Test method and device for perception fusion system
By acquiring sensor data and generating coded sequence comparisons, the problem of insufficient testing granularity in the perception fusion system was solved, enabling fine-grained testing and accurate location of abnormal modules.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- ALIBABA DAMO (HANGZHOU) TECH CO LTD
- Filing Date
- 2023-01-19
- Publication Date
- 2026-05-12
AI Technical Summary
In existing technologies, testing methods for perception fusion systems can only provide overall feedback on the results, resulting in overly coarse testing granularity that makes it difficult to achieve fine-grained testing and anomaly localization for perception fusion systems.
By acquiring sensor data and labeling obstacle information, the outputs of the perception module and fusion module are obtained using probes, and a coding sequence is generated for comparison. Based on the coding sequence, the test results and abnormal modules of the perception fusion system are determined.
It enables fine-grained testing of the perception fusion system, accurately identifies abnormal modules, and improves the accuracy and reliability of the test.
Smart Images
Figure CN116304970B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of autonomous driving technology, and in particular to a testing method and apparatus for a perception fusion system. Background Technology
[0002] Autonomous vehicles rely on the collaborative efforts of perception sensors, artificial intelligence, and global positioning systems to drive safely and autonomously. The system primarily comprises three main functions: perception, decision-making, and execution. The deployment of autonomous vehicles requires extensive testing, as does the iteration of algorithms. Among these, perception, as a crucial component of autonomous driving technology, directly impacts the performance of the entire system through testing.
[0003] In Level 3 and above autonomous driving systems, multi-sensor perception fusion solutions are gradually becoming the mainstream approach. Perception fusion refers to collecting data from multiple different types of sensors, processing each type of sensor data separately to obtain perception results, and then fusing these results to obtain the final perception result. Therefore, how to effectively test perception fusion systems has become a pressing issue. Summary of the Invention
[0004] In view of this, this application provides a testing method and apparatus for a perception fusion system, so as to facilitate effective testing of the perception fusion system.
[0005] This application provides the following solution:
[0006] Firstly, a method for testing a perception fusion system is provided, the perception fusion system including a fusion module and at least two perception modules, the method comprising:
[0007] Acquire test samples, the test samples including perception sensor data and obstacle information labeled with the perception sensor data;
[0008] The sensor data is input into the perception fusion system, and the outputs of each perception module and the fusion module in the perception fusion system are obtained respectively.
[0009] The outputs of each sensing module and the fusion module are compared with the labeled obstacle information to obtain the encoding sequence corresponding to the comparison results of each sensing module and the fusion module.
[0010] The test results for the perceptual fusion system are determined based on the encoded sequence.
[0011] According to one achievable method in this application embodiment, obtaining the outputs of each sensing module and the output of the fusion module in the sensing fusion system respectively includes:
[0012] The output of each sensing module is obtained by pre-inserting probes at the output end of each sensing module;
[0013] The output of the fusion module can be obtained by pre-inserting a probe at the output end of the fusion module, or by directly obtaining the output of the fusion module.
[0014] According to one achievable method in an embodiment of this application, determining the test result of the perceptual fusion system based on the encoded sequence includes:
[0015] Based on the pre-configured correspondence between the encoded sequence and the test result, the test result corresponding to the obtained encoded sequence is determined;
[0016] The correspondence between the encoded sequence and the test result is pre-configured based on the topology of the perceptual fusion system.
[0017] According to one achievable method in an embodiment of this application, determining the test result of the perceptual fusion system based on the encoded sequence includes:
[0018] The test result is determined based on the encoded sequence.
[0019] If the perception fusion system fails the test, the abnormal module in the perception fusion system is located based on the encoded sequence.
[0020] According to one achievable method in an embodiment of this application, determining whether the perceptual fusion system passes the test based on the encoded sequence includes:
[0021] If the encoded sequence obtained from test samples exceeding a preset first proportion or a first number indicates that the output of the fusion module is consistent with the labeled obstacle information, then the perception fusion system is determined to have passed the test; otherwise, the perception fusion system is determined to have failed the test.
[0022] According to one achievable method in an embodiment of this application, locating the abnormal module in the perception fusion system based on the encoded sequence includes:
[0023] If the encoded sequence obtained from test samples exceeding a preset second ratio or second number indicates that the output of the fusion module is inconsistent with the labeled obstacle information, then the fusion module is determined to be an abnormal module.
