WO2021134172A1 - 一种轨迹预测方法及相关设备 - Google Patents

一种轨迹预测方法及相关设备 Download PDF

Info

Publication number
WO2021134172A1
WO2021134172A1 PCT/CN2019/129828 CN2019129828W WO2021134172A1 WO 2021134172 A1 WO2021134172 A1 WO 2021134172A1 CN 2019129828 W CN2019129828 W CN 2019129828W WO 2021134172 A1 WO2021134172 A1 WO 2021134172A1
Authority
WO
WIPO (PCT)
Prior art keywords
information
target vehicle
trajectory
vehicle
input
Prior art date
Application number
PCT/CN2019/129828
Other languages
English (en)
French (fr)
Inventor
刘亚林
宋晓琳
曾艳兵
曹昊天
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Priority to EP19958672.8A priority Critical patent/EP4075407A4/en
Priority to PCT/CN2019/129828 priority patent/WO2021134172A1/zh
Priority to CN201980080199.4A priority patent/CN113261035B/zh
Publication of WO2021134172A1 publication Critical patent/WO2021134172A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0125—Traffic data processing
    • G08G1/0129—Traffic data processing for creating historical data or processing based on historical data
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W60/00—Drive control systems specially adapted for autonomous road vehicles
    • B60W60/001—Planning or execution of driving tasks
    • B60W60/0027—Planning or execution of driving tasks using trajectory prediction for other traffic participants
    • B60W60/00274—Planning or execution of driving tasks using trajectory prediction for other traffic participants considering possible movement changes
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W2554/00—Input parameters relating to objects
    • B60W2554/40—Dynamic objects, e.g. animals, windblown objects
    • B60W2554/404—Characteristics
    • B60W2554/4045—Intention, e.g. lane change or imminent movement
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W2554/00—Input parameters relating to objects
    • B60W2554/80—Spatial relation or speed relative to objects
    • B60W2554/801—Lateral distance
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W2554/00—Input parameters relating to objects
    • B60W2554/80—Spatial relation or speed relative to objects
    • B60W2554/802—Longitudinal distance
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W2554/00—Input parameters relating to objects
    • B60W2554/80—Spatial relation or speed relative to objects
    • B60W2554/803—Relative lateral speed
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W2554/00—Input parameters relating to objects
    • B60W2554/80—Spatial relation or speed relative to objects
    • B60W2554/804—Relative longitudinal speed
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W2556/00—Input parameters relating to data
    • B60W2556/10—Historical data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/16—Anti-collision systems
    • G08G1/166—Anti-collision systems for active traffic, e.g. moving vehicles, pedestrians, bikes

