Intersection pedestrian trajectory prediction method and device

Abstract

This application relates to the field of intelligent driving technology for automobiles, and particularly to a method and apparatus for predicting pedestrian trajectories at intersections. The method includes: acquiring the current location and historical trajectory of at least one pedestrian, and identifying the movement trend of each pedestrian; identifying the predicted movement state of each pedestrian based on the current location, historical trajectory, and movement trend; when the predicted movement state is a turning state, a crossing state, or a starting state, predicting at least one predicted target point for each pedestrian based on their movement state, and generating a predicted intersection trajectory for each pedestrian based on the at least one predicted target point. The embodiments of this application can generate a predicted intersection trajectory for each pedestrian based on their current location, historical trajectory, and movement trend, effectively adapting to all environments, improving the applicability and accuracy of pedestrian trajectory prediction, effectively ensuring the safety and reliability of user driving, and meeting user needs.

Description

Technical Field

[0001] This application relates to the field of intelligent driving technology for automobiles, and in particular to a method and device for predicting pedestrian trajectories at intersections. Background Technology

[0002] In related technologies, data-driven prediction methods are used, employing neural network methods such as RNN (Recurrent Neural Network) and LSTM (Long Short-Term Memory). The neural network parameters are obtained through adaptive learning of a large amount of data, which can better address the challenges of dynamic changes in pedestrian movement.

[0003] However, due to the large number of neural network parameters and the complex model structure, the related technologies cannot converge during training and cannot be effectively applied to all environments. This reduces the applicability and accuracy of pedestrian trajectory prediction, resulting in lower driving safety for users and failing to meet their vehicle usage needs, which urgently needs to be addressed. Summary of the Invention

[0004] This application provides a method and apparatus for predicting pedestrian trajectories at intersections, in order to solve the technical problems in related technologies where the large number of neural network parameters and the complex model structure lead to the inability to converge during training, and the inability to be effectively applied to all environments, thus reducing the applicability of pedestrian trajectory prediction, lowering the safety of driving, and failing to meet the needs of users.

[0005] The first aspect of this application provides a method for predicting pedestrian trajectories at intersections, comprising the following steps: obtaining the current location and historical trajectory of at least one pedestrian, and identifying the movement trend of each pedestrian; identifying the predicted movement state of each pedestrian based on the current location, the historical trajectory, and the movement trend; when the predicted movement state is a turning state, a crossing state, or a starting state, predicting at least one predicted target point for each pedestrian based on the pedestrian movement state, and generating the predicted intersection trajectory of each pedestrian based on the at least one predicted target point.

[0006] Based on the above technical means, the embodiments of this application can generate a predicted intersection trajectory for each pedestrian based on the pedestrian's current location, historical trajectory, and movement trend. This effectively adapts to all environments, improves the applicability and accuracy of pedestrian trajectory prediction, effectively ensures the safety and reliability of the user's driving, and meets the user's vehicle usage needs.

[0007] Optionally, in one embodiment of this application, obtaining the current location and historical trajectory of at least one pedestrian includes: at least one pedestrian coordinate point based on a preset historical time in the vehicle coordinate system; and fitting the historical trajectory of each pedestrian based on the at least one pedestrian coordinate point.

[0008] Based on the above technical means, the embodiments of this application can convert the coordinates of pedestrians into the historical trajectory of pedestrians, effectively improving the accuracy of each pedestrian's historical trajectory and increasing the feasibility of pedestrian trajectory prediction.

[0009] Optionally, in one embodiment of this application, obtaining the current location and historical trajectory of at least one pedestrian includes: obtaining lane information at the current intersection; and calculating the pedestrian's current location relative to the road based on the lane information.

[0010] Based on the above technical means, the embodiments of this application can accurately determine the current location of pedestrians, improve the reliability of pedestrian trajectory prediction, enhance vehicle driving safety, and reduce the occurrence of traffic accidents.

[0011] Optionally, in one embodiment of this application, before obtaining the current location and historical trajectory of the at least one pedestrian, the method further includes: filtering pedestrians from all pedestrians who are within a first preset distance from the vehicle and do not meet the conditions of being outside the road or being more than a second preset distance from the road edge, and determining the at least one pedestrian.