[0024] If the encoded sequence obtained from test samples exceeding a preset third proportion or third number indicates that the output of one of the sensing modules is inconsistent with the labeled obstacle information, then one of the sensing modules is determined to be an abnormal module.
[0025] According to one achievable method in an embodiment of this application, the method further includes:
[0026] The distribution information of the encoded sequence is output to adjust the test sample based on the distribution information.
[0027] Secondly, a testing apparatus for a perception fusion system is provided, the fusion system including a fusion module and at least two perception modules, the apparatus comprising:
[0028] The sample acquisition unit is configured to acquire test samples, the test samples including sensor data and obstacle information labeled on the sensor data;
[0029] The system interaction unit is configured to input the sensor data into the perception fusion system and acquire the outputs of each perception module and the fusion module in the perception fusion system, respectively.
[0030] The information comparison unit is configured to compare the outputs of each sensing module and the fusion module with the labeled obstacle information to obtain the encoding sequence corresponding to the comparison results of each sensing module and the fusion module.
[0031] The result determination unit is configured to determine the test result of the perceptual fusion system based on the encoded sequence.
[0032] According to a third aspect, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in any one of the first aspects.
[0033] According to the fourth aspect, an electronic device is provided, comprising:
[0034] One or more processors; and
[0035] A memory associated with the one or more processors, the memory being used to store program instructions that, when read and executed by the one or more processors, perform the steps of the method described in any one of the first aspects above.
[0036] According to the specific embodiments provided in this application, the following technical effects are disclosed:
[0037] 1) In this application, the outputs of each sensing module and the fusion model in the sensing fusion system are compared with the labeled obstacle information to obtain the coding sequence corresponding to the comparison results of each sensing module and the fusion module. Based on this formal expression of the coding sequence, the test results of the sensing fusion system are determined, thereby realizing the effective testing of the sensing fusion system.
[0038] 2) The method provided in this application can not only obtain the comparison results of the overall output of the perception fusion system, but also the comparison results of each module in the perception fusion system, thereby achieving a more granular testing effect and realizing accurate anomaly localization.
[0039] 3) This application uses a formalized expression of encoded sequences, which can reflect the rationality of the test samples by outputting the distribution information of the encoded sequences, thereby assisting in the adjustment of the test samples.
[0040] Of course, any product implementing this application does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description
[0041] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0042] Figure 1 This is a system architecture diagram applicable to the embodiments of this application;
[0043] Figure 2 This is a flowchart of a testing method for a perception fusion system provided in an embodiment of this application;
[0044] Figure 3 This is a schematic diagram of the topology of the perception fusion system provided in the embodiments of this application;
[0045] Figure 4 A schematic diagram illustrating the principle of the testing method provided in the embodiments of this application;
[0046] Figure 5 This is a schematic block diagram of a testing device for a perception fusion system provided in an embodiment of this application;
[0047] Figure 6 A schematic block diagram of an electronic device provided in an embodiment of this application. Detailed Implementation
[0048] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.
[0049] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. The singular forms “a,” “the,” and “the” as used in the embodiments of this invention and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.
[0050] It should be understood that the term "and / or" used in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.
[0051] Depending on the context, the word "if" as used here can be interpreted as "when," "when," "in response to determination," or "in response to detection." Similarly, depending on the context, the phrase "if determination" or "if detection (of the stated condition or event)" can be interpreted as "when determination," "in response to determination," "when detection (of the stated condition or event)," or "in response to detection (of the stated condition or event)."
[0052] Current testing methods for perceptual fusion systems compare the system's output with the labeled results to determine the overall performance. However, this approach only provides a general overview of the system's effectiveness, resulting in a coarse-grained test.
[0053] In view of this, this application provides a new approach to testing perceptual fusion systems. To facilitate understanding of the embodiments of this application, the system architecture on which the embodiments of this application are based will first be briefly described. Figure 1 An exemplary system architecture that can be applied to embodiments of this application is shown, such as Figure 1 As shown, the system mainly includes a perception testing device and a perception fusion system for autonomous vehicles.
[0054] The perception testing device is mainly used to test the perception fusion system of an autonomous vehicle using the perception testing method provided in the embodiments of this application. The perception test mainly involves the perception fusion system in the autonomous vehicle; other systems are not shown in the figure.