Definitions

  • This application relates to the field of artificial intelligence, and in particular to a trajectory prediction method and related equipment.
  • driverless cars a major application scenario of driverless cars is mixed traffic flow roads, that is, on roads where driverless cars and driver-driven cars coexist.
  • the driverless car will be affected by the driver driving the car. Therefore, in order to ensure user safety, the driverless car needs to obtain the trajectory of the surrounding drivers driving the car, and control the driverless car based on this trajectory, so as to ensure that the driverless car can reach the destination safely and efficiently.
  • the present application provides a trajectory prediction method and related equipment, which use the driving intention of a target vehicle as the input of trajectory prediction, so that the trajectory prediction of the target vehicle is more accurate.
  • the present application provides a trajectory prediction method, the method includes: acquiring relative motion information between a target vehicle and at least one first vehicle located within the detection range of a sensor, and the first vehicle of the target vehicle Historical trajectory information, wherein the sensor is set in the vehicle, and the sensor may be a radar sensor or a video acquisition sensor; based on the relative motion information, the target information representing the driving intention of the target vehicle is acquired; and the target information and the The first historical trajectory information is input, and the first predicted trajectory of the target vehicle is obtained through the first trajectory prediction model.
  • the relative motion information between the target vehicle and at least one first vehicle and the first historical trajectory information of the target vehicle are acquired; target information representing the driving intention of the target vehicle is acquired based on the relative motion information ; With the target information and the first historical trajectory information as input, the first predicted trajectory of the target vehicle is obtained through a first trajectory prediction model.
  • the driving intention of the target vehicle is used as the input of the trajectory prediction, so that the trajectory prediction of the target vehicle is more accurate.
  • the driving intention includes at least one of the following: changing lanes to the left, changing lanes to the right, or keeping straight.
  • the relative motion information represents relative distance information and/or relative speed information at multiple times between the target vehicle and each first vehicle
  • the first historical trajectory information includes The first historical track position of the target vehicle at multiple times.
  • the obtaining target information representing the driving intention of the target vehicle based on the relative motion information includes: taking the relative motion information as an input, and obtaining the The profit value of each driving intention of the target vehicle at each time; taking the profit value of each driving intention of the target vehicle at each time as input, through the cyclic neural network, the target information representing the driving intention of the target vehicle is obtained.
  • the interaction between the target vehicle and the surrounding first vehicle is modeled by the profit function, and then the profit value is used as the input of the driving intention recognition model, which can more accurately identify the driving intention of the target vehicle while reducing The complexity of the model, and has a higher model interpretability.
  • the first historical trajectory information and the first predicted trajectory are components of the driving trajectory of the target vehicle in a direction perpendicular to the lane, and the lane is the target vehicle The lane you are in.
  • the first predicted trajectory includes a plurality of predicted trajectory positions, and the target information and the first historical trajectory information are used as input through a first trajectory prediction model
  • Obtaining the first predicted trajectory of the target vehicle includes: taking the target information and the first historical trajectory information as input, and obtaining multiple pieces of coded information through an LSTM encoder, wherein each time corresponds to one piece of coded information;
  • the multiple encoded information is input, and the multiple predicted trajectory positions are obtained through an LSTM decoder, where the LSTM decoder is used to obtain the multiple predicted trajectory positions by decoding based on multiple input data, and each input data
  • the output information of the decoder is effectively and fully utilized, so that the prediction result is more accurate.
  • each input data is obtained by weighting the multiple pieces of coded information.
  • the method further includes: acquiring the motion state of the target vehicle and second historical trajectory information, where the second historical trajectory information is that the driving trajectory of the target vehicle is The component of the parallel direction of the lane, the lane is the lane where the target vehicle is located; taking the motion state of the target vehicle and the second historical trajectory information as input, the second trajectory prediction model is used to obtain the second trajectory of the target vehicle A predicted trajectory, and the second predicted trajectory is a component of the driving trajectory of the target vehicle in a direction parallel to the lane.
  • the lateral and longitudinal target vehicle trajectory prediction models are constructed separately, and the driving intention of the target vehicle is considered in the lateral and longitudinal directions, so as to obtain a more accurate target vehicle trajectory prediction.
  • this application provides an execution device, including:
  • An acquisition module for acquiring relative motion information between the target vehicle and at least one first vehicle located within the detection range of the sensor, and first historical trajectory information of the target vehicle;
  • the prediction module is configured to use the target information and the first historical trajectory information as input to obtain the first predicted trajectory of the target vehicle through a first trajectory prediction model.
  • the driving intention includes at least one of the following:
  • the relative motion information represents relative distance information and/or relative speed information at multiple times between the target vehicle and each first vehicle
  • the first historical trajectory information includes The first historical track position of the target vehicle at multiple times.
  • the acquisition module is specifically used for:
  • the target information representing the driving intention of the target vehicle is obtained.
  • the first historical trajectory information and the first predicted trajectory are components of the driving trajectory of the target vehicle in a direction perpendicular to the lane, and the lane is the target vehicle The lane you are in.
  • the first predicted trajectory includes multiple predicted trajectory positions
  • the prediction module is specifically configured to:
  • the multiple predicted trajectory positions are obtained through an LSTM decoder, where the LSTM decoder is used to obtain the multiple predicted trajectory positions by decoding based on multiple input data, and each input The data is related to the plurality of coded information.
  • each input data is obtained by weighting the multiple pieces of coded information.
  • the acquisition module is further used for:
  • the second historical trajectory information is the component of the driving trajectory of the target vehicle in a direction parallel to the lane, and the lane is the lane where the target vehicle is located;
  • the prediction module is also used for:
  • the second trajectory prediction model is used to obtain the second predicted trajectory of the target vehicle, and the second predicted trajectory is the driving trajectory of the target vehicle.
  • an embodiment of the present invention provides a terminal device, including a memory, a communication interface, and a processor coupled to the memory and the communication interface; the memory is used to store instructions, and the processor is used to execute the instructions
  • the communication interface is used to communicate with other devices under the control of the processor; wherein, when the processor executes the instructions, the method described in the first aspect or the possible embodiments of the first aspect is executed .
  • a computer-readable storage medium stores program codes for vehicle control.
  • the program code includes instructions for executing the method described in the foregoing first aspect or possible embodiments of the first aspect.
  • a computer program product including instructions, which when run on a computer, causes the computer to execute the method described in the first aspect or the possible embodiments of the first aspect.
  • the relative motion information between the target vehicle and at least one first vehicle located within the detection range of the sensor, and the first historical trajectory information of the target vehicle are acquired; the representation is acquired based on the relative motion information Target information of the driving intention of the target vehicle; taking the target information and the first historical trajectory information as input, and obtaining the first predicted trajectory of the target vehicle through a first trajectory prediction model.
  • the driving intention of the target vehicle is used as the input of the trajectory prediction, so that the trajectory prediction of the target vehicle is more accurate.
  • Figure 1A is a schematic structural diagram of the artificial intelligence main frame
  • FIG. 1B is a schematic structural diagram of a possible terminal device of this application.
  • FIG. 1C is a schematic structural diagram of another possible terminal device of this application.
  • FIG. 2 is a schematic diagram of an embodiment of a trajectory prediction method provided by an embodiment of this application.
  • FIG. 3 is a schematic diagram of a pedestrian behavior intention provided by an embodiment of the application.
  • FIG. 4 is a schematic diagram of historical trajectory information processing provided by an embodiment of this application.
  • Figure 5A is a schematic flow chart of a decoding process
  • Figure 5B is a schematic flow chart of a decoding process
  • FIG. 6 is a schematic diagram of a possible structure of the terminal device involved in this embodiment.
  • FIG. 7 is a schematic structural diagram of another terminal device 700 provided by an embodiment of this application.
  • FIG. 8 is a schematic structural diagram of an execution device provided by an embodiment of this application.
  • FIG. 9 is a schematic diagram of a structure of a chip provided by an embodiment of the application.
  • the embodiments of the present application provide a trajectory prediction method and related equipment, which are used to improve the accuracy of the trajectory prediction of a target vehicle.
  • Figure 1A shows a schematic diagram of the main framework of artificial intelligence.
  • the following section describes the "intelligent information chain” (horizontal axis) and “IT value chain” ( (Vertical axis)
  • the "intelligent information chain” reflects a series of processes from data acquisition to processing. For example, it can be the general process of intelligent information perception, intelligent information representation and formation, intelligent reasoning, intelligent decision-making, intelligent execution and output. In this process, the data has gone through the condensing process of "data-information-knowledge-wisdom”.
  • the "IT value chain” from the underlying infrastructure of human intelligence, information (providing and processing technology realization) to the industrial ecological process of the system, reflects the value that artificial intelligence brings to the information technology industry.
  • the infrastructure provides computing power support for the artificial intelligence system, realizes communication with the outside world, and realizes support through the basic platform.
  • smart chips hardware acceleration chips such as CPU, NPU, GPU, ASIC, FPGA
  • basic platforms include distributed computing frameworks and network related platform guarantees and support, which can include cloud storage and Computing, interconnection network, etc.
  • sensors communicate with the outside to obtain data, and these data are provided to the smart chip in the distributed computing system provided by the basic platform for calculation.
  • the data in the upper layer of the infrastructure is used to represent the data source in the field of artificial intelligence.
  • the data involves graphics, images, voice, text, and IoT data of traditional devices, including business data of existing systems and sensory data such as force, displacement, liquid level, temperature, and humidity.
  • Data processing usually includes data training, machine learning, deep learning, search, reasoning, decision-making and other methods.
  • machine learning and deep learning can symbolize and formalize data for intelligent information modeling, extraction, preprocessing, training, etc.
  • Reasoning refers to the process of simulating human intelligent reasoning in a computer or intelligent system, using formal information to conduct machine thinking and solving problems based on reasoning control strategies.
  • the typical function is search and matching.
  • Decision-making refers to the process of making decisions after intelligent information is reasoned, and usually provides functions such as classification, ranking, and prediction.
  • some general capabilities can be formed based on the results of the data processing, such as an algorithm or a general system, for example, translation, text analysis, computer vision processing, speech recognition, image Recognition and so on.
  • Intelligent products and industry applications refer to the products and applications of artificial intelligence systems in various fields. It is an encapsulation of the overall solution of artificial intelligence, productizing intelligent information decision-making and realizing landing applications. Its application fields mainly include: intelligent terminals, intelligent manufacturing, Intelligent transportation, smart home, smart medical, smart security, autonomous driving, safe city, etc.
  • Fig. 1B shows a schematic structural diagram of a possible terminal device of the present application.
  • the terminal device includes an environment perception module 102, a planning decision module 104, and a control processing module 106.
  • the environment perception module 102 collects obstacle information, surrounding environment information where the terminal device is located, and driving information of the vehicle where the terminal device is located mainly through peripheral systems (such as sensors, cameras, etc.).
  • the obstacle information includes, but is not limited to, information such as the geographic location of the obstacle, the movement speed of the obstacle, the movement direction of the obstacle, the movement acceleration of the obstacle, the variance of the movement direction of the obstacle, the variance of the movement speed of the obstacle, etc.
  • the obstacles include, but are not limited to, vehicles, pedestrians, animate living obstacles, inanimate obstacles, and the like. This application will take the obstacle as an example to describe some embodiments involved in this application.
  • the surrounding environment information includes but is not limited to information such as map information, weather information, intersection type, lane line, number of lanes, whether the road is congested, traffic flow speed, traffic flow acceleration, and the distance between the terminal device and the obstacle.
  • the driving information includes, but is not limited to, the geographic location, driving speed, driving direction, driving acceleration, distance between the vehicle and obstacles, and the like of the vehicle.
  • the terminal equipment includes, but is not limited to, vehicles such as automobiles, trains, trucks, and cars, and communication equipment installed on the vehicles, such as vehicle-mounted equipment.
  • the planning decision module 104 includes a behavior prediction module and a planning module.
  • the behavior prediction module is mainly used to predict the behavior intention of the obstacle (for example, the driving intention of the target vehicle described later in this application) and the historical movement trajectory according to the above-mentioned information collected by the environment perception module.
  • the planning module is used to obtain a corresponding control strategy under the premise of ensuring safety, so as to subsequently use the control strategy to control the vehicle for safe driving.
  • the control strategy is pre-customized on the user side or the terminal device side, or generated according to the behavior intention, which will be described in detail below.
  • the control strategy is used to instruct the vehicle to adjust the corresponding vehicle parameters, so as to realize the safe driving of the vehicle.
  • the control processing module is used to control and adjust the vehicle correspondingly according to the control strategy obtained by the planning decision module to avoid collision between the vehicle and the obstacle. For example, control the vehicle parameters such as the steering wheel angle of the vehicle, driving speed, whether to brake or not, and whether to press the accelerator pedal.
  • FIG. 1C shows a schematic structural diagram of another possible terminal device of the present application.
  • the terminal device 100 may include: a baseband chip 110, a memory 115, including one or more computer-readable storage media, a radio frequency (RF) module 116, and a peripheral system 117. These components may communicate on one or more communication buses 114.
  • RF radio frequency
  • the peripheral system 117 is mainly used to implement the interactive function between the terminal device 110 and the user/external environment, and mainly includes an input and output device of the terminal 100.
  • the peripheral system 117 may include: a touch screen controller 118, a camera controller 119, an audio controller 120, and a sensor management module 121.
  • each controller can be coupled with its corresponding peripheral devices, such as a touch screen 123, a camera 124, an audio circuit 125, and a sensor 126.
  • the gesture sensor in the sensor 126 may be used to receive gesture control operations input by the user.
  • the speed sensor in the sensor 126 may be used to collect the driving speed of the terminal device itself or to collect the movement speed of obstacles in the environment.
  • the touch screen 123 can be used as a prompting device, which is mainly used to prompt obstacles through screen display, projection, etc., for example, when a pedestrian is crossing a road, the display screen displays text to prompt the pedestrian to speed up walking.
  • the peripheral system 117 may also include other prompting devices such as lights, displays, etc., for interactive prompting between the vehicle and the pedestrian, so as to avoid collisions between the vehicle and the pedestrian. It should be noted that the peripheral system 117 may also include other I/O peripherals.
  • the baseband chip 110 may integrate: one or more processors 111, a clock module 112, and a power management module 113.
  • the clock module 112 integrated in the baseband chip 110 is mainly used to generate a clock required for data transmission and timing control for the processor 111.
  • the power management module 113 integrated in the baseband chip 110 is mainly used to provide a stable and high-precision voltage for the processor 111, the radio frequency module 116, and the peripheral system.
  • the radio frequency (RF) module 116 is used to receive and transmit radio frequency signals, and mainly integrates the receiver and transmitter of the terminal 100.
  • the radio frequency (RF) module 116 communicates with a communication network and other communication devices through radio frequency signals.
  • the radio frequency (RF) module 116 may include, but is not limited to: an antenna system, an RF transceiver, one or more amplifiers, a tuner, one or more oscillators, a digital signal processor, a CODEC chip, a SIM card, and Storage media, etc.
  • the radio frequency (RF) module 116 may be implemented on a separate chip.
  • the memory 115 is coupled with the processor 111, and is used to store various software programs and/or multiple sets of instructions.
  • the memory 115 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic disk storage devices, flash memory devices, or other non-volatile solid-state storage devices.
  • the memory 115 may store an operating system, such as an embedded operating system such as ANDROID, IOS, WINDOWS, or LINUX.
  • the memory 115 may also store a network communication program, which may be used to communicate with one or more additional devices, one or more terminal devices, one or more terminal devices, and so on.
  • This application can be applied to a driving scene, which includes the own vehicle and the target vehicle.
  • This vehicle is any driving vehicle.
  • the target vehicle is the vehicles around the vehicle.
  • the vehicle is used to predict the trajectory of the target vehicle.
  • the vehicle can be an unmanned car or a driver-driving car.
  • the vehicle can be a new energy vehicle or a fuel vehicle.
  • the new energy vehicle can be a pure electric vehicle or a hybrid vehicle.
  • the own vehicle is not specifically limited.
  • the target vehicle can be an unmanned car, or it can be a driver driving a car.
  • the target vehicle can be a new energy vehicle or a fuel vehicle.
  • the target vehicle is also not specifically limited.
  • the target vehicle may be located around the own vehicle, where the surroundings of the own vehicle refer to the area formed by the own vehicle as the center and within a preset radius.
  • the preset radius can be set and changed as needed.
  • the preset radius is not specifically limited.
  • the preset radius can be set based on different road attributes; for example, the preset radius corresponding to an expressway can be 100 meters or 200 meters; the preset radius corresponding to an ordinary road can be 20 meters or 30 meters, etc.
  • the target vehicle may be a vehicle in front of the host vehicle, that is, in the driving direction of the host vehicle.
  • the vehicle needs to predict the driving trajectory of surrounding vehicles, and control the vehicle based on the predicted trajectory, so as to ensure that the driverless car can safely reach the destination.
  • autonomous vehicles need to have the ability to predict the driving intentions and future trajectories of surrounding vehicles, and improve the efficiency and safety of the entire system.
  • the trajectory prediction of the target vehicle can be performed based on the historical trajectory.
  • the historical trajectory information of the target vehicle can be used as input to directly output the predicted trajectory of the target vehicle.
  • the trajectory prediction algorithm based on the historical trajectory can be, for example, Gaussian process regression (GPR), which first maps the low-dimensional state space to the high-dimensional space through the kernel function, and then uses the mean function and the covariance function to convert the history
  • the trajectory and predicted trajectory are modeled as a multi-dimensional Gaussian distribution. Given a historical trajectory, output the predicted future trajectory.
  • GPR Gaussian process regression
  • the Gaussian process regression method is a statistical model based on a large amount of historical trajectory information, and does not consider the current motion state of the target vehicle. In addition, the method does not consider the interaction between the target vehicle and its surrounding vehicles, which does not conform to the actual traffic situation. , The prediction accuracy rate is low.
  • FIG. 2 is a schematic diagram of an embodiment of a trajectory prediction method provided by an embodiment of the application.
  • the trajectory prediction method provided in this embodiment includes:
  • the relative motion information may represent relative distance information and/or relative speed information between the target vehicle and each first vehicle at multiple times
  • the first historical trajectory information includes the target vehicle The first historical track position at each moment.
  • the own vehicle may detect at least one first vehicle based on the sensor carried by itself, and relative motion information between the target vehicle and the at least one first vehicle within the detection range of the sensor.
  • the first vehicle may be a vehicle near the target vehicle.
  • the number of the first vehicle may be determined according to the vehicle conditions around the target vehicle.
  • the target vehicle may be Five cars in the front left, rear left, front right, rear right, and directly in front of the vehicle are the first vehicles.
  • the relative motion information of the five first vehicles relative to the target vehicle is extracted respectively, where the relative motion information can represent the relative distance information and/or relative speed information between the target vehicle and each first vehicle at multiple times, for example ,
  • the relative distance information may be the longitudinal distance d along the driving direction of the lane and the speed ⁇ v between the target vehicle and the first vehicle.
  • the relative motion information corresponding to the first vehicle on the left in FIG. 3 may be set as the following information:
  • the relative motion information corresponding to the first vehicle on the left front in FIG. 3 may be set as the following information:
  • the relative motion information corresponding to the first vehicle on the left rear in FIG. 3 may be set as the following information:
  • the relative motion information corresponding to the first vehicle in the front left and rear left in FIG. Set to the following information:
  • the corresponding relative motion information is set as the following information:
  • the setting of the relative motion information in the above embodiment is only an illustration, and does not constitute a limitation of the present application.
  • the relative motion information of the first vehicle located on the right side of the target vehicle can be set with reference to the above embodiment, which is not limited here.
  • the vehicle may obtain the relative motion information between the target vehicle and the at least one first vehicle through its own sensor for collecting motion information.
  • the sensor for collecting motion information may be a camera or a radar. Wait.
  • the radar may be millimeter wave radar or lidar, etc., which is not limited here.
  • the first historical trajectory information of the target vehicle may also be acquired, where the first historical trajectory information may include the first historical trajectory position of the target vehicle at multiple times, and the first historical trajectory information Represents trajectory information perpendicular to the direction of the lane where the target vehicle is located.
  • the first historical trajectory information may include the first historical trajectory position of the target vehicle at time A, time B, and time C.
  • second historical trajectory information of the target vehicle may also be acquired, where the second historical trajectory information is trajectory information consistent with the direction of the lane where the target vehicle is located.
  • the historical trajectory information of the target vehicle can be obtained.
  • the own vehicle may send a trajectory acquisition request to the target vehicle, and the trajectory acquisition request is used to request acquisition of the first historical trajectory information and the second historical trajectory information of the target vehicle.
  • the target vehicle receives the trajectory acquisition request sent by the own vehicle, and sends the first historical trajectory information and second historical trajectory information of the target vehicle to the own vehicle according to the trajectory acquisition request.
  • the vehicle Before the vehicle sends a trajectory acquisition request to the target vehicle, the vehicle establishes an information transmission channel with the target vehicle.
  • the host vehicle and the target vehicle perform information transmission through the information transmission channel.
  • the information transmission channel may be a vehicle of the Internet (V2I) transmission channel or a short-distance communication (for example, Bluetooth) transmission channel, etc.
  • the own vehicle may obtain the first historical trajectory information and the second historical trajectory information of the target vehicle from the monitoring device.
  • the step for the vehicle to obtain the historical trajectory information of the target vehicle may be: the vehicle sends a trajectory acquisition request to the monitoring device, and the trajectory acquisition request carries the vehicle identification of the target vehicle.
  • the monitoring device receives the trajectory acquisition request sent by the host vehicle, obtains the historical trajectory information of the target vehicle according to the vehicle identification of the target vehicle, and sends the historical trajectory information of the target vehicle to the host vehicle.
  • the host vehicle receives the first historical trajectory information and the second historical trajectory information of the target vehicle sent by the monitoring device.
  • the monitoring equipment is a sensor used to collect motion information, for example, a camera or a radar. In order to improve accuracy, the radar can be millimeter wave radar or lidar.
  • the vehicle identification may be the license plate number of the target vehicle, etc.
  • the driving intention may include at least one of the following: changing lanes to the left, changing lanes to the right, or staying straight.
  • the relative motion information can be used as input to obtain the profit value of each driving intention of the target vehicle at each time through a linear profit function model; and the profit value of each driving intention of the target vehicle at each time may be used as the input.
  • the income value of is input, and the target information representing the driving intention of the target vehicle is obtained through the cyclic neural network.
  • this application uses a linear function as a revenue function to determine the revenue values GL, GR, GK of each driving intention, and the matrix form can be as follows:
  • GL represents the profit value of the target vehicle changing lanes to the left
  • GR represents the profit value of the target vehicle changing lanes to the right
  • GK represents the profit value of the target vehicle keeping straight