[0012] Based on the above technical means, the embodiments of this application can filter and determine the location of pedestrians related to the vehicle, effectively preventing unimportant pedestrian trajectories from affecting the control of the vehicle's rear end and improving the reliability of pedestrian trajectory prediction.

[0013] Optionally, in one embodiment of this application, the step of predicting at least one pedestrian prediction target point for each pedestrian based on the pedestrian movement state, and generating a predicted intersection trajectory for each pedestrian based on the at least one pedestrian prediction target point, includes: when the pedestrian's movement state is a straight-ahead state, obtaining a predicted endpoint by translating the historical trajectory and the lane line closest to the pedestrian; when the pedestrian's movement state is a turning state, a starting state, or a crossing state, if the pedestrian is not within the zebra crossing expansion area, obtaining the predicted endpoint based on the intersection of the pedestrian's current speed direction and the lane line; if the pedestrian is within the zebra crossing expansion area, comparing the intersection of the pedestrian's current speed direction and the lane line with the intersection of the zebra crossing and the lane line, otherwise obtaining the predicted endpoint based on the intersection of the pedestrian's current speed direction and the lane line.

[0014] Based on the above technical means, the embodiments of this application can obtain the predicted destination according to the pedestrian's movement state, effectively improving the efficiency of pedestrian trajectory prediction, enhancing the applicability and reliability of pedestrian trajectory prediction, and meeting the user's experience.

[0015] Optionally, in one embodiment of this application, the step of predicting at least one pedestrian prediction target point for each pedestrian based on the pedestrian movement state of each pedestrian, and generating the intersection prediction trajectory for each pedestrian based on the at least one pedestrian prediction target point, further includes: fitting the prediction endpoint and the pedestrian historical trajectory points on the historical trajectory into a third-order curve to obtain the intersection prediction trajectory.

[0016] Based on the above technical means, the embodiments of this application can improve the accuracy and stability of intersection prediction trajectory by fitting a third-order curve, improve the applicability of pedestrian trajectory prediction, and effectively enhance the user's driving experience.

[0017] A second aspect of this application provides an intersection pedestrian trajectory prediction device, comprising: an acquisition module for acquiring the current location and historical trajectory of at least one pedestrian, and identifying the movement trend of each pedestrian; an identification module for identifying the predicted movement state of each pedestrian based on the current location, the historical trajectory, and the movement trend; and a generation module for predicting at least one pedestrian prediction target point for each pedestrian based on the pedestrian movement state when the predicted movement state is a turning state, a crossing state, or a starting state, and generating the intersection prediction trajectory of each pedestrian based on the at least one pedestrian prediction target point.

[0018] Optionally, in one embodiment of this application, the acquisition module includes: a first acquisition unit, used for at least one pedestrian coordinate point based on a preset historical time in the vehicle coordinate system; and a fitting unit, used for fitting the historical trajectory of each pedestrian based on the at least one pedestrian coordinate point.

[0019] Optionally, in one embodiment of this application, the acquisition module further includes: a second acquisition unit, used to acquire lane information of the current intersection; and a calculation unit, used to calculate the current position of the pedestrian relative to the road based on the lane information.

[0020] Optionally, in one embodiment of this application, the apparatus of this application embodiment further includes: a filtering module, used to filter pedestrians from all pedestrians who are within a first preset distance from the vehicle and do not meet the requirements of being outside the road or being more than a second preset distance from the road edge, before obtaining the current location and historical trajectory of the at least one pedestrian, to determine the at least one pedestrian.

[0021] Optionally, in one embodiment of this application, the generation module includes: a prediction unit, configured to, when the pedestrian's movement state is a straight-ahead state, obtain a predicted endpoint based on the historical trajectory and the lane line closest to the pedestrian; and a generation unit, configured to, when the pedestrian's movement state is a turning state, a starting state, or a crossing state, obtain the predicted endpoint based on the intersection of the pedestrian's current speed direction and the lane line if the pedestrian is not within the zebra crossing expansion area; if the pedestrian is within the zebra crossing expansion area, compare the intersection of the pedestrian's current speed direction and the lane line with the intersection of the zebra crossing and the lane line; otherwise, obtain the predicted endpoint based on the intersection of the pedestrian's current speed direction and the lane line.

