Control Method, Device, and Computer Equipment for Autonomous Driving

Abstract

The present disclosure relates to a control method, device, and computer device for autonomous driving. The method includes: acquiring current image data and historical image data of a road collected in real time; extracting a current drivable area according to the current image data, and obtaining current smooth data of the road according to the current drivable area and the historical image data; extracting a target area according to the current image data, and obtaining current feature data of the road according to the target area; fusing the current smooth data and the current feature data to generate an input signal; calculating the input signal using a preset algorithm, and outputting a control signal, where the control signal is used to instruct the vehicle to perform autonomous driving. The present disclosure can suppress the driving jitter problem of autonomous driving implemented directly based on an end-to-end algorithm, thereby improving driving comfort.

Description

Technical Field

[0001] The present disclosure relates to the field of autonomous driving technology, and particularly to a control method, device, and computer device for autonomous driving. Background Art

[0002] With the continuous development of artificial intelligence technology, it is more realistic for the automotive industry to apply artificial intelligence technology to solve the problem of autonomous driving. Currently, there are mainly two directions to achieve autonomous driving. One is to decompose the autonomous driving tasks and use different algorithms to solve the problems of the decomposed tasks respectively. For example, deep learning algorithms are used for the detection and recognition of road information, mapping and positioning modules are used for mapping and positioning, and decision-making planning and control are used to determine how to drive. The other direction is to directly use end-to-end algorithms to achieve autonomous driving without decomposing the autonomous driving module, such as reinforcement learning, imitation learning, etc.

[0003] The end-to-end algorithm can theoretically simplify the design of autonomous driving and achieve autonomous driving from the perception end to the control end. Compared with the way of decomposing problems, it has much less work content and also conforms to the human driving mode. Information is obtained from the images at the perception end and control is achieved at the control end without knowing the accurate surrounding information. Therefore, the end-to-end algorithm is more widely used in the field of autonomous driving. However, the actions output by the end-to-end algorithm to the environment are only related to the input of the algorithm. The input of the end-to-end algorithm has the limitation that the input state is relatively single. Looking at the time axis, the output results are not smooth and fluctuate greatly, resulting in obvious jitter in the driving trajectory of the vehicle. Summary of the Invention

[0004] Accordingly, in view of the above technical problems, it is necessary to provide a control method, device, computer device, storage medium, and computer program product for autonomous driving.

[0005] In a first aspect, the present disclosure provides a control method for autonomous driving. The method includes:

[0006] Obtain the current image data and historical image data of the road collected in real time; the historical image data includes the road image data of a preset duration closest to the current moment;

[0007] Obtain the current smooth data of the road according to the current image data and the historical image data;

[0008] Obtain the current feature data of the road according to the current image data;

[0009] Fuse the current smooth data and the current feature data to generate an input signal;

[0010] Calculate the input signal using a preset algorithm and output a control signal for instructing the vehicle to perform autonomous driving.

[0011] In one embodiment, the obtaining of the current smooth data of the road according to the current image data and historical image data includes:

[0012] Perform data processing on the current image data, set the pixels of non-drivable areas to a specified value, and extract the pixel values of the drivable area to obtain current first image data;

[0013] Obtain the historical first image data and historical smooth data of the historical image data;

[0014] Input the current first image data into a preset recurrent neural network, and the recurrent neural network outputs the current smooth data according to the current first image data, historical first image data, and historical smooth data.

[0015] In one embodiment, the obtaining of the current feature data of the road according to the current image data includes:

[0016] Extract a target area according to the current first image data to obtain second image data;

[0017] Process the second image data into a one-dimensional vector to obtain the current feature data.

[0018] In one embodiment, the obtaining of the current feature data of the road according to the current image data includes:

[0019] Perform data processing on the current image data, extract a target area, and obtain second image data;

[0020] Process the second image data into a one-dimensional vector to obtain the current feature data.

[0021] In one embodiment, the fusing of the current smooth data and current feature data to generate an input signal includes:

[0022] Obtain the current smooth data and current feature data;

[0023] Concatenate the current smooth data and current feature data or calculate the input signal according to a preset algorithm.

[0024] In one embodiment, the calculating of the input signal using a preset algorithm and outputting a control signal includes:

[0025] Input the input signal into a preset reinforcement learning algorithm;

[0026] Obtain the control signal output by the reinforcement learning algorithm.

