Scene similarity evaluation method, related equipment and vehicle

Abstract

The embodiment of the invention discloses a scene similarity evaluation method which is mainly applied to similarity analysis in a traffic scene. In the method, at least two pieces of traffic scene data are acquired, each piece of traffic scene data comprises multiple frames of data, and each frame of data comprises at least one of the following items of at least one object: position, driving speed, shape and timestamp corresponding to the position; determining an interaction coefficient of a sampling point corresponding to each object in each piece of traffic scene data, wherein the interaction coefficient is used for indicating the degree of influence of other objects on driving of the object; calculating the trajectory similarity between the two objects based on the interaction coefficient; determining at least one group of objects through the trajectory similarity; and determining the similarity of the at least two traffic scene data based on the at least one group of objects. The interaction coefficient is introduced to indicate the passing influence degree between the objects, and the interaction relation between the objects can be reflected in the scene similarity, so that the rationality and accuracy of quantitative scoring of the scene similarity are improved.

Description

Technical Field

[0001] The present application relates to the field of autonomous driving, and in particular to a scene similarity assessment method, related equipment and vehicles. Background Art

[0002] Artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. In other words, artificial intelligence is a branch of computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can respond in a similar way to human intelligence. Artificial intelligence is also the study of the design principles and implementation methods of various intelligent machines, so that machines have the functions of perception, reasoning and decision-making. Research in the field of artificial intelligence includes robotics, natural language processing, computer vision, decision-making and reasoning, human-computer interaction, recommendation and search, basic AI theory, etc.

[0003] Autonomous driving (or unmanned driving, intelligent driving, etc.) is a mainstream application in the field of artificial intelligence. Autonomous driving technology relies on the collaboration of computer vision, radar, monitoring devices and global positioning systems, and realizes intelligent information exchange with people, cars, roads, etc. through on-board sensor systems and information terminals, so that vehicles have intelligent environmental perception capabilities, and achieve the purpose of replacing human operation to reduce the burden of human driving. Among them, the similarity evaluation technology of regulatory scenes is a common technology in autonomous driving. For example, it can be applied to scene clustering / logical classification of problem scenes; data mining based on poorly performing scenes, looking for scenes of the same type, as training sets for neural network tuning, or providing them to the business side to find common problems, to improve the performance of the algorithm in the same type of scenes; and evaluating the coverage of the simulation scene set for road test problems, supplementing uncovered scenes in a targeted manner; searching for the optimal driving plan based on the current driving scene, etc. Existing technologies often use the distance between different vehicle trajectories to determine similarity.

[0004] However, the actual driving conditions of vehicles are very complex, and it is difficult to reasonably analyze the similarity between scenes based only on trajectory distance. Summary of the invention

[0005] The embodiments of the present application provide a scene similarity assessment method, related equipment and vehicle, which are used to consider the interaction relationship between objects in traffic scenes when evaluating the similarity of different traffic scenes, and reflect the interaction relationship between objects in the scene similarity, thereby improving the rationality and accuracy of the scene similarity assessment method.

[0006] The first aspect of the embodiment of the present application provides a scene similarity evaluation method, which is mainly used for similarity analysis in traffic scenes. The method can be executed by a data processing device, or by a component of the data processing device (such as a processor, a chip, or a chip system, etc.). In the method, at least two traffic scene data are obtained, each of which includes multiple frames of data, and each frame of data includes at least one of the following items of at least one object: position, driving speed, shape, and timestamp corresponding to the position; the interaction coefficient of the sampling point corresponding to each object in each traffic scene data is determined, and the interaction coefficient is used to indicate the degree to which the driving of the object is affected by other objects; the trajectory similarity between the two objects is calculated based on the interaction coefficient, wherein the at least two traffic scene data include the first traffic scene data and the second traffic scene data, and the two objects are any object in the first traffic scene data and any object in the second traffic scene data; then at least one group of objects is determined based on the trajectory similarity, wherein the trajectory similarity of two objects in each group of objects in at least one group of objects meets a preset condition; the similarity of at least two traffic scene data is determined based on at least one group of objects. The above-mentioned preset condition may be that the trajectory similarity of two objects in each group of objects is less than or equal to a threshold, or the preset condition may be that the trajectory similarity of two objects in each group of objects is greater than or equal to a threshold, etc.

[0007] In the embodiment of the present application, an interaction coefficient is introduced to indicate the degree of traffic influence between objects, and the interaction relationship between objects is reflected in the scene similarity, thereby improving the rationality and accuracy of the quantitative scoring of the scene similarity.

[0008] Optionally, in a possible implementation manner of the first aspect, the trajectory similarity includes at least one of the following: a first similarity, a second similarity, a third similarity and a fourth similarity; the first similarity is used to indicate the similarity between the trajectories of the first ego vehicle in the first traffic scene data and the second ego vehicle in the second traffic scene data, the second similarity is used to indicate the similarity between the trajectory of the first object in the first traffic scene data in the coordinate system of the first ego vehicle and the trajectory of the second object in the second traffic scene data in the coordinate system of the second ego vehicle, the third similarity is used to indicate the similarity between the orientation of the first object relative to the first ego vehicle and the orientation of the second object relative to the second ego vehicle, and the fourth similarity is used to indicate the similarity between the trajectories of the first object and the second object in the map coordinate system.

[0009] In this possible implementation, on the one hand, the trajectory similarity obtained by multiple similarities is more accurate. On the other hand, by introducing the first object trajectory similarity and the second object trajectory similarity and the orientation similarity, the similarity of the interactive information can be reflected, so that the trajectory similarity fully considers the interaction between different objects, thereby improving the rationality of calculating the trajectory similarity.

[0010] Optionally, in a possible implementation manner of the first aspect, the above-mentioned step: calculating the trajectory similarity between two objects based on the interaction coefficient includes: obtaining a first sampling point and a second sampling point, the first sampling point including any two sampling points of the object in the first traffic scene data, and the second sampling point including any two sampling points of the object in the second traffic scene data; calculating difference information between the first sampling point and the second sampling point, the difference information including angle difference and / or length difference; determining the trajectory similarity based on the difference information and the interaction coefficient.

[0011] In this possible implementation, two sampling points in different traffic scene data are obtained through the interaction coefficient of each sampling point, so that the trajectory similarity is calculated according to the difference between the sampling point groups. In this process, the angle and / or length between the sampling point groups can be taken into account, so that the trajectory similarity calculated based on the difference information can better reflect the similarity between the trajectories.

[0012] Optionally, in a possible implementation of the first aspect, the above step: determining the interaction coefficient of the sampling point corresponding to each object in each traffic scene data includes: obtaining a safety factor of the sampling point, the safety factor being used to indicate the collision risk of the sampling point; determining the interaction coefficient based on the safety factor and the relative distance, wherein the relative position is used to indicate the distance between the sampling point and other objects.

[0013] In this possible implementation, the interaction coefficient is determined according to the collision risk of the sampling points, so that the interaction coefficient can reflect the degree of interaction between the objects.

[0014] Optionally, in a possible implementation manner of the first aspect, before the above step of: determining the interaction coefficient of the sampling point corresponding to each object in each traffic scene data, the method further includes: determining that the classifications of at least two traffic scene data are the same.

[0015] In this possible implementation, the classification can be judged before calculating the scene similarity, thereby reducing the computing power loss of subsequent calculations of the scene similarity.

[0016] Optionally, in a possible implementation manner of the first aspect, the above classification includes at least one of the following: classification of object behavior, classification of topological information of the road where the object is located, and classification of interaction behavior between objects.

[0017] In this possible implementation, a preliminary screening may be performed on various classifications of scenes, so as to calculate the similarity of scenes that meet preset conditions.

[0018] Optionally, in a possible implementation manner of the first aspect, each frame of data further includes topological information of a road where at least one object is located.

[0019] In this possible implementation, topological information is introduced to classify traffic scene data to improve classification accuracy.

[0020] Optionally, in a possible implementation manner of the first aspect, the scene similarity is a sum of trajectory similarities between at least one group of objects or a weighted sum of trajectory similarities between at least one group of objects.

[0021] In this possible implementation, the scene similarity may be obtained based on the trajectory similarity between the most representative or most similar group of objects, so that the most representative or most similar group of objects may be reflected in the scene similarity.

[0022] Optionally, in a possible implementation of the first aspect, the above-mentioned scene similarity is the sum of the similarities of the trajectories corresponding to all objects in at least two traffic scene data or the weighted sum of the similarities of the trajectories corresponding to all objects in at least two traffic scene data, and the trajectory similarity of all objects except for objects in at least one group of objects is a preset value.

[0023] In this possible implementation, the scene similarity can be obtained according to the trajectory similarity of all objects, so that the influence between the objects can be considered, making the calculation of the scene similarity more accurate.

[0024] Optionally, in a possible implementation manner of the first aspect, the above method is applied to at least one of the following scenarios: a scene coverage scenario, a data mining scenario, a clustering scenario, and a driving guidance scenario.

[0025] In this possible implementation, by quantifying the similarity of two traffic scene data, it can be applied to upper-level applications based on similarity analysis, such as scene coverage scenarios, data mining scenarios, clustering scenarios, driving guidance scenarios, etc., and is universal.

[0026] Optionally, in a possible implementation of the first aspect, the at least two traffic scene data mentioned above include first traffic scene data and second traffic scene data, the first traffic scene data is road test data, and the second traffic scene data is simulation data; the method also includes: determining the simulation coverage based on the scene similarity, and the simulation coverage is used to indicate the proportion of the simulation data covering the road test data.

[0027] In this possible implementation, the coverage of the simulation scenario to the drive test scenario can be counted, that is, whether the simulation scenario fully covers the drive test scenario can be counted, so that the simulation scenario can be supplemented in a targeted manner.

[0028] Optionally, in a possible implementation manner of the first aspect, the above steps further include: acquiring multiple training data based on scene similarity, and the training data is used to train the network.