[0022] Optionally, in one embodiment of this application, the generation module is further configured to fit the predicted endpoint to the pedestrian historical trajectory points on the historical trajectory into a third-order curve to obtain the intersection predicted trajectory.

[0023] A third aspect of this application provides a vehicle comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the intersection pedestrian trajectory prediction method as described in the above embodiments.

[0024] A fourth aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for predicting pedestrian trajectories at intersections.

[0025] The beneficial effects of this application are:

[0026] (1) The embodiments of this application can generate the intersection prediction trajectory of each pedestrian based on the pedestrian's current location, historical trajectory and movement trend, effectively adapting to all environments, improving the applicability of pedestrian trajectory prediction, improving the safety of user driving, and effectively meeting the user's vehicle use needs.

[0027] (2) The embodiments of this application can filter and determine the location of pedestrians related to the vehicle, effectively preventing unimportant pedestrian trajectories from affecting the control of the vehicle's rear end and improving the reliability of pedestrian trajectory prediction.

[0028] (3) The embodiments of this application can generate the intersection prediction trajectory of each pedestrian based on the pedestrian's current location, historical trajectory and movement trend, effectively adapting to all environments, improving the applicability and accuracy of pedestrian trajectory prediction, effectively ensuring the safety and reliability of the user's driving, and meeting the user's vehicle use needs.

[0029] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0030] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0031] Figure 1 This is a flowchart of a pedestrian trajectory prediction method at an intersection according to an embodiment of this application;

[0032] Figure 2 A flowchart of a pedestrian trajectory prediction method at an intersection according to a specific embodiment of this application;

[0033] Figure 3 This is a schematic diagram illustrating pedestrian movement state determination according to a specific embodiment of this application;

[0034] Figure 4 This is a schematic diagram illustrating pedestrian state transitions according to a specific embodiment of this application;

[0035] Figure 5 This is a schematic diagram of the structure of a pedestrian trajectory prediction device at an intersection according to an embodiment of this application;

[0036] Figure 6 This is a structural schematic diagram of a vehicle provided according to an embodiment of this application.

[0037] Among them, 10-intersection pedestrian trajectory prediction device; 100-acquisition module, 200-identification module and 300-generation module; 601-memory, 602-processor and 603-communication interface. Detailed Implementation

[0038] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0039] The following description, with reference to the accompanying drawings, describes a method and apparatus for predicting pedestrian trajectories at intersections according to embodiments of this application. Addressing the problems mentioned in the background section regarding the related technologies where numerous neural network parameters and complex model structures lead to training failures and limited applicability to all environments, thus reducing the applicability of pedestrian trajectory prediction, lowering driving safety, and failing to meet user needs, this application provides a method for predicting pedestrian trajectories at intersections. This method identifies the movement state of each pedestrian based on their current location, historical trajectory, and movement trend, predicts the pedestrian prediction target point for each pedestrian, and generates a predicted intersection trajectory for each pedestrian. This effectively adapts to all environments, improves the applicability and accuracy of pedestrian trajectory prediction, effectively ensures driving safety and reliability, and meets user needs. Therefore, this solves the technical problems in the related technologies where numerous neural network parameters and complex model structures lead to training failures and limited applicability to all environments, reducing the applicability of pedestrian trajectory prediction, lowering driving safety, and failing to meet user needs.

[0040] Specifically, Figure 1 This is a flowchart illustrating a method for predicting pedestrian trajectories at intersections, as provided in an embodiment of this application.

[0041] like Figure 1 As shown, the pedestrian trajectory prediction method at this intersection includes the following steps:

[0042] In step S101, the current location and historical trajectory of at least one pedestrian are obtained, and the movement trend of each pedestrian is identified.

[0043] It is understood that the embodiments of this application can obtain the current location and historical trajectory of pedestrians in the following steps, for example, by means of fisheye cameras, lidar, etc., and identify the movement trend of each pedestrian, such as forward or backward, to ensure that vehicles can more accurately predict pedestrian trajectories and reduce the occurrence of traffic accidents.

