Pose determination method, apparatus, device, and medium

Abstract

Embodiments of the present application provide a pose determination method, device, equipment and medium, specifically comprising: obtaining sensor data of a target object at multiple time points; determining a reference object pose of the target object at any two time points among the multiple time points, the reference object pose being determined based on a pose variable; and solving the pose variable based on a first cost function, the multiple sensor data and the reference object pose, to obtain an estimated object pose of the target object at the any two time points, with a first generation cost being less than a first threshold value as a target. Embodiments of the present application can effectively improve the accuracy of obtaining an object pose.

Description

Technical Field

[0001] This application belongs to the field of state estimation technology, and in particular relates to a pose determination method, apparatus, device and medium. Background Technology

[0002] With the rapid development of science and technology, state estimation is the foundation for autonomous vehicles, robots and other objects to realize core functions such as autonomous navigation, environmental perception, path planning and precise control. Therefore, it is particularly important to accurately estimate the state of autonomous vehicles, robots and other objects. The core goal of state estimation is to accurately obtain the pose of the aforementioned autonomous vehicles, robots and other objects, but it is difficult to accurately obtain the pose of the objects in the existing technology. Summary of the Invention

[0003] This application provides a pose determination method, apparatus, device, and medium that can effectively improve the accuracy of obtaining the pose of an object.

[0004] In a first aspect, embodiments of this application provide a pose determination method, the method comprising:

[0005] Acquire sensor data of the target object at multiple time points;

[0006] Determine the pose of the reference object at any two times among multiple time points. The pose of the reference object is determined based on the pose variables.

[0007] With the goal of the first generation value being less than the first threshold, the pose variables are solved based on the first cost function, multiple sensor data and the pose of the reference object, so as to obtain the estimated pose of the target object at any two time points.

[0008] The first cost function includes a first function for constraining the relative pose error of the target object in the first time period and a second function for constraining the target error between the sensor data and the pre-built map data. The first generation value includes the sum of the relative pose error and the target error, which is used to characterize the error between the estimated object pose and the actual object pose of the target object. The first time period includes any two time periods.

[0009] Secondly, embodiments of this application provide a pose determination device, the device comprising:

[0010] The acquisition module is used to acquire sensor data of the target object at multiple times.

[0011] The determination module is used to determine the pose of the reference object at any two times among multiple times. The pose of the reference object is determined based on the pose variables.

[0012] The solution module is used to solve the pose variables based on the first cost function, multiple sensor data and the pose of the reference object, with the first generation value being less than the first threshold as the objective, and to obtain the estimated pose of the target object at any two time points.

[0013] The first cost function includes a first function for constraining the relative pose error of the target object in the first time period and a second function for constraining the target error between the sensor data and the pre-built map data. The first generation value includes the sum of the relative pose error and the target error, which is used to characterize the error between the estimated object pose and the actual object pose of the target object. The first time period includes any two time periods.

[0014] Thirdly, an electronic device is provided, comprising: a memory for storing computer program instructions; and a processor for reading and executing the computer program instructions stored in the memory to perform the pose determination method provided in the first aspect.

[0015] Fourthly, a computer storage medium is provided, on which computer program instructions are stored, and when the computer program instructions are executed by a processor, the pose determination method provided in the first aspect is implemented.

[0016] Fifthly, a computer program product is provided, comprising a computer program that, when executed by a processor, implements the pose determination method provided in the first aspect.

[0017] In this embodiment, sensor data of the target object at multiple times can be acquired, and reference object poses of the target object at any two times can be determined. The reference object poses are determined based on pose variables. Then, with the goal that the sum of the relative pose error and the target error included in the first generation value is less than a first threshold, the pose variables are solved based on a first function used to constrain the relative pose error of the target object in the first time period, a second function used to constrain the target error between the sensor data and the pre-built map data, multiple sensor data, and the reference object poses. Since the first generation value is used to characterize the error between the estimated object pose and the actual object pose of the target object, the smaller the first generation value, the closer the estimated object pose is to the actual object pose, thereby effectively improving the accuracy of acquiring the object pose. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is one of the flowcharts illustrating a pose determination method provided in an embodiment of this application;

[0020] Figure 2 This is a second schematic flowchart of a pose determination method provided in an embodiment of this application;

[0021] Figure 3 This is a schematic diagram of the structure of a pose determination device provided in an embodiment of this application.

