Autonomous driving method and system

Abstract

The application provides an automatic driving method and system, the method is applied to a first edge computing server, comprising: receiving a road condition image sent by an automatic driving vehicle in real time, and obtaining a hash value corresponding to the road condition image; respectively obtaining a Hamming distance between the hash value corresponding to the road condition image and each hash value stored in a training image hash value database; determining a corresponding automatic driving operation based on a comparison result between the Hamming distance and a preset threshold. Through the automatic driving method and system provided by the application, the safety of the automatic driving vehicle can be effectively improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automatic driving, in particular to an automatic driving method and system. BACKGROUND

[0002] At present, the safety risks of automatic driving vehicles mainly occur when the vehicles cannot make correct driving judgments on the current road conditions. There are mainly two types of methods to deal with the risks: one is that human drivers actively take over the vehicles before the danger comes; the other is that automatic driving vehicles actively ask human drivers to take over when they judge that the current road conditions are beyond their processing capacity, mainly in two cases: 1) the recognition accuracy (confidence) of the perception layer algorithms such as deep learning algorithms (CNN (Convolutional Neutral Network), RNN (Recurrent Neural Network)) on the surrounding environment is lower than a certain threshold, prompting human drivers to actively take over the vehicles; 2) the confidence of the decision layer algorithms such as fuzzy reasoning, Bayesian network, neural network and reinforcement learning is lower than a certain threshold, actively reminding human drivers to take over the vehicles.

[0003] However, the existing processing methods for dealing with the safety risks of automatic driving vehicles have the following disadvantages: 1) it is difficult for human drivers to always maintain full concentration and be ready to take over the vehicles at any time; 2) there is a problem of false negatives in the judgment of the current road conditions by the automatic driving vehicles, that is, the real road conditions have exceeded their processing capacity, but the automatic driving vehicles still drive according to a certain road condition mode learned by them, resulting in traffic accidents. SUMMARY

[0004] In view of the problems existing in the prior art, the embodiments of the present application provide an automatic driving method and system.

[0005] In a first aspect, the present application provides an automatic driving method applied to a first edge computing server, comprising:

[0006] receiving a road condition image sent by an automatic driving vehicle in real time, and obtaining a hash value corresponding to the road condition image;

[0007] respectively obtaining a Hamming distance between the hash value corresponding to the road condition image and each hash value stored in a training image hash value database;

[0008] determining a corresponding automatic driving operation based on a comparison result between the Hamming distance and a preset threshold.

[0009] Optionally, the determining of the corresponding automatic driving operation based on the comparison result between the Hamming distance and the preset threshold comprises:

[0010] If it is determined that the minimum value in the Hamming distance is less than or equal to a preset threshold, the autonomous vehicle is allowed to automatically drive according to a pre-trained driving strategy; or

[0011] If it is determined that the minimum value in the Hamming distance is greater than the preset threshold, an instruction is sent to the autonomous vehicle that a human driver needs to take over the vehicle or automatically park the vehicle.

[0012] Optionally, the training image hash value database stores a first hash value dataset corresponding to training images in a first base station coverage area, and the first base station is a base station connected to the first edge computing server.

[0013] Optionally, the training image hash value database also stores a second hash value dataset corresponding to training images in a second base station coverage area; wherein the second hash value dataset is obtained from a second edge computing server corresponding to the second base station and sent to the first edge computing server by the data center based on the determination that the second base station has similar road conditions to the first base station coverage area based on the road fixed attributes of each base station coverage area.

[0014] Optionally, the road fixed attributes include any one or a combination of lane lines, road edges, road dividers, and traffic lights.

[0015] Optionally, after determining that the minimum value in the Hamming distance is greater than the preset threshold, the method further comprises:

[0016] sending the road condition image to an autonomous driving company for scenario supplementary training, and supplementally updating the training image hash value database based on the hash value corresponding to the road condition image.

[0017] In a second aspect, the present application also provides an autonomous driving system, comprising:

[0018] a data center and a plurality of edge computing servers;

[0019] The data center comprises a high-precision map database, a land matching server, and a training image exchange server.

[0020] The edge computing server comprises an image transmission interface server, a training image hash value database, and an image similarity calculation server.