[0044] Optionally, in one embodiment of this application, obtaining the current location and historical trajectory of at least one pedestrian includes: at least one pedestrian coordinate point based on a preset historical time in the vehicle coordinate system; and fitting the historical trajectory of each pedestrian based on at least one pedestrian coordinate point.

[0045] In actual implementation, such as Figure 2As shown, this embodiment of the application can cache pedestrian coordinate points based on the vehicle's coordinate system for a certain historical period in the pedestrian historical trajectory point caching device, such as pedestrian coordinate points in the historical 3-second period. And based on the pedestrian historical coordinate points, the historical trajectory of each pedestrian is fitted. Since the pedestrian historical coordinate points are based on the historical vehicle coordinate system, it is necessary to convert the pedestrian historical coordinate points to points in the current vehicle coordinate system through a coordinate transformation device, thereby effectively improving the accuracy of each pedestrian historical trajectory and increasing the feasibility of pedestrian trajectory prediction.

[0046] It should be noted that the preset historical duration is set by those skilled in the art based on the actual situation, and no specific limitation is made here.

[0047] Optionally, in one embodiment of this application, obtaining the current location and historical trajectory of at least one pedestrian includes: obtaining lane information at the current intersection; and calculating the pedestrian's current location relative to the road based on the lane information.

[0048] In some embodiments, such as Figure 2 As shown, this application embodiment can obtain lane information of the current intersection through a high-precision map, and calculate the current position of the pedestrian relative to the road based on the lane information. This can accurately determine the current position of the pedestrian, improve the reliability of pedestrian trajectory prediction, enhance vehicle driving safety, and reduce the occurrence of traffic accidents.

[0049] Optionally, in one embodiment of this application, before obtaining the current location and historical trajectory of at least one pedestrian, the method further includes: filtering pedestrians from all pedestrians who are within a first preset distance from the vehicle and do not meet the conditions of being outside the road or being more than a second preset distance from the road edge, thereby determining at least one pedestrian.

[0050] In some embodiments, the present application embodiments can use a pedestrian screening device to predict pedestrians based on screening principles. The screening principles include the target type being a pedestrian, the pedestrian's lateral distance being within a certain distance of the vehicle, and the pedestrian not meeting the requirements of being outside the road and being more than a certain distance from the road edge. This effectively prevents unimportant pedestrian trajectories from affecting the vehicle's rear-end control and improves the reliability of pedestrian trajectory prediction.

[0051] In actual implementation, such as Figure 2 As shown, this application embodiment can filter pedestrians from all pedestrians who are within a certain distance of the vehicle, for example, within 3 meters of the vehicle, and who do not meet the requirements of being outside the road or more than a certain distance from the road edge, such as being inside the road or more than 1 meter from the road edge. This allows for trajectory prediction of pedestrians, thereby narrowing the range of pedestrian trajectory prediction, preventing unimportant pedestrian trajectories from affecting the vehicle's rear-end control, improving the vehicle's automation level, and meeting user needs.

[0052] It should be noted that the first preset distance and the second preset distance are set by those skilled in the art according to the actual situation, and are not specifically limited here.

[0053] In step S102, the predicted motion state of each pedestrian is identified based on their current location, historical trajectory, and motion trend.

[0054] It is understood that the embodiments of this application can identify the predicted movement state of each pedestrian based on the pedestrian's current location, historical trajectory and movement trend. For example, the pedestrian's movement state can be straight, turning, stopping, starting or crossing, etc. It can accurately determine the pedestrian's movement state, improve the accuracy of pedestrian trajectory prediction, improve vehicle driving safety and reduce the occurrence of traffic accidents.

[0055] For example, such as Figure 3 As shown in the embodiment of this application, the movement state of a pedestrian can be determined by a pedestrian movement state determination device, as detailed below:

[0056] (1) Turning: When a pedestrian turns from going straight to the side of the lane, it is considered that the pedestrian has the tendency to cut into the lane.

[0057] (2) Crossing: Pedestrians have entered the lane, including stopping and starting during the crossing process.

[0058] (3) Start: The pedestrian has not fully entered the lane but has a tendency to cut in. The historical status is stopped.

[0059] (4) Stop: The pedestrian has not fully entered the lane but has a history of cutting in, and is currently in a waiting state.