  • d lf represents the distance between the target vehicle and the first vehicle in front of the left
  • ⁇ v lf represents the relative speed between the target vehicle and the first vehicle in front of the left
  • d lr represents the distance between the target vehicle and the first vehicle behind the left
  • ⁇ v lr represents the relative speed between the target vehicle and the first vehicle behind the left
  • d rf represents the distance between the target vehicle and the first vehicle on the right front
  • ⁇ v rf represents the relative speed between the target vehicle and the first vehicle ahead on the right
  • d rr represents the distance between the target vehicle and the first vehicle on the right rear
  • ⁇ v rr represents the relative speed between the target vehicle and the first vehicle on the right rear side
  • d ef represents the distance between the target vehicle and the first vehicle directly in front of the vehicle
  • ⁇ v ef represents the relative speed between the target vehicle and the first vehicle directly in front of the vehicle.
  • the embodiment of the present application can use the actual driving intention of the target vehicle to learn the various parameters included in the above matrix.
  • the profit value of each driving intention of the target vehicle can be calculated (the leftward lane change)
  • the profit value is obtained from the relative motion information of the target vehicle and at least one first vehicle around it, and therefore, can represent the interaction relationship between them to a large extent.
  • the linear revenue function model is used to obtain the revenue value of each driving intention of the target vehicle at each time
  • the driving intention of the target vehicle at each time The income value of is input, and the target information representing the driving intention of the target vehicle is obtained through the cyclic neural network.
  • the recurrent neural network may be a long short-term memory network (LSTM) or a gated recurrent unit GRU.
  • LSTM long short-term memory network
  • GRU gated recurrent unit
  • LSTM algorithm is a specific form of recurrent neural network (recurrent neural network, RNN), and RNN is a general term for a series of neural networks capable of processing sequence data.
  • RNN recurrent neural network
  • RNN is a general term for a series of neural networks capable of processing sequence data.
  • RNN will encounter huge difficulties when dealing with long-term dependencies (nodes that are far away in the time series), because the calculation of the links between nodes that are far away involves multiple multiplications of the Jacobian matrix.
  • the most widespread is gated RNN (Gated RNN), and LSTM is the most famous kind of threshold RNN.
  • Leaky units allow RNNs to accumulate long-term connections between nodes that are far away by designing the weight coefficients between connections. Threshold RNNs generalize this idea, allowing the coefficient to be changed at different times, and allowing the network to forget the current accumulation. Information.
  • LSTM is such a threshold RNN. LSTM makes the weight of the self-loop change by increasing the input threshold, forgetting threshold and output threshold. In this way, when the model parameters are fixed, the integration scale at different times can be dynamically changed, thereby avoiding the problem of gradient disappearance or gradient expansion .
  • the distance (dl, dr) and lateral speed vd between the target vehicle and the left lane line and the right lane line can also be obtained, and the target vehicle
  • the revenue value of each driving intention at each moment, the distance (dl, dr) between the target vehicle and the left lane line and the right lane line, and the lateral speed vd are input.
  • the target information representing the driving intention of the target vehicle is output.
  • the information may be the probability (the probability of changing lanes to the left, the probability of changing lanes to the right, and the probability of staying straight) that indicate each driving intention.
  • the interaction between the target vehicle and the surrounding first vehicle is modeled by the profit function, and then the profit value is used as the input of the driving intention recognition model, which can more accurately identify the driving intention of the target vehicle while reducing The complexity of the model, and has a higher model interpretability.
  • the motion state of the target vehicle and second historical trajectory information may also be obtained.
  • the second historical trajectory information is trajectory information consistent with the driving direction of the target vehicle, and the motion state of the target vehicle may be Indicates the current motion state of the target vehicle, such as driving speed or driving acceleration.
  • the motion state of the target vehicle and the second historical trajectory information may also be used as input, and the second trajectory prediction model is used to obtain the second predicted trajectory of the target vehicle, wherein the second prediction The trajectory is the component of the driving trajectory of the target vehicle in a direction parallel to the lane.
  • the predicted trajectory of the target vehicle can be determined.
  • the embodiments of the present application can establish horizontal and vertical trajectory prediction models respectively, where the horizontal direction can be understood as being perpendicular to the direction of the lane where the target vehicle is located, and the longitudinal direction can be understood. It is consistent with the direction of the lane where the target vehicle is located. Among them, the direction of the lane can be understood as the direction of the lane line.
  • the lateral and longitudinal target vehicle trajectory prediction models are constructed separately, and the driving intention of the target vehicle is considered in the lateral and longitudinal directions, so as to obtain a more accurate target vehicle trajectory prediction.
  • FIG. 4 is a schematic diagram of the encoding process of the trajectory prediction model in the embodiment of the application.
  • the lateral encoder considers the historical lateral trajectory (first trajectory information) of the target vehicle and driving Intent (the probability of changing lanes to the right, changing lanes to the right and staying straight)
  • the longitudinal encoder considers the historical longitudinal trajectory of the target vehicle (second trajectory information) and the current state of motion (such as speed, acceleration) to obtain the transverse and longitudinal Encoding information.
  • the first predicted trajectory includes a plurality of predicted trajectory positions.
  • multiple encoded information is obtained through the LSTM encoder.
  • Each time corresponds to one encoded information, and the The multiple encoded information is input, and the multiple predicted trajectory positions are obtained through an LSTM decoder, where the LSTM decoder is used to obtain the multiple predicted trajectory positions by decoding based on multiple input data, and each input data is The multiple encoding information is related.
  • each input data is obtained by weighting the multiple pieces of coded information.
  • the function of the decoder is to parse the encoding information of the encoder to obtain the predicted trajectory of the target vehicle for a period of time in the future (for example, 5s).
  • ordinary decoders only use the hidden layer output at the last moment of the encoder as the input of the decoder
  • h1-h11 are the outputs of the encoder at each moment, as shown in Figure 5A
  • Figure 5A is a flow diagram of the decoding process.
  • the trajectory decoder only uses h11 as the input of the decoder, and does not make full use of historical information.
  • FIG. 5B is a flow chart of a decoding process provided by this application.
  • FIG. Output Take the input input1 at the first moment of the decoder as an example, through the attention mechanism, get the weights of the encoder output h1 to h11 at all moments Then there are:
  • the output information of the decoder is effectively and fully utilized, so that the prediction result is more accurate.
  • the first step data collection and processing.
  • NGSIM next-generation-simulation
  • NGSIM next-generation-simulation
  • the historical trajectory information of the target vehicle was processed, and the longitudinal distance and longitudinal speed of the surrounding 5 first vehicles relative to the target vehicle were extracted and standardized. For example, a total of 13,023 samples were extracted as a training set, 6511 A sample is used as a test set.
  • the following table shows one of the samples:
  • Table 1 Data extraction and processing example information table
  • the second step Calculate the profit value of the target vehicle to change lanes to the left, change lanes to the right and keep the execution.
  • the third step Identify the driving intention of the target vehicle.
  • the above profit value and the distance of the target vehicle relative to the left and right lanes and the lateral speed are used as the input of the LSTM network.
  • Step 4 Predict the lateral and longitudinal trajectories of the target vehicle (the first predicted trajectory and the second predicted trajectory);
  • the historical lateral trajectory of the target vehicle 2s is used as the input of the lateral encoder to obtain the initial encoding information, and then the initial encoding information is spliced with the driving intention probability obtained in the previous step to obtain the final lateral encoding information.
  • the encoded information is input to a decoder based on the attention mechanism to obtain the predicted trajectory of the target vehicle in the next 5s.
  • the historical longitudinal trajectory of the target vehicle 2s is used as the input of the longitudinal encoder to obtain the initial encoding information, and then the initial encoding information is spliced with the current motion state to obtain the final longitudinal encoding information.
  • the encoded information is input to a decoder based on the attention mechanism to obtain the longitudinal predicted trajectory of the target vehicle in the next 5s.
  • the relative motion information between the target vehicle and at least one first vehicle and the first historical trajectory information of the target vehicle are acquired; target information representing the driving intention of the target vehicle is acquired based on the relative motion information ; With the target information and the first historical trajectory information as input, the first predicted trajectory of the target vehicle is obtained through a first trajectory prediction model.
  • the driving intention of the target vehicle is used as the input of the trajectory prediction, so that the trajectory prediction of the target vehicle is more accurate.
  • the terminal device includes a hardware structure and/or software module corresponding to each function.
  • the embodiments of the present invention can be implemented in the form of hardware or a combination of hardware and computer software. Whether a certain function is executed by hardware or computer software-driven hardware depends on the specific application and design constraint conditions of the technical solution. Those skilled in the art can use different methods for each specific application to implement the described functions, but such implementation should not be considered as going beyond the scope of the technical solutions of the embodiments of the present invention.
  • the embodiment of the present invention may divide the functional units of the terminal device according to the foregoing method examples.
  • each functional unit may be divided corresponding to each function, or two or more functions may be integrated into one processing unit.
  • the above-mentioned integrated unit can be implemented in the form of hardware or software functional unit. It should be noted that the division of units in the embodiment of the present invention is illustrative, and is only a logical function division, and there may be other division methods in actual implementation.
  • FIG. 6 shows a schematic diagram of a possible structure of the terminal device involved in the foregoing embodiment.
  • the terminal device 600 includes: an acquisition module 601 for acquiring relative motion information between a target vehicle and at least one first vehicle within the detection range of the sensor, and first historical trajectory information of the target vehicle;
  • the motion information acquires target information representing the driving intention of the target vehicle;
  • the prediction module 602 is configured to use the target information and the first historical trajectory information as input, and obtain the first prediction of the target vehicle through the first trajectory prediction model Trajectory.
  • the driving intention includes at least one of the following: changing lanes to the left, changing lanes to the right, or keeping straight.
  • the relative motion information represents relative distance information and/or relative speed information between the target vehicle and each first vehicle at multiple times
  • the first historical trajectory information includes information about the target vehicle at multiple times. The first historical track position.
  • the acquisition module 601 is specifically configured to: use the relative motion information as input and obtain the profit value of each driving intention of the target vehicle at each moment through a linear profit function model; The profit value of each driving intention at each moment is input, and the target information representing the driving intention of the target vehicle is obtained through the cyclic neural network.
  • the first historical trajectory information and the first predicted trajectory are components of the driving trajectory of the target vehicle in a direction perpendicular to the lane, and the lane is the lane where the target vehicle is located.
  • the first predicted trajectory includes multiple predicted trajectory positions
  • the prediction module 602 is specifically configured to: take the target information and the first historical trajectory information as input, and obtain multiple predicted trajectory positions through an LSTM encoder. Pieces of coded information, each of which corresponds to one piece of coded information; taking the multiple pieces of coded information as input, through an LSTM decoder, to obtain the multiple predicted trajectory positions, wherein the LSTM decoder is used to obtain the multiple predicted trajectory positions based on multiple input data
  • the multiple predicted track positions are obtained by decoding, and each input data is related to the multiple encoded information.
  • each input data is obtained by weighting the multiple pieces of coded information.
  • the acquisition module 601 is further configured to: acquire the motion state of the target vehicle and second historical trajectory information, where the second historical trajectory information is the driving trajectory of the target vehicle in a direction parallel to the lane Component, the lane is the lane in which the target vehicle is located; the prediction module 602 is also used to: take the motion state of the target vehicle and the second historical trajectory information as input, and obtain the result through the second trajectory prediction model The second predicted trajectory of the target vehicle, where the second predicted trajectory is a component of the driving trajectory of the target vehicle in a direction parallel to the lane.
  • the terminal device 600 may further include a storage unit for storing program codes and data of the terminal device 600.
  • the aforementioned prediction module 601 may be integrated in a processing module, where the processing module may be a processor or a controller, for example, a central processing unit (CPU), a general-purpose processor, or digital signal processing.
  • DSP digital signal processor
  • ASIC application-specific integrated circuit
  • FPGA field programmable gate array
  • the processor may also be a combination for realizing computing functions, for example, including a combination of one or more microprocessors, a combination of a DSP and a microprocessor, and so on.
  • FIG. 7 is a schematic structural diagram of another terminal device 700 provided by an embodiment of this application.
  • the terminal device 700 includes: a processor 712, a communication interface 713, and a memory 77.
  • the terminal device 700 may further include a bus 714.
  • the communication interface 713, the processor 712, and the memory 77 may be connected to each other through a bus 714;
  • the bus 714 may be a peripheral component interconnect standard (PCI) bus or an extended industry standard architecture (EISA) Bus and so on.
  • PCI peripheral component interconnect standard
  • EISA extended industry standard architecture
  • the bus 714 can be divided into an address bus, a data bus, a control bus, and so on. For ease of presentation, only one thick line is used in FIG. 7B, but it does not mean that there is only one bus or one type of bus.
  • processor 712 may perform the following steps:
  • the first predicted trajectory of the target vehicle is obtained through a first trajectory prediction model.
  • the driving intention includes at least one of the following:
  • the relative motion information represents relative distance information and/or relative speed information between the target vehicle and each first vehicle at multiple times
  • the first historical trajectory information includes information about the target vehicle at multiple times. The first historical track position.
  • processor 712 may perform the following steps:
  • the target information representing the driving intention of the target vehicle is obtained through the cyclic neural network.
  • the first historical trajectory information and the first predicted trajectory indicate trajectory information perpendicular to the direction of the lane where the target vehicle is located.
  • the first predicted trajectory includes multiple predicted trajectory positions
  • the processor 712 may perform the following steps:
  • the multiple predicted trajectory positions are obtained through an LSTM decoder, where the LSTM decoder is used to obtain the multiple predicted trajectory positions by decoding based on multiple input data, and each input The data is related to the plurality of coded information.
  • each input data is obtained by weighting the multiple pieces of coded information.
  • processor 712 may also perform the following steps:
  • the second trajectory prediction model is used to obtain the second predicted trajectory of the target vehicle, and the second predicted trajectory is the driving trajectory of the target vehicle.
  • FIG. 8 is a schematic structural diagram of an execution device provided by an embodiment of this application.
  • the terminal device described in the embodiment corresponding to FIG. 6 or FIG. 7 may be deployed on the execution device 800 to implement the function of the terminal device in the embodiment corresponding to FIG. 6 and FIG. 7.
  • the execution device 800 includes: a receiver 801, a transmitter 802, a processor 803, and a memory 804 (the number of processors 803 in the execution device 800 may be one or more, and one processor is taken as an example in FIG. 8) , Where the processor 803 may include an application processor 8031 and a communication processor 8032.
  • the receiver 801, the transmitter 802, the processor 803, and the memory 804 may be connected by a bus or other methods.
  • the memory 804 may include a read-only memory and a random access memory, and provides instructions and data to the processor 803. A part of the memory 804 may also include a non-volatile random access memory (NVRAM).
  • NVRAM non-volatile random access memory
  • the memory 804 stores a processor and operating instructions, executable modules or data structures, or a subset of them, or an extended set of them.
  • the operating instructions may include various operating instructions for implementing various operations.
  • the processor 803 controls the operation of the execution device.
  • the various components of the execution device are coupled together through a bus system, where the bus system may include a power bus, a control bus, and a status signal bus in addition to a data bus.
  • bus system may include a power bus, a control bus, and a status signal bus in addition to a data bus.
  • various buses are referred to as bus systems in the figure.
  • the method disclosed in the foregoing embodiment of the present application may be applied to the processor 803 or implemented by the processor 803.
  • the processor 803 may be an integrated circuit chip with signal processing capability. In the implementation process, the steps of the foregoing method can be completed by an integrated logic circuit of hardware in the processor 803 or instructions in the form of software.
  • the aforementioned processor 803 may be a general-purpose processor, a digital signal processing (DSP), a microprocessor or a microcontroller, and may further include an application specific integrated circuit (ASIC), field programmable Field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components.
  • DSP digital signal processing
  • ASIC application specific integrated circuit
  • FPGA field programmable Field-programmable gate array
  • the processor 803 may implement or execute the methods, steps, and logical block diagrams disclosed in the embodiments of the present application.
  • the general-purpose processor may be a microprocessor or the processor may also be any conventional processor or the like.
  • the steps of the method disclosed in the embodiments of the present application may be directly embodied as being executed and completed by a hardware decoding processor, or executed and completed by a combination of hardware and software modules in the decoding processor.
  • the software module can be located in a mature storage medium in the field, such as random access memory, flash memory, read-only memory, programmable read-only memory, or electrically erasable programmable memory, registers.
  • the storage medium is located in the memory 804, and the processor 803 reads the information in the memory 804, and completes the steps of the foregoing method in combination with its hardware.
  • the receiver 801 can be used to receive input digital or character information, and to generate signal input related to the relevant settings and function control of the execution device.
  • the transmitter 802 can be used to output digital or character information through the first interface; the transmitter 802 can also be used to send instructions to the disk group through the first interface to modify the data in the disk group.
  • the embodiment of the present application also provides a product including a computer program, which when running on a computer, causes the computer to execute the steps performed by the terminal device in the method described in the embodiment shown in FIG. 6 or FIG. 7.
  • the embodiment of the present application also provides a computer-readable storage medium, the computer-readable storage medium stores a program for signal processing, and when it runs on a computer, the computer executes the program shown in Figure 6 or Figure 7 above. The steps performed by the terminal device in the method described in the embodiment are shown.
  • the execution device or terminal device provided by the embodiment of the present application may specifically be a chip.
  • the chip includes a processing unit and a communication unit.
  • the processing unit may be a processor, for example, and the communication unit may be an input/output interface, a pin, or Circuit etc.
  • the processing unit can execute the computer-executable instructions stored in the storage unit, so that the chip in the execution device executes the trajectory prediction method described in the embodiment shown in FIG. 2 above.
  • the storage unit is a storage unit in the chip, such as a register, a cache, etc., and the storage unit may also be a storage unit located outside the chip in the wireless access device, such as Read-only memory (ROM) or other types of static storage devices that can store static information and instructions, random access memory (RAM), etc.
  • ROM Read-only memory
  • RAM random access memory
  • Figure 9 is a schematic structural diagram of a chip provided by an embodiment of the application.
  • the chip may be represented as a neural network processor NPU 900, which is mounted as a coprocessor to the main CPU (Host On the CPU), the Host CPU assigns tasks.
  • the core part of the NPU is an arithmetic circuit, and the arithmetic circuit 903 is controlled by the controller 904 to extract matrix data from the memory and perform multiplication operations.
  • the arithmetic circuit 903 includes multiple processing units (Process Engine, PE). In some implementations, the arithmetic circuit 903 is a two-dimensional systolic array. The arithmetic circuit 903 may also be a one-dimensional systolic array or other electronic circuits capable of performing mathematical operations such as multiplication and addition. In some implementations, the arithmetic circuit 903 is a general-purpose matrix processor.
  • the arithmetic circuit fetches the data corresponding to matrix B from the weight memory 902 and caches it on each PE in the arithmetic circuit.
  • the arithmetic circuit fetches matrix A data and matrix B from the input memory 901 to perform matrix operations, and the partial results or final results of the obtained matrix are stored in an accumulator 908.
  • the unified memory 906 is used to store input data and output data.
  • the weight data directly passes through the memory unit access controller (Direct Memory Access Controller, DMAC) 905, and the DMAC is transferred to the weight memory 902.
  • the input data is also transferred to the unified memory 906 through the DMAC.
  • DMAC Direct Memory Access Controller
  • the BIU is the Bus Interface Unit, that is, the bus interface unit 910, which is used for the interaction of the AXI bus with the DMAC and the instruction fetch buffer (IFB) 909.
  • IFB instruction fetch buffer
  • the bus interface unit 910 (Bus Interface Unit, BIU for short) is used for the instruction fetch memory 909 to obtain instructions from an external memory, and is also used for the storage unit access controller 905 to obtain the original data of the input matrix A or the weight matrix B from the external memory.
  • BIU Bus Interface Unit
  • the DMAC is mainly used to transfer the input data in the external memory DDR to the unified memory 906 or to transfer the weight data to the weight memory 902 or to transfer the input data to the input memory 901.
  • the vector calculation unit 907 includes multiple arithmetic processing units, and further processes the output of the arithmetic circuit if necessary, such as vector multiplication, vector addition, exponential operation, logarithmic operation, size comparison, and so on. It is mainly used in the calculation of non-convolutional/fully connected layer networks in neural networks, such as Batch Normalization, pixel-level summation, and upsampling of feature planes.
  • the vector calculation unit 907 can store the processed output vector to the unified memory 906.
  • the vector calculation unit 907 may apply a linear function and/or a nonlinear function to the output of the arithmetic circuit 903, such as linearly interpolating the feature plane extracted by the convolutional layer, and for example, a vector of accumulated values, to generate the activation value.
  • the vector calculation unit 907 generates normalized values, pixel-level summed values, or both.
  • the processed output vector can be used as an activation input to the arithmetic circuit 903, for example for use in a subsequent layer in a neural network.
  • the instruction fetch buffer 909 connected to the controller 904 is used to store instructions used by the controller 904;
  • the unified memory 906, the input memory 901, the weight memory 902, and the fetch memory 909 are all On-Chip memories.
  • the external memory is private to the NPU hardware architecture.
  • processor mentioned in any of the foregoing may be a general-purpose central processing unit, a microprocessor, an ASIC, or one or more integrated circuits used to control the execution of the program of the method in the first aspect.
  • the device embodiments described above are only illustrative, and the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physically separate.
  • the physical unit can be located in one place or distributed across multiple network units. Some or all of the modules can be selected according to actual needs to achieve the objectives of the solutions of the embodiments.
  • the connection relationship between the modules indicates that they have a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines.
  • this application can be implemented by means of software plus necessary general hardware.
  • it can also be implemented by dedicated hardware including dedicated integrated circuits, dedicated CPUs, dedicated memory, Dedicated components and so on to achieve.
  • all functions completed by computer programs can be easily implemented with corresponding hardware.
  • the specific hardware structures used to achieve the same function can also be diverse, such as analog circuits, digital circuits or special-purpose circuits. Circuit etc.
  • software program implementation is a better implementation in more cases.
  • the technical solution of this application essentially or the part that contributes to the existing technology can be embodied in the form of a software product, and the computer software product is stored in a readable storage medium, such as a computer floppy disk. , U disk, mobile hard disk, ROM, RAM, magnetic disk or optical disk, etc., including several instructions to make a computer device (which can be a personal computer, training device, or network device, etc.) execute the various embodiments described in this application method.
  • a computer device which can be a personal computer, training device, or network device, etc.
  • the computer program product includes one or more computer instructions.
  • the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.
  • the computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium.
  • the computer instructions may be transmitted from a website, computer, training device, or data.
  • the center transmits to another website, computer, training equipment, or data center through wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.).
  • wired such as coaxial cable, optical fiber, digital subscriber line (DSL)
  • wireless such as infrared, wireless, microwave, etc.
  • the computer-readable storage medium may be any available medium that can be stored by a computer or a data storage device such as a training device or a data center integrated with one or more available media.
  • the usable medium may be a magnetic medium (for example, a floppy disk, a hard disk, and a magnetic tape), an optical medium (for example, a DVD), or a semiconductor medium (for example, a solid state disk (SSD)).