[0029] In this possible implementation, the same type of scenes can be searched in batches through scene similarity and the template scene to be mined (also called the original scene) to implement the generalization test of the problem scene.

[0030] Optionally, in a possible implementation manner of the first aspect, the above steps also include: clustering at least two traffic scene data based on scene similarity.

[0031] In this possible implementation, the drive test problem scenario set can be clustered according to the scenario similarity, thereby dividing different logical scenarios to facilitate the management of the scenario set.

[0032] Optionally, in a possible implementation of the first aspect, the at least two traffic scene data mentioned above include third traffic scene data and fourth traffic scene data, the third traffic scene data belongs to road test data, and the fourth traffic scene data belongs to database data, and the method also includes: determining the driving route of the first object in the third traffic scene data based on the scene similarity and the fourth traffic scene data.

[0033] In this possible implementation, similar or future safer scenes can be found in the scene library through scene similarity, so as to give the best driving advice.

[0034] The second aspect of the embodiment of the present application provides a data processing device, which is mainly used for similarity analysis in traffic scenes. The data processing device includes: an acquisition unit, which is used to acquire at least two traffic scene data, each traffic scene data includes multiple frames of data, and each frame of data includes at least one of the following items of at least one object: position, driving speed, shape, and timestamp corresponding to the position; a determination unit, which is used to determine the interaction coefficient of the sampling point corresponding to each object in each traffic scene data, and the interaction coefficient is used to indicate the degree to which the driving of the object is affected by other objects; a calculation unit, which is used to calculate the trajectory similarity between two objects based on the interaction coefficient, and the at least two traffic scene data include first traffic scene data and second traffic scene data, and the two objects include any object in the first traffic scene data and any object in the second traffic scene data; the determination unit is also used to determine at least one group of objects through trajectory similarity, and the trajectory similarity of two objects in each group of objects in at least one group of objects meets a preset condition; the determination unit is also used to determine the similarity of at least two traffic scene data based on at least one group of objects.

[0035] Optionally, in a possible implementation manner of the second aspect, the trajectory similarity between the above-mentioned two objects includes at least one of the following: a first similarity, a second similarity, a third similarity and a fourth similarity; the first similarity is used to indicate the similarity between the trajectories of the first vehicle in the first traffic scene data and the second vehicle in the second traffic scene data, the second similarity is used to indicate the similarity between the trajectory of the first object in the first traffic scene data in the coordinate system of the first vehicle and the trajectory of the second object in the second traffic scene data in the coordinate system of the second vehicle, the third similarity is used to indicate the similarity between the position of the first object relative to the first vehicle and the position of the second object relative to the second vehicle, and the fourth similarity is used to indicate the similarity between the trajectories of the first object and the second object in the map coordinate system.

[0036] Optionally, in a possible implementation manner of the second aspect, the above-mentioned calculation unit is specifically used to obtain a first sampling point and a second sampling point, the first sampling point includes any two sampling points of an object in the first traffic scene data, and the second sampling point includes any two sampling points of an object in the second traffic scene data; the calculation unit is specifically used to calculate difference information between the first sampling point and the second sampling point, the difference information includes angle difference and / or length difference; the calculation unit is specifically used to calculate trajectory similarity based on the difference information and the interaction coefficient.

[0037] Optionally, in a possible implementation manner of the second aspect, the above-mentioned determination unit is specifically used to obtain a safety factor of the sampling point, and the safety factor is used to indicate the collision risk of the sampling point; the determination unit is specifically used to determine the interaction coefficient based on the safety factor and the relative distance, and the relative distance is used to indicate the distance between the sampling point and other objects.

[0038] Optionally, in a possible implementation manner of the second aspect, the above-mentioned determination unit is further used to determine that the classifications of at least two traffic scene data are the same.

[0039] Optionally, in a possible implementation manner of the second aspect, the above classification includes at least one of the following: classification of object behavior, classification of topological information of the road where the object is located, and classification of interaction behavior between objects.

[0040] Optionally, in a possible implementation manner of the second aspect, each frame of data further includes topological information of a road where at least one object is located.

[0041] Optionally, in a possible implementation manner of the second aspect, the scene similarity is a sum of trajectory similarities between at least one group of objects or a weighted sum of trajectory similarities between at least one group of objects.

[0042] Optionally, in a possible implementation manner of the second aspect, the above-mentioned scene similarity is the sum of the similarities of trajectories corresponding to all objects in at least two traffic scene data or the weighted sum of the similarities of trajectories corresponding to all objects in at least two traffic scene data.

[0043] Optionally, in a possible implementation manner of the second aspect, the above-mentioned device is applied to at least one of the following scenarios: a scene coverage scenario, a data mining scenario, a clustering scenario, and a driving guidance scenario.

[0044] A third aspect of an embodiment of the present application provides a data processing device, including: a processor, the processor is coupled to a memory, the memory is used to store programs or instructions, when the program or instructions are executed by the processor, the data processing device implements the method in the above-mentioned first aspect or any possible implementation of the first aspect.

[0045] A fourth aspect of an embodiment of the present application provides a vehicle, which includes the device in the second aspect or any possible implementation of the second aspect; or the vehicle includes the device in the third aspect.

[0046] A fifth aspect of an embodiment of the present application provides a computer-readable storage medium on which a computer program or computer instructions are stored. When the computer program or computer is executed by a processor, the method in the first aspect or any possible implementation of the first aspect is implemented.

[0047] A sixth aspect of an embodiment of the present application provides a computer program product, which includes computer instructions. When the computer instructions are executed by a processor, the method in the first aspect or any possible implementation of the first aspect is implemented.

[0048] Among them, the technical effects brought about by the second, third, fourth, fifth, and sixth aspects or any possible implementation methods thereof can refer to the technical effects brought about by the first aspect or different possible implementation methods of the first aspect, and will not be repeated here.

[0049] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages: by introducing the interaction coefficient to consider the strength of the traffic impact between objects, the interaction relationship between objects can be reflected in the scene similarity, thereby improving the rationality and accuracy of the quantitative scoring of the scene similarity. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 A schematic diagram of a vehicle system architecture provided in an embodiment of the present application;

[0051] Figure 2 A schematic diagram of an application scenario provided for an embodiment of the present application;

[0052] Figure 3A A schematic diagram of the structure of a data processing system provided in an embodiment of the present application;

[0053] Figure 3B A schematic diagram of the structure of another data processing system provided in an embodiment of the present application;

[0054] Figure 4 A schematic diagram of a flow chart of a scene similarity assessment method provided in an embodiment of the present application;

[0055] Figure 5A An example diagram of a first traffic scenario provided in an embodiment of the present application;

[0056] Figure 5B An example diagram of a second traffic scenario provided in an embodiment of the present application;

[0057] Fig. 6A An example diagram of first traffic scene data provided in an embodiment of the present application;

[0058] Figure 6B An example diagram of second traffic scene data provided in an embodiment of the present application;

[0059] Fig. 7A An example diagram of calculating similarity provided in an embodiment of the present application;

[0060] Figure 7B Another example diagram for calculating similarity provided in an embodiment of the present application;

[0061] Figure 8 A schematic diagram of a process for determining at least one group of objects by trajectory similarity provided in an embodiment of the present application;

[0062] Fig. 9 An example diagram of a traffic scene provided in an embodiment of the present application;

[0063] Fig.10 A schematic diagram of a flow chart of a scene similarity assessment method provided in an embodiment of the present application;

[0064] Fig.11 A flowchart of another scene similarity assessment method provided in an embodiment of the present application;

[0065] Fig.12 A schematic diagram of scene classification provided in an embodiment of the present application;

[0066] Fig.13 A schematic diagram of the structure of a data processing device provided in an embodiment of the present application;

[0067] Fig.14 A schematic diagram of the structure of another data processing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0068] The embodiments of the present application provide a scene similarity assessment method, related equipment and vehicle, which are used to consider the strength of the traffic impact between objects in the traffic scene when evaluating the similarity of different traffic scenes, that is, the interaction relationship between the objects can be reflected in the scene similarity, thereby improving the rationality and accuracy of the scene similarity assessment method.

[0069] The terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and need not be used to describe a specific order or sequential order. It should be understood that the terms used in this way can be interchangeable under appropriate circumstances, which is only to describe the distinction mode adopted by the objects of the same attributes when describing in the embodiments of the present application. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, so that the process, method, system, product or equipment comprising a series of units need not be limited to those units, but may include other units that are not clearly listed or inherent to these processes, methods, products or equipment.

[0070] In order to better understand the scheme of the embodiment of the present application, the following first introduces the relevant terms and concepts that may be involved in the embodiment of the present application. It should be understood that the interpretation of the relevant concepts may be limited due to the specific circumstances of the embodiment of the present application, but it does not mean that the present application is limited to the specific circumstances. There may be differences in the specific circumstances of different embodiments, which are not specifically limited here.

[0071] 1. Object

[0072] The objects involved in the embodiments of the present application can be understood as objects that may be encountered in traffic scenes, such as vehicles, pedestrians, obstacles, etc.

[0073] 2. Data processing equipment

[0074] The data processing device refers to a device that applies the scene similarity assessment method provided in the embodiments of the present application, and the data processing device may be a cloud device or a terminal device.

[0075] Among them, the cloud device may refer to a device or server with data processing function such as a cloud server, a network server, an application server, a management server or a remote controller. The embodiment of the present application is only described by taking the terminal device as a vehicle as an example. In practical applications, the terminal device may also be a processing device on a vehicle (such as a vehicle-mounted media player on a vehicle, a wearable electronic device for controlling the vehicle, a virtual reality (VR) terminal device for controlling the vehicle, an augmented reality (AR) for controlling the vehicle, a vehicle-mounted terminal), an intelligent robot, an electric cart, a balance car, a boat, an airplane and other equipment. Among them, the vehicle may be an internal combustion engine vehicle using an engine as a power source, a hybrid vehicle using an engine and an electric motor as a power source, an electric vehicle using an electric motor as a power source, and other types of vehicles.