[0027] In a second aspect, the present disclosure also provides a control device for autonomous driving. The device includes:

[0028] An image acquisition module, configured to obtain current image data and historical image data of the road collected in real time; the historical image data includes road image data of a preset duration closest to the current moment;

[0029] A smoothing data module, configured to obtain current smoothing data of the road according to the current image data and the historical image data;

[0030] A feature data module, configured to obtain current feature data of the road according to the current image data;

[0031] A fusion module, configured to fuse the current smoothing data and the current feature data to generate an input signal;

[0032] An output module, configured to calculate the input signal using a preset algorithm and output a control signal, where the control signal is used to instruct the vehicle to perform autonomous driving.

[0033] In one embodiment, the smoothing data module includes:

[0034] A current first image data unit, configured to perform data processing on the current image data, set the pixels of the non-drivable area to a specified value, and extract the pixel values of the drivable area to obtain current first image data;

[0035] A historical first image data unit, configured to obtain historical first image data and historical smoothing data of the historical image data;

[0036] A current smoothing data unit, configured to input the current first image data into a preset recurrent neural network, and the recurrent neural network outputs the current smoothing data according to the current first image data, the historical first image data, and the historical smoothing data.

[0037] In one embodiment, the feature data module includes:

[0038] A second image data unit, configured to extract a target area according to the current first image data to obtain second image data;

[0039] A current feature data unit, configured to process the second image data into a one-dimensional vector to obtain the current feature data.

[0040] In one embodiment, the feature data module includes:

[0041] A second image data unit for processing the current image data, extracting a target area, and obtaining second image data;

[0042] A current feature data unit for processing the second image data into a one-dimensional vector to obtain the current feature data.

[0043] In one embodiment, the fusion module includes:

[0044] A data acquisition unit for acquiring current smoothed data and current feature data;

[0045] A data fusion unit for splicing the current smoothed data and the current feature data or calculating the input signal according to a preset algorithm.

[0046] In one embodiment, the output module includes:

[0047] An algorithm input unit for inputting the input signal into a preset reinforcement learning algorithm;

[0048] An algorithm output unit for obtaining a control signal output by the reinforcement learning algorithm.

[0049] In a third aspect, the present disclosure also provides a computer device. The computer device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps of the above control method for autonomous driving are implemented.

[0050] In a fourth aspect, the present disclosure also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above control method for autonomous driving are implemented.

[0051] In a fifth aspect, the present disclosure also provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the steps of the above control method for autonomous driving are implemented.

[0052] The above control method, device, computer device, storage medium, and computer program product for autonomous driving at least include the following beneficial effects:

[0053] By fusing the current smoothed data and the current feature data to generate an input signal, the present disclosure enables the subsequent output control signal to be jointly determined by the current smoothed data and the current feature data, introduces the current smoothed data into the current feature data, and can suppress the driving jitter problem of autonomous driving implemented directly based on an end-to-end algorithm, thereby improving driving comfort. Description of the Drawings

[0054] To more clearly illustrate the technical solutions in the embodiments of the present disclosure or in the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only some embodiments of the present disclosure. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0055] Figure 1 It is an application environment diagram of a control method for autonomous driving in an embodiment;

[0056] Figure 2 It is a flowchart of a control method for autonomous driving in an embodiment;

[0057] Figure 3 It is a flowchart of controlling a vehicle for autonomous driving in an embodiment;

[0058] Figure 4 It is a flowchart of the step of obtaining current smoothed data in an embodiment;

[0059] Figure 5 It is a flowchart of the step of obtaining current feature data in an embodiment;

[0060] Figure 6 It is another flowchart of the step of obtaining current feature data in an embodiment;

[0061] Figure 7 It is a flowchart of the step of obtaining an input signal in an embodiment;

[0062] Figure 8 It is a flowchart of the step of obtaining an output signal in an embodiment;

[0063] Figure 9 It is a structural block diagram of a control device for autonomous driving in an embodiment;

[0064] Figure 10 It is a structural block diagram of a smoothed data module in an embodiment;

[0065] Figure 11 It is a structural block diagram of a feature data module in an embodiment;

[0066] Figure 12 It is a structural block diagram of a fusion module in an embodiment;

[0067] Figure 13 It is a structural block diagram of an output module in an embodiment;

[0068] Figure 14 It is an internal structural block diagram of a computer device in an embodiment. Detailed implementation manners

[0069] In order to make the objectives, technical solutions and advantages of the present application more clearly understood, the following further describes the present application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0070] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which this disclosure belongs. The terms used in the specification of this disclosure herein are only for the purpose of describing specific embodiments and are not intended to limit this disclosure.