[0055] The term "autonomous vehicle" in this application is used in a broad sense and can refer to either driverless vehicles or driver-assisted vehicles.
[0056] Perception sensors in autonomous vehicles can include image sensors, radar, infrared sensors, ultrasonic sensors, etc. Image sensors can include cameras, camcorders, etc. Radar can include lidar, millimeter-wave radar, etc.
[0057] The perception fusion system processes data collected by multiple types of perception sensors (referred to as perception sensor data in this embodiment) separately using perception algorithms to obtain individual perception results; then, these results are fused to obtain the final perception result. The perception result is the detection result of environmental data, which may include road information, obstacle information, etc. In this embodiment, the perception mainly refers to the perception of obstacle information. Obstacles may include vehicles, pedestrians, traffic facilities, etc., and may also be any other objects that pose a safety hazard to vehicle traffic, such as trees, animals, etc.
[0058] Perception testing equipment can be any device with computing capabilities, such as a laptop or PC (Personal Computer). It can even be set up on a server to test the perception fusion system of autonomous vehicles.
[0059] It should be understood that Figure 1 The number of autonomous vehicles, perception fusion systems, and perception testing devices shown is merely illustrative. Depending on implementation needs, any number of autonomous vehicles, perception fusion systems, and perception testing devices can be included.
[0060] Figure 2 This is a flowchart of a testing method for a perceptual fusion system provided in an embodiment of this application. This method can be performed by… Figure 1 The perception testing device in the system architecture shown is executed. For example... Figure 2 As shown, the method may include the following steps:
[0061] Step 202: Obtain test samples, which include sensor data and obstacle information labeled on the sensor data.
[0062] Step 204: Input the sensor data into the perception fusion system, and obtain the output of each perception module and the output of the fusion module in the perception fusion system.
[0063] Step 206: Compare the outputs of each sensing module and the fusion module with the labeled obstacle information to obtain the encoding sequence corresponding to the comparison results of each sensing module and the fusion module.
[0064] Step 208: Determine the test results of the perceptual fusion system based on the coding sequence.
[0065] As can be seen from the above process, this application compares the outputs of each sensing module and the fusion model in the perception fusion system with the labeled obstacle information to obtain the encoding sequences corresponding to the comparison results of each sensing module and the fusion module. The test results of the perception fusion system are then determined based on the encoding sequences. This method can achieve a more granular testing effect.
[0066] The method provided in the embodiments of this application will be described in detail below. Step 202, namely "obtaining test samples," will be described in detail below with reference to the embodiments.
[0067] In this embodiment, the perception sensor data in the test sample is data collected by the perception sensors of the autonomous vehicle, which can be images, point cloud data, etc. This sensor data can be collected by various types of sensors of the autonomous vehicle under test, or by various types of sensors of other autonomous vehicles. For example, a test sample can be constructed by collecting perception sensor data in various scenarios using a single data collection vehicle.
[0068] During testing, obstacle information can be manually labeled on the sensor data, including obstacle location information, obstacle type information, etc.
[0069] The various perception sensors of an autonomous vehicle collect data on the same scene. The collected perception sensor data can be multiple frames, each labeled with obstacle information. For example, when an autonomous vehicle approaches an intersection and is about to turn right, perception sensors such as cameras and lidar will collect data on the surrounding environment, acquiring multiple frames of image data and multiple frames of point cloud data. In this embodiment, the perception testing device acquires this perception sensor data for perception testing.
[0070] When constructing test samples, it is necessary to cover as many scenarios as possible. For example, autonomous vehicles may be driving straight, turning right, turning left, making a U-turn, experiencing traffic jams, slowing down at intersections, encountering pedestrians crossing the road, or encountering various types of obstacles during the journey.
[0071] The following describes step 204, namely, "inputting sensor data into the perception fusion system and obtaining the outputs of each perception module and the fusion module in the perception fusion system," in conjunction with an embodiment.
[0072] Typically, a perception fusion system in an autonomous vehicle includes at least two perception modules and a fusion module. One perception module corresponds to one type of perception sensor and is responsible for processing the data collected by that type of sensor using perception algorithms to output perception results, primarily obstacle information, such as the location and type of obstacles. The fusion module is responsible for fusing the perception results from all the perception modules and outputting the final perception result.