[0076] 3. Self-driving

[0077] An autonomous driving vehicle, that is, a vehicle in which the method of the embodiment of the present application works.

[0078] 4. Other cars (also called social cars)

[0079] Taking the object as a vehicle as an example, the concept of other vehicles, relative to the own vehicle, can also be understood as vehicles that may affect the passage of the own vehicle or vehicles that have the ability to affect the passage of the own vehicle.

[0080] In order to facilitate the understanding of this solution, the present application embodiment first combines Figure 1 For an introduction to the structure of the vehicle, please refer to Figure 1 , Figure 1 A schematic diagram of a structure of a vehicle provided in an embodiment of the present application. The vehicle may be an example diagram of a terminal device or an object in an embodiment of the present application.

[0081] The vehicle may include various subsystems, such as a travel system 102, a sensor system 104, a control system 106, one or more peripheral devices 108, and a power source 110 and a user interface 116. Optionally, the vehicle may include more or fewer subsystems, and each subsystem may include multiple components. In addition, each subsystem and component of the vehicle may be interconnected by wire or wirelessly (e.g., Bluetooth).

[0082] The travel system 102 may include components that provide powered movement for the vehicle. In one embodiment, the travel system 102 may include an engine 118 , a power source 119 , a transmission 120 , and wheels 121 .

[0083] Among them, the engine 118 can be an internal combustion engine, an electric motor, an air compression engine, or a combination of other types of engines, for example, a hybrid engine consisting of a gasoline engine and an electric motor, and a hybrid engine consisting of an internal combustion engine and an air compression engine. The engine 118 converts the energy source 119 into mechanical energy. Examples of energy sources 119 include gasoline, diesel, other petroleum-based fuels, propane, other compressed gas-based fuels, ethanol, solar panels, batteries, and other sources of electricity. The energy source 119 can also provide energy for other systems of the vehicle. The transmission 120 can transmit the mechanical power from the engine 118 to the wheels 121. The transmission 120 may include a gearbox, a differential, and a drive shaft. In one embodiment, the transmission 120 may also include other devices, such as a clutch. Among them, the drive shaft may include one or more shafts that can be coupled to the wheels 121.

[0084] The sensor system 104 may include several sensors that sense information about the vehicle's location. For example, the sensor system 104 may include a positioning system 122 (the positioning system may be a global positioning GPS system, or a Beidou system or other positioning systems), an inertial measurement unit (IMU) 124, a radar 126, a laser rangefinder 128, and a camera 130. The sensor system 104 may also include sensors of the internal systems of the monitored vehicle (e.g., an in-vehicle air quality monitor, a fuel gauge, an oil temperature gauge, etc.). The sensing data from one or more of these sensors may be used to detect objects and their corresponding characteristics (e.g., position, shape, direction, speed, etc.). Such detection and recognition are key functions for the safe operation of autonomous vehicles.

[0085] Among them, the positioning system 122 can be used to estimate the geographic location of the vehicle, such as the latitude and longitude information of the vehicle's location. The IMU 124 is used to sense the position and orientation changes of the vehicle based on the inertial acceleration rate. In one embodiment, the IMU 124 can be a combination of an accelerometer and a gyroscope. The radar 126 can use radio signals to sense objects in the surrounding environment of the vehicle, and can be specifically expressed as a millimeter wave radar or a laser radar. In some embodiments, in addition to sensing objects, the radar 126 can also be used to sense the speed and / or forward direction of the object. The laser rangefinder 128 can use lasers to sense objects in the environment where the vehicle is located. In some embodiments, the laser rangefinder 128 may include one or more laser sources, a laser scanner, and one or more detectors, as well as other system components. The camera 130 can be used to capture multiple images of the surrounding environment of the vehicle. The camera 130 can be a static camera or a video camera.

[0086] The control system 106 controls the operation of the vehicle and its components and may include various components, including a steering system 132 , an accelerator 134 , a brake unit 136 , an electronic control unit (ECU) 138 , and a body control module (BCM) 140 .

[0087] Among them, the steering system 132 is operable to adjust the forward direction of the vehicle. For example, in one embodiment, it can be a steering wheel system. The throttle 134 is used to control the operating rate of the engine 118 and thus control the speed of the vehicle. The brake unit 136 is used to control the deceleration of the vehicle. The brake unit 136 can use friction to slow down the wheel 121. In other embodiments, the brake unit 136 can convert the kinetic energy of the wheel 121 into electric current. The brake unit 136 can also take other forms to slow down the rotation speed of the wheel 121 to control the speed of the vehicle. The vehicle electronic control unit 138 can be implemented as a single ECU or multiple ECUs on the vehicle, and the single ECU or multiple ECUs are configured to communicate with the peripheral device 108 and the sensor system 104. The vehicle electronic control unit 138 may include at least one processor 1381 and at least one memory 1382 (read-only memory, ROM). One or more general-purpose processors, content addressable memory, digital signal processor, application-specific integrated circuit, field programmable gate array, any appropriate programmable logic device, discrete gate or transistor logic, discrete hardware components or any combination designed to perform the functions described here can be used to implement or execute at least one processor. In particular, at least one processor may be implemented as one or more microprocessors, controllers, microcontroller units (MCUs), or state machines. In addition, at least one processor may be implemented as a combination of computing devices, such as a digital signal processor or a microprocessor, a plurality of microprocessors, one or more microprocessors combined with a digital signal processor core, or any other combination of such configurations. The ROM may provide storage of data, including storage of addresses, routes, and driving directions in the present application.

[0088] The vehicle interacts with external sensors, other vehicles, other computer systems, or users through the peripheral device 108. The peripheral device 108 may include a wireless communication system 146, a navigation system 148, a microphone 150, and / or a speaker 152. In some embodiments, the peripheral device 108 provides a means for the user of the vehicle to interact with the user interface 116. For example, the navigation system 148 may be implemented as part of an in-vehicle entertainment system, an in-vehicle display system, an in-vehicle instrument cluster, etc. In a practical embodiment, the navigation system 148 is implemented to include or cooperate with the sensor system 104, which derives the current geographic location of the vehicle in real time or substantially in real time. The navigation system 148 is configured to provide navigation data to the driver of the vehicle. The navigation data may include the location data of the vehicle, driving instructions for suggested route planning, and visible map information for the vehicle operator. The navigation system 148 may present the location data to the driver of the vehicle through a display element or other presentation device. The current location of the vehicle may be described by one or more of the following: a triangulated location, a latitude / longitude location, an x ​​and y coordinate, or any other symbol or measurement that indicates the geographic location of the vehicle.

[0089] The user interface 116 can also operate the navigation system 148 to receive the user's input. The navigation system 148 can be operated by a touch screen. The navigation system 148 provides the ability to plan routes and the ability to navigate when the user enters the geographic location values ​​of the starting point and the end point. In other cases, the peripheral device 108 can provide a means for the vehicle to communicate with other devices located in the vehicle. For example, the microphone 150 can receive audio (e.g., voice commands or other audio input) from the user of the vehicle. Similarly, the speaker 152 can output audio to the user of the vehicle. The wireless communication system 146 can communicate wirelessly with one or more devices directly or via a communication network. For example, the wireless communication system 146 can use 3G cellular communication, such as code division multiple access (code division multiple access, CDMA), EVDO, global system for mobile communications (global system for mobile communications, GSM) / general packet radio service technology (general packet radio service, GPRS), or 4G cellular communication, such as long term evolution (long term evolution, LTE), or 5G cellular communication, etc. The wireless communication system 146 can use WiFi to communicate with a wireless local area network (wireless local area network, WLAN). In some embodiments, the wireless communication system 146 may utilize infrared links, Bluetooth, or ZigBee to communicate directly with the device. Other wireless protocols, such as various vehicle communication systems, for example, the wireless communication system 146 may include one or more dedicated short range communications (DSRC) devices, which may include public and / or private data communications between vehicles and / or roadside stations.

[0090] The power source 110 can provide power to various components of the vehicle. In one embodiment, the power source 110 can be a rechargeable lithium ion or lead acid battery. One or more battery packs of such batteries can be configured as a power source to provide power to various components of the vehicle. In some embodiments, the power source 110 and the energy source 119 can be implemented together, such as in some all-electric vehicles.

[0091] Optionally, one or more of the above-mentioned components may be installed or associated separately from the vehicle. For example, the memory 1382 may be partially or completely separate from the vehicle. The above-mentioned components may be communicatively coupled together in a wired and / or wireless manner.

[0092] Optionally, the above components are only examples. In actual applications, the components in the above modules may be added or deleted according to actual needs. Figure 1It should not be understood as limiting the embodiments of the present application.

[0093] The above-mentioned vehicle can be a car, a truck, a motorcycle, a bus, a boat, a lawn mower, an entertainment vehicle, an amusement park vehicle, construction equipment, a tram, a golf cart, and a cart, etc., and the embodiments of the present application are not particularly limited.

[0094] Combination of the above Figure 1 The architecture diagram of the system applicable to the embodiment of the present application is explained. In order to better understand the technical solution provided by the embodiment of the present application, the research ideas of the technical solution described in the present application are elaborated in detail below.

