Data debugging method, device, equipment and readable storage medium

Abstract

The present invention provides a data debugging method, device, equipment and readable storage medium. The data debugging method includes: obtaining a key target from problem data; setting the attribute values ​​of other targets in the problem data to the same attribute values ​​as the key target; inputting the set problem data into the debugged algorithm model for data re-injection in the time sequence of the problem data; and visually outputting the simulation results of the data re-injection. The present invention forms a unified observable output result by setting the attribute values ​​of other targets to the same attribute values ​​as the key target during the data re-injection process, and intuitively displays it in the curve chart of the coordinate axis, which is convenient for algorithm personnel to analyze and locate problems. Furthermore, the present invention performs multiple assignments when multiple problem time points appear in the problem data, and outputs multiple visualized results, which allows algorithm personnel to quickly analyze and locate problems.

Description

Technical Field

[0001] The present invention relates to the field of intelligent driving, and in particular to a data debugging method, device, equipment and readable storage medium. Background Art

[0002] The intelligentization of automobiles is a future trend, and intelligent driving is an important part of automobile intelligence. The typical architecture of automobile intelligent driving is generally composed of three modules: perception, decision-making and control. The perception module uses multiple sensors such as cameras, lidar, millimeter-wave radar, ultrasonic radar, gyroscope, accelerometer, etc. to perceive the surrounding environment information and vehicle status information. The perceived information includes the shape, direction, curvature, slope and lane of the road, traffic signs, signal lights, and the position, size, direction and speed of other vehicles or pedestrians.

[0003] In the process of developing intelligent driving products, it is necessary to collect a large amount of real vehicle test data and then input the test data into the data injection simulation environment for analysis and debugging to verify product performance, accelerate product development cycle, reduce the demand for real vehicle road testing, and improve the safety of intelligent driving products. Among them, the target screening algorithm is mainly used to screen dynamic traffic participants such as vehicles and pedestrians. The target screening model is an important algorithm module that connects the perception fusion and decision-making modules. For L2 intelligent driving, the perception fusion module generally outputs 8-32 dynamic and static road traffic participants, generally referred to as targets. Then the target screening model selects and outputs a main target based on the relative position, speed, acceleration and movement trend of the target model and the vehicle, and provides it to the subsequent decision-making planning model for planning and decision-making.

[0004] The target screening algorithm needs to perform numerical calculations on multiple different targets. Usually, a loop method is used to batch process the data of multiple targets. When observing and analyzing the output results, there is a problem that multiple output results are projected to the same time point in one operation cycle, such as Figure 1 As shown, Figure 1 For the key output signal curve diagram in the data re-injection simulation result visualization output curve diagram, refer to Figure 1 For example, at the 5th second, the different output results of 32 targets are projected onto the same vertical axis. It is impossible to visually observe the specific value of the target of interest, or it is difficult to determine the specific target of interest based on the value, which is not convenient for problem analysis and location. Summary of the invention

[0005] The main purpose of the present invention is to provide a data debugging method, device, equipment and readable storage medium, aiming to solve the technical problem in the field of intelligent driving that the degree of visualization is not high during the target screening and debugging of problem data, which is not convenient for problem analysis and positioning.

[0006] In a first aspect, the present invention provides a data debugging method, the data debugging method comprising:

[0007] Get key goals from the problem data;

[0008] Setting the attribute values ​​of other targets in the problem data to the same attribute value as the key target;

[0009] Input the set problem data into the debugged algorithm model in the order of the time when the problem data occurred to perform data reinjection;

[0010] The simulation results of data re-injection are output visually.

[0011] Optionally, the step of obtaining key objectives from the problem data includes:

[0012] Input the problem data into the debugged algorithm model in the order of the time when the problem data occurred, and output the simulation results of the data injection to form multiple visual curve graphs;

[0013] The horizontal axis of the key target curve graph represents time, and the vertical axis represents the key target;

[0014] A key target corresponding to a certain time point in the curve graph is obtained as the key target.

[0015] Optionally, the step of obtaining key objectives from the problem data includes:

[0016] When a problem data has multiple problem time points, obtain each problem time point and its corresponding key target.