Landscapes

  • Engineering & Computer Science (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Automation & Control Theory (AREA)
  • Human Computer Interaction (AREA)
  • Transportation (AREA)
  • Mechanical Engineering (AREA)
  • Traffic Control Systems (AREA)

Abstract

一种轨迹预测方法，应用于人工智能领域下的自动驾驶，轨迹预测方法包括：获取目标车辆与位于传感器的检测范围内的至少一辆第一车辆之间的相对运动信息，以及所述目标车辆的第一历史轨迹信息（201）；基于所述相对运动信息获取表示目标车辆行车意图的目标信息（202）；以所述目标信息和所述第一历史轨迹信息为输入，通过第一轨迹预测模型，得到所述目标车辆的第一预测轨迹（203）。该方法提高了对车辆的轨迹预测准确性。

Description

一种轨迹预测方法及相关设备 技术领域
本申请涉及人工智能领域，尤其涉及一种轨迹预测方法及相关设备。
背景技术
随着社会经济高速发展，汽车越来越普及，并逐渐成为人们日常工作生活中不可或缺的一部分。同时，随着汽车以及通信技术的发展，汽车逐渐朝着智能化方向发展，未来，通过利用机器智能化引导驾驶取代人工驾驶的自动化汽车(即智能汽车)的普及程度可能会越来越高。
目前，无人驾驶汽车的一大应用场景是混合交通流道路，即无人驾驶汽车和驾驶人驾驶汽车并存的道路上。在该场景中，无人驾驶汽车会受到来自驾驶人驾驶汽车的影响。因此，为了保证用户安全，无人驾驶汽车需要获取周围驾驶人驾驶汽车的轨迹，基于这个轨迹对无人驾驶汽车进行控制，从而保证无人驾驶汽车能够安全高效的抵达目的地。
然而，无论是通过人工驾驶的传统汽车，还是通过机器智能化引导驾驶的智能汽车，其行驶安全性问题都是人们十分的关注的问题。
因此，如何保证汽车行驶安全成为一个亟待解决的技术问题。
发明内容
本申请提供了一种轨迹预测方法及相关设备，将目标车辆的行车意图作为轨迹预测的输入，使得对目标车辆的轨迹预测更加准确。
第一方面，本申请提供了一种轨迹预测方法，所述方法包括：获取目标车辆与位于传感器的检测范围内的至少一辆第一车辆之间的相对运动信息，以及所述目标车辆的第一历史轨迹信息，其中，传感器设置于本车中，传感器具体可以是雷达传感器或者是视频获取传感器；基于所述相对运动信息获取表示目标车辆行车意图的目标信息；以所述目标信息和所述第一历史轨迹信息为输入，通过第一轨迹预测模型，得到所述目标车辆的第一预测轨迹。本申请实施例中，获取目标车辆与至少一辆第一车辆之间的相对运动信息，以及所述目标车辆的第一历史轨迹信息；基于所述相对运动信息获取表示目标车辆行车意图的目标信息；以所述目标信息和所述第一历史轨迹信息为输入，通过第一轨迹预测模型，得到所述目标车辆的第一预测轨迹。通过上述方式，将目标车辆的行车意图作为轨迹预测的输入，使得对目标车辆的轨迹预测更加准确。
在第一方面的一种可选设计中，所述行车意图至少包括如下的一种：向左变道、向右变道或保持直行。
在第一方面的一种可选设计中，所述相对运动信息表示目标车辆与每个第一车辆之间多个时刻的相对距离信息和/或相对速度信息，所述第一历史轨迹信息包括所述目标车辆多个时刻的第一历史轨迹位置。
在第一方面的一种可选设计中，所述基于所述相对运动信息获取表示目标车辆行车意图的目标信息，包括：以所述相对运动信息为输入，通过线性收益函数模型，得到所述目 标车辆每个时刻各行车意图的收益值；以所述目标车辆每个时刻各行车意图的收益值为输入，通过循环神经网络，得到表示目标车辆行车意图的目标信息。本申请实施例中，通过收益函数建模目标车辆及其周围第一车辆间的交互作用，再把收益值作为行车意图识别模型的输入，能够更加准确地识别出目标车辆的行车意图，同时减少模型的复杂度，并具有更高的模型可解释性。
在第一方面的一种可选设计中，所述第一历史轨迹信息和所述第一预测轨迹为所述目标车辆的行车轨迹在与车道垂直方向的分量，所述车道为所述目标车辆所在的车道。
在第一方面的一种可选设计中，所述第一预测轨迹包括多个预测轨迹位置，所述以所述目标信息和所述第一历史轨迹信息为输入，通过第一轨迹预测模型，得到所述目标车辆的第一预测轨迹，包括：以所述目标信息和所述第一历史轨迹信息为输入，通过LSTM编码器，得到多个编码信息，其中每个时刻对应一个编码信息；以所述多个编码信息为输入，通过LSTM解码器，得到所述多个预测轨迹位置，其中，所述LSTM解码器用于基于多个输入数据解码得到所述多个预测轨迹位置，每个输入数据与所述多个编码信息有关。本申请实施例中，对解码器的输出信息进行有效充分地利用，使得预测结果更加准确。
在第一方面的一种可选设计中，每个输入数据为对所述多个编码信息进行加权得到的。
在第一方面的一种可选设计中，所述方法还包括：获取所述目标车辆的运动状态和第二历史轨迹信息，所述第二历史轨迹信息为所述目标车辆的行车轨迹在与车道平行方向的分量，所述车道为所述目标车辆所在的车道；以所述目标车辆的运动状态和第二历史轨迹信息为输入，通过第二轨迹预测模型，得到所述目标车辆的第二预测轨迹，所述第二预测轨迹为所述目标车辆的行车轨迹在与车道平行方向的分量。本申请实施例中，分别构建横向和纵向的目标车辆轨迹预测模型，并在横向和纵向分别考虑目标车辆的行车意图，能够得到更加准确的目标车辆预测轨迹。
第二方面，本申请提供了一种执行设备，包括：
获取模块，用于获取目标车辆与位于传感器的检测范围内的至少一辆第一车辆之间的相对运动信息，以及所述目标车辆的第一历史轨迹信息；
基于所述相对运动信息获取表示目标车辆行车意图的目标信息；
预测模块，用于以所述目标信息和所述第一历史轨迹信息为输入，通过第一轨迹预测模型，得到所述目标车辆的第一预测轨迹。
在第二方面的一种可选设计中，所述行车意图至少包括如下的一种：
向左变道、向右变道或保持直行。
在第二方面的一种可选设计中，所述相对运动信息表示目标车辆与每个第一车辆之间多个时刻的相对距离信息和/或相对速度信息，所述第一历史轨迹信息包括所述目标车辆多个时刻的第一历史轨迹位置。
在第二方面的一种可选设计中，所述获取模块，具体用于：
以所述相对运动信息为输入，通过线性收益函数模型，得到所述目标车辆每个时刻各行车意图的收益值；
以所述目标车辆每个时刻各行车意图的收益值为输入，通过循环神经网络，得到表示 目标车辆行车意图的目标信息。
在第二方面的一种可选设计中，所述第一历史轨迹信息和所述第一预测轨迹为所述目标车辆的行车轨迹在与车道垂直方向的分量，所述车道为所述目标车辆所在的车道。
在第二方面的一种可选设计中，所述第一预测轨迹包括多个预测轨迹位置，所述预测模块，具体用于：
以所述目标信息和所述第一历史轨迹信息为输入，通过LSTM编码器，得到多个编码信息，其中每个时刻对应一个编码信息；
以所述多个编码信息为输入，通过LSTM解码器，得到所述多个预测轨迹位置，其中，所述LSTM解码器用于基于多个输入数据解码得到所述多个预测轨迹位置，每个输入数据与所述多个编码信息有关。
在第二方面的一种可选设计中，每个输入数据为对所述多个编码信息进行加权得到的。
在第二方面的一种可选设计中，所述获取模块，还用于：
所述第二历史轨迹信息为所述目标车辆的行车轨迹在与车道平行方向的分量，所述车道为所述目标车辆所在的车道；
所述预测模块，还用于：
以所述目标车辆的运动状态和第二历史轨迹信息为输入，通过第二轨迹预测模型，得到所述目标车辆的第二预测轨迹，所述第二预测轨迹为所述目标车辆的行车轨迹在与车道平行方向的分量。
第三方面，本发明实施例提供了一种终端设备，包括存储器、通信接口及与所述存储器和通信接口耦合的处理器；所述存储器用于存储指令，所述处理器用于执行所述指令，所述通信接口用于在所述处理器的控制下与其他设备进行通信；其中，所述处理器执行所述指令时执行上述第一方面或第一方面可能的实施例中所描述的方法。
第四方面，提供了一种计算机可读存储介质，所述计算机可读存储介质存储了用于车辆控制的程序代码。所述程序代码包括用于执行上述第一方面或第一方面可能的实施例中所描述的方法的指令。
第五方面，提供了一种包括指令的计算机程序产品，当其在计算机上运行时，使得计算机执行上述第一方面或第一方面可能的实施例中所描述的方法。
本申请实施例中，获取目标车辆与位于传感器的检测范围内的至少一辆第一车辆之间的相对运动信息，以及所述目标车辆的第一历史轨迹信息；基于所述相对运动信息获取表示目标车辆行车意图的目标信息；以所述目标信息和所述第一历史轨迹信息为输入，通过第一轨迹预测模型，得到所述目标车辆的第一预测轨迹。通过上述方式，将目标车辆的行车意图作为轨迹预测的输入，使得对目标车辆的轨迹预测更加准确。
附图说明
图1A为人工智能主体框架的一种结构示意图；
图1B为本申请的一种可能的终端设备的结构示意图；
图1C为本申请的又一种可能的终端设备的结构示意图；
图2为本申请实施例提供的一种轨迹预测方法的实施例示意图；
图3为本申请实施例提供的一种行人行为意图的示意；
图4为本申请实施例提供的一种历史轨迹信息处理示意；
图5A为一种解码过程的流程示意；
图5B为一种解码过程的流程示意；
图6为本实施例中所涉及的终端设备的一种可能的结构示意图；
图7为本申请实施例提供的另一种终端设备700的结构示意；
图8为本申请实施例提供的执行设备的一种结构示意图；
图9为本申请实施例提供的芯片的一种结构示意图。
具体实施方式
本申请实施例提供了一种轨迹预测方法方法及相关设备，用于提高对目标车辆的轨迹预测的准确性。
下面结合附图，对本申请的实施例进行描述。本领域普通技术人员可知，随着技术的发展和新场景的出现，本申请实施例提供的技术方案对于类似的技术问题，同样适用。
本申请的说明书和权利要求书及上述附图中的术语“第一”、“第二”等是用于区别类似的对象，而不必用于描述特定的顺序或先后次序。应该理解这样使用的术语在适当情况下可以互换，这仅仅是描述本申请的实施例中对相同属性的对象在描述时所采用的区分方式。此外，术语“包括”和“具有”以及他们的任何变形，意图在于覆盖不排他的包含，以便包含一系列单元的过程、方法、系统、产品或设备不必限于那些单元，而是可包括没有清楚地列出的或对于这些过程、方法、产品或设备固有的其它单元。
首先对人工智能系统总体工作流程进行描述，请参见图1A，图1A示出的为人工智能主体框架的一种结构示意图，下面从“智能信息链”(水平轴)和“IT价值链”(垂直轴)两个维度对上述人工智能主题框架进行阐述。其中，“智能信息链”反映从数据的获取到处理的一列过程。举例来说，可以是智能信息感知、智能信息表示与形成、智能推理、智能决策、智能执行与输出的一般过程。在这个过程中，数据经历了“数据—信息—知识—智慧”的凝练过程。“IT价值链”从人智能的底层基础设施、信息(提供和处理技术实现)到系统的产业生态过程，反映人工智能为信息技术产业带来的价值。
(1)基础设施
基础设施为人工智能系统提供计算能力支持，实现与外部世界的沟通，并通过基础平台实现支撑。通过传感器与外部沟通；计算能力由智能芯片(CPU、NPU、GPU、ASIC、FPGA等硬件加速芯片)提供；基础平台包括分布式计算框架及网络等相关的平台保障和支持，可以包括云存储和计算、互联互通网络等。举例来说，传感器和外部沟通获取数据，这些数据提供给基础平台提供的分布式计算系统中的智能芯片进行计算。
(2)数据
基础设施的上一层的数据用于表示人工智能领域的数据来源。数据涉及到图形、图像、语音、文本，还涉及到传统设备的物联网数据，包括已有系统的业务数据以及力、位 移、液位、温度、湿度等感知数据。
(3)数据处理
数据处理通常包括数据训练，机器学习，深度学习，搜索，推理，决策等方式。
其中，机器学习和深度学习可以对数据进行符号化和形式化的智能信息建模、抽取、预处理、训练等。
推理是指在计算机或智能系统中，模拟人类的智能推理方式，依据推理控制策略，利用形式化的信息进行机器思维和求解问题的过程，典型的功能是搜索与匹配。
决策是指智能信息经过推理后进行决策的过程，通常提供分类、排序、预测等功能。
(4)通用能力
对数据经过上面提到的数据处理后，进一步基于数据处理的结果可以形成一些通用的能力，比如可以是算法或者一个通用系统，例如，翻译，文本的分析，计算机视觉的处理，语音识别，图像的识别等等。
(5)智能产品及行业应用
智能产品及行业应用指人工智能系统在各领域的产品和应用，是对人工智能整体解决方案的封装，将智能信息决策产品化、实现落地应用，其应用领域主要包括：智能终端、智能制造、智能交通、智能家居、智能医疗、智能安防、自动驾驶、平安城市等。
下面将结合本发明的附图，对本发明实施例中的技术方案进行详细描述。
首先，介绍本申请适用的终端设备的结构示意图。如图1B示出本申请的一种可能的终端设备的结构示意图。如图1B，该终端设备包括环境感知模块102、规划决策模块104以及控制处理模块106。其中，所述环境感知模块102主要通过外围系统(如传感器、摄像头等)采集障碍物信息、终端设备所处的周围环境信息以及终端设备所在车辆的行驶信息。所述障碍物信息包括但不限于障碍物的地理位置、障碍物的运动速度、障碍物的运动方向、障碍物的运动加速度、障碍物的运动方向的方差、障碍物的运动速度的方差等信息。所述障碍物包括但不限于车辆、行人、有生命的活体障碍物以及无生命的障碍物等等。本申请将以所述障碍物为行车辆为例，具体阐述本申请所涉及的一些实施例。
所述周围环境信息包括但不限于地图信息、天气信息、路口类型、车道线、车道数量、道路是否拥塞、车流速度、车流加速度以及终端设备与障碍物之间的距离等等信息。
所述行驶信息包括但不限于车辆的地理位置、行驶速度、行驶方向、行驶加速度、车辆与障碍物之间的距离等等。所述终端设备包括但不限于汽车、火车、货车、小轿车等车辆以及安装在车辆上的通讯设备，如车载设备等。
所述规划决策模块104包括行为预测模块和规划模块。其中，所述行为预测模块主要用于根据环境感知模块所采集的上述信息预测障碍物的行为意图(如，本申请后文所述的目标车辆的行车意图)以及历史运动轨迹。所述规划模块用于在保证安全的前提下，获得对应的控制策略，以便后续利用该控制策略控制车辆进行安全行驶。所述控制策略为用户侧或终端设备侧预先自定义设置的，或者根据所述行为意图生成的，具体在下文中进行详述。所述控制策略用于指示对所述车辆进行相应车辆参数的调整，以实现车辆安全驾驶。
所述控制处理模块用于根据所述规划决策模块所获得的控制策略，对所述车辆进行相应地控制和调整，以避免车辆与障碍物发生碰撞。例如对车辆的方向盘转角、行驶速度、是否制动刹车、是否按压加速踏板等车辆参数进行控制。
如图1C示出本申请的又一种可能的终端设备的结构示意图。如图1C所示，终端设备100可包括：基带芯片110、存储器115，包括一个或多个计算机可读存储介质、射频(RF)模块116、外围系统117。这些部件可在一个或多个通信总线114上通信。
外围系统117主要用于实现终端设备110和用户/外部环境之间的交互功能，主要包括终端100的输入输出装置。具体实现中，外围系统117可包括：触摸屏控制器118、摄像头控制器119、音频控制器120以及传感器管理模块121。其中，各个控制器可与各自对应的外围设备，例如触摸屏123、摄像头124、音频电路125以及传感器126，耦合。在一些实施例中，传感器126中的手势传感器可用于接收用户输入的手势控制操作。传感器126中的速度传感器可用于采集终端设备自身的行驶速度或用于采集环境中障碍物的运动速度等。触摸屏123可作为提示装置，主要用于通过屏幕显示、投影等方式来提示障碍物，例如在行人横穿马路时通过显示屏显示文字的方式来提示行人加速行走等。可选的，外围系统117还可包括灯光、显示器等其他提示装置，以用于车辆与行人之间的交互提示，避免车辆与行人发生碰撞。需要说明的，外围系统117还可以包括其他I/O外设。
基带芯片110可集成包括：一个或多个处理器111、时钟模块112以及电源管理模块113。集成于基带芯片110中的时钟模块112主要用于为处理器111产生数据传输和时序控制所需要的时钟。集成于基带芯片110中的电源管理模块113主要用于为处理器111、射频模块116以及外围系统提供稳定的、高精确度的电压。
射频(RF)模块116用于接收和发送射频信号，主要集成了终端100的接收器和发射器。射频(RF)模块116通过射频信号与通信网络和其他通信设备通信。具体实现中，射频(RF)模块116可包括但不限于：天线系统、RF收发器、一个或多个放大器、调谐器、一个或多个振荡器、数字信号处理器、CODEC芯片、SIM卡和存储介质等。在一些实施例中，可在单独的芯片上实现射频(RF)模块116。
存储器115与处理器111耦合，用于存储各种软件程序和/或多组指令。具体实现中，存储器115可包括高速随机存取的存储器，并且也可包括非易失性存储器，例如一个或多个磁盘存储设备、闪存设备或其他非易失性固态存储设备。存储器115可以存储操作系统，例如ANDROID，IOS，WINDOWS，或者LINUX等嵌入式操作系统。存储器115还可以存储网络通信程序，该网络通信程序可用于与一个或多个附加设备，一个或多个终端设备，一个或多个终端设备进行通信等。