[0073] Therefore, the topology of the perception fusion system can be obtained first. This topology mainly includes the modular composition of the perception fusion system and the connection relationships between the modules, that is, which perception modules, fusion modules, and how the perception modules and fusion modules are connected. Figure 3 Taking a fusion perception system as an example, suppose it can include perception module a, perception module b, and fusion module c. Perception module a is responsible for processing image data, and perception module b is responsible for processing point cloud data. The perception results output by perception modules a and b are both output to fusion module c, which fuses the perception results from the two modules and outputs the final perception result.
[0074] In this step, probes can be pre-installed in the sensing fusion system according to the topology. Probes can be installed at the output end of each sensing module. These probes can acquire the output results of the corresponding sensing module and send them to the sensing test device. The probes typically copy the output results and send the copied data to the sensing test device, without affecting the normal transmission of the output results to the fusion module.
[0075] The perception testing device can directly acquire the output of the fusion module, or it can pre-install a probe at the output end of the fusion module to acquire the output result of the fusion module and send it to the perception testing device.
[0076] like Figure 4 As shown, the sensor data is input to the perception fusion system, and the perception test device obtains the outputs of perception module a, perception module b and fusion module c respectively through probes (black dots in the figure).
[0077] The following describes in detail step 206, namely, "comparing the outputs of each sensing module and the fusion module with the labeled obstacle information to obtain the encoding sequence corresponding to the comparison results of each sensing module and the fusion module," with reference to the embodiments.
[0078] The topology of the perceptual fusion system has been mentioned above. Based on this topology, the meaning of each bit in the encoding sequence can be pre-configured. The encoding sequence is obtained by encoding the comparison results of each perceptual module and the fusion module. Assuming the perceptual fusion system has N-1 perceptual modules and 1 fusion module, then the encoding sequence is an N-bit sequence, where each bit corresponds to the comparison result of one module.
[0079] Still with Figure 3 Taking the topology shown as an example, the encoding sequence is a 3-bit sequence. The first to third bits correspond to the comparison results of perception module a, perception module b, and fusion module c, respectively. For example, "0" can be used to indicate that the comparison result is consistent, that is, the output of the corresponding module is consistent with the labeled obstacle information; "1" can be used to indicate that the comparison result is inconsistent, that is, the output of the corresponding module is inconsistent with the labeled obstacle information.
[0080] For example, if the output of perception module a matches the labeled obstacle information, the output of perception module b does not match the labeled obstacle information, and the output of fusion module c matches the labeled obstacle information, then the corresponding encoding sequence is "101". As another example, if the output of perception module a does not match the labeled obstacle information, the output of perception module b matches the labeled obstacle information, and the output of fusion module c matches the labeled obstacle information, then the corresponding encoding sequence is "011".
[0081] Figure 3 The possible coded sequences output by the topology shown are shown in Table 1:
[0082] Table 1
[0083] Perception module a Perception module b Fusion module c sequence type 0 0 1 N1 0 0 0 N2 0 1 0 F1 1 0 0 F2 1 1 0 F3 1 1 1 T1 1 0 1 T2 0 1 1 T3
[0084] The following describes step 208, namely "determining the test results of the perception fusion system based on the coding sequence", in detail with reference to the embodiments.
[0085] One feasible approach is to directly output the encoded sequence, which testers can then analyze to obtain test results, i.e., whether each module in the perception fusion system meets the test requirements.
[0086] As another possible approach, in the application embodiment, the correspondence between the encoded sequences and test results can be pre-configured based on the topology of the perceptual fusion system. Since the topology of the perceptual fusion system is known in advance, all possible encoded sequences can also be pre-enumerated. A correspondence is established between all possible encoded sequences and test results, and this correspondence is stored for querying during testing.
[0087] During the testing in this step, the pass / fail status of the perception fusion system can be determined based on the encoded sequence. For example, if the encoded sequence obtained from more than a preset first proportion or a first number of test samples indicates that the output of the fusion module is consistent with the labeled obstacle information, then the perception fusion system is determined to have passed the test; otherwise, the perception fusion system is determined to have failed the test.