[0017] Optionally, after the step of obtaining each problem time point and the key target corresponding thereto, the method further includes:

[0018] Set other target attributes before each problem time point to the same attribute values ​​as the key target corresponding to each problem time point;

[0019] Input the set problem data into the debugged algorithm model in the order of the time when the problem data occurred to perform data reinjection;

[0020] The simulation results of data re-injection are output visually.

[0021] Optionally, the step of visually outputting the simulation results of the data re-injection includes:

[0022] Output the simulation results of data re-injection to form multiple visual curve graphs;

[0023] The horizontal axis of the key output signal curve graph represents time, and the vertical axis represents the key output signal.

[0024] In a second aspect, the present invention further provides a data debugging device, the data debugging device comprising:

[0025] The acquisition module is used to obtain key objectives from the problem data;

[0026] A setting module, used to set the attribute values ​​of other targets in the problem data to the same attribute values ​​as the key target;

[0027] The re-injection module is used to input the set problem data into the debugged algorithm model in the chronological order of the problem data for data re-injection;

[0028] The output module is used to visualize the simulation results of data feedback.

[0029] Optionally, the acquisition module is used to:

[0030] When a problem data has multiple problem time points, obtain each problem time point and its corresponding key target.

[0031] Optionally, the setting module is used to:

[0032] Set other target attributes before each problem time point to the same attribute values ​​as the key target corresponding to each problem time point;

[0033] Input the set problem data into the debugged algorithm model in the order of the time when the problem data occurred to perform data reinjection;

[0034] The simulation results of data re-injection are output visually.

[0035] In a third aspect, the present invention also provides a data debugging device, comprising a processor, a memory, and a data debugging program stored in the memory and executable by the processor, wherein when the data debugging program is executed by the processor, the steps of the data debugging method described above are implemented.

[0036] In a fourth aspect, the present invention further provides a readable storage medium, on which a data debugging program is stored, wherein when the data debugging program is executed by a processor, the steps of the data debugging method as described above are implemented.

[0037] In the present invention, a key target is obtained from the problem data; the attribute values ​​of other targets in the problem data are set to the same attribute values ​​as the key target; the problem data after setting is input into the debugged algorithm model for data re-injection in the time sequence of the problem data; and the simulation results of the data re-injection are output in a visual manner. The present invention sets the attribute values ​​of other targets to the same attribute values ​​as the key target during the data re-injection process to form a unified observable output result, which is intuitively displayed in the curve chart of the coordinate axis, so that algorithm personnel can analyze and locate problems easily, and avoid the situation where it is impossible to observe which target is wrong at a certain problem time point due to different output results caused by different attribute values ​​of multiple targets when output to the curve chart. Furthermore, the present invention performs multiple assignments when multiple problem time points appear in the problem data and outputs multiple visualized results, which allows algorithm debuggers to quickly analyze and locate problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 Key output signal curves in the data re-injection simulation result visualization output curve graph;

[0039] Figure 2 A schematic diagram of the hardware structure of a data debugging device according to an embodiment of the present invention;

[0040] Figure 3 A schematic diagram of a flow chart of an embodiment of a data debugging method of the present invention;

[0041] Figure 4 A comparison diagram of the display effects of the key output signal curves before and after the present invention is used;

[0042] Figure 5 A schematic diagram of a method for locating a problem time of problem data according to an embodiment of a data debugging method of the present invention;

[0043] Figure 6 A schematic diagram of a method for locating a problem time of another problem data according to an embodiment of a data debugging method of the present invention;

[0044] Figure 7 It is a functional module diagram of an embodiment of a data debugging device of the present invention.

[0045] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0046] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0047] In a first aspect, an embodiment of the present invention provides a data debugging device, which may be a device having a data processing function, such as a personal computer (PC), a notebook computer, or a server.

[0048] Reference Figure 2 , Figure 2 It is a schematic diagram of the hardware structure of the data debugging device involved in the embodiment of the present invention. In the embodiment of the present invention, the data debugging device may include a processor 1001 (such as a central processing unit, CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these components; the user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard); the network interface 1004 may optionally include a standard wired interface, a wireless interface (such as wireless fidelity WIreless-FIdelity, WI-FI interface); the memory 1005 may be a high-speed random access memory (random access memory, RAM), or a stable memory (non-volatile memory), such as a disk storage, and the memory 1005 may optionally be a storage device independent of the aforementioned processor 1001. Those skilled in the art will understand that Figure 2 The hardware structure shown in the figure does not constitute a limitation of the present invention, and may include more or less components than those shown in the figure, or combine certain components, or arrange the components differently.