本申请可以应用于行车场景中，该行车场景中包括本车和目标车辆。本车为任一行驶的车辆。目标车辆为本车周围的车辆。本车用于预测目标车辆的运动轨迹。其中，本车可以为无人驾驶汽车，也可以为驾驶员驾驶汽车。并且，本车可以为新能源汽车，也可以为燃油汽车。其中，新能源汽车可以为纯电动汽车或者混动动力汽车。在本发明实施例中，对本车不作具体限定。同样，该目标车辆可以为无人驾驶汽车，也可以为驾驶员驾驶汽车。该目标车辆可以为新能源汽车，也可以为燃油汽车。在本公开实施例中，对目标车辆同样 不作具体限定。
本申请实施例中，目标车辆可以位于本车的周围，其中本车的周围是指本车为中心，预设半径范围内组成的区域。预设半径可以根据需要进行设置并更改，在本公开实施例中，对预设半径不作具体限定。并且，预设半径可以根据基于不同的道路属性进行设置；例如，高速公路对应的预设半径可以为100米或者200米；普通道路对应的预设半径可以为20米或者30米等。在一种实施例中，目标车辆可以为本车前方的车辆，即位于本车的行车方向上。
在上述场景中，本车需要对周围的车辆进行行车轨迹的预测，基于预测轨迹对本车进行控制，从而保证无人驾驶汽车能够安全的抵达目的地。为了能够做出合理的决策规划，自动驾驶汽车需要具备预测周围车辆行车意图和未来轨迹的能力，提高整个系统的效率和安全性。
现有技术中，可以基于历史轨迹进行目标车辆的轨迹预测，具体的，可以以目标车辆的历史轨迹信息作为输入，直接输出目标车辆的预测轨迹。基于历史轨迹的轨迹预测算法例如可以是，高斯过程回归(gaussian process regression，GPR)，其先通过核函数将低维的状态空间映射到高维空间，再利用均值函数和协方差函数，将历史轨迹和预测轨迹建模为多维高斯分布，给定一段历史轨迹，输出预测的未来轨迹。然而，高斯过程回归方法是基于大量历史轨迹信息得到的统计学模型，没有考虑目标车辆当前的运动状态，此外，该方法没有考虑目标车辆及其周围车辆之间的交互作用，不符合实际交通情况，预测准确率较低。
为了解决上述技术问题，参照图2，图2为本申请实施例提供的一种轨迹预测方法的实施例示意，参照图2，本实施例提供的轨迹预测方法，包括：
201、获取目标车辆与位于传感器的检测范围内的至少一辆第一车辆之间的相对运动信息，以及所述目标车辆的第一历史轨迹信息。
本申请实施例中，所述相对运动信息可以表示目标车辆与每个第一车辆之间多个时刻的相对距离信息和/或相对速度信息，所述第一历史轨迹信息包括所述目标车辆多个时刻的第一历史轨迹位置。
本申请实施例中，本车可以基于自身携带的传感器检测到至少一辆第一车辆，以及目标车辆与位于传感器的检测范围内的至少一辆第一车辆之间的相对运动信息。
本申请实施例中，第一车辆可以为目标车辆附近的车辆，在一种实施例中，第一车辆的数量可以根据目标车辆周围的车况来定，例如，如图3所示，可以将目标车辆左前方、左后方、右前方、右后方和正前方五辆车作为第一车辆。分别提取出上述五辆第一车辆相对于目标车辆的相对运动信息，其中，相对运动信息可以表示目标车辆与每个第一车辆之间多个时刻的相对距离信息和/或相对速度信息，例如，相对距离信息可以是沿车道行车方向的纵向距离d和目标车辆与第一车辆之间的速度Δv。
特别的，在一种实施例中，当目标车辆位于最左侧车道，即目标车辆左侧车道不存在时，可以将图3中与左侧第一车辆对应的相对运动信息设置为如下信息：
d lf＝0，d lr＝0，Δv lf＝0；
特别的，在一种实施例中，当目标车辆左侧车道存在但左前方没有第一车辆时，可以 将图3中与左前方的第一车辆对应的相对运动信息设置为如下信息：
d lf＝+1000，Δv lf＝0；
特别的，在一种实施例中，当目标车辆左侧车道存在但左后方没有第一车辆时，可以将图3中与左后方的第一车辆对应的相对运动信息设置为如下信息：
d lf＝-1000，Δv lf＝0；
特别的，在一种实施例中，当目标车辆左侧车道存在但左前方、左后方均没有第一车辆时，可以将图3中与左前方以及左后方的第一车辆对应的相对运动信息设置为如下信息：
d lf＝+1000，d lr＝-1000，Δv lf＝0；特别的，在一种实施例中，当目标车辆正前方没有第一车辆时，可以将图3中与正前方的第一车辆对应的相对运动信息设置为如下信息：
d ef＝+1000，Δv ef＝0；
以上实施例中相对运动信息的设置仅为一种示意，并不构成对本申请的限定，位于目标车辆右侧的第一车辆的相对运动信息可以参照上述实施例来设置，这里并不限定。
在一种实施例中，本车可以通过自身携带的用于采集运动信息的传感器来获取目标车辆与至少一个第一车辆之间的相对运动信息，例如，采集运动信息的传感器可以为摄像头或者雷达等。该雷达可以为毫米波雷达或者激光雷达等，这里并不限定。
本申请实施例中，还可以获取所述目标车辆的第一历史轨迹信息，其中，所述第一历史轨迹信息可以包括所述目标车辆多个时刻的第一历史轨迹位置，第一历史轨迹信息表示与所述目标车辆所在车道的方向相垂直的轨迹信息，例如，所述第一历史轨迹信息可以包括所述目标车辆A时刻、B时刻、C时刻的第一历史轨迹位置。
本申请实施例中，还可以获取所述目标车辆的第二历史轨迹信息，所述第二历史轨迹信息为与所述目标车辆所在车道的方向一致的轨迹信息。通过第一历史轨迹信息和第二历史轨迹信息，可以获取到目标车辆的历史轨迹信息。
在一个可能的实现方式中，本车可以向目标车辆发送轨迹获取请求，轨迹获取请求用于请求获取该目标车辆的第一历史轨迹信息和第二历史轨迹信息。该目标车辆接收本车发送的该轨迹获取请求，根据轨迹获取请求向本车发送目标车辆的第一历史轨迹信息和第二历史轨迹信息。本车向该目标车辆发送轨迹获取请求之前，本车与该目标车辆建立信息传输通道。相应的，本车与该目标车辆通过该信息传输通道进行信息传输。其中，该信息传输通道可以为车联网(vehicles of internet，V2I)传输通道或者近距离通信(例如，蓝牙)传输通道等。
在另一个可能的实现方式中，本车可以从监控设备中获取目标车辆的第一历史轨迹信息和第二历史轨迹信息。相应的，本车获取目标车辆的历史轨迹信息的步骤可以为：本车向监控设备发送轨迹获取请求，轨迹获取请求携带目标车辆的车辆标识。监控设备接收本车发送的轨迹获取请求，根据目标车辆的车辆标识，获取目标车辆的历史轨迹信息，并向本车发送目标车辆的历史轨迹信息。本车接收监控设备发送的目标车辆的第一历史轨迹信息和第二历史轨迹信息。监控设备为用于采集运动信息的传感器，例如，摄像头或者雷达等。为了提高精度，该雷达可以为毫米波雷达或者激光雷达等。车辆标识可以为目标车辆的车牌号码等。
以上关于如何获取目标对象的第一历史轨迹信息和第二历史轨迹信息、以及目标车辆与至少一辆第一车辆之间的相对运动信息的方式仅为一种示意，本申请并不限定。
202、基于所述相对运动信息获取表示目标车辆行车意图的目标信息。
本申请实施例中，所述行车意图可以至少包括如下的一种：向左变道、向右变道或保持直行。
在一种实施例中，可以以所述相对运动信息为输入，通过线性收益函数模型，得到所述目标车辆每个时刻各行车意图的收益值；并以所述目标车辆每个时刻各行车意图的收益值为输入，通过循环神经网络，得到表示目标车辆行车意图的目标信息。
且在，本申请利用线性函数作为收益函数来确定各个行车意图的收益值GL,GR,GK，其矩阵形式可以为如下：
Figure PCTCN2019129828-appb-000001
其中，GL表示目标车辆向左变道的收益值，GR表示目标车辆向右变道的收益值，GK表示目标车辆保持直行的收益值；
d lf表示目标车辆与左前方的第一车辆之间的距离；
Δv lf表示目标车辆与左前方的第一车辆之间的相对速度；
d lr表示目标车辆与左后方的第一车辆之间的距离；
Δv lr表示目标车辆与左后方的第一车辆之间的相对速度；
d rf表示目标车辆与右前方的第一车辆之间的距离；
Δv rf表示目标车辆与右前方的第一车辆之间的相对速度；
d rr表示目标车辆与右后方的第一车辆之间的距离；
Δv rr表示目标车辆与右后方的第一车辆之间的相对速度；
d ef表示目标车辆与行车正前方的第一车辆之间的距离；
Δv ef表示目标车辆与行车正前方的第一车辆之间的相对速度。
本实施例中，上式中线性收益函数中共有11x3＝33个参数：
Figure PCTCN2019129828-appb-000002
本申请实施例可以利用目标车辆实际的行车意图对上述矩阵中包括的各个参数进行学习，示例性的，首先可以初始化参数
Figure PCTCN2019129828-appb-000003
i＝0，1,2,3,...,10，然后进行迭代运算：
1、计算收益值GL、GR、GK；
2、利用softmax函数分别计算向右变道、向右变道和保持直行的概率；
3、利用目标车辆实际的行车意图计算分类的交叉熵损失；
4、利用梯度下降算法更新参数
Figure PCTCN2019129828-appb-000004
i＝0,1,2,3,...,10；
5、达到收敛条件，停止迭代。
本申请实施例中，在完成上述线性收益函数模型的训练后，给定目标车辆和其周围的第一车辆的相对运动信息，便可计算目标车辆各个行车意图的收益值(向左变道的收益值GL、向右变道GR的收益值和保持直行的收益值GK)。该收益值是由目标车辆及其周围的至少一个第一车辆的相对运动信息得到的，因此，能够很大程度地表征它们之间的交互作用关系。
本申请实施例中，在以所述相对运动信息为输入，通过线性收益函数模型，得到所述目标车辆每个时刻各行车意图的收益值之后，可以以所述目标车辆每个时刻各行车意图的收益值为输入，通过循环神经网络，得到表示目标车辆行车意图的目标信息。
本申请实施例中，循环神经网络可以是长短期记忆网络(long short-term memory，LSTM)或门控循环单元GRU，接下来以循环神经网络模型为长短期记忆网络LSTM为例进行说明。
关于LSTM：
LSTM算法是一种特定形式的循环神经网络(recurrent neural network，RNN)，而RNN是一系列能够处理序列数据的神经网络的总称。RNN还有许多变形，例如双向RNN(bidirectional RNN)等。然而，RNN在处理长期依赖(时间序列上距离较远的节点)时会遇到巨大的困难，因为计算距离较远的节点之间的联系时会涉及雅可比矩阵的多次相乘，这会带来梯度消失(经常发生)或者梯度膨胀(较少发生)的问题，为了解决该问题，最广泛的就是门限RNN(Gated RNN)，而LSTM就是门限RNN中最著名的一种。有漏单元通过设计连接间的权重系数，从而允许RNN累积距离较远节点间的长期联系；而门限RNN则泛化了这样的思想，允许在不同时刻改变该系数，且允许网络忘记当前已经累积的信息。LSTM就是这样的门限RNN。LSTM通过增加输入门限，遗忘门限和输出门限，使得自循环的权重是变化的，这样，在模型参数固定的情况下，不同时刻的积分尺度可以动态改变，从而避免了梯度消失或者梯度膨胀的问题。
需要说明的是，除了获取所述目标车辆每个时刻各行车意图的收益值，还可以获取目标车辆与左车道线以及右车道线的距离(dl，dr)、横向速度vd，并以目标车辆每个时刻各行车意图的收益值、目标车辆与左车道线以及右车道线的距离(dl，dr)以及横向速度vd为输入，通过LSTM模型，输出表示目标车辆行车意图的目标信息，该目标信息可以是 表示各个行车意图的概率(左变道的概率、右变道的概率和保持直行的概率)。
本申请实施例中，通过收益函数建模目标车辆及其周围第一车辆间的交互作用，再把收益值作为行车意图识别模型的输入，能够更加准确地识别出目标车辆的行车意图，同时减少模型的复杂度，并具有更高的模型可解释性。
203、以所述目标信息和所述第一历史轨迹信息为输入，通过第一轨迹预测模型，得到所述目标车辆的第一预测轨迹。
本申请实施例中，还可以获取所述目标车辆的运动状态和第二历史轨迹信息，所述第二历史轨迹信息为与所述目标车辆行车方向一致的轨迹信息，目标车辆的运动状态可以为表示目标车辆当前的运动状态，例如行车速度或行车加速度。
本申请实施例中，还可以以所述目标车辆的运动状态和第二历史轨迹信息为输入，通过第二轨迹预测模型，得到所述目标车辆的第二预测轨迹，其中，所述第二预测轨迹为所述目标车辆的行车轨迹在与车道平行方向的分量。
本申请实施例中，基于上述获取到的第一预测轨迹和第二预测轨迹，可以确定目标车辆的预测轨迹。
本申请实施例中，考虑到目标车辆在横向和纵向的损失相差巨大(2-3个数量级)，在优化过程中会出现横向损失无法下降的问题。同时考虑到行车意图和运动状态对横向及纵向的影响不同，本申请实施例可以分别建立横向和纵向的轨迹预测模型，其中横向可以理解为与所述目标车辆所在车道的方向垂直，纵向可以理解为与所述目标车辆所在车道的方向一致。其中，车道的方向可以理解为车道线的方向。
本申请实施例中，分别构建横向和纵向的目标车辆轨迹预测模型，并在横向和纵向分别考虑目标车辆的行车意图，能够得到更加准确的目标车辆预测轨迹。
其中，横向和纵向的目标车辆轨迹预测模型均为编码解码器。如图4所示，图4为本申请实施例中轨迹预测模型的编码过程示意，在图4示出的编码过程中，横向编码器考虑目标车辆的历史横向轨迹(第一轨迹信息)和行车意图(向右变道、向右变道和保持直行的概率)，纵向编码器考虑目标车辆的历史纵向轨迹(第二轨迹信息)和当前的运动状态(例如速度、加速度)，得到横向和纵向的编码信息。
第一预测轨迹包括多个预测轨迹位置，以所述目标信息和所述第一历史轨迹信息为输入，通过LSTM编码器，得到多个编码信息，其中每个时刻对应一个编码信息，并以所述多个编码信息为输入，通过LSTM解码器，得到所述多个预测轨迹位置，其中，所述LSTM解码器用于基于多个输入数据解码得到所述多个预测轨迹位置，每个输入数据与所述多个编码信息有关。可选的，每个输入数据为对所述多个编码信息进行加权得到的。
解码器的作用是将编码器的编码信息解析，得到目标车辆未来一段时间(例如5s)的预测轨迹。但普通的解码器仅将编码器最后时刻的隐藏层输出作为解码器的输入，h1-h11为编码器各个时刻的输出，如图5A所示，图5A为一种解码过程的流程示意，普通轨迹解码器仅将h11作为解码器的输入，没有充分利用历史信息。
本申请实施例中，如图5B所示，图5B为本申请提供的一种解码过程的流程示意，如图5B中示出的那样，解码器每个时刻的输入取决于编码器所有时刻的输出。以解码器第一 个时刻的输入input1为例，通过注意力机制，得到关于编码器输出h1到h11，所有时刻的权值
Figure PCTCN2019129828-appb-000005
则有：
Figure PCTCN2019129828-appb-000006
本申请实施例中，对解码器的输出信息进行有效充分地利用，使得预测结果更加准确。
接下来结合一个应用实例对本申请实施例提供的轨迹预测方法进行描述：
第一步：数据采集和处理。
利用NGSIM(next-generation-simulation)开源数据库，提取目标车辆及其周围的第一车辆过去2s的历史轨迹，采样时间为0.2s，共11个采样点。对目标车辆的历史轨迹信息进行处理，提取出周围5辆第一车辆相对于目标车辆的纵向距离和纵向速度，并对其进行标准化，示例性的，总共提取了13023个样本做训练集，6511个样本做测试集。下表示出了其中一个样本：
表1：数据提取及处理实例信息表
Figure PCTCN2019129828-appb-000007
第二步：计算目标车辆向左变道、向右变道和保持执行的收益值。
计算收益值(GL,GR,GK)并进行归一化，其中表2为其中一个样本的收益值。
表2 目标车辆收益值表
Figure PCTCN2019129828-appb-000008
Figure PCTCN2019129828-appb-000009
第三步：识别目标车辆的行车意图。
将上述收益值与目标车辆相对于左右车道线的距离，以及横向速度，作为LSTM网络的输入。网络超参数可以示例性的设置为：批量大小batch_size＝128,迭代次数Epoch＝150,学习率＝0.001，采用交叉熵损失，利用Adam优化器进行参数更新。网络训练完成后离线保存，测试时直接调用该网络，对6511个测试样本进行行车意图的识别。
第四步：预测目标车辆横向和纵向轨迹(第一预测轨迹和第二预测轨迹)；
将目标车辆2s的历史横向轨迹作为横向编码器的输入，得到初始编码信息，再将初始编码信息将上一步得到的行车意图概率拼接，得到最终的横向编码信息。将该编码信息输入到基于注意力机制的解码器，得到目标车辆未来5s横向的预测轨迹。
同理，将目标车辆2s的历史纵向轨迹作为纵向编码器的输入，得到初始编码信息，再将初始编码信息将当前的运动状态拼接，得到最终的纵向编码信息。将该编码信息输入到基于注意力机制的解码器，得到目标车辆未来5s纵向的预测轨迹。
其中，编码解码器的网络结构示例性的可以为：输入长度input_length＝11，输出长度output_length＝25，编码器隐藏层大小encoder_size＝64，解码器隐藏层大小decoder_size＝64，在输入层和LSTM编码层之后分别加入一个全连接层，隐藏层大小均为32，批量大小batch_size＝128,迭代次数Epoch＝150,学习率＝0.001，采用均方误差损失，利用优化器进行参数更新。