[0060] (5) Going straight: This includes pedestrians stopping, starting, and turning away from the side of the road while going straight.

[0061] In summary, the embodiments of this application, based on the aforementioned pedestrian movement states, can effectively improve the accuracy of pedestrian trajectory prediction, enhance vehicle driving safety, and reduce the occurrence of traffic accidents.

[0062] In step S103, when the predicted motion state is a turning state, a crossing state, or a starting state, at least one predicted target point for each pedestrian is predicted based on the pedestrian motion state of each pedestrian, and the intersection prediction trajectory for each pedestrian is generated based on at least one predicted target point.

[0063] It is understood that, in the embodiments of this application, when the predicted motion state is a turning state, a crossing state, or a starting state, at least one pedestrian prediction target point can be predicted for each pedestrian based on the pedestrian motion state, thereby improving the accuracy of the prediction target. Furthermore, the intersection prediction trajectory of each pedestrian can be generated based on at least one pedestrian prediction target point, which can effectively improve the accuracy of pedestrian trajectory prediction and reduce the occurrence of traffic accidents.

[0064] Optionally, in one embodiment of this application, at least one predicted target point for each pedestrian is predicted based on the pedestrian's movement state, and a predicted intersection trajectory for each pedestrian is generated based on the at least one predicted target point. This includes: when the pedestrian's movement state is straight, the predicted endpoint is obtained by translating the historical trajectory and the lane line closest to the pedestrian; when the pedestrian's movement state is turning, starting, or crossing, if the pedestrian is not within the zebra crossing expansion area, the predicted endpoint is obtained based on the intersection of the pedestrian's current speed direction and the lane line; if the pedestrian is within the zebra crossing expansion area, the predicted endpoint is obtained by comparing the intersection of the pedestrian's current speed direction and the lane line with the intersection of the zebra crossing and the lane line; otherwise, the predicted endpoint is obtained based on the intersection of the pedestrian's current speed direction and the lane line.

[0065] In actual implementation, such as Figure 2 As shown, when the pedestrian is moving in a straight line, the present application embodiment can obtain the predicted endpoint based on the historical trajectory and the lane line closest to the pedestrian, such as the predicted endpoint being the pedestrian's current longitudinal speed * t1s.

[0066] In addition, such as Figure 2 As shown, when a pedestrian is turning, starting, or crossing, the predicted endpoint is determined as follows: if the pedestrian is not within the zebra crossing expansion area, the predicted endpoint is the intersection of the pedestrian's current speed direction and the lane line; if the pedestrian is within the zebra crossing expansion area, the intersection of the pedestrian's current speed direction and the lane line is further away from the vehicle than the intersection of the zebra crossing and the lane line, so the midpoint of the two points is taken; otherwise, the intersection of the pedestrian's current speed direction and the lane line is taken. The predicted endpoint is then processed using Kalman filtering based on historical predicted endpoints to increase the stability of the predicted trajectory. This effectively improves the efficiency, applicability, and reliability of pedestrian trajectory prediction, thus enhancing the user experience.

[0067] Optionally, in one embodiment of this application, predicting at least one pedestrian prediction target point for each pedestrian based on the pedestrian movement state, and generating an intersection prediction trajectory for each pedestrian based on the at least one pedestrian prediction target point, further includes: fitting the predicted endpoint and the pedestrian historical trajectory points on the historical trajectory into a third-order curve to obtain the intersection prediction trajectory.

[0068] In some embodiments, such as Figure 2 As shown, the embodiments of this application can fit the predicted endpoint and the historical pedestrian trajectory points on the historical trajectory into a third-order curve to obtain the intersection prediction trajectory, thereby improving the accuracy and stability of the intersection prediction trajectory, increasing the applicability of pedestrian trajectory prediction, and effectively improving the user's driving experience.

[0069] For example, the following is combined with Figure 2 and Figure 4 The working principle of a specific embodiment of this application will be described in detail below.

[0070] like Figure 4 As shown below, the working principle of pedestrian motion state transition according to a specific embodiment of this application is as follows:

[0071] T0: In this embodiment of the application, pedestrians are assumed to be walking straight when they are present.