[0095] Autonomous driving (or unmanned driving, intelligent driving, etc.) is a mainstream application in the field of artificial intelligence. Autonomous driving technology relies on the collaboration of computer vision, radar, monitoring devices and global positioning systems, and realizes intelligent information exchange with people, cars, roads, etc. through on-board sensor systems and information terminals, so that vehicles have intelligent environmental perception capabilities, and achieve the purpose of replacing human operation to reduce the burden of human driving. Among them, the similarity evaluation technology of regulatory scenes is a common technology in autonomous driving. For example, it can be applied to scene clustering / logical classification of problem scenes; data mining based on poorly performing scenes, looking for scenes of the same type, as training sets for neural network tuning, or providing them to the business side to find common problems, to improve the performance of the algorithm in the same type of scenes; and evaluating the coverage of simulation scene sets for road test problems, supplementing uncovered scenes in a targeted manner; searching for the optimal driving plan based on the current driving scene, etc. Existing technologies often use the distance between different vehicle trajectories to determine the similarity. However, the actual driving conditions of autonomous vehicles are very complex, and it is difficult to reasonably analyze the similarity between scenes based only on the trajectory distance. Therefore, how to accurately determine the similarity of vehicle scenes is a technical problem that needs to be solved urgently.

[0096] In response to the above problems, the present application provides a scene similarity evaluation method, related equipment and vehicle. In the process of calculating the traffic scene similarity, multi-frame data containing multiple objects are considered, and a comprehensive score of the scene similarity can be quantified and given, which can be effectively applied to upper-level applications such as data mining and scene clustering.

[0097] The following first describes the application scenario of the scene similarity evaluation method provided in the embodiment of the present application. Figure 2 shown.

[0098] Scenario 1: Scenario clustering.

[0099] In this scenario, the scene similarity obtained by the scene similarity evaluation method provided in the embodiment of the present application can be used to cluster the road test problem scene set, thereby dividing different logical scenes to facilitate the management of the scene set.

[0100] Scenario 2: Data mining.

[0101] In this scenario, the scene similarity obtained by the scene similarity evaluation method provided in the embodiment of the present application and the template scene to be mined (also referred to as the original scene) can be used to batch search for scenes of the same type to implement the generalization test of the problem scenario.

[0102] Scenario 3: Coverage statistics.

[0103] In this scenario, the coverage of the simulation scenario to the road test scenario can be statistically calculated by the scenario similarity evaluation method provided in the embodiment of the present application, that is, whether the simulation scenario fully covers the road test scenario can be statistically calculated, so that the simulation scenario can be supplemented in a targeted manner.

[0104] Scenario 4: Driving guidance.

[0105] In this scenario, the scene similarity obtained by the scene similarity evaluation method provided in the embodiment of the present application can be used to find similar or future safer scenes in the scene library, so as to give the best driving advice.

[0106] It is understandable that the above-mentioned scenarios are just examples. In actual applications, there may be other scenarios, which are not specifically limited here.

[0107] Next, several deployment scenarios of this application are introduced.

[0108] Figure 3A A structural diagram of a data processing system provided in an embodiment of the present application, the data processing system includes a terminal device ( Figure 3A Only the terminal device is a vehicle as an example) and a data processing device. It can be understood that in addition to being a vehicle, the terminal device can also be a processing device on a vehicle (such as a vehicle-mounted media player on a vehicle, a wearable electronic device for controlling the vehicle, a virtual reality (VR) terminal device for controlling the vehicle, an augmented reality (AR) for controlling the vehicle, a vehicle-mounted terminal), an intelligent robot, an electric cart, a balance car, a ship, an airplane and other terminal devices. The terminal device is the initiator of data processing. As the initiator of the data processing request, the request is usually initiated by the user through the terminal device.

[0109] The above-mentioned data processing device can be a device or server with data processing function such as a cloud server, a network server, an application server, a management server or a remote controller. The data processing device receives data processing requests from the terminal device through an interactive interface, and then performs data processing in the form of machine learning, deep learning, search, reasoning, decision-making, etc. through the memory that stores data and the processor link of data processing. The memory in the data processing device can be a general term, including local storage and databases that store historical data. The database can be on the data processing device or on other network servers.

[0110] exist Figure 3A In the data processing system shown, the terminal device can receive the user's instruction. For example, the terminal device can obtain at least two traffic scene data (such as drive test data, simulation data, etc.) input / selected by the user, and then initiate a request to the data processing device so that the data processing device executes a data processing application (such as the aforementioned Figure 2 Perform scene similarity judgment, scene clustering, data mining, coverage statistics, driving guidance, etc. in the above-mentioned several scenarios, so as to obtain corresponding processing results for at least two traffic scene data.

[0111] exist Figure 3A In the embodiment of the present application, the data processing device can execute the scene similarity evaluation method.

[0112] Figure 3B Another structural diagram of the data processing system provided in the embodiment of the present application is shown in FIG. Figure 3B In the terminal device ( Figure 3B In the example, the terminal device is a vehicle. The terminal device can directly obtain at least two traffic scene data (such as road test data, simulation data, etc.), and directly process them by the hardware of the terminal device itself. The specific process is the same as Figure 3A Similarly, please refer to the above description and will not repeat them here.

[0113] Optionally, in Figure 3B In the data processing system shown, the terminal device can receive instructions from the user. For example, the terminal device can obtain at least two traffic scene data (such as road test data, simulation data, etc.) selected by the user in the terminal device, and then the terminal device itself executes a data processing application for the image (for example, computer vision tasks such as classification, segmentation, detection, and image generation), thereby obtaining the corresponding processing results for the image, and displaying the processing results for the user to view and use.

[0114] Optionally, in Figure 3BIn the data processing system shown in the figure, the terminal device can collect images in real time or periodically, and then the terminal device itself executes a data processing application (such as the aforementioned Figure 2 The above-mentioned several scenarios include scene similarity judgment, scene clustering, data mining, coverage statistics, driving guidance, etc.), so as to obtain the processing results.

[0115] exist Figure 3B In the embodiment of the present application, the terminal device itself can execute the scene similarity evaluation method.

[0116] The following describes the scene similarity assessment method provided in the embodiment of the present application. The method can be executed by a data processing device, or by a component of the data processing device (such as a processor, a chip, or a chip system, etc.). The data processing device can be a cloud device (such as the aforementioned Figure 3A as shown), or a terminal device (such as Figure 3B Of course, the method can also be executed by a system consisting of a cloud device and a terminal device (as described above). Figure 3A shown).

[0117] The method provided in the embodiment of the present application is mainly applicable to traffic scenarios. Figure 2 Scene clustering, data mining, coverage statistics, driving guidance, etc. are shown.

[0118] See also Figure 4 , a flow chart of a scene similarity evaluation method provided in an embodiment of the present application, the method may include steps 401 to 405. Steps 401 to 405 are described in detail below.

[0119] Step 401, obtaining at least two traffic scene data.

[0120] The data processing device acquires at least two traffic scene data, each traffic scene data includes multiple frames of data, and each frame of data includes at least one of the following items of at least one object: position, driving speed, shape, timestamp corresponding to the position, direction, etc.

[0121] In the embodiment of the present application, there are multiple ways for the data processing device to obtain data of at least two traffic scenes, which may be through collection / photography, or by receiving data sent by other devices (such as image acquisition devices), or by selecting from a database, etc., which are not limited here.

[0122] In one possible implementation, the data processing device is a self-vehicle. For example, the self-vehicle can collect the traffic scene data of the self-vehicle through an image acquisition device installed on the outside of the self-vehicle body, and can also receive the traffic scene data of other vehicles sent by cloud devices or other vehicles. It should be noted that multiple image acquisition devices, such as cameras, laser radars, millimeter wave radars, ultrasonic radars, etc., can be installed around the body of the self-vehicle. For example, but not limited to 4 image acquisition devices can be installed around the body of the self-vehicle, and the 4 image acquisition devices can be used as front view image acquisition devices, rear view image acquisition devices, left view image acquisition devices, and right view image acquisition devices, respectively. For example, the front view image acquisition device is installed at the center of the front of the vehicle, the rear view image acquisition device is installed at the center of the rear of the vehicle, the left view image acquisition device is installed at the midpoint of the left side of the vehicle along the length direction, and the right view image acquisition device is installed at the midpoint of the right side of the vehicle along the length direction. It is worth noting that the above is only described by taking 4 image acquisition devices as an example. In actual applications, more or fewer image acquisition devices can be installed around the body of the self-vehicle. The image acquisition device can capture images of the road conditions around the vehicle while the vehicle is driving, and transmit the acquisition results to the vehicle or a cloud server.

[0123] In another possible implementation, the data processing device is a cloud device, and the cloud device receives at least two traffic scene data sent by the image acquisition device.

[0124] The objects involved in the embodiments of the present application can be understood as objects that may be encountered in traffic scenes, such as vehicles, obstacles, etc. Obstacles can include dynamic obstacles (such as pedestrians, other vehicles, etc.) and static obstacles (such as cones, signs, etc.). For example, in a scenario where the terminal device is a vehicle, the object can include the vehicle itself, and can also include other vehicles or other vehicles, pedestrians, etc. in the traffic scene.

[0125] Among them, the traffic scene data can be directly collected by the vehicle, or it can be obtained after processing the collected data. The traffic scene data includes multiple frames of data, and each frame of data includes the position, driving speed, direction, etc. of the object in the traffic scene. The above-mentioned position may include the position of the object in a certain coordinate system (such as the coordinates of GPS, Beidou system or other positioning systems, etc.), or it may refer to the relative position between objects. The driving speed is used to indicate the running speed of the object. Each frame of data may also include the shape of the object, such as the outline size and body shape of the object (for example, when the object is a pedestrian), etc. Each frame of data may also include the time information corresponding to the frame data (such as a timestamp), and the timestamp may correspond to the position of the above-mentioned object.

[0126] Optionally, each frame of data may also include topological information of roads in the traffic scene.

[0127] Optionally, the data processing device can also process the acquired traffic scene data. For example, the trajectory points of each object in the traffic scene can be periodically sampled, and the corresponding traffic scene data can be obtained after a preset time. Exemplarily, the data processing device samples each object in the traffic scene every 0.5 seconds to obtain a sampling point, and after 6 seconds, the information of 12 sampling points will be obtained. The information of each sampling point may include at least one of the following: the corresponding object identification, the location of the object, the driving speed of the object, the direction of the object, the timestamp of the sampling point, etc.