[0049] Continue to refer to Figure 2 , Figure 2 The memory 1005 as a computer storage medium may include an operating system, a network communication module, a user interface module and a data debugging program. The processor 1001 may call the data debugging program stored in the memory 1005 and execute the data debugging method provided in the embodiment of the present invention.

[0050] In a second aspect, an embodiment of the present invention provides a data debugging method.

[0051] In order to more clearly demonstrate the data debugging method provided in the embodiment of the present application, the application scenario of the data debugging method provided in the embodiment of the present application is first introduced.

[0052] The data debugging method provided in the embodiment of the present application is applied in the process of analyzing and debugging problem data in the field of intelligent driving. After the data acquisition environment is configured, the acquisition model collects input signals and key output signals generated in the actual vehicle test, marks the problem data in the signal data, and analyzes and locates the problem by re-injecting the problem data.

[0053] In one embodiment, referring to Figure 3 , Figure 3 FIG. 1 is a flow chart of an embodiment of a data debugging method of the present invention. Figure 3 As shown, the data debugging method includes:

[0054] Step S10, obtaining key objectives from the problem data.

[0055] In this embodiment, the problem data is data that is marked when it is found from the collected data that the actual output result of the algorithm is different from the expected result. For example, at the 10th second from the start of data collection, the value of a key output signal should change from 0 to 1, but in fact there is no change, and the actual output value is still 0. In this case, the problem data is marked, and the 10th second is the problem time point of the problem data.

[0056] In this embodiment, the target refers to 8-32 dynamic and static road traffic participants output by the perception fusion module in L2 intelligent driving, where dynamic targets include surrounding vehicles, pedestrians, traffic signals, etc., and static targets include road facilities, traffic rules, etc. The target screening algorithm compares and analyzes multiple targets with the relative position, speed, acceleration, motion trend and other conditions of the vehicle to select a main target as the key target. The key target is the target object that we should focus on in the process of debugging and analyzing the problem data.

[0057] Step S20: setting the attribute values ​​of other targets in the problem data to the same attribute value as the key target.

[0058] In this embodiment, the attribute values ​​of the target include relative position, speed, acceleration, etc. The problem data usually includes multiple targets. The attribute values ​​of the relative position, speed, acceleration, etc. of other targets in the problem data are set to be the same as the attribute values ​​of the relative position, speed, acceleration, etc. of the key target. In this way, the attribute values ​​of all targets are consistent and can be used to input the target attribute values ​​of the debugged algorithm model.

[0059] Step S30, input the set problem data into the debugged algorithm model in the chronological order of the problem data to perform data reinjection.

[0060] In this embodiment, by building a data re-injection simulation analysis environment, the problem data is reproduced in a desktop environment to facilitate the analysis and location of the problem. The data re-injection simulation analysis environment consists of a data reading module, a debugged algorithm model, and a data monitoring module. The debugged algorithm model includes algorithm models such as perception fusion, target screening, and decision planning. The collected problem data is read out in chronological order through the data reading module and then input into the debugged algorithm model to drive the algorithm model to run again. This process is called data re-injection.

[0061] In this embodiment, through the aforementioned step S20, the attribute values ​​of other targets in the problem data are set to the same attribute values ​​as the key target, so that the same target attribute values ​​are obtained for all targets. The modified problem data is input into the debugged algorithm model in the chronological order of the occurrence of the problem data for data reinjection.

[0062] Step S40: Visually output the simulation results of the data re-injection.

[0063] In this embodiment, the key output signals of the algorithm model and the process quantity data generated during the algorithm operation are input into the data monitoring module through data feedback simulation. The data monitoring module is responsible for outputting the signals of concern to form a visual curve graph, which is provided to algorithm personnel for problem analysis and location.