[0072] T1: The pedestrian is in a straight-ahead position and has a tendency to cut in, but does not encroach on the lane.

[0073] T2, T5, T7: The pedestrian target has entered this lane (T2 exists to provide fault tolerance and avoid missing the cross-traffic detection after the target distance changes).

[0074] T3, T6: The pedestrian's target speed is less than the first preset threshold and lasts for t2 hours, and the distance traveled during t2 hours is less than the preset distance.

[0075] T4: Pedestrian target speed > second preset threshold and pedestrian target continues to approach within c1(20, TBD) period.

[0076] T8: Pedestrian targets cross to the other side of the road, or return to the roadside after intruding.

[0077] T9: The pedestrian target did not intrude into the lane and did not meet the requirement of having a tendency to cut in.

[0078] It should be noted that the first preset threshold, the second preset threshold, and the preset distance are set by those skilled in the art according to the actual situation, and are not specifically limited here.

[0079] Among them, the method for judging the pedestrian's approach trend is that if the pedestrian target has a period of approach-no approach > c2(50, TBD) within the past 3 seconds, and the pedestrian's movement distance within the past 3 seconds is > d2m, then the pedestrian has an approach trend.

[0080] Specifically, if any one of the three conditions in the following steps exists, then:

[0081] c2 = c2 - n * 10

[0082] Among them, n is the number of three conditions in the following steps.

[0083] Among them, in combination Figure 2 and Figure 4 as shown, the three conditions can be that the zebra crossing traffic light is green, the pedestrian is within the zebra crossing area, and the current road segment is an intersection.

[0084] For example, the embodiment of the present application can obtain the lane traffic light information from the perception device. If any current lane traffic light is red, it is inferred that the zebra crossing traffic light is green, otherwise it is red.

[0085] For another example, the embodiment of the present application can obtain the longitude and latitude points of the zebra crossing from the high-precision map, calculate the zebra crossing coordinate points in the vehicle coordinate system according to the vehicle's longitude and latitude information, and determine whether the pedestrian is within the zebra crossing area according to the zebra crossing coordinate points in the vehicle coordinate system.

[0086] For yet another example, the embodiment of the present application can obtain the intersection information of the lane segment from the high-precision map. If the current lane Link is an intersection, or the next lane Link is an intersection and the distance of the vehicle from the end point of the current lane Link in the current lane Link < d1m, it is considered that the current road segment is an intersection.

[0087] According to the intersection pedestrian trajectory prediction method proposed by the embodiment of the present application, the pedestrian movement state of each pedestrian can be recognized according to the current position, historical trajectory and movement trend of the pedestrian, so as to predict the pedestrian prediction target point of each pedestrian, generate the intersection prediction trajectory of each pedestrian, and then effectively adapt to all environments, improve the applicability and accuracy of pedestrian trajectory prediction, effectively ensure the safety and reliability of user driving, and meet the user's vehicle use requirements. Thus, it solves the technical problems in the related art that due to a large number of neural network parameters and a complex model structure, it is impossible to converge during training, and it is impossible to effectively apply to all environments, reducing the applicability of pedestrian trajectory prediction, with low safety of user driving and unable to meet the user's vehicle use requirements.

[0088] Secondly, a description is given of an intersection pedestrian trajectory prediction device according to an embodiment of the present application with reference to the accompanying drawings.

[0089] Figure 5 is a block diagram of an intersection pedestrian trajectory prediction device according to an embodiment of the present application.

[0090] As Figure 5 shown, the intersection pedestrian trajectory prediction device 10 includes: an acquisition module 100, an identification module 200, and a generation module 300.

[0091] Specifically, the acquisition module 100 is configured to acquire the current position and historical trajectory of at least one pedestrian, and identify the movement trend of each pedestrian.

[0092] The identification module 200 is used to identify the pedestrian movement status of each pedestrian based on their current location, historical trajectory, and movement trend.

[0093] The generation module 300 is used to predict at least one pedestrian prediction target point for each pedestrian based on the pedestrian movement state when the predicted motion state is turning, crossing, or starting, and to generate the intersection prediction trajectory for each pedestrian based on the at least one pedestrian prediction target point.

[0094] Optionally, in one embodiment of this application, the acquisition module 100 includes: a first acquisition unit and a fitting unit.