[0021] The data center and the edge computing server transmit data through the image transmission interface server.

[0022] The land matching server is configured to read different geographical areas in the high-precision map database with a single base station coverage area as a basic reading unit and compare road conditions.

[0023] The training image exchange server is configured to exchange the training image hash value databases corresponding to the two base stations with each other through the image transmission interface server if the location matching server determines that there is a similar road condition between the coverage areas of the two base stations.

[0024] The image similarity calculation server is configured to perform the steps of the automatic driving method according to the first aspect.

[0025] In a third aspect, the present application further provides an automatic driving device applied to the first edge computing server, comprising:

[0026] The receiving unit is configured to receive a road condition image sent by an automatic driving vehicle in real time and obtain a hash value corresponding to the road condition image.

[0027] The obtaining unit is configured to obtain a Hamming distance between the hash value corresponding to the road condition image and each hash value stored in the training image hash value database, respectively.

[0028] The determining unit is configured to determine a corresponding automatic driving operation based on a comparison result between the Hamming distance and a preset threshold.

[0029] In a fourth aspect, the present application further provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the automatic driving method according to the first aspect when executing the program.

[0030] In a fifth aspect, the present application further provides a non-transitory computer readable storage medium having a computer program stored thereon, wherein the computer program is executable on a processor to implement the steps of the automatic driving method according to the first aspect.

[0031] The automatic driving method and system provided by the present application calculate the hash value of the road condition image collected by the automatic driving vehicle through the edge computing server, and calculate the Hamming distance between each hash value in the training image hash value database, respectively, determine whether the automatic driving vehicle is currently in an unsafe state in an untrained scene based on the comparison result between the Hamming distance and the preset threshold, and determine the corresponding automatic driving operation, thereby effectively improving the safety of the automatic driving vehicle. BRIEF DESCRIPTION OF DRAWINGS

[0032] In order to more clearly illustrate the technical solutions in the present application or the prior art, the following will briefly introduce the drawings needed in the embodiments or the prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and those skilled in the art can obtain other drawings according to these drawings without creative labor.

[0033] Figure 1 is a flowchart of an automatic driving method provided by the present application;

[0034] Figure 2 is a basic architecture diagram of an automatic driving system provided by the present application;

[0035] Figure 3 is a topology diagram of a data center provided by the present application;

[0036] Figure 4 is a topology diagram of an edge computing server provided by the present application;

[0037] Figure 5 is a structural diagram of an automatic driving device provided by the present application;

[0038] Figure 6 is a physical structure diagram of an electronic device provided by the present application. DETAILED DESCRIPTION

[0039] In order to make the objectives, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be described clearly and completely below with reference to the drawings in the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0040] In order to reduce the safety risk of an automatic driving vehicle, the present application provides a solution which can significantly reduce traffic accidents caused by false negative judgments of the automatic driving vehicle when encountering road conditions not trained in advance, and realize sharing training on similar road sections. At the same time, the automatic driving capability of the vehicle is significantly improved by supplementing training on discovered untrained scenes.

[0041] Figure 1 is a flowchart of an automatic driving method provided by the present application, which is applied to a first edge computing server. As shown in Figure 1 the method comprises the following steps:

[0042] Step 100, receiving a road condition image sent by an automatic driving vehicle in real time, and obtaining a hash value corresponding to the road condition image;

[0043] Specifically, in the embodiments of the present application, the road condition image can be a real-time road condition image collected by a vehicle-mounted sensor during driving of the automatic driving vehicle. The road condition image can be one or more, and the vehicle-mounted sensor can be a camera, a laser radar or a millimeter wave radar, etc.

[0044] When the autonomous vehicle is in the coverage area of a base station (for example, a 5G base station gNB (next generation Node B)), the two-dimensional image or depth map image and other image information collected by the vehicle-mounted sensor can be uploaded in real time to an edge computing server MEC (Mobile Edge Computing) server, that is, a first edge computing server, which is in communication connection with the base station through the 5G network RTMP (Real Time Messaging Protocol) protocol. After the first edge computing server receives the above-mentioned image, the hash value corresponding to the image can be calculated by a hash algorithm.