[0128] For example, the at least two traffic scenes include a first traffic scene and a second traffic scene. The first traffic scene is as follows: Figure 5A As shown, an example of the second traffic scene is as follows Figure 5B Correspondingly, an example of the first traffic scene data is as follows Fig. 6A As shown, an example of the second traffic scene data is as follows Figure 6B As shown. Wherein, the horizontal axis and the vertical axis are distances, and the points or vertical lines represent sampling points. The first traffic scene data includes the self-vehicle (i.e., the first self-vehicle), three social vehicles (i.e., the first object), and the sampling point information of each object in the first traffic scene data. The second traffic scene data includes the self-vehicle (i.e., the second self-vehicle), four social vehicles (i.e., the second object), and the sampling point information of each object in the second traffic scene data. The above-mentioned sampling point information may include at least one of the following: the object identification of the sampling point (not shown in the figure), the location of the sampling point, the sampling point time, and the driving speed at the sampling point time (not shown in the figure), the direction of the object at the sampling point time, etc.

[0129] In addition, the sampling points can also be understood as the trajectory points of the collected object movement. The number of sampling points can be set according to actual needs. The number of sampling points corresponding to different objects can be the same or different, which is not specifically limited here.

[0130] Step 402: Determine the interaction coefficient of the sampling point corresponding to each object in each traffic scene data.

[0131] After the data processing device acquires at least two traffic scene data, it can determine the interaction coefficient of the sampling point corresponding to each object in each traffic scene data.

[0132] The interaction coefficient is used to indicate the degree to which the driving of an object is affected by other objects, or it can be understood as indicating the importance of a sampling point.

[0133] It is understandable that, in the process of calculating the interaction coefficient, the driving conditions of the object may also be determined in combination with traffic rules, so that the interaction coefficient is more in line with traffic rules.

[0134] There are many ways to calculate the interaction coefficient in the embodiments of the present application. It can be calculated based on a safety factor, or it can be calculated based on the interaction point between objects (such as the distance or time to reach the interaction point, etc.), or it can be calculated based on the speed, direction and distance between objects, etc., and the specifics are not limited here.

[0135] For example, taking the time interval of each object reaching the interaction point as an example to calculate the interaction coefficient, the smaller the time interval, the higher the degree of interaction, and the larger the interaction coefficient.

[0136] Optionally, the process of the data processing device acquiring the interaction coefficient includes: first acquiring the safety factor of the sampling point, and then determining the interaction coefficient according to the safety factor and the relative distance, wherein the relative distance is used to indicate the distance between the sampling point and other objects.

[0137] Specifically, the data processing device first obtains the safety factor of the sampling point, which is used to indicate the collision risk of the sampling point. There are many ways to obtain the safety factor, which can be a collision time interval (TTC), a collision risk field, or a reachable set, etc., which are not limited here.

[0138] The data processing device then determines the interaction coefficient based on the safety factor and the relative distance. For example, the safety factor and the mapping table determine the interaction coefficient, and the mapping table is used to indicate the association between the safety factor, the interaction coefficient and the relative distance. Table 1 gives an example of the mapping table:

[0139] Table 1

[0140]

[0141] It can be seen that for a certain sampling point, the farther other objects are from the sampling point, the smaller the corresponding safety factor weight is, indicating that the influence on the passage of the sampling point is weaker. The closer other objects are to the sampling point, the larger the corresponding safety factor weight is, indicating that the influence on the passage of the sampling point is stronger. Therefore, the interaction coefficient obtained based on the safety factor and the relative distance can reflect the degree to which the passage of the sampling point is affected by other objects, thereby reflecting the interaction relationship or degree of interaction between objects.

[0142] Exemplarily, the first traffic scene data is recorded as represents the trajectory of the i-th object in the first traffic scene data, and n is the number of objects in the first traffic scene data. Represents the trajectory of the i-th object in the first traffic scene data The interaction coefficient corresponding to the a-th sampling point b represents the number of sampling points included in the i-th object in the first traffic scene data. The second traffic scene data is recorded as represents the trajectory of the jth object in the second traffic scene data, and m is the number of objects in the second traffic scene data. Represents the trajectory of the jth object in the second traffic scene data The interaction coefficient corresponding to the cth sampling point d represents the number of sampling points included in the j-th object in the second traffic scene data.

[0143] Step 403: Calculate the trajectory similarity between the two objects based on the interaction coefficient.

[0144] After the data processing device obtains the interaction coefficient, it can calculate the trajectory similarity between the two objects based on the interaction coefficient.

[0145] The two objects include any one object in the first traffic scene data and any one object in the second traffic scene data.

[0146] It is understandable that the selection of any of the above objects can be set according to actual needs. For example, when it is necessary to calculate the trajectory similarity between self-vehicles in different traffic scenes, any object in the above traffic scenes is the self-vehicle. For another example, when it is necessary to calculate the trajectory similarity between social vehicles in different traffic scenes, any object in the above traffic scenes is the social vehicle. Of course, social vehicles can also be considered when calculating the similarity of the self-vehicle trajectory. Self-vehicles, etc. can also be considered when calculating the similarity of social vehicle trajectories. Any object can be determined according to actual needs, and is not limited here.

[0147] The trajectory similarity in the embodiment of the present application includes at least one of the following: a first similarity, a second similarity, a third similarity, and a fourth similarity.

[0148] The first similarity is used to indicate the similarity between the trajectory of the first ego vehicle in the first traffic scene data and the trajectory of the second ego vehicle in the second traffic scene data; the second similarity is used to indicate the similarity between the trajectory of the first object in the first traffic scene data in the coordinate system of the first ego vehicle and the trajectory of the second object in the second traffic scene data in the coordinate system of the second ego vehicle; the third similarity is used to indicate the similarity between the orientation of the first object relative to the first ego vehicle and the orientation of the second object relative to the second ego vehicle; the fourth similarity is used to indicate the similarity between the trajectories of the first object and the second object in the map coordinate system.

[0149] In order to facilitate the understanding of the meaning of the above similarities, Fig. 6A The first traffic scene data shown is Figure 6BThe example of the second traffic scene data shown in FIG. 1 is used to describe the above similarities. The above first similarity and fourth similarity are as follows: Fig. 7A As shown, the second similarity and the third similarity are as follows Figure 7B shown.

[0150] Optionally, the data processing device may calculate any of the above similarities in a manner that includes: first obtaining a first sampling point and a second sampling point, the first sampling point including any two sampling points of an object in the first traffic scene data, and the second sampling point including any two sampling points of an object in the second traffic scene data; calculating difference information between the first sampling point and the second sampling point, the difference information including an angle difference and / or a length difference; and determining the similarity based on the difference information and an interaction coefficient of the sampling points. This will be described in conjunction with the formula later and will not be expanded here.

[0151] It can be understood that the above is a description of calculating any of the above similarities from the perspective of sampling points. In practical applications, the sampling points can also be used to calculate trajectory similarity in the form of vectors. For example, the first vector and the second vector are obtained, the difference information between the first vector and the second vector is calculated, and then the similarity is calculated based on the difference information and the interaction coefficient. Among them, the difference information includes angle difference and / or length difference, the first vector is the vector between any two sampling points of the object in the first traffic scene data, and the second vector is the vector between any two sampling points of the object in the second traffic scene data. For the calculation of different similarities, the above objects used are different.

[0152] For example, in the case of calculating the first similarity, the first vector corresponds to the first ego vehicle in the first traffic scene data, and the second vector corresponds to the second ego vehicle in the second traffic scene data. For another example, in the case of calculating the second similarity, the first vector corresponds to the vector of the first object in the first traffic scene data in the coordinate system of the first ego vehicle, and the second vector corresponds to the vector of the second object in the second traffic scene data in the coordinate system of the second ego vehicle. For another example, in the case of calculating the third similarity, the first vector corresponds to the orientation vector of the first object in the first traffic scene data relative to the first ego vehicle, and the second vector corresponds to the orientation vector of the second object in the second traffic scene data relative to the second ego vehicle. For another example, in the case of calculating the fourth similarity, the first vector corresponds to the vector of the first object in the first traffic scene data in the map coordinate system, and the second vector corresponds to the vector of the second object in the second traffic scene data in the map coordinate system.

[0153] Exemplarily, the function used to calculate the similarity based on the difference information and the interaction coefficient is expressed as follows:

[0154]

[0155]

[0156] in, represents the interaction coefficient of the i+1th sampling point in the first traffic scene data, represents the interaction coefficient of the i-th sampling point in the first traffic scene data, represents the interaction coefficient of the j+1th sampling point in the second traffic scene data, represents the interaction coefficient of the jth sampling point in the second traffic scene data, and v1 represents and The vector between, v2 represents and yaw_cost represents the cosine distance between v1 and v2, which is used to measure the difference in vector angles (i.e., angle difference). length_cost represents the length distance between v1 and v2, which is used to measure the difference in vector lengths (i.e., length difference). The length difference can also be understood as the length multiple of the two vectors, such as 1 or 3.

[0157] Optionally, i and j in the above formula may be equal or unequal (or understood as, the number of sampling points of each object is the same or different). For the convenience of calculation, i=j is often used.

[0158] It is understandable that the above formula is only an example, and in practical applications, there may be other situations, which are not specifically limited here.

[0159] Based on the above four similarities, the trajectory similarity of the objects in the first traffic scene data and the second traffic scene data can be obtained.

[0160] Exemplarily, the trajectory similarity between the first ego vehicle and the second ego vehicle is:

[0161]

[0162] Among them, the above is the first similarity, which is used to measure the similarity of the vehicle trajectory.