本申请实施例中，获取目标车辆与至少一辆第一车辆之间的相对运动信息，以及所述目标车辆的第一历史轨迹信息；基于所述相对运动信息获取表示目标车辆行车意图的目标信息；以所述目标信息和所述第一历史轨迹信息为输入，通过第一轨迹预测模型，得到所述目标车辆的第一预测轨迹。通过上述方式，将目标车辆的行车意图作为轨迹预测的输入，使得对目标车辆的轨迹预测更加准确。
上述主要从终端设备的角度出发对本发明实施例提供的方案进行了介绍。可以理解的是，终端设备为了实现上述功能，其包含了执行各个功能相应的硬件结构和/或软件模块。结合本发明中所公开的实施例描述的各示例的单元及算法步骤，本发明实施例能够以硬件或硬件和计算机软件的结合形式来实现。某个功能究竟以硬件还是计算机软件驱动硬件的方式来执行，取决于技术方案的特定应用和设计约束条件。本领域技术人员可以对每个特定的应用来使用不同的方法来实现所描述的功能，但是这种实现不应认为超出本发明实施例的技术方案的范围。
本发明实施例可以根据上述方法示例对终端设备进行功能单元的划分，例如，可以对应各个功能划分各个功能单元，也可以将两个或两个以上的功能集成在一个处理单元中。上述集成的单元既可以采用硬件的形式实现，也可以采用软件功能单元的形式实现。需要 说明的是，本发明实施例中对单元的划分是示意性的，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式。
在采用集成的单元的情况下，图6示出了上述实施例中所涉及的终端设备的一种可能的结构示意图。终端设备600包括：获取模块601，用于获取目标车辆与位于传感器检测范围内的至少一辆第一车辆之间的相对运动信息，以及所述目标车辆的第一历史轨迹信息；基于所述相对运动信息获取表示目标车辆行车意图的目标信息；预测模块602，用于以所述目标信息和所述第一历史轨迹信息为输入，通过第一轨迹预测模型，得到所述目标车辆的第一预测轨迹。
可选的，所述行车意图至少包括如下的一种：向左变道、向右变道或保持直行。
可选的，所述相对运动信息表示目标车辆与每个第一车辆之间多个时刻的相对距离信息和/或相对速度信息，所述第一历史轨迹信息包括所述目标车辆多个时刻的第一历史轨迹位置。
可选的，所述获取模块601，具体用于：以所述相对运动信息为输入，通过线性收益函数模型，得到所述目标车辆每个时刻各行车意图的收益值；以所述目标车辆每个时刻各行车意图的收益值为输入，通过循环神经网络，得到表示目标车辆行车意图的目标信息。
可选的，所述第一历史轨迹信息和所述第一预测轨迹为所述目标车辆的行车轨迹在与车道垂直方向的分量，所述车道为所述目标车辆所在的车道。
可选的，所述第一预测轨迹包括多个预测轨迹位置，所述预测模块602，具体用于：以所述目标信息和所述第一历史轨迹信息为输入，通过LSTM编码器，得到多个编码信息，其中每个时刻对应一个编码信息；以所述多个编码信息为输入，通过LSTM解码器，得到所述多个预测轨迹位置，其中，所述LSTM解码器用于基于多个输入数据解码得到所述多个预测轨迹位置，每个输入数据与所述多个编码信息有关。
可选的，每个输入数据为对所述多个编码信息进行加权得到的。
可选的，所述获取模块601，还用于：获取所述目标车辆的运动状态和第二历史轨迹信息，所述第二历史轨迹信息为所述目标车辆的行车轨迹在与车道平行方向的分量，所述车道为所述目标车辆所在的车道；所述预测模块602，还用于：以所述目标车辆的运动状态和第二历史轨迹信息为输入，通过第二轨迹预测模型，得到所述目标车辆的第二预测轨迹，所述第二预测轨迹为所述目标车辆的行车轨迹在与车道平行方向的分量。
可选的，终端设备600还可以包括存储单元，用于存储终端设备600的程序代码和数据。
本申请实施例中，上述预测模块601可以集成在处理模块中，其中，处理模块可以是处理器或控制器，例如可以是中央处理器(central processing unit，CPU)，通用处理器，数字信号处理器(digital signal processor，DSP)，专用集成电路(application-specific integrated circuit，ASIC)，现场可编程门阵列(field programmable gate array，FPGA)或者其他可编程逻辑器件、晶体管逻辑器件、硬件部件或者其任意组合。其可以实现或执行结合本发明公开内容所描述的各种示例性的逻辑方框，模块和电路。所述处理器也可以是实现计算功能的组合，例如包含一个或多个微处理器组合，DSP和微处理器的组合等等。
参照图7，图7为本申请实施例提供的另一种终端设备700的结构示意。
参阅图7所示，该终端设备700包括：处理器712、通信接口713、存储器77。可选地，终端设备700还可以包括总线714。其中，通信接口713、处理器712以及存储器77可以通过总线714相互连接；总线714可以是外设部件互连标准(peripheral component interconnect，PCI)总线或扩展工业标准结构(extended industry standard architecture，EISA)总线等。所述总线714可以分为地址总线、数据总线、控制总线等。为便于表示，图7B中仅用一条粗线表示，但并不表示仅有一根总线或一种类型的总线。
其中，处理器712可以执行如下步骤：
获取目标车辆与至少一辆第一车辆之间的相对运动信息，以及所述目标车辆的第一历史轨迹信息；
基于所述相对运动信息获取表示目标车辆行车意图的目标信息；
以所述目标信息和所述第一历史轨迹信息为输入，通过第一轨迹预测模型，得到所述目标车辆的第一预测轨迹。
可选的，所述行车意图至少包括如下的一种：
向左变道、向右变道或保持直行。
可选的，所述相对运动信息表示目标车辆与每个第一车辆之间多个时刻的相对距离信息和/或相对速度信息，所述第一历史轨迹信息包括所述目标车辆多个时刻的第一历史轨迹位置。
可选的，处理器712可以执行如下步骤：
以所述相对运动信息为输入，通过线性收益函数模型，得到所述目标车辆每个时刻各行车意图的收益值；
以所述目标车辆每个时刻各行车意图的收益值为输入，通过循环神经网络，得到表示目标车辆行车意图的目标信息。
可选的，所述第一历史轨迹信息和所述第一预测轨迹表示与所述目标车辆所在车道的方向相垂直的轨迹信息。
可选的，所述第一预测轨迹包括多个预测轨迹位置，处理器712可以执行如下步骤：
以所述目标信息和所述第一历史轨迹信息为输入，通过LSTM编码器，得到多个编码信息，其中每个时刻对应一个编码信息；
以所述多个编码信息为输入，通过LSTM解码器，得到所述多个预测轨迹位置，其中，所述LSTM解码器用于基于多个输入数据解码得到所述多个预测轨迹位置，每个输入数据与所述多个编码信息有关。
可选的，每个输入数据为对所述多个编码信息进行加权得到的。
可选的，处理器712还可以执行如下步骤：
获取所述目标车辆的运动状态和第二历史轨迹信息，所述第二历史轨迹信息为与所述目标车辆行车方向一致的轨迹信息；
以所述目标车辆的运动状态和第二历史轨迹信息为输入，通过第二轨迹预测模型，得 到所述目标车辆的第二预测轨迹，所述第二预测轨迹为所述目标车辆的行车轨迹在与车道平行方向的分量。
上述图7所示的终端设备的具体实现还可以对应参照前述所述实施例的相应描述，此处不再赘述。
接下来介绍本申请实施例提供的一种执行设备，请参阅图8，图8为本申请实施例提供的执行设备的一种结构示意图。其中，执行设备800上可以部署有图6或图7对应实施例中所描述的终端设备，用于实现图6和图7对应实施例中终端设备的功能。具体的，执行设备800包括：接收器801、发射器802、处理器803和存储器804(其中执行设备800中的处理器803的数量可以一个或多个，图8中以一个处理器为例)，其中，处理器803可以包括应用处理器8031和通信处理器8032。在本申请的一些实施例中，接收器801、发射器802、处理器803和存储器804可通过总线或其它方式连接。
存储器804可以包括只读存储器和随机存取存储器，并向处理器803提供指令和数据。存储器804的一部分还可以包括非易失性随机存取存储器(non-volatile random access memory，NVRAM)。存储器804存储有处理器和操作指令、可执行模块或者数据结构，或者它们的子集，或者它们的扩展集，其中，操作指令可包括各种操作指令，用于实现各种操作。
处理器803控制执行设备的操作。具体的应用中，执行设备的各个组件通过总线系统耦合在一起，其中总线系统除包括数据总线之外，还可以包括电源总线、控制总线和状态信号总线等。但是为了清楚说明起见，在图中将各种总线都称为总线系统。
上述本申请实施例揭示的方法可以应用于处理器803中，或者由处理器803实现。处理器803可以是一种集成电路芯片，具有信号的处理能力。在实现过程中，上述方法的各步骤可以通过处理器803中的硬件的集成逻辑电路或者软件形式的指令完成。上述的处理器803可以是通用处理器、数字信号处理器(digital signal processing，DSP)、微处理器或微控制器，还可进一步包括专用集成电路(application specific integrated circuit，ASIC)、现场可编程门阵列(field-programmable gate array，FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件。该处理器803可以实现或者执行本申请实施例中的公开的各方法、步骤及逻辑框图。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。结合本申请实施例所公开的方法的步骤可以直接体现为硬件译码处理器执行完成，或者用译码处理器中的硬件及软件模块组合执行完成。软件模块可以位于随机存储器，闪存、只读存储器，可编程只读存储器或者电可擦写可编程存储器、寄存器等本领域成熟的存储介质中。该存储介质位于存储器804，处理器803读取存储器804中的信息，结合其硬件完成上述方法的步骤。
接收器801可用于接收输入的数字或字符信息，以及产生与执行设备的相关设置以及功能控制有关的信号输入。发射器802可用于通过第一接口输出数字或字符信息；发射器802还可用于通过第一接口向磁盘组发送指令，以修改磁盘组中的数据。
本申请实施例中还提供一种包括计算机程序产品，当其在计算机上运行时，使得计算机执行如前述图6或图7所示实施例描述的方法中终端设备所执行的步骤。
本申请实施例中还提供一种计算机可读存储介质，该计算机可读存储介质中存储有用于进行信号处理的程序，当其在计算机上运行时，使得计算机执行如前述图6或图7所示实施例描述的方法中终端设备所执行的步骤。
本申请实施例提供的执行设备或终端设备具体可以为芯片，芯片包括：处理单元和通信单元，所述处理单元例如可以是处理器，所述通信单元例如可以是输入/输出接口、管脚或电路等。该处理单元可执行存储单元存储的计算机执行指令，以使执行设备内的芯片执行上述图2所示实施例描述的轨迹预测方法。可选地，所述存储单元为所述芯片内的存储单元，如寄存器、缓存等，所述存储单元还可以是所述无线接入设备端内的位于所述芯片外部的存储单元，如只读存储器(read-only memory，ROM)或可存储静态信息和指令的其他类型的静态存储设备，随机存取存储器(random access memory，RAM)等。
具体的，请参阅图9，图9为本申请实施例提供的芯片的一种结构示意图，所述芯片可以表现为神经网络处理器NPU 900，NPU 900作为协处理器挂载到主CPU(Host CPU)上，由Host CPU分配任务。NPU的核心部分为运算电路，通过控制器904控制运算电路903提取存储器中的矩阵数据并进行乘法运算。
在一些实现中，运算电路903内部包括多个处理单元(Process Engine，PE)。在一些实现中，运算电路903是二维脉动阵列。运算电路903还可以是一维脉动阵列或者能够执行例如乘法和加法这样的数学运算的其它电子线路。在一些实现中，运算电路903是通用的矩阵处理器。
举例来说，假设有输入矩阵A，权重矩阵B，输出矩阵C。运算电路从权重存储器902中取矩阵B相应的数据，并缓存在运算电路中每一个PE上。运算电路从输入存储器901中取矩阵A数据与矩阵B进行矩阵运算，得到的矩阵的部分结果或最终结果，保存在累加器(accumulator)908中。
统一存储器906用于存放输入数据以及输出数据。权重数据直接通过存储单元访问控制器(Direct Memory Access Controller，DMAC)905，DMAC被搬运到权重存储器902中。输入数据也通过DMAC被搬运到统一存储器906中。
BIU为Bus Interface Unit即，总线接口单元910，用于AXI总线与DMAC和取指存储器(Instruction Fetch Buffer，IFB)909的交互。
总线接口单元910(Bus Interface Unit，简称BIU)，用于取指存储器909从外部存储器获取指令，还用于存储单元访问控制器905从外部存储器获取输入矩阵A或者权重矩阵B的原数据。
DMAC主要用于将外部存储器DDR中的输入数据搬运到统一存储器906或将权重数据搬运到权重存储器902中或将输入数据数据搬运到输入存储器901中。
向量计算单元907包括多个运算处理单元，在需要的情况下，对运算电路的输出做进一步处理，如向量乘，向量加，指数运算，对数运算，大小比较等等。主要用于神经网络中非卷积/全连接层网络计算，如Batch Normalization(批归一化)，像素级求和，对特征平面进行上采样等。
在一些实现中，向量计算单元907能将经处理的输出的向量存储到统一存储器906。 例如，向量计算单元907可以将线性函数和/或非线性函数应用到运算电路903的输出，例如对卷积层提取的特征平面进行线性插值，再例如累加值的向量，用以生成激活值。在一些实现中，向量计算单元907生成归一化的值、像素级求和的值，或二者均有。在一些实现中，处理过的输出的向量能够用作到运算电路903的激活输入，例如用于在神经网络中的后续层中的使用。
控制器904连接的取指存储器(instruction fetch buffer)909，用于存储控制器904使用的指令；
统一存储器906，输入存储器901，权重存储器902以及取指存储器909均为On-Chip存储器。外部存储器私有于该NPU硬件架构。
其中，上述任一处提到的处理器，可以是一个通用中央处理器，微处理器，ASIC，或一个或多个用于控制上述第一方面方法的程序执行的集成电路。
另外需说明的是，以上所描述的装置实施例仅仅是示意性的，其中所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部模块来实现本实施例方案的目的。另外，本申请提供的装置实施例附图中，模块之间的连接关系表示它们之间具有通信连接，具体可以实现为一条或多条通信总线或信号线。
通过以上的实施方式的描述，所属领域的技术人员可以清楚地了解到本申请可借助软件加必需的通用硬件的方式来实现，当然也可以通过专用硬件包括专用集成电路、专用CPU、专用存储器、专用元器件等来实现。一般情况下，凡由计算机程序完成的功能都可以很容易地用相应的硬件来实现，而且，用来实现同一功能的具体硬件结构也可以是多种多样的，例如模拟电路、数字电路或专用电路等。但是，对本申请而言更多情况下软件程序实现是更佳的实施方式。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来，该计算机软件产品存储在可读取的存储介质中，如计算机的软盘、U盘、移动硬盘、ROM、RAM、磁碟或者光盘等，包括若干指令用以使得一台计算机设备(可以是个人计算机，训练设备，或者网络设备等)执行本申请各个实施例所述的方法。
在上述实施例中，可以全部或部分地通过软件、硬件、固件或者其任意组合来实现。当使用软件实现时，可以全部或部分地以计算机程序产品的形式实现。
所述计算机程序产品包括一个或多个计算机指令。在计算机上加载和执行所述计算机程序指令时，全部或部分地产生按照本申请实施例所述的流程或功能。所述计算机可以是通用计算机、专用计算机、计算机网络、或者其他可编程装置。所述计算机指令可以存储在计算机可读存储介质中，或者从一个计算机可读存储介质向另一计算机可读存储介质传输，例如，所述计算机指令可以从一个网站站点、计算机、训练设备或数据中心通过有线(例如同轴电缆、光纤、数字用户线(DSL))或无线(例如红外、无线、微波等)方式向另一个网站站点、计算机、训练设备或数据中心进行传输。所述计算机可读存储介质可以 是计算机能够存储的任何可用介质或者是包含一个或多个可用介质集成的训练设备、数据中心等数据存储设备。所述可用介质可以是磁性介质，(例如，软盘、硬盘、磁带)、光介质(例如，DVD)、或者半导体介质(例如固态硬盘(Solid State Disk，SSD))等。