[0071] It should be noted that the terms "first", "second", etc. in the specification and claims of this disclosure and the above-mentioned drawings are used to distinguish similar objects and do not necessarily have to be used to describe a specific order or sequence. It should be understood that such used data can be interchanged under appropriate circumstances so that the embodiments of this disclosure described herein can be implemented in an order other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. On the contrary, they are merely examples of devices and methods consistent with some aspects of this disclosure as detailed in the appended claims. The term "comprising", "including" or any other variant thereof is intended to cover a non-exclusive inclusion, such that a process, method, product or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or further includes elements inherent to such process, method, product or device. Without further limitation, there is no exclusion of the presence of additional identical or equivalent elements in the process, method, product or device comprising the said elements. For example, if the terms first, second, etc. are used to denote names, they do not denote any specific order.

[0072] As used herein, the singular forms "a", "an" and "the" may also include the plural forms unless the context clearly dictates otherwise. It should also be understood that the terms "including" / "comprising" or "having" etc. specify the presence of the stated features, wholes, steps, operations, components, parts or combinations thereof, but do not preclude the presence or addition of one or more other features, wholes, steps, operations, components, parts or combinations thereof. At the same time, in this specification, the term "and / or" includes any and all combinations of the related listed items.

[0073] The control method for autonomous driving provided by the embodiments of the present application can be applied to, for example Figure 1In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or placed on the cloud or other network servers. Among them, the terminal 102 can be equipped on an autonomous vehicle, and the terminal 102 can be integrated with an image acquisition device, a positioning device, etc. The autonomous vehicle does not depend on the driving operation of the driver and autonomously travels on the driving path from the departure location to the destination location of the autonomous vehicle. The autonomous vehicle can include driverless taxis, driverless buses, etc. The server 104 can be implemented by an independent server or a server cluster composed of multiple servers.

[0074] In some embodiments of the present disclosure, as Figure 2 shown, a control method for autonomous driving is provided, and this method is applied to Figure 1 the terminal 102 or the server 104 in it as an example for illustration. The method includes the following steps:

[0075] Step S10: Obtain the current image data and historical image data of the road collected in real time; the historical image data includes the road image data of a preset duration closest to the current moment.

[0076] Specifically, the image data of the surrounding environment of the autonomous vehicle can be collected in real time through an image acquisition device such as an in-vehicle camera and stored. Obtain the collected current image data and historical image data. Among them, the historical image data can be the road image data of a preset duration closest to the current moment. For example, the historical image data is all the image data within a 15-second time period before the current moment.

[0077] Step S20: Obtain the current smooth data of the road according to the current image data and the historical image data.

[0078] Specifically, the collected current image data is processed using an artificial algorithm and stored for subsequent acquisition of historical image data. In this embodiment, the current image data and the historical image data can be used as the initial data for data processing. For example, a visual perception algorithm can be used for data processing. The current image processing collects the image information of the surrounding environment of the autonomous vehicle, including at least the environmental information in the forward direction of the vehicle head of the autonomous vehicle.

[0079] Further, the current image data and the historical image data can be directly used as inputs, or the current image data and the historical image data after data processing can be used as inputs, and smooth data is calculated and output through a preset algorithm. The smooth data can be used to characterize the shape of the road and the occupied space of other vehicles and obstacles. That is, the smooth data is related to the road shape information and the road condition information. For example, on a road without other vehicles and obstacles, the smooth data is mainly determined by the road shape information; if there are other vehicles or obstacles on the road, it is determined by the road shape information and the occupied space of other vehicles or obstacles.

[0080] Step S30: Obtain the current feature data of the road according to the current image data.

[0081] Specifically, the collected current image data is processed using an artificial algorithm to obtain the current feature data. In this embodiment, the current image data can be used as the initial data for data processing. For example, a visual perception algorithm can be used for data processing. The feature data is calculated and extracted from the current image data through a preset algorithm. The feature data can be information that is effective for the subsequent output control signal.

[0082] Step S40: Fuse the current smooth data and the current feature data to generate an input signal.