[0045] The hash algorithm can be any one of a mean hash (ahash) algorithm, a difference hash algorithm (dhash), or a perceptual hash algorithm (phash), and is not specifically limited herein, as long as it is consistent with the hash value acquisition algorithm in the training image hash value database.

[0046] Optionally, considering that the ahash algorithm has a faster calculation speed than other hash algorithms, the ahash algorithm can be used to quickly process the obtained road condition image.

[0047] Step 101, respectively obtaining the Hamming distance between the hash value corresponding to the road condition image and each hash value stored in the training image hash value database;

[0048] Specifically, after the first edge computing server receives the road condition image and calculates the hash value corresponding to the road condition image, the Hamming distance between the hash value and each hash value stored in the training image hash value database of the first edge computing server can be calculated. The size of the obtained Hamming distance can be used to represent the difference between the road condition image and each training image.

[0049] The training image is a road condition image of a historical training scene of the autonomous vehicle, and the hash value corresponding to the training image is stored in the training image hash value database.

[0050] Step 102, determining the corresponding autonomous driving operation based on the comparison result between the Hamming distance and the preset threshold.

[0051] Specifically, the larger the Hamming distance between two hash values, the lower the similarity between the images corresponding to the two hash values. Therefore, the first edge computing server can determine the similarity between the current road environment of the autonomous vehicle and the road environment previously trained by comparing the calculated Hamming distance with a preset threshold, thereby assessing the safety risk of the autonomous vehicle and determining the corresponding autonomous driving operation, such as whether a human driver needs to take over the vehicle or automatically park the vehicle. The preset threshold can be set according to actual needs and is not limited here.

[0052] The autonomous driving method provided by an embodiment of the present invention obtains a hash value for the road condition image collected by the autonomous driving vehicle through an edge computing server, and calculates the Hamming distance with each hash value in the training image hash value database. Based on the comparison result between the Hamming distance and the preset threshold, it determines whether the autonomous driving vehicle is currently in an unsafe state of an untrained scenario, thereby determining the corresponding autonomous driving operation, effectively improving the safety of the autonomous driving vehicle.

[0053] Optionally, determining a corresponding autonomous driving operation based on a comparison result between the Hamming distance and a preset threshold includes:

[0054] If it is determined that the minimum value in the Hamming distance is less than or equal to the preset threshold, the autonomous driving vehicle is allowed to perform autonomous driving according to the pre-trained driving strategy; or,

[0055] If it is determined that the minimum value in the Hamming distance is greater than a preset threshold, an instruction is sent to the autonomous driving vehicle requiring the human driver to take over the vehicle or automatically park.

[0056] Specifically, after calculating the hash value of the image, the Hamming distance between the hash values ​​of the two images is calculated to compare the similarity of the two images. In the embodiment of the present invention, if the Hamming distance between the two images is less than or equal to a preset threshold, the two images are determined to be similar images.

[0057] If there are images in the historical training images that are similar to the road condition images uploaded by the autonomous driving vehicle in real time, that is, after calculating the Hamming distance between the hash value corresponding to the road condition image and each hash value stored in the training image hash value database, the minimum value of the multiple Hamming distances obtained is less than or equal to the preset threshold, then it indicates that the road conditions previously learned by the autonomous driving vehicle are highly similar to the current road conditions, and autonomous driving can be performed according to the driving strategy learned in previous training.

[0058] Otherwise, if there is no image similar to the real-time uploaded road condition image of the autonomous vehicle in the historical training image, that is, the minimum value of the plurality of hamming distances obtained by calculating the hamming distance between the hash value corresponding to the road condition image and each hash value stored in the training image hash value database is greater than the preset threshold, it indicates that the road condition where the vehicle is currently located is obviously different from the road condition encountered during previous training, and there is a safety risk in continuous driving. In this case, the first edge computing server can issue an instruction to the autonomous vehicle through the base station, requiring a human driver to take over the vehicle or automatically park the vehicle.

[0059] Optionally, the first edge computing server can also compare each hamming distance with the preset threshold after calculating each hamming distance. If it is determined that a certain hamming distance is less than or equal to the preset threshold, the calculation of the subsequent hamming distance is stopped, it is determined that the road condition where the autonomous vehicle has previously learned is highly similar to the current road condition, and the autonomous driving can be performed according to the driving strategy learned during previous training; otherwise, until all hamming distances are calculated, it is determined that all hamming distances are greater than the preset threshold, it is determined that the road condition where the vehicle is currently located is obviously different from the road condition encountered during previous training, and there is a safety risk in continuous driving.