[0163] Exemplarily, the trajectory similarity of the first object in the first traffic scene data and the second object in the second traffic scene data is any one of the following:

[0164] (i.e., the trajectory similarity in this example = the second similarity); (i.e., trajectory similarity in this example = first similarity + second similarity);

[0165] (i.e., trajectory similarity in this example = first similarity + third similarity);

[0166] (i.e., trajectory similarity in this example = second similarity + third similarity);

[0167] (i.e., trajectory similarity in this example = first similarity + second similarity + third similarity);

[0168] in, is the first similarity, which is used to indicate the similarity of the vehicle trajectory. is the second similarity, which is used to indicate the trajectory similarity of the social car in the ego-car coordinate system. The third item is the third similarity, which is used to indicate the similarity of the social car in the orientation of the ego car. is the fourth similarity, which is used to indicate the trajectory similarity of the social car in the map coordinate system.

[0169] It is understandable that the above-mentioned several cases of calculating the trajectory similarity of the first object in the first traffic scene data and the second object in the second traffic scene data are just examples. In practical applications, there may be other cases, such as the trajectory similarity in this example = the third similarity + the fourth similarity. For another example, the trajectory similarity in this example = the first similarity + the fourth similarity. For another example, the trajectory similarity in this example = the second similarity + the fourth similarity. For another example, the trajectory similarity in this example = the third similarity + the fourth similarity. For another example, the trajectory similarity in this example = the first similarity + the second similarity + the fourth similarity. For another example, the trajectory similarity in this example = the first similarity + the third similarity + the fourth similarity, etc., which are not specifically limited here.

[0170] Exemplarily, the trajectory similarity of the first object in the first traffic scene data and the second object in the second traffic scene data may also be:

[0171]

[0172] Among them, the first item is the first similarity, which is used to indicate the similarity of the ego vehicle's trajectory, the second item is the second similarity, which is used to indicate the similarity of the trajectory of the social vehicle in the ego vehicle's coordinate system, the third item is the third similarity, which is used to indicate the similarity of the social vehicle in the ego vehicle's orientation, and the fourth item is the fourth similarity, which is used to indicate the similarity of the trajectory of the social vehicle in the map coordinate system.

[0173] It is understandable that in practical applications, the above formula for calculating trajectory similarity may have other situations. For example, trajectory similarity includes at least one of the above four items. Specifically, it may include only the first item, or only the second item, or only the third item, or only the fourth item, or include the first item and the second item, or include the first item, the second item, and the third item, etc.

[0174] Step 404: determine at least one group of objects by trajectory similarity.

[0175] After the data processing device obtains the trajectory similarity, it can determine at least one group of objects through the trajectory similarity, each group of objects includes two objects, and the trajectory similarity of the two objects in each group of objects meets the preset condition. This step 404 can also be understood as a process of determining key obstacles.

[0176] The preset condition includes any one of the following: the trajectory similarity of two objects in each group of objects is less than or equal to a threshold, the trajectory similarity of two objects in each group of objects is greater than or equal to a threshold, etc. Specifically, if a higher trajectory similarity value indicates a more similar trajectory, then the preset condition is: greater than or equal to a threshold; otherwise, then the preset condition is less than or equal to a threshold.

[0177] Optionally, the data processing device first calculates the trajectory similarity between each trajectory in pairs, and then searches for the corresponding best matching trajectory for each trajectory according to the optimal solution method (that is, the two objects corresponding to each trajectory and its corresponding best matching trajectory can be called a group of objects).

[0178] It should be noted that the above pairwise calculation of the trajectory similarity between each trajectory may include the self-vehicle or exclude the self-vehicle. In other words, only the trajectory similarity between social vehicles may be calculated, or the self-vehicle may be considered when calculating the trajectory similarity between social vehicles, or social vehicles may be considered when calculating the trajectory similarity between self-vehicles, etc., and the specifics are not limited here.

[0179] For example, Figure 8 The flowchart of step 404 is shown. The data processing device first calculates the trajectory similarity between each trajectory to obtain a trajectory similarity matrix, and then searches for the corresponding best matching trajectory (ie, determines at least one group of objects) according to the optimal solution method.

[0180] In addition, the above-mentioned optimal solution algorithm may be a Kuhn-Munkres algorithm (KM) or the like. In practical applications, other optimal solution algorithms may also be used, which are not specifically limited here.

[0181] Of course, before the optimal solution algorithm is used for matching, the trajectories can also be preliminarily screened, and the trajectories left after the screening are processed by the optimal solution algorithm. Among them, the preliminary screening rules can be to eliminate trajectories with trajectory similarity values ​​less than a threshold, or to screen based on user operations (for example, trajectories determined by user operations are eliminated or retained), etc., which are not limited here. Through the preliminary screening of trajectories, the computing power of the subsequent use of the optimal solution algorithm can be reduced.

[0182] For example, continuing the above Fig. 6A and Figure 6B For example, Fig. 9 As shown, scene 1 includes the self-car, social car 1, social car 2 and social car 3. Scene 2 includes the self-car, social car a, social car b, social car c and social car d.

[0183] Exemplarily, at least one group of objects determined in step 404 includes social car 3 and social car b, and the trajectory similarity between social car 3 and social car b is 15.

[0184] Step 405: determine scene similarity based on at least one group of objects.

[0185] After the data processing device determines at least one group of objects, it may determine the similarity (which may be referred to as scene similarity) of at least two traffic scene data based on the at least one group of objects.

[0186] After the device obtains at least one group of objects in the data, the scene similarity can be determined in a variety of ways. For example, the trajectory similarity of each group of objects is weighted averaged to obtain the scene similarity. For another example, the trajectory similarity of each group of objects is directly summed to obtain the scene similarity, etc. For another example, the trajectory similarity of all objects (for example, a preset value can be set if no matching object is found) is weighted averaged to obtain the scene similarity, etc., which is not limited here.

[0187] Since the interaction coefficient of each sampling point is taken into account in the above scene similarity calculation process, the scene similarity can reflect the degree of interaction between objects.

[0188] In the above embodiments, the similarity of the scene is obtained by weighted averaging the trajectory similarity of each group of objects or weighted averaging the trajectory similarity of all objects. The data processing device first determines the trajectory interaction coefficient of the trajectory to which the sampling point belongs based on the interaction coefficient of the sampling point, and then uses the trajectory interaction coefficient of the object as the weight coefficient of the trajectory similarity.

[0189] For example, for the first traffic scene data The data processing device sums the interaction coefficients of all sampling points contained in the ith object in the first traffic scene data to obtain the trajectory interaction coefficient of the ith object, that is, For another example, the data processing device selects the interaction coefficients of the sampling points greater than the threshold from the interaction coefficients of all the sampling points contained in the i-th object in the first traffic scene data and sums them to obtain the trajectory interaction coefficient of the i-th object. For another example, all the interaction coefficients of the i-th object are sorted by size, and then the interaction coefficients of the first K sampling points are selected and summed to obtain the trajectory interaction coefficient of the i-th object, where K is an integer greater than 0, etc. The specific calculation method of the trajectory interaction coefficient is not limited here.

[0190] Example 1: Continuing with the above scenario example, taking the weighted average of the trajectory similarities of each group of objects to obtain the scene similarity, the scene similarity is as follows:

[0191]

[0192] Among them, G 3b represents the trajectory similarity between social car 3 and social car b, G b3 It represents the trajectory similarity between social car b and social car 3, A represents the trajectory interaction coefficient corresponding to social car 3, and B represents the trajectory interaction coefficient corresponding to social car b. A and B can also be understood as the weight coefficients of trajectory similarity.

[0193] Example 2: Continuing with the above scene example, taking the weighted average of the trajectory similarities of all objects to obtain the scene similarity, the scene similarity is as follows:

[0194]

[0195] Among them, G1 represents the trajectory similarity of social car 1, G2 represents the trajectory similarity of social car 2, and G 3b represents the trajectory similarity between social car 3 and social car b, G a represents the trajectory similarity of social car a, G b3 represents the trajectory similarity between social car b and social car 3, G c represents the trajectory similarity of social car c, G d represents the trajectory similarity of the social vehicle d, A, B, C, D, E, F, and G represent the trajectory interaction coefficients corresponding to each object, which can also be understood as the weight coefficients of trajectory similarity.

[0196] In the embodiment of the present application, the scene similarity can be expressed in the form of a score value or a degree description, etc., which is not specifically limited here. In a possible implementation method, the scene similarity is expressed in the form of a score value, and the value range of the score value can be set according to actual needs. For example, it can be a 1-point system (for example, 0.4 points, 0.8 points, etc.), or a 10-point system (for example, 4.9 points, 5.8 points, etc.), or a percentage system (for example, 60 points, 80 points, etc.), etc. The value range and precision (or understood as decimal point precision) are not limited here. It can be understood that the higher the score, the more similar it is, or the lower the score, the more similar it is, which is not specifically limited here.

[0197] For example, taking the 10-point scale as an example, the above Figure 5A and Figure 5B The scene similarity is 4.5 points, where 0-5 points represent very similar, 5-8 points represent relatively similar, and 8-10 points represent not very similar.

[0198] In another possible implementation, the scene similarity is expressed in the form of a degree description, such as very similar, relatively similar, not very similar, 50% similarity, 80% similarity, etc., which is not specifically limited here.

[0199] In the embodiment of the present application, on the one hand, in the process of calculating the similarity of traffic scenes, multi-frame data containing multiple objects are considered, and a comprehensive score of scene similarity can be quantified, so that it can be effectively applied to upper-level applications such as data mining and scene clustering. On the other hand, the strength of the traffic impact between objects can be considered by introducing an interaction coefficient, so that the scene similarity can be calculated based on at least one group of objects determined in the traffic scene according to the interaction coefficient, so that the interaction relationship between objects can be reflected in the scene similarity, and the rationality of the quantitative scoring of scene similarity can be improved.