[0083] Specifically, the current smooth data obtained in step S20 and the current feature data obtained in step S30 are fused to obtain an input signal, that is, the subsequent output control signal is jointly determined by the current smooth data and the current feature data.

[0084] Step S50: Calculate the input signal using a preset algorithm and output a control signal, where the control signal is used to instruct the vehicle to perform autonomous driving.

[0085] Specifically, the generated input signal is input into a preset algorithm. In this embodiment, the final control signal is output based on an end-to-end algorithm. The vehicle is instructed to perform autonomous driving according to the control signal. Combining the process schematic Figure 3 , where the first AI algorithm and the second AI algorithm can be different algorithms or the same algorithm, and are adjusted according to the actual situation.

[0086] In the above control method for autonomous driving, by fusing the current smooth data and the current feature data to generate an input signal, the subsequent output control signal is jointly determined by the current smooth data and the current feature data, introducing the current smooth data into the current feature data, which can suppress the driving jitter problem of autonomous driving directly based on the end-to-end algorithm, and thus improve the driving comfort.

[0087] In some embodiments of the present disclosure, as Figure 4 shown, step S20 includes:

[0088] Step S22: Perform data processing on the current image data, set the pixels of the non-drivable area to a specified value, and extract the pixel values of the drivable area to obtain the current first image data.

[0089] Specifically, when performing data processing on the current image data, the current image data can be divided into a drivable area and a non-drivable area. The drivable area may refer to the area where an autonomous vehicle can safely travel along a set route without colliding with surrounding objects, and the drivable area does not contain obstacles. The non-drivable area may refer to the area that is not in the drivable area, such as an area with obstacles or an area that is not drivable on the road.

[0090] When determining the drivable area, it can be processed based on the pixel values in the current image data; or based on a deep learning semantic segmentation network or instance segmentation network, segment the current image data and mark the drivable area; or obtain the drivable area by determining the boundary of the drivable area.

[0091] Furthermore, in this embodiment, processing is performed based on the pixel values in the current image data, and the pixels of the non-drivable area in the current image data are set to a specified value. The choice of the specified value is not unique, and a specified value that is different from the pixels of the drivable area can be selected. By setting the specified value for the pixels of the non-drivable area, the drivable area and the non-drivable area in the current image data can be distinguished by the pixel values. The current first image data is obtained by extracting the pixel values of the drivable area.

[0092] Step S24: Obtain the historical first image data and historical smoothing data of the historical image data.

[0093] Specifically, by storing the current first image data and the current smoothing data, it is convenient to obtain the historical first image data and historical smoothing data subsequently. The historical first image data and historical smoothing data can be stored in a database or cache.

[0094] Step S26: Input the current first image data into a preset recurrent neural network, and the recurrent neural network outputs the current smoothing data according to the current first image data, historical first image data, and historical smoothing data.

[0095] Specifically, a recurrent neural network generally refers to a type of recursive neural network that takes sequential data as input, recurs in the evolution direction of the sequence, and all nodes (recurrent units) are connected in a chain. In this embodiment, the current first image data is used as the input of the recurrent neural network. When the recurrent neural network performs computational processing on the current first image data, it also refers to the historical first image data, and calculates and outputs the current smoothed data by combining the current first image data and the historical first image data. That is, the current smoothed data depends not only on the current first image data but also on the historical first image data.

[0096] In this embodiment, the current smoothed data is calculated by combining the current first image data and the historical first image data through a recurrent neural network. The recurrent neural network has memory and can calculate the current smoothed data according to the road conditions.

[0097] In some embodiments of the present disclosure, as Figure 5 shown, step S30 includes:

[0098] Step S312: Extract the target area according to the current first image data to obtain the second image data.

[0099] Specifically, when obtaining the second image data, it can be processed based on the current first image data in step S20 to extract the target area. Based on the first image data, the target area is extracted according to actual needs. The target area can be for the end-to-end algorithm and can be the region of interest of the end-to-end algorithm. The data required by the end-to-end algorithm is obtained by extracting the target area. The method of extracting the target area can refer to the method of extracting the drivable area and will not be elaborated here.

[0100] Step S314: Process the second image data into a one-dimensional vector to obtain the current feature data.

[0101] Specifically, the second image data is processed into a one-dimensional vector to obtain the current feature data, so that the input data generated after introducing the smoothed data is a sequence that can be processed by the recurrent neural network.