[0060] It should be noted that if the real-time sent road condition image of the autonomous vehicle is a group of road condition images, for example, images of the front, left and right of the vehicle driving direction collected by multiple vehicle-mounted cameras at a certain time, it is necessary to determine that there are similar images of the three images in the historical training image, so as to determine that the road condition where the autonomous vehicle has previously learned is highly similar to the current road condition, so that the autonomous driving can be performed according to the driving strategy learned during previous training, otherwise, it is determined that the autonomous vehicle is currently in an unsafe state of an untrained scene, and there is a safety risk in continuous driving.

[0061] The autonomous driving method provided by the embodiment of the application uploads the road condition image uploaded by the autonomous vehicle driving under the coverage of the base station to the edge computing server, and determines the similarity with the training image of the autonomous vehicle before. Only in the case of similarity with the training image before can the autonomous driving state be maintained, otherwise, the human driver is prompted to take over the vehicle or automatically park the vehicle, thereby realizing sufficient coverage of various driving scene training categories and reducing the error decision of the autonomous driving algorithm for a certain untrained scene caused by overfitting.

[0062] Optionally, the training image hash value database stores a first hash value data set corresponding to the training image of the first base station coverage area, and the first base station is a base station connected to the first edge computing server.

[0063] Specifically, in the embodiment of the present application, the training image hash value database can only store the hash value dataset (i.e., the first hash value dataset) corresponding to the training image in the coverage area of the base station (i.e., the first base station) connected to the edge computing server (i.e., the first edge computing server) to reduce the response delay caused by image similarity calculation.

[0064] Optionally, the training image hash value database also stores a second hash value dataset corresponding to the training image in the coverage area of the second base station; wherein the second hash value dataset is obtained from the second edge computing server corresponding to the second base station and sent to the first edge computing server by the data center based on the determination that the second base station has similar road conditions with the first base station coverage area.

[0065] Specifically, in the embodiment of the present application, the training image hash value database can also store a second hash value dataset corresponding to the training image in the coverage area of the second base station, wherein the second base station is a base station with similar road conditions to the first base station, thereby achieving the purpose of improving the driving continuity of the autonomous vehicle based on sharing training image hash values with similar road conditions.

[0066] Optionally, the data center can determine the base station with similar road conditions and share the training image hash value by the following method steps:

[0067] Step one, count the geographical area covered by each base station;

[0068] The land matching server of the data center regularly counts the latitude and longitude margin and base station ID (Identifier, identifier) covered by each base station, and stores it in queue A with base station as granularity.

[0069] Step two, find the similar road conditions between different base station coverage areas;

[0070] The land matching server sequentially matches from the head to the tail of queue A, and extracts the blocks covered by each base station in the high-precision map database of the data center according to the latitude and longitude margin information, and compares whether there are similar combinations of road fixed attributes in the two blocks, wherein the road fixed attributes include any one or combination of lane lines, road edges, road dividers and traffic lights.

[0071] For example, if the road fixed attribute combination in two blocks meets one or more of the following: lane lines, road edges, road dividers, and traffic lights are similar in two regions belonging to two blocks; lane lines, road edges, and traffic lights are similar in two regions belonging to two blocks, and there is no road divider; lane lines and road edges are similar in two regions belonging to two blocks, and there is no road divider and traffic light; lane lines, road edges, and road dividers are similar in two regions belonging to two blocks, and there is no traffic light, it is determined that there is similar road conditions between the two base stations, and the ID information of the two base stations is sent to the training image exchange server of the data center.

[0072] Step three, the training image hash value sharing is performed for the two base stations with similar road conditions.

[0073] The training image exchange server finds the image transmission interface server of the corresponding edge computing server according to the base station ID A information, obtains the training image hash value data set stored by the edge computing server corresponding to the base station, and transmits the hash value data set to the image transmission interface server of the edge computing server corresponding to the base station ID B information, and stores the aforementioned hash value data set into the training image hash value database of the edge computing server by the image transmission interface server. Correspondingly, the same process is used to store the training image hash value data set stored by the edge computing server corresponding to the base station ID B into the training image hash value database of the edge computing server corresponding to the base station ID A.