[0200] Optionally, Figure 4 Another process expression form of the embodiment shown can be as follows Fig.10 As shown, first obtain the two scenes to be compared, and then perform multi-frame multi-vehicle similarity matching to obtain a quantitative score of scene similarity. Multi-frame multi-vehicle similarity matching includes: calculating the interaction coefficient, calculating the scene similarity (including: key obstacle screening, single track similarity matching, and calculating the weighted average of track distances, etc.).

[0201] In addition, in order to improve the processing efficiency of scene similarity, before determining the interaction coefficient of the sampling point corresponding to each object in each traffic scene data, the classification of at least two traffic scene data can be judged, and the scene similarity can be calculated when the classification of at least two traffic scene data is the same. The corresponding method flow is as follows Fig.11 As shown, the method may include steps 1101 to 1103. Steps 1101 to 1103 are described in detail below. Or it can be understood that the aforementioned Figure 4 The illustrated embodiment may further include a step of determining that the classifications of at least two traffic scene data are the same.

[0202] Step 1101, obtaining at least two traffic scene data.

[0203] This step 1101 is similar to the above Figure 4 Step 401 in the illustrated embodiment is similar, and reference may be made to the relevant description, which will not be repeated here.

[0204] Step 1102, determining that at least two traffic scene data have the same classification.

[0205] After the data processing device acquires at least two traffic scene data, it first determines whether the at least two traffic scene data are classified the same. Alternatively, it can be understood that the classification of the at least two traffic scene data is preliminarily determined, and if the classification is the same, step 1103 is executed.

[0206] In the embodiment of the present application, the method for determining classification can be single-frame classification or multi-frame classification, etc., which is not limited here.

[0207] For example, in a single-frame dimension, topological information includes at least one of the following: intersection, roadside edge, distance from an object to a lateral lane (a lane perpendicular to the lane where the object is located is called a lateral lane) (also called a vertical edge), traffic lights, lane information, etc. For another example, in a multi-frame dimension, topological information includes at least one of the following: highway, main road / branch road, urban village, etc.

[0208] The above classification includes at least one of the following: classification of object behavior, classification of topological information of the road where the object is located, classification of interactive behavior between objects, etc. Object behavior can also be understood as lane changing behavior, driving behavior, etc. of objects such as vehicles on traffic roads. Topological information can also be understood as the type of road or lane where the object is located, etc. Interactive behavior can be understood as the correlation between driving routes between objects, etc.

[0209] For example, the above categories are as follows: Fig.12 As shown, the classification of object behavior includes at least one of the following: initial state (or understood as a state other than subsequent state), left turn state, right turn state, left lane change, right lane change, left turn state, right turn state, left turn intermediate state (or understood as a state of possible left turn), right turn intermediate state (or understood as a state of possible right turn), etc. Topological information classification includes global topological classification and single-frame topological classification. Global topological classification includes at least one of the following: expressway, main road, branch road, etc.; single-frame topological classification includes at least one of the following: road intersection, road test edge, vertical edge, red light, etc. Interaction behavior classification includes at least one of the following: cut-in, merge, intersection interaction, reverse, etc.

[0210] It is understandable that a traffic scene data may have one or more classification labels. Figure 5A The categories are: the vehicle goes straight, other vehicles turn left, intersection interaction, main road, etc.

[0211] In addition, the classification method in this step can be determined based on a neural network or a classifier, or based on manual experience, etc., which is not limited here.

[0212] Step 1103: Determine the scene similarity between at least two traffic scene data.

[0213] This step 1103 is similar to the above Figure 4 Steps 402 to 405 in the illustrated embodiment are similar, and reference may be made to the relevant descriptions, which will not be repeated here.

[0214] In a possible implementation, the at least two traffic scene data include first traffic scene data and second traffic scene data, the first traffic scene data is drive test data, and the second traffic scene data is simulation data. The data processing device can also determine the simulation coverage rate based on the scene similarity, and the simulation coverage rate is used to indicate the proportion of the simulation data covering the drive test data. That is, it is applied to the scene covered by the aforementioned scene.

[0215] In another possible implementation, the data processing device can also obtain multiple training data based on scene similarity, and the training data is used to train a neural network, and the neural network is applied to decision-making planning and control in the field of intelligent driving, that is, applied to the aforementioned data mining scenario.

[0216] In another possible implementation manner, the data processing device may also cluster at least two traffic scene data based on scene similarity, that is, the scenario applied to the aforementioned scene clustering.

[0217] In another possible implementation, the at least two traffic scene data include third traffic scene data and fourth traffic scene data, the third traffic scene data belongs to the road test data, and the fourth traffic scene data belongs to the database data. The data processing device can also determine the driving route of the first object in the third traffic scene data based on the scene similarity and the fourth traffic scene data, that is, the scene applied to the aforementioned driving guidance.

[0218] In the embodiments of the present application, on the one hand, multiple vehicles, multiple frames and topological scenes are comprehensively considered in the similarity matching, and a comprehensive score of similarity is given quantitatively, so that this function can be more effectively applied to upper-level applications such as data mining and scene clustering. On the other hand, the concept of interaction coefficient is introduced to quantify the degree of interaction between social vehicles and self-vehicles, and the strength of the influence of social vehicles on self-vehicles, so that the algorithm can pay more attention to the differences in grabbing and giving way relationships, interaction patterns, etc. in the field of regulation and control. On the other hand, before calculating the scene similarity, the classification is judged to reduce the computing power loss of subsequent calculations of scene similarity.

[0219] The scene similarity evaluation method in the embodiment of the present application is described above. The data processing device in the embodiment of the present application is described below. Fig.13 , an embodiment of the data processing device in the embodiment of the present application includes:

[0220] An acquisition unit 1301 is used to acquire at least two traffic scene data, each of which includes multiple frames of data, and each frame of data includes at least one of the following items of at least one object: position, driving speed, shape, and a timestamp corresponding to the position;

[0221] A determination unit 1302 is used to determine an interaction coefficient of a sampling point corresponding to each object in each traffic scene data, where the interaction coefficient is used to indicate the degree to which the travel of the object is affected by other objects;

[0222] A calculation unit 1303 is used to calculate the trajectory similarity between two objects based on the interaction coefficient, the at least two traffic scene data include first traffic scene data and second traffic scene data, and the two objects include any object in the first traffic scene data and any object in the second traffic scene data;

[0223] The determining unit 1302 is further configured to determine at least one group of objects by trajectory similarity, wherein trajectory similarity of two objects in each group of objects in the at least one group of objects meets a preset condition;

[0224] The determining unit 1302 is further configured to determine the similarity of at least two traffic scene data based on at least one group of objects.

[0225] Optionally, the trajectory similarity includes at least one of the following: a first similarity, a second similarity, a third similarity, and a fourth similarity;

[0226] The first similarity is used to indicate the similarity between the trajectory of the first ego vehicle in the first traffic scene data and the trajectory of the second ego vehicle in the second traffic scene data.

[0227] The second similarity is used to indicate the trajectory similarity between the trajectory of the first object in the first traffic scene data in the coordinate system of the first vehicle and the trajectory of the second object in the second traffic scene data in the coordinate system of the second vehicle.

[0228] The third similarity is used to indicate the similarity between the orientation of the first object relative to the first ego vehicle and the orientation of the second object relative to the second ego vehicle,

[0229] The fourth similarity is used to indicate the similarity of the trajectories of the first object and the second object in the map coordinate system.

[0230] Optionally, the calculation unit 1303 is specifically used to obtain a first sampling point and a second sampling point, the first sampling point includes any two sampling points of an object in the first traffic scene data, and the second sampling point includes any two sampling points of an object in the second traffic scene data; the calculation unit 1303 is specifically used to calculate difference information between the first sampling point and the second sampling point, the difference information includes angle difference and / or length difference; the calculation unit 1303 is specifically used to calculate trajectory similarity based on the difference information and the interaction coefficient.

[0231] Optionally, the determination unit 1302 is specifically used to obtain a safety factor of the sampling point, where the safety factor is used to indicate the collision risk of the sampling point; the determination unit 1302 is specifically used to determine an interaction coefficient based on the safety factor and a relative distance, where the relative distance is used to indicate the distance between the sampling point and other objects.

[0232] Optionally, the determination unit 1302 is further configured to determine that the classifications of at least two traffic scene data are the same.

[0233] Optionally, the classification includes at least one of the following: classification of object behavior, classification of topological information of the road where the object is located, and classification of interaction behavior between objects.

[0234] Optionally, each frame of data further includes topological information of a road where at least one object is located.

[0235] Optionally, the scene similarity is the sum of trajectory similarities between at least one group of objects or the weighted sum of trajectory similarities between at least one group of objects.

[0236] Optionally, the scene similarity is the sum of similarities of trajectories corresponding to all objects in at least two traffic scene data or the weighted sum of similarities of trajectories corresponding to all objects in at least two traffic scene data.

[0237] Optionally, the device is applied to at least one of the following scenarios: a scene coverage scenario, a data mining scenario, a clustering scenario, and a driving guidance scenario.

[0238] In this embodiment, the operations performed by each unit in the data processing device are the same as those described above. Figures 1 to 12 The description in the illustrated embodiment is similar and will not be repeated here.

[0239] In this embodiment, on the one hand, in the process of calculating the similarity of traffic scenes, considering the multi-frame data containing multiple objects, the determination unit 1302 can quantify and give a comprehensive score for the scene similarity, so that it can be effectively applied to upper-level applications such as data mining and scene clustering. On the other hand, the strength of the traffic impact between objects can be considered by introducing an interaction coefficient, so that the scene similarity can be calculated based on at least one group of objects determined in the traffic scene according to the interaction coefficient, so that the interaction relationship between objects, such as the grab and give relationship or interaction mode, can be reflected in the scene similarity, thereby improving the rationality of the quantitative scoring of the scene similarity. On the other hand, before determining the scene similarity, the determination unit 1302 can also judge the classification, so as to reduce the computing power loss of the subsequent calculation of the scene similarity.