[0102] In this embodiment, by reusing the current first image data in the foregoing steps when obtaining the current feature data and obtaining the target area based on the current first image data, the extraction steps of the target area are simplified, and the data processing efficiency and accuracy are improved.

[0103] In some embodiments of the present disclosure, as Figure 6 shown, step S30 includes:

[0104] Step S322: Perform data processing on the current image data, extract the target area, and obtain the second image data.

[0105] Specifically, when obtaining the second image data, it is directly processed based on the currently acquired image data to extract the target area. Based on the currently acquired image data, the target area is extracted according to actual requirements.

[0106] Step S324: Process the second image data into a one-dimensional vector to obtain the current feature data.

[0107] Specifically, the second image data is processed into a one-dimensional vector to obtain the current feature data, so that the input data generated after introducing the smoothing data into the current feature data is a sequence that can be processed by a recurrent neural network.

[0108] In this embodiment, when obtaining the current feature data, it is directly processed based on the currently acquired image data to extract the target area, so that obtaining the current smoothing data and obtaining the current feature data are independent of each other, which is convenient for their respective optimization and expansion.

[0109] In some embodiments of the present disclosure, as Figure 7 shown, step S40 includes:

[0110] Step S42: Obtain the current smoothing data and the current feature data.

[0111] Specifically, to ensure the timeliness of the current smoothing data and the current feature data, the current smoothing data and the current feature data are read in real time.

[0112] Step S42: Concatenate the current smoothing data and the current feature data or calculate the input signal according to a preset algorithm.

[0113] Specifically, the current smoothing data and the current feature data are fused. For example, the current smoothing data and the current feature data can be directly concatenated, or weighted summation, multiplication, etc. can be performed to calculate the input signal.

[0114] In this embodiment, by fusing the current smoothing data and the current feature data and introducing the smoothing data into the current feature data, the obtained input signal is not only related to the current image data but also depends on the historical image data, improving the smoothness of the input signal.

[0115] In some embodiments of the present disclosure, as Figure 8 shown, step S50 includes:

[0116] Step S52: Input the input signal into a preset reinforcement learning algorithm.

[0117] Specifically, a preset reinforcement learning algorithm can be obtained through training. By inputting an input signal into the trained reinforcement learning algorithm, a control signal output by the reinforcement learning algorithm can be obtained. For example, tags can be added to the target regions in the currently collected image data, and a deep learning semantic segmentation network can be trained using the tagged image set. The currently collected image data is segmented using the trained deep learning semantic segmentation network, post-processed for the segmented image, and current feature data is extracted. The current feature data is used to train a deep reinforcement learning algorithm model with the input signal after current smoothing data fusion, so as to enable the vehicle to travel in a specified lane in a simulation environment. The current feature data is useful information for training the deep reinforcement learning algorithm. On the one hand, it can reduce the training amount and calculation amount by reducing the data, thereby improving the efficiency. On the other hand, after reducing the inefficient and invalid information, the learning and exploration scope during the training of the deep reinforcement learning algorithm is greatly reduced, which can also improve the training efficiency.

[0118] Step S54: Obtain the control signal output by the reinforcement learning algorithm.

[0119] Specifically, the image data collected in the real environment is processed to obtain current feature data, and the input signal after current smoothing data fusion of the current feature data is input into the reinforcement learning algorithm to obtain the control signal output by the reinforcement learning algorithm.

[0120] In this embodiment, end-to-end autonomous driving is achieved through a reinforcement learning algorithm, and the smoothness of autonomous driving is controlled using smooth data, introducing front-to-back smoothness in autonomous driving, effectively suppressing the driving jitter problem of autonomous driving directly implemented based on the end-to-end algorithm, and thus improving driving comfort.

[0121] It should be understood that although the steps in the flowcharts involved in the above-described embodiments are shown in sequence according to the arrows, these steps do not necessarily need to be executed in the order indicated by the arrows. Unless there is a clear indication in this article, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, at least some of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages do not necessarily need to be executed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential either, but can be executed alternately or in turn with at least some of the steps or stages in other steps or other steps.

[0122] According to the same inventive concept, embodiments of the present disclosure further provide a control device for autonomous driving for implementing the control method for autonomous driving involved above. The solution provided by this device for solving problems is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the control device for autonomous driving provided below can refer to the limitations for the control method for autonomous driving in the foregoing text, and will not be repeated here.