Claims (18)

  1. 一种轨迹预测方法，其特征在于，所述方法包括：
    获取目标车辆与位于传感器的检测范围内的至少一辆第一车辆之间的相对运动信息，以及所述目标车辆的第一历史轨迹信息；
    基于所述相对运动信息获取表示目标车辆行车意图的目标信息；
    以所述目标信息和所述第一历史轨迹信息为输入，通过第一轨迹预测模型，得到所述目标车辆的第一预测轨迹。
  2. 根据权利要求1所述的方法，其特征在于，所述行车意图至少包括如下的一种：
    向左变道、向右变道或保持直行。
  3. 根据权利要求1或2所述的方法，其特征在于，所述相对运动信息表示目标车辆与每个第一车辆之间多个时刻的相对距离信息和/或相对速度信息，所述第一历史轨迹信息包括所述目标车辆多个时刻的第一历史轨迹位置。
  4. 根据权利要求3所述的方法，其特征在于，所述基于所述相对运动信息获取表示目标车辆行车意图的目标信息，包括：
    以所述相对运动信息为输入，通过线性收益函数模型，得到所述目标车辆每个时刻各行车意图的收益值；
    以所述目标车辆每个时刻各行车意图的收益值为输入，通过循环神经网络，得到表示目标车辆行车意图的目标信息。
  5. 根据权利要求1至4任一所述的方法，其特征在于，所述第一历史轨迹信息和所述第一预测轨迹为所述目标车辆的行车轨迹在与车道垂直方向的分量，所述车道为所述目标车辆所在的车道。
  6. 根据权利要求3至5任一所述的方法，其特征在于，所述第一预测轨迹包括多个预测轨迹位置，所述以所述目标信息和所述第一历史轨迹信息为输入，通过第一轨迹预测模型，得到所述目标车辆的第一预测轨迹，包括：
    以所述目标信息和所述第一历史轨迹信息为输入，通过LSTM编码器，得到多个编码信息，其中每个时刻对应一个编码信息；
    以所述多个编码信息为输入，通过LSTM解码器，得到所述多个预测轨迹位置，其中，所述LSTM解码器用于基于多个输入数据解码得到所述多个预测轨迹位置，每个输入数据与所述多个编码信息有关。
  7. 根据权利要求6所述的方法，其特征在于，每个输入数据为对所述多个编码信息进行加权得到的。
  8. 根据权利要求1至7任一所述的方法，其特征在于，所述方法还包括：
    获取所述目标车辆的运动状态和第二历史轨迹信息，所述第二历史轨迹信息为所述目标车辆的行车轨迹在与车道平行方向的分量，所述车道为所述目标车辆所在的车道；
    以所述目标车辆的运动状态和第二历史轨迹信息为输入，通过第二轨迹预测模型，得到所述目标车辆的第二预测轨迹，所述第二预测轨迹为所述目标车辆的行车轨迹在与车道平行方向的分量。
  9. 一种执行设备，其特征在于，包括：
    获取模块，用于获取目标车辆与位于传感器的检测范围内的至少一辆第一车辆之间的相对运动信息，以及所述目标车辆的第一历史轨迹信息；
    基于所述相对运动信息获取表示目标车辆行车意图的目标信息；
    预测模块，用于以所述目标信息和所述第一历史轨迹信息为输入，通过第一轨迹预测模型，得到所述目标车辆的第一预测轨迹。
  10. 根据权利要求9所述的执行设备，其特征在于，所述行车意图至少包括如下的一种：
    向左变道、向右变道或保持直行。
  11. 根据权利要求9或10所述的执行设备，其特征在于，所述相对运动信息表示目标车辆与每个第一车辆之间多个时刻的相对距离信息和/或相对速度信息，所述第一历史轨迹信息包括所述目标车辆多个时刻的第一历史轨迹位置。
  12. 根据权利要求11所述的执行设备，其特征在于，所述获取模块，具体用于：
    以所述相对运动信息为输入，通过线性收益函数模型，得到所述目标车辆每个时刻各行车意图的收益值；
    以所述目标车辆每个时刻各行车意图的收益值为输入，通过循环神经网络，得到表示目标车辆行车意图的目标信息。
  13. 根据权利要求9至12任一所述的执行设备，其特征在于，所述第一历史轨迹信息和所述第一预测轨迹为所述目标车辆的行车轨迹在与车道垂直方向的分量，所述车道为所述目标车辆所在的车道。
  14. 根据权利要求11至13任一所述的执行设备，其特征在于，所述第一预测轨迹包括多个预测轨迹位置，所述预测模块，具体用于：
    以所述目标信息和所述第一历史轨迹信息为输入，通过LSTM编码器，得到多个编码信息，其中每个时刻对应一个编码信息；
    以所述多个编码信息为输入，通过LSTM解码器，得到所述多个预测轨迹位置，其中，所述LSTM解码器用于基于多个输入数据解码得到所述多个预测轨迹位置，每个输入数据与所述多个编码信息有关。
  15. 根据权利要求14所述的执行设备，其特征在于，每个输入数据为对所述多个编码信息进行加权得到的。
  16. 根据权利要求9至15任一所述的执行设备，其特征在于，所述获取模块，还用于：
    获取所述目标车辆的运动状态和第二历史轨迹信息，所述第二历史轨迹信息为所述目标车辆的行车轨迹在与车道平行方向的分量，所述车道为所述目标车辆所在的车道；
    所述预测模块，还用于：
    以所述目标车辆的运动状态和第二历史轨迹信息为输入，通过第二轨迹预测模型，得到所述目标车辆的第二预测轨迹，所述第二预测轨迹为所述目标车辆的行车轨迹在与车道平行方向的分量。
  17. 一种终端设备，其特征在于，包括存储器、通信接口及与所述存储器和通信接口 耦合的处理器；所述存储器用于存储指令，所述处理器用于执行所述指令，所述通信接口用于在所述处理器的控制下与目标车辆进行通信；其中，所述处理器执行所述指令时执行如上权利要求1-8中任一项所述的方法。
  18. 一种计算机可读存储介质，所述计算机可读存储介质存储有计算机程序，其特征在于，所述计算机程序被处理器执行时实现如权利要求1至8中任一项所述方法。
PCT/CN2019/129828 2019-12-30 2019-12-30 一种轨迹预测方法及相关设备 WO2021134172A1 (zh)

Priority Applications (3)

Application Number Priority Date Filing Date Title
EP19958672.8A EP4075407A4 (en) 2019-12-30 2019-12-30 TRAJECTORY PREDICTION METHOD AND ASSOCIATED DEVICE
PCT/CN2019/129828 WO2021134172A1 (zh) 2019-12-30 2019-12-30 一种轨迹预测方法及相关设备
CN201980080199.4A CN113261035B (zh) 2019-12-30 2019-12-30 一种轨迹预测方法及相关设备

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2019/129828 WO2021134172A1 (zh) 2019-12-30 2019-12-30 一种轨迹预测方法及相关设备

Publications (1)

Publication Number Publication Date
WO2021134172A1 true WO2021134172A1 (zh) 2021-07-08

Family

ID=76685789

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/129828 WO2021134172A1 (zh) 2019-12-30 2019-12-30 一种轨迹预测方法及相关设备

Country Status (3)

Country Link
EP (1) EP4075407A4 (zh)
CN (1) CN113261035B (zh)
WO (1) WO2021134172A1 (zh)

Cited By (28)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113911129A (zh) * 2021-11-23 2022-01-11 吉林大学 一种基于驾驶行为生成机制的交通车意图识别方法
CN113954863A (zh) * 2021-09-07 2022-01-21 江苏大学 基于双注意力机制和改进Social Gan的车辆轨迹预测方法
CN113989330A (zh) * 2021-11-03 2022-01-28 中国电信股份有限公司 车辆轨迹预测方法、装置、电子设备和可读存储介质
CN114115247A (zh) * 2021-11-10 2022-03-01 北京三快在线科技有限公司 一种无人设备控制方法、装置、存储介质及电子设备
CN114228746A (zh) * 2022-01-17 2022-03-25 北京经纬恒润科技股份有限公司 一种预测车辆运动轨迹的方法及装置
CN114291116A (zh) * 2022-01-24 2022-04-08 广州小鹏自动驾驶科技有限公司 周围车辆轨迹预测方法、装置、车辆及存储介质
CN114312769A (zh) * 2022-02-11 2022-04-12 清华大学 考虑周车横纵向运动意图的智能车辆紧急制动方法及系统
CN114312831A (zh) * 2021-12-16 2022-04-12 浙江零跑科技股份有限公司 一种基于空间注意力机制的车辆轨迹预测方法
CN114368387A (zh) * 2021-12-21 2022-04-19 吉林大学 一种基于注意力机制的驾驶员意图识别及车辆轨迹预测方法
CN114426032A (zh) * 2022-01-05 2022-05-03 重庆长安汽车股份有限公司 一种基于自动驾驶的本车轨迹预测方法、系统、车辆及计算机可读存储介质
CN114506344A (zh) * 2022-03-10 2022-05-17 福瑞泰克智能系统有限公司 一种车辆轨迹的确定方法及装置
CN114565191A (zh) * 2022-03-30 2022-05-31 吉林大学 一种基于长短记忆网络的双向交互车辆的轨迹预测方法
CN114613144A (zh) * 2022-04-07 2022-06-10 重庆大学 一种基于Embedding-CNN的混合车群运动演化规律的刻画方法
CN114781758A (zh) * 2022-05-27 2022-07-22 吉林大学 一种考虑持续交互信息的车辆轨迹预测方法
CN114872718A (zh) * 2022-04-11 2022-08-09 清华大学 车辆轨迹预测方法、装置、计算机设备和存储介质
CN114898558A (zh) * 2022-05-23 2022-08-12 东风悦享科技有限公司 车辆协作通行的方法、系统、电子设备和存储介质
CN115086645A (zh) * 2022-06-10 2022-09-20 湖南师范大学 一种面向全景视频的视点预测方法、装置及介质
CN115131393A (zh) * 2021-08-16 2022-09-30 北京百度网讯科技有限公司 轨迹预测方法、碰撞检测方法、装置、电子设备及介质
CN115158331A (zh) * 2022-07-12 2022-10-11 东风柳州汽车有限公司 防止乘客眩晕的方法、装置、设备及存储介质
CN115440087A (zh) * 2022-06-06 2022-12-06 北京罗克维尔斯科技有限公司 车辆检测方法、装置、设备、介质及车辆
CN115805947A (zh) * 2022-11-21 2023-03-17 清华大学 一种个性化车辆轨迹预测方法、系统、设备和存储介质
CN115965657A (zh) * 2023-02-28 2023-04-14 安徽蔚来智驾科技有限公司 目标跟踪方法、电子设备、存储介质及车辆
CN116110216A (zh) * 2022-10-21 2023-05-12 中国第一汽车股份有限公司 车辆跨线时间确定方法、装置、存储介质及电子装置
CN116662788A (zh) * 2023-07-27 2023-08-29 太平金融科技服务(上海)有限公司深圳分公司 一种车辆轨迹处理方法、装置、设备和存储介质
CN116890881A (zh) * 2023-09-08 2023-10-17 摩尔线程智能科技(北京)有限责任公司 一种车辆变道决策生成方法及装置、电子设备和存储介质
CN116989818A (zh) * 2023-09-26 2023-11-03 毫末智行科技有限公司 一种轨迹生成方法、装置、电子设备及可读存储介质
CN117251748A (zh) * 2023-10-10 2023-12-19 中国船舶集团有限公司第七〇九研究所 一种基于历史规律挖掘的航迹预测方法、设备及存储介质
CN117775078A (zh) * 2024-02-28 2024-03-29 山西阳光三极科技股份有限公司 一种基于深度学习的矿内货运列车行驶方向判断方法