[0240] See also Fig.14 , a schematic diagram of the structure of another data processing device provided by the present application. The data processing device may include a processor 1401, a memory 1402 and a communication port 1403. The processor 1401, the memory 1402 and the communication port 1403 are interconnected via a line. The memory 1402 stores program instructions and data.

[0241] The memory 1402 stores the aforementioned Figures 1 to 12 In the corresponding embodiment shown, the steps executed by the data processing device correspond to program instructions and data.

[0242] Processor 1401, used to execute the above Figures 1 to 12 The steps shown in any of the illustrated embodiments are performed by a data processing device.

[0243] Communication port 1403 can be used to receive and send data, and to perform the above Figures 1 to 12 The steps related to obtaining, sending and receiving in any of the embodiments shown.

[0244] In one implementation, the data processing device may include Fig.14 More or fewer components are merely illustrative in this application and are not limiting.

[0245] An embodiment of the present application further provides a computer-readable storage medium storing one or more computer-executable instructions. When the computer-executable instructions are executed by a processor, the method described in the possible implementation manner of the data processing device in the aforementioned embodiment is implemented.

[0246] The embodiment of the present application also provides a vehicle, the vehicle comprising the aforementioned Figure 4 The data processing device in the embodiment shown in the figure may include the aforementioned Fig.13 The data processing device shown, or comprising the aforementioned Fig.14 The data processing equipment shown.

[0247] An embodiment of the present application also provides a computer program product (or computer program) storing one or more computers, wherein the computer program product includes computer instructions, and when the computer instructions are executed by the processor, a method is implemented as a possible implementation method of the above-mentioned data processing device.

[0248] The embodiment of the present application also provides a chip system, which includes at least one processor for supporting a terminal device to implement the functions involved in the possible implementation of the above-mentioned data processing device. Optionally, the chip system also includes an interface circuit, which provides program instructions and / or data for the at least one processor. In one possible design, the chip system may also include a memory, which is used to store the necessary program instructions and data for the terminal device. The chip system may be composed of chips, or may include chips and other discrete devices.

[0249] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0250] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0251] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0252] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0253] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, read-only memory), random access memory (RAM, random access memory), disk or optical disk and other media that can store program code.

Claims

1. A scene similarity assessment method, characterized in that: The method comprises: Acquire at least two traffic scene data, each traffic scene data includes multiple frames of data, and each frame of data includes at least one of the following items of at least one object: position, driving speed, shape, and a timestamp corresponding to the position; Determine an interaction coefficient of a sampling point corresponding to each object in each traffic scene data, wherein the interaction coefficient is used to indicate the degree to which the travel of the object is affected by other objects; Calculating the trajectory similarity between two objects based on the interaction coefficient, the at least two traffic scene data include first traffic scene data and second traffic scene data, and the two objects include any object in the first traffic scene data and any object in the second traffic scene data; Determine at least one group of objects by using the trajectory similarity, wherein the trajectory similarity of two objects in each group of objects in the at least one group of objects meets a preset condition; The similarity of the at least two traffic scene data is determined based on the at least one group of objects.

2. The method according to claim 1, characterized in that: The trajectory similarity includes at least one of the following: a first similarity, a second similarity, a third similarity, and a fourth similarity; The first similarity is used to indicate the similarity between the trajectory of the first ego vehicle in the first traffic scene data and the trajectory of the second ego vehicle in the second traffic scene data. The second similarity is used to indicate the similarity between the trajectory of the first object in the first traffic scene data in the coordinate system of the first vehicle and the trajectory of the second object in the second traffic scene data in the coordinate system of the second vehicle. The third similarity is used to indicate a similarity between the position of the first object relative to the first ego vehicle and the position of the second object relative to the second ego vehicle, The fourth similarity is used to indicate the similarity between the trajectories of the first object and the second object in the map coordinate system.

3. The method according to claim 1 or 2, characterized in that: The calculating the trajectory similarity between the two objects based on the interaction coefficient includes: Acquire a first sampling point and a second sampling point, wherein the first sampling point includes any two sampling points of an object in the first traffic scene data, and the second sampling point includes any two sampling points of an object in the second traffic scene data; Calculating difference information between the first sampling point and the second sampling point, the difference information including an angle difference and / or a length difference; The trajectory similarity is determined based on the difference information and the interaction coefficient.

4. The method according to any one of claims 1 to 3, characterized in that: The determining of the interaction coefficient of the sampling point corresponding to each object in each traffic scene data comprises: Acquire a safety factor of the sampling point, where the safety factor is used to indicate a collision risk of the sampling point; The interaction coefficient is determined based on the safety factor and a relative distance, where the relative distance is used to indicate the distance between the sampling point and other objects.

5. The method according to any one of claims 1 to 4, characterized in that Before determining the interaction coefficient of the sampling point corresponding to each object in each traffic scene data, the method further includes: It is determined that the classifications of the at least two traffic scene data are the same.

6. The method according to claim 5, characterized in that The classification includes at least one of the following: classification of object behavior, classification of topological information of the road where the object is located, and classification of interaction behavior between objects.

7. The method according to any one of claims 1 to 6, characterized in that Each frame of data also includes topological information of the road where the at least one object is located.

8. The method according to any one of claims 1 to 7, characterized in that The scene similarity is the sum of the trajectory similarities between the at least one group of objects or the weighted sum of the trajectory similarities between the at least one group of objects.

9. The method according to any one of claims 1 to 7, characterized in that The scene similarity is the sum of similarities of trajectories corresponding to all objects in the at least two traffic scene data or the weighted sum of similarities of trajectories corresponding to all objects in the at least two traffic scene data.

10. The method according to any one of claims 1 to 9, characterized in that The method is applied to at least one of the following scenarios: The scenarios include coverage scenarios, data mining scenarios, clustering scenarios, and driving guidance scenarios.

11. A data processing device, characterized in that: The device comprises: An acquisition unit, used to acquire at least two traffic scene data, each traffic scene data includes multiple frames of data, and each frame of data includes at least one of the following items of at least one object: position, driving speed, shape, and a timestamp corresponding to the position; A determination unit, used to determine an interaction coefficient of a sampling point corresponding to each object in each traffic scene data, wherein the interaction coefficient is used to indicate the degree to which the travel of the object is affected by other objects; a calculation unit, configured to calculate a trajectory similarity between two objects based on the interaction coefficient, wherein the at least two traffic scene data include first traffic scene data and second traffic scene data, and the two objects include any one object in the first traffic scene data and any one object in the second traffic scene data; The determining unit is further configured to determine at least one group of objects by using the trajectory similarity, wherein the trajectory similarity of two objects in each group of objects in the at least one group of objects satisfies a preset condition; The determining unit is further configured to determine the similarity of the at least two traffic scene data based on the at least one group of objects.

12. The device according to claim 11, characterized in that The trajectory similarity includes at least one of the following: a first similarity, a second similarity, a third similarity, and a fourth similarity; The first similarity is used to indicate the similarity between the trajectory of the first ego vehicle in the first traffic scene data and the trajectory of the second ego vehicle in the second traffic scene data. The second similarity is used to indicate the similarity between the trajectory of the first object in the first traffic scene data in the coordinate system of the first vehicle and the trajectory of the second object in the second traffic scene data in the coordinate system of the second vehicle. The third similarity is used to indicate a similarity between the position of the first object relative to the first ego vehicle and the position of the second object relative to the second ego vehicle, The fourth similarity is used to indicate the similarity between the trajectories of the first object and the second object in the map coordinate system.

13. The device according to claim 11 or 12, characterized in that The computing unit is specifically used for: Acquire a first sampling point and a second sampling point, wherein the first sampling point includes any two sampling points of an object in the first traffic scene data, and the second sampling point includes any two sampling points of an object in the second traffic scene data; Calculating difference information between the first sampling point and the second sampling point, the difference information including an angle difference and / or a length difference; The trajectory similarity is calculated based on the difference information and the interaction coefficient.

14. The device according to any one of claims 11 to 13, characterized in that The determining unit is specifically configured to: Acquire a safety factor of the sampling point, where the safety factor is used to indicate a collision risk of the sampling point; The interaction coefficient is determined based on the safety factor and a relative distance, where the relative distance is used to indicate the distance between the sampling point and other objects.

15. The device according to any one of claims 11 to 14, characterized in that The determining unit is further configured to determine that the classifications of the at least two traffic scene data are the same.

16. The device according to claim 15, characterized in that The classification includes at least one of the following: classification of object behavior, classification of topological information of the road where the object is located, and classification of interaction behavior between objects.

17. The device according to any one of claims 11 to 16, characterized in that Each frame of data also includes topological information of the road where the at least one object is located.

18. The device according to any one of claims 11 to 17, characterized in that The scene similarity is the sum of the trajectory similarities between the at least one group of objects or the weighted sum of the trajectory similarities between the at least one group of objects.

19. The device according to any one of claims 11 to 17, characterized in that The scene similarity is the sum of similarities of trajectories corresponding to all objects in the at least two traffic scene data or the weighted sum of similarities of trajectories corresponding to all objects in the at least two traffic scene data.

20. The device according to any one of claims 11 to 19, characterized in that The device is applied to at least one of the following scenarios: The scenarios include coverage scenarios, data mining scenarios, clustering scenarios, and driving guidance scenarios.

21. A data processing device, characterized in that: include: A processor, the processor is coupled to a memory, the memory is used to store programs or instructions, when the program or instructions are executed by the processor, the data processing device executes the method according to any one of claims 1 to 10.

22. A vehicle, characterized in that: A device comprising the apparatus of any one of claims 11 to 21.

23. A computer-readable storage medium, characterized in that: The method comprises computer instructions which, when executed by a processor, implement the method according to any one of claims 1 to 10.

24. A computer program product, characterized in that The computer program product comprises computer instructions, which, when executed by a processor, implement the method according to any one of claims 1 to 10.

Images