[0074] The automatic driving method provided by the embodiment of the application can improve the continuity of vehicle automatic driving in the case that the road conditions of the coverage areas of two base stations are similar, that is, the driving scene contained in the training image hash value data set stored by the edge computing servers corresponding to the two base stations in any similar road conditions is completely adapted to the algorithm used by the autonomous vehicle, and therefore the mutual transmission and mutual backup of the training image hash value data sets stored by the two edge computing servers can improve the continuity of vehicle automatic driving.

[0075] Optionally, after determining that the minimum value in the Hamming distance is greater than the preset threshold, the method further includes:

[0076] The road condition image is sent to an autonomous driving company for scene supplementary training, and the training image hash value database is supplemented and updated based on the hash value corresponding to the road condition image.

[0077] Specifically, if it is determined that there is no image similar to the real-time uploaded road condition image of the autonomous vehicle in the historical training image, that is, the minimum value of the plurality of hamming distances obtained by calculating the hamming distance between the hash value corresponding to the road condition image and each hash value stored in the training image hash value database is greater than the preset threshold, it indicates that the road condition where the vehicle is currently located is obviously different from the road condition encountered during previous training. In addition to the first edge computing server issuing an instruction to the autonomous vehicle through the base station to require a human driver to take over the vehicle or automatically park, the first edge computing server can also subsequently send the road condition image to the autonomous driving company through the fixed network for subsequent scene supplement training and online upgrade of the program package to improve the safety of vehicle autonomous driving. Moreover, the training image hash value database is synchronously supplemented and updated based on the hash value corresponding to the road condition image.

[0078] The automatic driving method provided by the embodiment of the present application can significantly improve the vehicle autonomous driving capability and fully cover various driving scene training categories by supplementing training of the discovered untrained scene and supplementing and updating the training image hash value database.

[0079] The automatic driving system provided by the present application is described below, and the automatic driving system described below can be correspondingly referred to the automatic driving method described above.

[0080] Figure 2 The basic architecture diagram of the automatic driving system provided by the present application is shown in FIG. 1, which comprises a data center 200 and a plurality of edge computing servers 210, wherein the edge computing server 210 and the base station 220 are connected to each other to provide autonomous driving operation support for the autonomous vehicle 230 within the coverage range of the base station 220. Figure 2

[0081] Figure 3 The topology structure diagram of the data center 200 provided by the present application is shown in FIG. 2, which comprises a high-precision map database 300, a land section matching server 310 and a training image exchange server 320. Figure 3

[0082] The topology structure diagram of the edge computing server 210 provided by the present application is shown in FIG. 3, which comprises an image transmission interface server 400, a training image hash value database 410 and an image similarity calculation server 420, and the data center 200 and the edge computing server 210 are connected to each other for data transmission through the image transmission interface server 400. Figure 4 Figure 4

[0083] ​​​Specifically, the high-precision map database 300 is used to store street view information (such as traffic lights, lane lines, road edges, or road dividers) and geographic information (such as latitude and longitude) of each street in each city in China.

[0084] The land matching server 310 is used to read different geographic regions in different cities or the same city in the high-precision map database with street view information as a comparison keyword and a single base station 220 coverage area as a basic reading unit, and perform a road condition similarity comparison.

[0085] The training image exchange server 320 is used to exchange the training image hash value databases 410 of the two base stations 220 through the image transmission interface server 400 if the land matching server 310 determines that there is similar road conditions between the two base station 220 coverage areas.

[0086] The image transmission interface server 400 is used to upload the training image hash value dataset stored in the edge computing server 210 to which it belongs to the training image exchange server 320 in response to the request of the training image exchange server 320, and receive the training image hash value dataset stored in the edge computing server 210 to which another base station 220 belongs issued by the training image exchange server 320.

[0087] The training image hash value database 410 is used to store the hash values corresponding to the training images of the base station 220 coverage area connected to the edge computing server 210.