[0123] The device may include a system (including a distributed system), software (application), module, component, server, client, etc. that uses the method described in the embodiments of this specification and combines the necessary implementation hardware. According to the same innovative concept, the devices in one or more embodiments provided by the embodiments of the present disclosure are as described in the following embodiments. Since the solution for the device to solve problems is similar to the method, the implementation of the specific device in the embodiments of this specification can refer to the implementation of the foregoing method, and the repeated parts will not be described again. As used below, the term "unit" or "module" may be a combination of software and / or hardware that can implement a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation in hardware, or a combination of software and hardware is also possible and contemplated.

[0124] In some embodiments of the present disclosure, as Figure 9 shown, a control device for autonomous driving is provided, and the device may be the aforementioned terminal. The device Z00 may include:

[0125] An image acquisition module Z10 for acquiring current image data and historical image data of the road collected in real time; the historical image data includes road image data of a preset duration closest to the current moment;

[0126] A smoothing data module Z20 for obtaining current smoothing data of the road according to the current image data and the historical image data;

[0127] A feature data module Z30 for obtaining current feature data of the road according to the current image data;

[0128] A fusion module Z40 for fusing the current smoothing data and the current feature data to generate an input signal;

[0129] An output module Z50 for calculating the input signal using a preset algorithm and outputting a control signal, where the control signal is used to instruct the vehicle to perform autonomous driving.

[0130] In some embodiments of the present disclosure, as Figure 10 shown, the smoothing data module Z20 includes:

[0131] The current first image data unit Z22 is used to perform data processing on the current image data, set the pixels of the non-drivable area to a specified value, and extract the pixel values of the drivable area to obtain the current first image data;

[0132] The historical first image data unit Z24 is used to obtain the historical first image data and historical smoothing data of the historical image data;

[0133] The current smoothing data unit Z26 is used to input the current first image data into a preset recurrent neural network, and the recurrent neural network outputs the current smoothing data according to the current first image data, historical first image data, and historical smoothing data.

[0134] In some embodiments of the present disclosure, as Figure 11 shown, the feature data module Z30 includes:

[0135] The second image data unit Z32 is used to extract a target area according to the current first image data to obtain second image data;

[0136] The current feature data unit Z34 is used to process the second image data into a one-dimensional vector to obtain the current feature data.

[0137] In some embodiments of the present disclosure, as Figure 11 shown, the feature data module Z30 includes:

[0138] The second image data unit Z32 is used to perform data processing on the current image data, extract a target area, and obtain second image data;

[0139] The current feature data unit Z34 is used to process the second image data into a one-dimensional vector to obtain the current feature data.

[0140] In some embodiments of the present disclosure, as Figure 12 shown, the fusion module Z40 includes:

[0141] The data acquisition unit Z42 is used to acquire the current smoothing data and current feature data;

[0142] The data fusion unit Z44 is used to splice the current smoothing data and current feature data or calculate the input signal according to a preset algorithm.

[0143] In some embodiments of the present disclosure, as Figure 13 shown, the output module Z50 includes:

[0144] The algorithm input unit Z52 is used to input the input signal into a preset reinforcement learning algorithm;

[0145] The algorithm output unit Z54 is used to obtain the control signal output by the reinforcement learning algorithm.

[0146] Each module in the above-mentioned control device for automatic driving can be implemented in whole or in part by software, hardware and a combination thereof. Each of the above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules. It should be noted that the division of modules in the embodiments of the present disclosure is schematic and is only a logical function division. There may be other division methods in actual implementation.

[0147] According to the aforementioned embodiment description of the control method for autonomous driving, in another embodiment provided by the present disclosure, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 14 As shown. The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a control method for automatic driving is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a button, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0148] Those skilled in the art will understand that the structure shown in the figure is merely a block diagram of a partial structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0149] According to the aforementioned description of the embodiment of the control method for autonomous driving, in another embodiment provided in the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0150] According to the description of the foregoing embodiments of the control method for autonomous driving, in another embodiment provided by the present disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the foregoing method embodiments.

[0151] Those of ordinary skill in the art can understand that all or part of the processes in the above method embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the foregoing method embodiments. Among them, any reference to a memory, database, or other medium used in the embodiments provided by the present application can include at least one of non-volatile and volatile memories. Non-volatile memory can include Read-Only Memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided by the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided by the present application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., without limitation.