Families Citing this family (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114153207B (zh) * 2021-11-29 2024-02-27 北京三快在线科技有限公司 一种无人驾驶设备的控制方法及控制装置
CN113997955B (zh) * 2021-12-13 2024-06-14 中国第一汽车股份有限公司 轨迹预测方法、装置、电子设备以及存储介质
CN114664051B (zh) * 2022-02-25 2023-09-29 长安大学 一种高速公路弯道临时施工区预警方法
CN114758502B (zh) * 2022-04-29 2023-08-18 北京百度网讯科技有限公司 双车联合轨迹预测方法及装置、电子设备和自动驾驶车辆
CN117799633A (zh) * 2022-09-26 2024-04-02 华为技术有限公司 一种预测方法、装置和智能驾驶设备
CN118067142A (zh) * 2022-11-22 2024-05-24 华为技术有限公司 一种预测方法、装置和移动载体
CN116561437A (zh) * 2023-07-10 2023-08-08 荣耀终端有限公司 用户行为预测方法、终端设备及存储介质
CN117330063B (zh) * 2023-12-01 2024-03-22 华南理工大学 一种提升imu和轮速计组合定位算法精度的方法

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2018090661A1 (en) * 2016-11-18 2018-05-24 Huawei Technologies Co., Ltd. Path planning for autonomous vehicle using bidirectional search
CN109583151A (zh) * 2019-02-20 2019-04-05 百度在线网络技术（北京）有限公司 车辆的行驶轨迹预测方法及装置
CN109801508A (zh) * 2019-02-26 2019-05-24 百度在线网络技术（北京）有限公司 路口处障碍物的运动轨迹预测方法及装置
CN109885066A (zh) * 2019-03-26 2019-06-14 北京经纬恒润科技有限公司 一种运动轨迹预测方法及装置
CN110293968A (zh) * 2019-06-18 2019-10-01 百度在线网络技术（北京）有限公司 自动驾驶车辆的控制方法、装置、设备及可读存储介质

Family Cites Families (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP5472238B2 (ja) * 2011-08-29 2014-04-16 株式会社デンソー 車両用無線通信装置及び通信システム
CN104680838B (zh) * 2013-11-28 2017-10-10 三星电子（中国）研发中心 用于汽车的安全辅助方法和系统
US10479373B2 (en) * 2016-01-06 2019-11-19 GM Global Technology Operations LLC Determining driver intention at traffic intersections for automotive crash avoidance
EP3548845B1 (en) * 2017-01-12 2021-10-13 Mobileye Vision Technologies Ltd. Navigation based on vehicle activity
CN108437986B (zh) * 2017-02-16 2020-07-03 上海汽车集团股份有限公司 车辆驾驶辅助系统及辅助方法
CN110267856B (zh) * 2017-03-01 2022-09-23 本田技研工业株式会社 车辆控制装置、车辆控制方法及存储介质
DE102017219377A1 (de) * 2017-10-27 2019-05-02 Bayerische Motoren Werke Aktiengesellschaft Verfahren zur Optimierung einer Pfadplanung für ein Fahrzeug
CN109413572A (zh) * 2018-10-31 2019-03-01 惠州市德赛西威汽车电子股份有限公司 车辆碰撞预警及车速引导的优化方法及系统
CN110223318A (zh) * 2019-04-28 2019-09-10 驭势科技（北京）有限公司 一种多目标轨迹的预测方法、装置、车载设备及存储介质
CN110264721B (zh) * 2019-07-01 2021-04-20 北京理工大学 一种城市交叉口周边车辆轨迹预测方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2018090661A1 (en) * 2016-11-18 2018-05-24 Huawei Technologies Co., Ltd. Path planning for autonomous vehicle using bidirectional search
CN109583151A (zh) * 2019-02-20 2019-04-05 百度在线网络技术（北京）有限公司 车辆的行驶轨迹预测方法及装置
CN109801508A (zh) * 2019-02-26 2019-05-24 百度在线网络技术（北京）有限公司 路口处障碍物的运动轨迹预测方法及装置
CN109885066A (zh) * 2019-03-26 2019-06-14 北京经纬恒润科技有限公司 一种运动轨迹预测方法及装置
CN110293968A (zh) * 2019-06-18 2019-10-01 百度在线网络技术（北京）有限公司 自动驾驶车辆的控制方法、装置、设备及可读存储介质

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP4075407A4 *

Cited By (47)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115131393B (zh) * 2021-08-16 2024-04-23 北京百度网讯科技有限公司 轨迹预测方法、碰撞检测方法、装置、电子设备及介质
CN115131393A (zh) * 2021-08-16 2022-09-30 北京百度网讯科技有限公司 轨迹预测方法、碰撞检测方法、装置、电子设备及介质
CN113954863A (zh) * 2021-09-07 2022-01-21 江苏大学 基于双注意力机制和改进Social Gan的车辆轨迹预测方法
CN113954863B (zh) * 2021-09-07 2024-06-11 江苏大学 基于双注意力机制和改进Social Gan的车辆轨迹预测方法
CN113989330A (zh) * 2021-11-03 2022-01-28 中国电信股份有限公司 车辆轨迹预测方法、装置、电子设备和可读存储介质
CN114115247A (zh) * 2021-11-10 2022-03-01 北京三快在线科技有限公司 一种无人设备控制方法、装置、存储介质及电子设备
CN114115247B (zh) * 2021-11-10 2024-03-26 北京三快在线科技有限公司 一种无人设备控制方法、装置、存储介质及电子设备
CN113911129A (zh) * 2021-11-23 2022-01-11 吉林大学 一种基于驾驶行为生成机制的交通车意图识别方法
CN113911129B (zh) * 2021-11-23 2023-02-24 吉林大学 一种基于驾驶行为生成机制的交通车意图识别方法
CN114312831A (zh) * 2021-12-16 2022-04-12 浙江零跑科技股份有限公司 一种基于空间注意力机制的车辆轨迹预测方法
CN114312831B (zh) * 2021-12-16 2023-10-03 浙江零跑科技股份有限公司 一种基于空间注意力机制的车辆轨迹预测方法
CN114368387A (zh) * 2021-12-21 2022-04-19 吉林大学 一种基于注意力机制的驾驶员意图识别及车辆轨迹预测方法
CN114426032A (zh) * 2022-01-05 2022-05-03 重庆长安汽车股份有限公司 一种基于自动驾驶的本车轨迹预测方法、系统、车辆及计算机可读存储介质
CN114228746B (zh) * 2022-01-17 2024-05-07 北京经纬恒润科技股份有限公司 一种预测车辆运动轨迹的方法及装置
CN114228746A (zh) * 2022-01-17 2022-03-25 北京经纬恒润科技股份有限公司 一种预测车辆运动轨迹的方法及装置
CN114291116B (zh) * 2022-01-24 2023-05-16 广州小鹏自动驾驶科技有限公司 周围车辆轨迹预测方法、装置、车辆及存储介质
CN114291116A (zh) * 2022-01-24 2022-04-08 广州小鹏自动驾驶科技有限公司 周围车辆轨迹预测方法、装置、车辆及存储介质
CN114312769A (zh) * 2022-02-11 2022-04-12 清华大学 考虑周车横纵向运动意图的智能车辆紧急制动方法及系统
CN114312769B (zh) * 2022-02-11 2023-11-10 清华大学 考虑周车横纵向运动意图的智能车辆紧急制动方法及系统
CN114506344B (zh) * 2022-03-10 2024-03-08 福瑞泰克智能系统有限公司 一种车辆轨迹的确定方法及装置
CN114506344A (zh) * 2022-03-10 2022-05-17 福瑞泰克智能系统有限公司 一种车辆轨迹的确定方法及装置
CN114565191A (zh) * 2022-03-30 2022-05-31 吉林大学 一种基于长短记忆网络的双向交互车辆的轨迹预测方法
CN114613144A (zh) * 2022-04-07 2022-06-10 重庆大学 一种基于Embedding-CNN的混合车群运动演化规律的刻画方法
CN114613144B (zh) * 2022-04-07 2024-04-30 重庆大学 一种基于Embedding-CNN的混合车群运动演化规律的刻画方法
CN114872718A (zh) * 2022-04-11 2022-08-09 清华大学 车辆轨迹预测方法、装置、计算机设备和存储介质
CN114898558A (zh) * 2022-05-23 2022-08-12 东风悦享科技有限公司 车辆协作通行的方法、系统、电子设备和存储介质
CN114781758A (zh) * 2022-05-27 2022-07-22 吉林大学 一种考虑持续交互信息的车辆轨迹预测方法
CN115440087A (zh) * 2022-06-06 2022-12-06 北京罗克维尔斯科技有限公司 车辆检测方法、装置、设备、介质及车辆
CN115086645A (zh) * 2022-06-10 2022-09-20 湖南师范大学 一种面向全景视频的视点预测方法、装置及介质
CN115086645B (zh) * 2022-06-10 2024-06-07 湖南师范大学 一种面向全景视频的视点预测方法、装置及介质
CN115158331A (zh) * 2022-07-12 2022-10-11 东风柳州汽车有限公司 防止乘客眩晕的方法、装置、设备及存储介质
CN116110216B (zh) * 2022-10-21 2024-04-12 中国第一汽车股份有限公司 车辆跨线时间确定方法、装置、存储介质及电子装置
CN116110216A (zh) * 2022-10-21 2023-05-12 中国第一汽车股份有限公司 车辆跨线时间确定方法、装置、存储介质及电子装置
CN115805947B (zh) * 2022-11-21 2023-06-30 清华大学 一种个性化车辆轨迹预测方法、系统、设备和存储介质
CN115805947A (zh) * 2022-11-21 2023-03-17 清华大学 一种个性化车辆轨迹预测方法、系统、设备和存储介质
CN115965657A (zh) * 2023-02-28 2023-04-14 安徽蔚来智驾科技有限公司 目标跟踪方法、电子设备、存储介质及车辆
CN115965657B (zh) * 2023-02-28 2023-06-02 安徽蔚来智驾科技有限公司 目标跟踪方法、电子设备、存储介质及车辆
CN116662788B (zh) * 2023-07-27 2024-04-02 太平金融科技服务(上海)有限公司深圳分公司 一种车辆轨迹处理方法、装置、设备和存储介质
CN116662788A (zh) * 2023-07-27 2023-08-29 太平金融科技服务(上海)有限公司深圳分公司 一种车辆轨迹处理方法、装置、设备和存储介质
CN116890881B (zh) * 2023-09-08 2023-12-08 摩尔线程智能科技(北京)有限责任公司 一种车辆变道决策生成方法及装置、电子设备和存储介质
CN116890881A (zh) * 2023-09-08 2023-10-17 摩尔线程智能科技(北京)有限责任公司 一种车辆变道决策生成方法及装置、电子设备和存储介质
CN116989818B (zh) * 2023-09-26 2024-01-19 毫末智行科技有限公司 一种轨迹生成方法、装置、电子设备及可读存储介质
CN116989818A (zh) * 2023-09-26 2023-11-03 毫末智行科技有限公司 一种轨迹生成方法、装置、电子设备及可读存储介质
CN117251748A (zh) * 2023-10-10 2023-12-19 中国船舶集团有限公司第七〇九研究所 一种基于历史规律挖掘的航迹预测方法、设备及存储介质
CN117251748B (zh) * 2023-10-10 2024-04-19 中国船舶集团有限公司第七〇九研究所 一种基于历史规律挖掘的航迹预测方法、设备及存储介质
CN117775078A (zh) * 2024-02-28 2024-03-29 山西阳光三极科技股份有限公司 一种基于深度学习的矿内货运列车行驶方向判断方法
CN117775078B (zh) * 2024-02-28 2024-05-07 山西阳光三极科技股份有限公司 一种基于深度学习的矿内货运列车行驶方向判断方法

Also Published As

Publication number Publication date
CN113261035B (zh) 2022-09-16
CN113261035A (zh) 2021-08-13
EP4075407A1 (en) 2022-10-19
EP4075407A4 (en) 2022-12-28

Similar Documents

Publication Publication Date Title
WO2021134172A1 (zh) 一种轨迹预测方法及相关设备
WO2021134169A1 (zh) 一种轨迹预测方法及相关设备
US11899411B2 (en) Hybrid reinforcement learning for autonomous driving
Li et al. Survey on artificial intelligence for vehicles
Ma et al. Artificial intelligence applications in the development of autonomous vehicles: A survey
US20220105961A1 (en) Method for automatic control of vehicle and method for training lane change intention prediction network
CN111971574B (zh) 用于自动驾驶车辆的lidar定位的基于深度学习的特征提取
KR102539942B1 (ko) 궤적 계획 모델을 훈련하는 방법, 장치, 전자 기기, 저장 매체 및 프로그램
EP4152204A1 (en) Lane line detection method, and related apparatus
CN113168708B (zh) 车道线跟踪方法和装置
CN111771135B (zh) 自动驾驶车辆中使用rnn和lstm进行时间平滑的lidar定位
CN111771141A (zh) 自动驾驶车辆中使用3d cnn网络进行解决方案推断的lidar定位
CN111476079A (zh) 将用于对象检测的地图特征与lidar合并的全面且有效的方法
CN110390240A (zh) 自动驾驶车辆中的车道后处理
WO2022178858A1 (zh) 一种车辆行驶意图预测方法、装置、终端及存储介质
US11203348B2 (en) System and method for predicting and interpreting driving behavior
CN115214708A (zh) 一种车辆意图预测方法及其相关装置
WO2024008086A1 (zh) 轨迹预测方法及其装置、介质、程序产品和电子设备
US20240246575A1 (en) Autonomous driving method
WO2023060963A1 (zh) 道路信息的识别方法、装置、电子设备、车辆及介质
CN114882457A (zh) 一种模型的训练方法、车道线的检测方法及设备
WO2023123172A1 (zh) 一种驾驶辅助的方法以及相关设备
Azadani et al. Toward driver intention prediction for intelligent vehicles: A deep learning approach
Khanum et al. Involvement of deep learning for vision sensor-based autonomous driving control: a review
CN116701586A (zh) 一种数据处理方法及其相关装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19958672

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2019958672

Country of ref document: EP

Effective date: 20220713

NENP Non-entry into the national phase

Ref country code: DE