[0088] The image similarity calculation server 420 is used to perform the steps of any of the automatic driving methods provided in the above embodiments, for example: receiving a road condition image sent by an automatic driving vehicle in real time, and obtaining a hash value corresponding to the road condition image; obtaining a Hamming distance between the hash value corresponding to the road condition image and each hash value stored in the training image hash value database; and determining a corresponding automatic driving operation based on a comparison result between the Hamming distance and a preset threshold.

[0089] It should be noted that the above automatic driving system provided by the present application can realize all the method steps realized by the above automatic driving method embodiments, and can achieve the same technical effects. Therefore, the same parts and beneficial effects of the method embodiments in the present embodiment will not be described in detail.

[0090] The automatic driving device provided by the present application will be described below. The automatic driving device described below can be referred to in correspondence with the automatic driving method described above.

[0091] Figure 5 The structure schematic diagram of the automatic driving device provided by the present application is shown in the figure. The device is applied to a first edge computing server, and the device comprises: Figure 5 ​

[0092] The receiving unit 500 is configured to receive a road condition image sent by the autonomous vehicle in real time and acquire a hash value corresponding to the road condition image.

[0093] The acquiring unit 510 is configured to acquire a Hamming distance between the hash value corresponding to the road condition image and each hash value stored in the training image hash value database.

[0094] The determining unit 520 is configured to determine the corresponding autonomous driving operation based on a comparison result between the Hamming distance and a preset threshold.

[0095] Optionally, the determining unit 520 is configured to: if it is determined that the minimum value in the Hamming distance is less than or equal to the preset threshold, allow the autonomous vehicle to perform autonomous driving according to the pre-trained driving strategy; or if it is determined that the minimum value in the Hamming distance is greater than the preset threshold, send an instruction to the autonomous vehicle to require a human driver to take over the vehicle or to automatically park the vehicle.

[0096] Optionally, the training image hash value database stores a first hash value dataset corresponding to a training image of a first base station coverage area, and the first base station is a base station connected to the first edge computing server.

[0097] Optionally, the training image hash value database further stores a second hash value dataset corresponding to a training image of a second base station coverage area; wherein the second hash value dataset is acquired from a second edge computing server corresponding to the second base station and sent to the first edge computing server by the data center based on the determination that the second base station has similar road conditions to the first base station coverage area according to road fixed attributes of each base station coverage area.

[0098] Optionally, the road fixed attributes include any one or a combination of lane lines, road edges, road dividers, and traffic lights.

[0099] Optionally, the determining unit 520 is further configured to send the road condition image to an autonomous driving company for scene supplementary training, and perform supplementary update on the training image hash value database based on the hash value corresponding to the road condition image.

[0100] It should be noted that the above-mentioned autonomous driving device provided by the present application can realize all the method steps realized by the above-mentioned autonomous driving method embodiment, and can achieve the same technical effects. Therefore, the same parts and beneficial effects of the method embodiment in this embodiment will not be described in detail.

[0101] Figure 6 The entity structure schematic diagram of the electronic device provided by the present application is as follows Figure 6As shown, the electronic device can include a processor 610, a communications interface 620, a memory 630, and a communications bus 640, wherein the processor 610, the communications interface 620, and the memory 630 complete mutual communication through the communications bus 640. The processor 610 can invoke a logic instruction in the memory 630 to execute the steps of any of the automatic driving methods provided by the above-mentioned embodiments, for example: receiving a real-time road condition image sent by an autonomous vehicle, and obtaining a hash value corresponding to the road condition image; respectively obtaining a Hamming distance between the hash value corresponding to the road condition image and each hash value stored in a training image hash value database; and determining a corresponding autonomous driving operation based on a comparison result between the Hamming distance and a preset threshold.

[0102] In addition, the logic instruction in the memory 630 described above can be implemented in the form of a software functional unit and sold or used as an independent product, which can be stored in a computer-readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the methods described in the embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0103] On the other hand, the present application also provides a non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to implement the steps of any of the automatic driving methods provided by the above-mentioned embodiments, for example: receiving a real-time road condition image sent by an autonomous vehicle, and obtaining a hash value corresponding to the road condition image; respectively obtaining a Hamming distance between the hash value corresponding to the road condition image and each hash value stored in a training image hash value database; and determining a corresponding autonomous driving operation based on a comparison result between the Hamming distance and a preset threshold.