[0064] In this embodiment, after the previous steps, the output results of all targets will be consistent through data re-injection. By visualizing the simulation results of data re-injection, algorithm personnel can intuitively observe the output results. Figure 4 , Figure 4 This is a comparison diagram of the display effects of the key output signal curve graphs before and after the use of the present invention. Compared with before the use of the method of the present invention, the method of the present invention can avoid the situation where different output results of 32 targets at a certain problem time point are projected onto the same vertical axis when debugging problem data, and the specific values ​​of the targets of interest cannot be observed intuitively, or it is difficult to determine the specific targets of interest based on the values, which is not convenient for analyzing and locating the problem.

[0065] Furthermore, in one embodiment, step S10 includes:

[0066] Input the problem data into the debugged algorithm model in the order of the time when the problem data occurred, and output the simulation results of the data injection to form multiple visual curve graphs;

[0067] The horizontal axis of the key target curve graph represents time, and the vertical axis represents the key target;

[0068] A key target corresponding to a certain time point in the curve graph is obtained as the key target.

[0069] In this embodiment, the simulation results of data re-injection are outputted by the data monitoring module to form a plurality of visual curve graphs, among which the key target curve graphs are as follows: Figure 5 As shown, Figure 5 This is a schematic diagram of a method for locating the problem time of problem data in an embodiment of a data debugging method of the present invention, referring to Figure 5 In the key target curve chart, the horizontal axis represents time and the vertical axis represents the key target, for example Figure 5 The 10S time point corresponds to key target 1, and key target 1 is taken as the key target.

[0070] Furthermore, in one embodiment, step S10 includes:

[0071] When a problem data has multiple problem time points, obtain each problem time point and its corresponding key target.

[0072] In this embodiment, the output results of a certain problem data at multiple problem time points may be different from the expected value, such as Figure 6 As shown, Figure 6 This is a schematic diagram of a method for locating a problem time of another problem data in an embodiment of a data debugging method of the present invention, referring to Figure 6 From the key target curve chart, we can obtain each problem time point and its corresponding key target, and obtain key target 1 corresponding to the 2S, key target 2 corresponding to the 5S, and key target 3 corresponding to the 10S.

[0073] Furthermore, in one embodiment, after step S10, the following steps are further included:

[0074] Set other target attributes before each problem time point to the same attribute values ​​as the key target corresponding to each problem time point;

[0075] Input the set problem data into the debugged algorithm model in the order of the time when the problem data occurred to perform data reinjection;

[0076] The simulation results of data re-injection are output visually.

[0077] In this embodiment, based on the situation when multiple problem time points appear in a certain problem data, each problem time point and the key target corresponding to it are obtained respectively, and other target attributes before each problem time point are repeatedly set to the same attribute value as the key target corresponding to each problem time point; the problem data after setting is repeatedly input into the debugged algorithm model in the time sequence of the occurrence of the problem data for data reinjection; and the simulation results of the data reinjection are repeatedly visualized and output.

[0078] In this embodiment, when a problem data has multiple problem time points, the process is to loop through key target acquisition, target attribute value setting, data re-injection, and simulation result visualization output. In this way, the problem can be analyzed and located at multiple problem time points of the problem data in a loop.

[0079] In a third aspect, an embodiment of the present invention further provides a data debugging device.

[0080] Reference Figure 7 , Figure 7 It is a functional module diagram of an embodiment of a data debugging device of the present invention.

[0081] In this embodiment, the data debugging device includes:

[0082] An acquisition module 10, for acquiring key objectives from problem data;

[0083] A setting module 20, used to set the attribute values ​​of other targets in the problem data to the same attribute values ​​as the key target;

[0084] The re-injection module 30 is used to input the set problem data into the debugged algorithm model in the time sequence of the problem data to perform data re-injection;

[0085] The output module 40 is used to visually output the simulation results of the data injection.

[0086] Furthermore, in one embodiment, the acquisition module 10 is used to:

[0087] When a problem data has multiple problem time points, obtain each problem time point and its corresponding key target.