[0095] The first acquisition unit is used to obtain at least one pedestrian coordinate point based on a preset historical time in the vehicle coordinate system.

[0096] A fitting unit is used to fit the historical trajectory of each pedestrian based on at least one pedestrian coordinate point.

[0097] Optionally, in one embodiment of this application, the acquisition module 100 further includes a second acquisition unit and a calculation unit.

[0098] The second acquisition unit is used to acquire lane information at the current intersection.

[0099] The calculation unit is used to calculate the current position of a pedestrian relative to the road based on lane information.

[0100] Optionally, in one embodiment of this application, the apparatus 10 of this application embodiment further includes a screening module.

[0101] The filtering module is used to filter pedestrians from all pedestrians who are within a first preset distance from the vehicle and do not meet the conditions of being outside the road or more than a second preset distance from the road edge, before obtaining the current location and historical trajectory of at least one pedestrian, thereby identifying at least one pedestrian.

[0102] Optionally, in one embodiment of this application, the generation module 300 includes a prediction unit and a generation unit.

[0103] The prediction unit is used to predict the endpoint based on the historical trajectory and the lane line closest to the pedestrian when the pedestrian is in a straight-ahead state.

[0104] The generation unit is used to determine the predicted endpoint based on the intersection of the pedestrian's current speed direction and the lane line when the pedestrian is in a turning, starting, or crossing state. If the pedestrian is not within the zebra crossing expansion area, the predicted endpoint is obtained by comparing the intersection of the pedestrian's current speed direction and the lane line with the intersection of the zebra crossing and the lane line. Otherwise, the predicted endpoint is obtained based on the intersection of the pedestrian's current speed direction and the lane line.

[0105] Optionally, in one embodiment of this application, the generation module 300 is further configured to fit the predicted endpoint and the pedestrian historical trajectory points on the historical trajectory into a third-order curve to obtain the intersection predicted trajectory.

[0106] It should be noted that the foregoing explanation of the embodiment of the pedestrian trajectory prediction method at intersections also applies to the pedestrian trajectory prediction device at intersections in this embodiment, and will not be repeated here.

[0107] The pedestrian trajectory prediction device for intersections proposed in this application can identify the movement state of each pedestrian based on their current location, historical trajectory, and movement trend, thereby predicting the pedestrian prediction target point for each pedestrian and generating the intersection prediction trajectory for each pedestrian. This effectively adapts to all environments, improving the applicability and accuracy of pedestrian trajectory prediction, effectively ensuring the safety and reliability of user driving, and meeting user needs. Therefore, it solves the technical problem in related technologies where the large number of neural network parameters and complex model structure lead to convergence failure during training, making it unsuitable for all environments, reducing the applicability of pedestrian trajectory prediction, resulting in lower user driving safety, and failing to meet user needs.

[0108] Figure 6 A schematic diagram of the structure of a vehicle provided in an embodiment of this application. The vehicle may include:

[0109] The memory 601, the processor 602, and the computer program stored on the memory 601 and capable of running on the processor 602.

[0110] When the processor 602 executes the program, it implements the intersection pedestrian trajectory prediction method provided in the above embodiments.

[0111] Furthermore, the vehicle also includes:

[0112] Communication interface 603 is used for communication between memory 601 and processor 602.

[0113] The memory 601 is used to store computer programs that can run on the processor 602.

[0114] The memory 601 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0115] If the memory 601, processor 602, and communication interface 603 are implemented independently, then the communication interface 603, memory 601, and processor 602 can be interconnected via a bus to complete communication between them. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, Figure 6 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0116] Optionally, in a specific implementation, if the memory 601, processor 602, and communication interface 603 are integrated on a single chip, then the memory 601, processor 602, and communication interface 603 can communicate with each other through an internal interface.

[0117] The processor 602 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application.

[0118] This embodiment also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for predicting pedestrian trajectories at intersections.

[0119] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0120] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0121] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or N executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0122] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0123] It should be understood that the various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0124] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0125] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0126] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of this application.