[0088] Referring to the encoded sequences T1, T2, and T3 in Table 1, the outputs of the perception fusion system in these sequences are consistent with the labeled obstacle information. The T1 encoded sequence indicates that both the perception module and the fusion module are consistent, meaning their outputs are all correct. Clearly, if the perception fusion system obtains the T1 encoded sequence for a large number of test samples, then it passes the test, and the performance of both the perception and fusion modules is excellent. The T2 and T3 encoded sequences indicate that some outputs from the perception module are correct, while others are incorrect, but the fusion module outputs correctly, indicating that the perception fusion system can output normally. If the perception fusion system obtains both T2 and T3 encoded sequences for a large number of test samples, then it passes the test, and the robustness of the fusion module is excellent.
[0089] If the perception fusion system fails the test, the abnormal modules in the perception fusion system can be located based on the encoded sequence. For example, if the encoded sequence obtained from more than a preset second proportion or a second number of test samples indicates that the output of the fusion module is inconsistent with the labeled obstacle information, then the fusion module is determined to be an abnormal module; if the encoded sequence obtained from more than a preset third proportion or a third number of test samples indicates that the output of one of the perception modules is inconsistent with the labeled obstacle information, then one of the perception modules is determined to be an abnormal module.
[0090] If the output of the perception module is correct (i.e., the output matches the labeled obstacle information), but the output of the fusion module is incorrect (i.e., the output does not match the labeled obstacle information), corresponding to cases F1, F2, and F3 in Table 1, then the fusion module may be abnormal. If the perception fusion system outputs the encoded sequences F1, F2, and F3 for a large number of test samples, then the fusion module is an abnormal module.
[0091] Specifically, for the encoded sequence shown in F3, the outputs of the sensing module are all correct, while the outputs of the fusion module are incorrect, indicating that the performance of the fusion module is particularly poor.
[0092] If the encoded sequence obtained from more than a preset third proportion or third number of test samples indicates that the output of one of the sensing modules is inconsistent with the labeled obstacle information, then one of the sensing modules is determined to be an abnormal module.
[0093] Referring to Table 1, if the perception fusion system outputs T2 and / or F2 for a large number of test samples, then perception module b is an abnormal module. If the perception fusion system outputs T3 and / or F1 for a large number of test samples, then perception module a is an abnormal module.
[0094] Additionally, the encoded sequences N1 and N2 shown in Table 1 above can be considered invalid results and do not need to be included in the statistics. However, if the proportion or number of invalid results is high, for example, exceeding the preset fourth proportion or fourth number, an anomaly can be output to prompt the testers to investigate.
[0095] It should be noted that the terms "first," "second," etc., used in this disclosure do not imply any limitation on size, order, or quantity; they are merely used to distinguish the terms by name. For example, "first ratio," "second ratio," and "third ratio" are used to distinguish three ratios by name, and these three ratios can be the same or different. Similarly, "first quantity," "second quantity," and "third quantity" are used to distinguish three quantities by name, and these three quantities can be the same or different.
[0096] Furthermore, the distribution information of the encoded sequences can be output to adjust the test samples based on this information. For example, if the encoded sequences output by each perceptual fusion system for the test samples are concentrated in F1 to F3, or in T1 to T3, it indicates that the distribution of the test samples is poor and the scenarios are not comprehensive enough, requiring adjustment. Testing artificial intelligence algorithms like perception and fusion requires a large number of challenging samples to cover the entire dataset. If the encoded sequences output by each perceptual fusion system for the test samples are concentrated in the distribution of T1 to T3, with almost no F1 to F3, it is very likely that the sample scenarios are too simple. Such test results are not scientifically sound, and the perceptual fusion system may fail to correctly perceive uncommon scenarios, requiring adjustment of the test samples to include broader scenario coverage.
[0097] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.
[0098] According to another embodiment, a test apparatus for a perception fusion system is provided. Figure 5 This diagram shows a schematic block diagram of a test apparatus for a perceptual fusion system according to one embodiment. Figure 1The perception testing device in the architecture shown. Figure 5 As shown, the device 500 includes: a sample acquisition unit 501, a system interaction unit 502, an information comparison unit 503, and a result determination unit 504. The main functions of each component are as follows:
[0099] The sample acquisition unit 501 is configured to acquire test samples, which include perception sensor data and obstacle information labeled on the perception sensor data.