[0104] The device embodiments described above are merely illustrative, wherein the units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, i.e., can be located in one place, or can be distributed to multiple network units. Part or all of the modules can be selected to achieve the purposes of the embodiments according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0105] Through the description of the above embodiments, those skilled in the art can clearly understand that the embodiments can be realized by means of software and the necessary general hardware platform, and of course can also be realized by hardware. Based on such understanding, the above technical solutions can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods described in each embodiment or some parts of the embodiments.

[0106] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An automatic driving method, characterized by, Applied to a first edge computing server, comprising: receiving a real-time road condition image sent by an autonomous vehicle, and obtaining a hash value corresponding to the road condition image; respectively obtaining the Hamming distance between the hash value corresponding to the road condition image and each hash value stored in a training image hash value database; based on the comparison result between the Hamming distance and the preset threshold, determining the corresponding autonomous driving operation, the comparison result is used to judge the similarity between the current road condition environment of the autonomous vehicle and the previously trained and learned road condition environment; the training image hash value database stores a first hash value data set corresponding to the training image of the first base station coverage area, and the first base station is the base station connected to the first edge computing server; the training image hash value database also stores a second hash value data set corresponding to the training image of the second base station coverage area; wherein the second hash value data set is obtained from the second edge computing server corresponding to the second base station and sent to the first edge computing server by the data center based on the road fixed attribute of each base station coverage area to determine that there is a similar road condition between the second base station and the first base station coverage area.

2. The automatic driving method of claim 1, wherein, the comparison result between the Hamming distance and the preset threshold, including: if it is determined that the minimum value in the Hamming distance is less than or equal to the preset threshold, the autonomous vehicle is allowed to drive according to the pre-trained driving strategy; or, if it is determined that the minimum value in the Hamming distance is greater than the preset threshold, an instruction is sent to the autonomous vehicle to require a human driver to take over the vehicle or to automatically stop the vehicle.

3. The automatic driving method of claim 1, wherein, The road fixed attribute includes any one or combination of lane lines, road edges, road dividers, and traffic lights.

4. The automatic driving method of claim 2, wherein, After determining that the minimum value in the Hamming distance is greater than the preset threshold, the method further comprises: sending the road condition image to an autonomous driving company for scene supplementary training, and supplementally updating the training image hash value database based on the hash value corresponding to the road condition image.

5. An autonomous driving system, characterized by, including: a data center and a plurality of edge computing servers; the data center includes a high-precision map database, a land matching server, and a training image exchange server; the edge computing server includes an image transmission interface server, a training image hash value database, and an image similarity calculation server; the data center and the edge computing server transmit data through the image transmission interface server; the land matching server is used to read different geographical areas in the high-precision map database with a single base station coverage area as a basic reading unit and compare the road conditions; the training image exchange server is used to exchange the training image hash value databases corresponding to the two base stations through the image transmission interface server if the land matching server determines that there is a similar road condition between the two base stations; the image similarity calculation server is used to execute the steps of the autonomous driving method according to any one of claims 1 to 4.

6. An automatic driving device characterized by comprising: applied to a first edge computing server, comprising: A receiving unit is configured to receive a road condition image sent by an autonomous vehicle in real time and obtain a hash value corresponding to the road condition image; A obtaining unit is configured to obtain a Hamming distance between the hash value corresponding to the road condition image and each hash value stored in a training image hash value database; A determining unit is configured to determine a corresponding autonomous driving operation based on a comparison result between the Hamming distance and a preset threshold, the comparison result being used to judge a similarity between a current road condition environment of the autonomous vehicle and a previously trained and learned road condition environment; The training image hash value database stores a first hash value dataset corresponding to a training image of a first base station coverage area, the first base station being a base station connected to the first edge computing server. The training image hash value database also stores a second hash value dataset corresponding to a training image of a second base station coverage area; wherein the second hash value dataset is obtained and sent by a second edge computing server corresponding to the second base station to the first edge computing server after the data center determines that the second base station exists in a similar road condition with the first base station coverage area based on fixed road properties of each base station coverage area.

7. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor implements the steps of the autonomous driving method according to any one of claims 1 to 4 when executing the program.

8. A non-transitory computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program implements the steps of the autonomous driving method according to any one of claims 1 to 4 when executed by the processor.

Images