Claims

1. A method for intersection pedestrian trajectory prediction, characterized in that, The method comprises the following steps: acquiring a current position and a historical trajectory of each pedestrian, and identifying a motion trend of each pedestrian; identifying a predicted motion state of each pedestrian according to the current position, the historical trajectory and the motion trend; and when the predicted motion state is a turning state, a crossing state or a starting state, predicting at least one predicted target point of each pedestrian based on a pedestrian motion state of each pedestrian, and generating a crossroad predicted trajectory of each pedestrian according to the at least one predicted target point. The step of predicting at least one predicted target point of each pedestrian based on a pedestrian motion state of each pedestrian, and generating a crossroad predicted trajectory of each pedestrian according to the at least one predicted target point comprises: when the pedestrian motion state is a straight-going state, obtaining a predicted end point based on a translation of the historical trajectory and a lane line closest to the pedestrian; when the pedestrian motion state is a turning state, a starting state or a crossing state, if the pedestrian is not in a zebra crossing inflation area, obtaining the predicted end point based on an intersection of a current speed direction of the pedestrian and the lane line; if the pedestrian is in the zebra crossing inflation area, obtaining the predicted end point based on an intersection of the current speed direction of the pedestrian and the lane line compared with an intersection of a zebra crossing and the lane line; otherwise, obtaining the predicted end point based on the intersection of the current speed direction of the pedestrian and the lane line.

2. The method of claim 1, wherein, The step of acquiring a current position and a historical trajectory of each pedestrian comprises: at least one pedestrian coordinate point in a preset historical time length based on a vehicle coordinate system; fitting the historical trajectory of each pedestrian based on the at least one pedestrian coordinate point.

3. The method of claim 1, wherein, The step of acquiring a current position and a historical trajectory of each pedestrian comprises: acquiring lane information of a current crossroad; calculating a current position of the pedestrian relative to a road based on the lane information.

4. The method of claim 1, wherein, Before the step of acquiring a current position and a historical trajectory of each pedestrian, the method further comprises: selecting pedestrians within a first preset distance from the vehicle and not meeting a condition of being outside the road and being greater than a second preset distance from a road edge from all pedestrians to determine the at least one pedestrian.

5. The method of claim 1, wherein, The step of predicting at least one predicted target point of each pedestrian based on a pedestrian motion state of each pedestrian, and generating a crossroad predicted trajectory of each pedestrian according to the at least one predicted target point further comprises: fitting the predicted end point and a pedestrian historical trajectory point on the historical trajectory into a third-order curve to obtain the crossroad predicted trajectory. 6.A crossing pedestrian trajectory prediction device characterized by comprising: The method comprises: an acquiring module, configured to acquire a current position and a historical trajectory of each pedestrian, and identify a motion trend of each pedestrian; an identifying module, configured to identify a predicted motion state of each pedestrian according to the current position, the historical trajectory and the motion trend; and a generating module, configured to, when the predicted motion state is a turning state, a crossing state or a starting state, predict at least one predicted target point of each pedestrian based on a pedestrian motion state of each pedestrian, and generate a crossroad predicted trajectory of each pedestrian according to the at least one predicted target point. The method comprises: when the pedestrian motion state is a straight-going state, a predicted end point is obtained based on a translation of the historical trajectory and a lane line closest to the pedestrian; when the pedestrian motion state is a turning state, a starting state or a crossing state, if the pedestrian is not in a zebra crossing expansion area, the predicted end point is obtained based on an intersection of a current speed direction of the pedestrian and the lane line; if the pedestrian is in the zebra crossing expansion area, the predicted end point is obtained based on an intersection of the current speed direction of the pedestrian and the lane line compared with an intersection of a zebra crossing and the lane line; otherwise, the predicted end point is obtained based on the intersection of the current speed direction of the pedestrian and the lane line.

7. The apparatus of claim 6, wherein, The obtaining module comprises: a first obtaining unit configured to obtain at least one pedestrian coordinate point of a preset historical time length in a vehicle coordinate system; a fitting unit configured to fit a historical trajectory of each pedestrian based on the at least one pedestrian coordinate point.

8. A vehicle characterized by comprising: The method comprises: a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the intersection pedestrian trajectory prediction method according to any one of claims 1-5.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the intersection pedestrian trajectory prediction method according to any one of claims 1-5.

Images