[0100] The system interaction unit 502 is configured to input sensor data into the perception fusion system and acquire the outputs of each perception module and the fusion module in the perception fusion system.
[0101] The information comparison unit 503 is configured to compare the outputs of each sensing module and the fusion module with the labeled obstacle information to obtain the encoding sequence corresponding to the comparison results of each sensing module and the fusion module.
[0102] Result determination unit 504 is configured to determine the test results of the perceptual fusion system based on the coding sequence.
[0103] As one possible implementation method, the system interaction unit 502 can be specifically configured to: acquire the output of each sensing module through probes pre-installed at the output end of each sensing module; acquire the output of the fusion module through probes pre-installed at the output end of the fusion module, or directly acquire the output of the fusion module.
[0104] As one possible implementation, the result determination unit 504 can be specifically configured to: determine the test result corresponding to the obtained encoding sequence based on the pre-configured correspondence between the encoding sequence and the test result; wherein the correspondence between the encoding sequence and the test result is pre-configured based on the topology of the perception fusion system.
[0105] As one possible implementation, the result determination unit 504 can be specifically configured to: determine whether the perception fusion system passes the test based on the encoding sequence; if the perception fusion system fails the test, locate the abnormal module in the perception fusion system based on the encoding sequence.
[0106] As one possible implementation, the result determination unit 504, when determining whether the perceptual fusion system passes the test based on the encoded sequence, can be specifically configured as follows:
[0107] If the encoded sequence obtained from test samples exceeding a preset first proportion or a first number indicates that the output of the fusion module is consistent with the labeled obstacle information, then the perception fusion system is determined to have passed the test; otherwise, the perception fusion system is determined to have failed the test.
[0108] As one possible implementation, when the result determination unit 504 locates anomaly modules in the perceptual fusion system based on the encoded sequence, it can be specifically configured as follows:
[0109] If the encoded sequence obtained from test samples exceeding the preset second ratio or second number indicates that the output of the fusion module is inconsistent with the labeled obstacle information, then the fusion module is determined to be an abnormal module.
[0110] If the encoded sequence obtained from more than a preset third proportion or third number of test samples indicates that the output of one of the sensing modules is inconsistent with the labeled obstacle information, then one of the sensing modules is determined to be an abnormal module.
[0111] Furthermore, the result determination unit 504 can be further configured to output the distribution information of the encoded sequence, so as to adjust the test samples based on the distribution information.
[0112] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the device embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without creative effort.
[0113] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation portals are provided for users to choose to authorize or refuse.
[0114] In addition, embodiments of this application also provide a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the method described in any of the foregoing method embodiments.
[0115] And an electronic device, comprising:
[0116] One or more processors; and
[0117] A memory associated with the one or more processors, the memory being used to store program instructions that, when read and executed by the one or more processors, perform the steps of the method described in any of the foregoing method embodiments.
[0118] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the method described in any of the foregoing method embodiments.
[0119] in, Figure 6 An exemplary architecture of an electronic device is shown, which may include a processor 610, a video display adapter 611, a disk drive 612, an input / output interface 613, a network interface 614, and a memory 620. The processor 610, video display adapter 611, disk drive 612, input / output interface 613, network interface 614, and memory 620 can communicate with each other via a communication bus 630.
[0120] The processor 610 can be implemented using a general-purpose CPU, microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits to execute relevant programs and implement the technical solution provided in this application.
[0121] The memory 620 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 620 can store the operating system 621 for controlling the operation of the electronic device 600, and the basic input / output system (BIOS) 622 for controlling the low-level operations of the electronic device 600. Additionally, it can store a web browser 623, a data storage management system 624, and a perception testing device 625, etc. The aforementioned perception testing device 625 can be the application program that specifically implements the aforementioned steps in this embodiment. In summary, when implementing the technical solution provided in this application through software or firmware, the relevant program code is stored in the memory 620 and is called and executed by the processor 610.
[0122] Input / output interface 613 is used to connect input / output modules to realize information input and output. Input / output modules can be configured as components in the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Input devices may include keyboards, mice, touch screens, microphones, various sensors, etc., and output devices may include displays, speakers, vibrators, indicator lights, etc.
[0123] Network interface 614 is used to connect a communication module (not shown in the figure) to enable communication between this device and other devices. The communication module can communicate via wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).
