according to the performance evaluation information; and An optimization module, configured to optimize the autonomous driving system according to the performance optimization strategy.

18. An electronic device, comprising: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method according to any one of claims 1 to 15.

19. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to execute the method according to any one of claims 1 to 15.

20. A computer program product, comprising a computer program which, when executed by a processor, implements the method according to any one of claims 1 to 15.

21. An autonomous driving vehicle, comprising the electronic device according to claim 18.

Images