[0152] In the description of this specification, the description with reference to terms such as "some embodiments", "other embodiments", "ideal embodiments", etc. means that the specific features, structures, materials, or features described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic description of the above terms does not necessarily refer to the same embodiment or example.

[0153] It can be understood that the various embodiments of the above methods in this specification are all described in a progressive manner. For the same / similar parts among the various embodiments, reference can be made to each other, and the key points of each embodiment are the differences from other embodiments. For the relevant parts, reference can be made to the descriptions of other method embodiments.

[0154] The technical features of the above-described embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features of the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as within the scope described in this specification.

[0155] The above-described embodiments only represent several implementation manners of the present disclosure. The description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the patent application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present disclosure, several modifications and improvements can still be made, and these all belong to the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure patent shall be subject to the appended claims.

Claims

1. A control method for autonomous driving, characterized in that, The method includes: Obtaining current image data and historical image data of a road collected in real time; the historical image data includes road image data of a preset duration closest to the current moment; Obtaining current smooth data of the road according to the current image data and the historical image data; wherein, the smooth data is used to represent the shape of the road and the occupied spaces of other vehicles and obstacles; The obtaining of the current smooth data of the road according to the current image data and the historical image data includes: performing data processing on the current image data, setting the pixels of non-drivable areas to a specified value, and extracting the pixel values of the drivable areas to obtain current first image data; obtaining historical first image data and historical smooth data of the historical image data; inputting the current first image data into a preset recurrent neural network, and the recurrent neural network outputs the current smooth data according to the current first image data, the historical first image data, and the historical smooth data; Obtaining current feature data of the road according to the current image data; Fusing the current smooth data and the current feature data to generate an input signal; Calculating the input signal by using a preset algorithm, and outputting a control signal, where the control signal is used to instruct the vehicle to perform autonomous driving.

2. The method according to claim 1, wherein The obtaining of the current feature data of the road according to the current image data includes: Extracting a target area according to the current first image data to obtain second image data; Processing the second image data into a one-dimensional vector to obtain the current feature data.

3. The method according to claim 1, wherein The obtaining of the current feature data of the road according to the current image data includes: Performing data processing on the current image data, extracting a target area, and obtaining second image data; Processing the second image data into a one-dimensional vector to obtain the current feature data.

4. The method according to claim 1, wherein The fusing of the current smooth data and the current feature data to generate an input signal includes: Obtaining the current smooth data and the current feature data; Performing splicing on the current smooth data and the current feature data or calculating the input signal according to a preset algorithm.

5. The method according to claim 1, wherein The calculating of the input signal by using a preset algorithm and outputting a control signal includes: Inputting the input signal into a preset reinforcement learning algorithm; Obtaining the control signal output by the reinforcement learning algorithm.

6. A control device for autonomous driving, characterized in that, The device includes: An image acquisition module, configured to obtain current image data and historical image data of a road collected in real time; the historical image data includes road image data of a preset duration closest to the current moment; A smooth data module, configured to obtain current smooth data of the road according to the current image data and the historical image data; wherein, the smooth data is used to represent the shape of the road and the occupied spaces of other vehicles and obstacles; The smooth data module includes: a current first image data unit for processing the current image data, setting the pixels of the non-drivable area to a specified value, and extracting the pixel values of the drivable area to obtain the current first image data; a historical first image data unit for acquiring the historical first image data and historical smooth data of the historical image data; a current smooth data unit for inputting the current first image data into a preset recurrent neural network, and the recurrent neural network outputs the current smooth data according to the current first image data, historical first image data, and historical smooth data; A feature data module for extracting and acquiring the current feature data of the road according to the current image data; A fusion module for fusing the current smooth data and the current feature data to generate an input signal; An output module for calculating the input signal using a preset algorithm and outputting a control signal, and the control signal is used to instruct the vehicle to perform autonomous driving.

7. The device according to claim 6, characterized in that, The feature data module includes: A second image data unit for extracting a target area according to the current first image data to obtain second image data; A current feature data unit for processing the second image data into a one-dimensional vector to obtain the current feature data.

8. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, the steps of the method according to any one of claims 1 to 5 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the steps of the method according to any one of claims 1 to 5 are implemented.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images