[0124] Bus 630 includes a pathway for transmitting information between various components of the device, such as processor 610, video display adapter 611, disk drive 612, input / output interface 613, network interface 614, and memory 620.
[0125] It should be noted that although the above-described device only shows the processor 610, video display adapter 611, disk drive 612, input / output interface 613, network interface 614, memory 620, bus 630, etc., in specific implementations, the device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the solution of this application, and does not necessarily include all the components shown in the figures.
[0126] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a computer program product. This computer program product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of this application.
[0127] The technical solutions provided in this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.
Claims
1. A method for testing a perception fusion system, the perception fusion system comprising a fusion module and at least two perception modules, characterized in that, The method includes: Acquire test samples, the test samples including perception sensor data and obstacle information labeled with the perception sensor data; The sensor data is input into the perception fusion system, and the outputs of each perception module and the fusion module in the perception fusion system are obtained respectively. The outputs of each sensing module and the fusion module are compared with the labeled obstacle information to obtain the encoding sequence corresponding to the comparison results of each sensing module and the fusion module. Based on the pre-configured correspondence between the encoding sequence and the test results, the test result corresponding to the obtained encoding sequence is determined, and the test result of the perception fusion system is determined based on the test result corresponding to the obtained encoding sequence; wherein the correspondence between the encoding sequence and the test result is pre-configured based on the topology of the perception fusion system.
2. The method according to claim 1, characterized in that, The step of acquiring the outputs of each sensing module and the fusion module in the sensing fusion system includes: The output of each sensing module is obtained by pre-inserting probes at the output end of each sensing module; The output of the fusion module can be obtained by using a probe pre-installed at the output end of the fusion module, or by directly obtaining the output of the fusion module.
3. The method according to claim 1 or 2, characterized in that, Based on the test results corresponding to the obtained encoded sequence, the test results of the perceptual fusion system are determined as follows: Based on the test results corresponding to the obtained encoded sequence, it is determined whether the perceptual fusion system passes the test; If the perception fusion system fails the test, the abnormal module in the perception fusion system is located based on the encoded sequence.
4. The method according to claim 3, characterized in that, Determining whether the perceptual fusion system passes the test based on the test results corresponding to the obtained encoded sequence includes: If the encoded sequence obtained from more than a preset first proportion or a first number of test samples indicates that the output of the fusion module is consistent with the labeled obstacle information, then the perception fusion system is determined to have passed the test; otherwise, the perception fusion system is determined to have failed the test.
5. The method according to claim 4, characterized in that, Locating the abnormal module in the perceptual fusion system based on the encoded sequence includes: If the encoded sequence obtained from test samples exceeding a preset second proportion or second number indicates that the output of the fusion module is inconsistent with the labeled obstacle information, then the fusion module is determined to be an abnormal module. If the encoded sequence obtained from test samples exceeding a preset third proportion or third number indicates that the output of one of the sensing modules is inconsistent with the labeled obstacle information, then one of the sensing modules is determined to be an abnormal module.
6. The method according to claim 1 or 2, characterized in that, The method further includes: The distribution information of the encoded sequence is output to adjust the test sample based on the distribution information.
7. A testing apparatus for a sensing fusion system, the fusion system comprising a fusion module and at least two sensing modules, characterized in that, The device includes: The sample acquisition unit is configured to acquire test samples, the test samples including sensor data and obstacle information labeled on the sensor data; The system interaction unit is configured to input the sensor data into the perception fusion system and acquire the outputs of each perception module and the fusion module in the perception fusion system, respectively. The information comparison unit is configured to compare the outputs of each sensing module and the fusion module with the labeled obstacle information to obtain the encoding sequence corresponding to the comparison results of each sensing module and the fusion module. The result determination unit is configured to determine the test result corresponding to the obtained encoding sequence based on the pre-configured correspondence between the encoding sequence and the test result, and to determine the test result of the perception fusion system based on the test result corresponding to the obtained encoding sequence; wherein the correspondence between the encoding sequence and the test result is pre-configured based on the topology of the perception fusion system.
8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps of the method according to any one of claims 1 to 6.
9. An electronic device, characterized in that, include: One or more processors; as well as A memory associated with the one or more processors, the memory being used to store program instructions that, when read and executed by the one or more processors, perform the steps of the method according to any one of claims 1 to 6.