[0088] Furthermore, in one embodiment, a module 20 is provided for:

[0089] Set the attribute values ​​of each problem time point and other corresponding targets before it to the same attribute values ​​as the key target corresponding to each problem time point;

[0090] Input the set problem data into the debugged algorithm model in the order of the time when the problem data occurred to perform data reinjection;

[0091] The simulation results of data re-injection are output visually.

[0092] Furthermore, in one embodiment, the acquisition module 10 is used to:

[0093] Input the problem data into the debugged algorithm model in the order of the time when the problem data occurred, and output the simulation results of the data injection to form multiple visual curve graphs;

[0094] The horizontal axis of the key target curve graph represents time, and the vertical axis represents the key target;

[0095] A key target corresponding to a certain time point in the curve graph is obtained as the key target.

[0096] Furthermore, in one embodiment, the output module 40 is used to:

[0097] Output the simulation results of data re-injection to form multiple visual curve graphs;

[0098] The horizontal axis of the key output signal curve graph represents time, and the vertical axis represents the key output signal.

[0099] Among them, the functional implementation of each module in the above-mentioned data debugging device corresponds to each step in the above-mentioned data debugging method embodiment, and its functions and implementation processes are no longer repeated here one by one.

[0100] In a fourth aspect, an embodiment of the present invention further provides a readable storage medium.

[0101] The readable storage medium of the present invention stores a data debugging program, wherein when the data debugging program is executed by a processor, the steps of the above-mentioned data debugging method are implemented.

[0102] Among them, the method implemented when the data debugging program is executed can refer to the various embodiments of the data debugging method of the present invention, and will not be repeated here.

[0103] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0104] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0105] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes a number of instructions for a terminal device to execute the methods described in each embodiment of the present invention.

[0106] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A data debugging method, characterized in that: The data debugging method comprises: Get key goals from the problem data; Setting the attribute values ​​of other targets in the problem data to the same attribute value as the key target; Input the set problem data into the debugged algorithm model in the order of the time when the problem data occurred to perform data reinjection; Visualize the simulation results of data re-injection; The key objectives of obtaining problem data include: When a problem data has multiple problem time points, obtain each problem time point and its corresponding key target; After the step of obtaining each problem time point and the key target corresponding thereto, the following step is further included: Set other target attributes before each problem time point to the same attribute values ​​as the key target corresponding to each problem time point; Input the set problem data into the debugged algorithm model in the order of the time when the problem data occurred to perform data reinjection; The simulation results of data re-injection are output visually.

2. The data debugging method according to claim 1, characterized in that: The key objectives of obtaining problem data include: Input the problem data into the debugged algorithm model in the order of the time when the problem data occurred, and output the simulation results of the data injection to form multiple visual curve graphs; The horizontal axis of the key target curve graph represents time, and the vertical axis represents the key target; A key target corresponding to a certain time point in the curve graph is obtained as the key target.

3. The data debugging method according to claim 1, characterized in that: The visual output of the simulation results of data re-injection includes: Output the simulation results of data re-injection to form multiple visual curve graphs; The horizontal axis of the key output signal curve graph represents time, and the vertical axis represents the key output signal.

4. A data debugging device, characterized in that: The data debugging device comprises: The acquisition module is used to obtain key objectives from the problem data; A setting module, used to set the attribute values ​​of other targets in the problem data to the same attribute values ​​as the key target; The re-injection module is used to input the set problem data into the debugged algorithm model in the chronological order of the problem data for data re-injection; Output module, used to visualize the simulation results of data injection; The acquisition module is used to: When a problem data has multiple problem time points, obtain each problem time point and its corresponding key target; The setting module is used to: Set other target attributes before each problem time point to the same attribute values ​​as the key target corresponding to each problem time point; Input the set problem data into the debugged algorithm model in the order of the time when the problem data occurred to perform data reinjection; The simulation results of data re-injection are output visually.

5. A data debugging device, characterized in that: The data debugging device comprises a processor, a memory, and a data debugging program stored in the memory and executable by the processor, wherein when the data debugging program is executed by the processor, the steps of the data debugging method according to any one of claims 1 to 3 are implemented.

6. A readable storage medium, characterized in that: The readable storage medium stores a data debugging program, wherein when the data debugging program is executed by a processor, the steps of the data debugging method according to any one of claims 1 to 3 are implemented.

Images