[0022] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0023] The features and exemplary embodiments of various aspects of this application will be described in detail below. To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain this application and not to limit it. For those skilled in the art, this application can be implemented without some of these specific details. The following description of the embodiments is merely to provide a better understanding of this application by illustrating examples.

[0024] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes the element.

[0025] In this article, the term "and / or" is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.

[0026] To address the issue of low accuracy in acquiring object pose in existing technologies, this application provides a pose determination method, apparatus, device, and medium. The pose determination method provided in this application can acquire sensor data of a target object at multiple time points and determine the reference object pose of the target object at any two of these time points. This reference object pose is determined based on pose variables. Furthermore, with the goal that the sum of the relative pose error and the target error included in the first-generation value is less than a first threshold, the pose variables are solved based on a first function constraining the relative pose error of the target object in a first time period, a second function constraining the target error between the sensor data and pre-constructed map data, multiple sensor data, and the reference object pose. Since the first-generation value characterizes the error between the estimated object pose and the actual object pose, the smaller the first-generation value, the closer the estimated object pose is to the actual object pose, thereby effectively improving the accuracy of acquiring the object pose.

[0027] The pose determination method provided in this application will be described in detail below with reference to specific embodiments and accompanying drawings.

[0028] Figure 1 This is one of the flowcharts illustrating a pose determination method provided in the embodiments of this application.

[0029] like Figure 1 As shown, the execution subject of this method can be a pose determination device. Based on this, the pose determination method provided in this application embodiment can specifically include the following steps:

[0030] S110, acquire sensor data of the target object at multiple times.

[0031] The target object mentioned above can be any object requiring pose determination, such as an autonomous vehicle or a robot, and is not specifically limited here. Furthermore, the sensor data mentioned above can be data related to the target object collected using sensors. These sensors can be devices deployed inside the target object to collect pose-related data. In the pose determination method provided in this application embodiment, the sensors can include devices such as lidar, inertial measurement units, and millimeter-wave radar. It should be noted that if the sensor includes lidar, since it can directly measure the distance to the target object and the laser emission angle relative to the target object, and this data can be converted into the target object's coordinate position in the sensor's Cartesian coordinate system, the sensor data can include the target object's coordinate position in the sensor's Cartesian coordinate system. If the sensor includes an inertial measurement unit, the sensor data can include the target object's acceleration and angular velocity. If the sensor includes millimeter-wave radar, the sensor data can include the target object's Doppler velocity, and in addition, the sensor data can also include the target object's coordinate position.

[0032] Additionally, it should be noted that the aforementioned sensors, in addition to lidar, inertial measurement units, and millimeter-wave radar, may also include sensors such as wheel speedometers and altimeters that can acquire data related to state estimation or pose determination; no specific limitations are made here.

[0033] S120, determine the reference object pose of the target object at any two times among multiple times.

[0034] The pose of the aforementioned reference object can be determined based on pose variables. It should be noted that the pose of the reference object can include reference attitude angles and reference positions; correspondingly, the pose variables can include attitude angle variables and position variables. More specifically, the aforementioned reference attitude angles can include roll, pitch, and yaw angles; the aforementioned reference positions can include the object's positions on the X, Y, and Z axes, respectively; and correspondingly, the pose variables can include variables corresponding to each of the aforementioned dimensions, such as roll, pitch, yaw angles, and the object's positions on the X, Y, and Z axes.

[0035] S130, with the goal of the first generation value being less than the first threshold, the pose variables are solved based on the first cost function, multiple sensor data and the pose of the reference object, to obtain the estimated pose of the target object at any two time points.

[0036] The first threshold mentioned above can be preset based on actual experience or circumstances, and is not specifically limited here.

[0037] In some embodiments, the first cost function may include a first function for constraining the relative pose error of the target object in a first time interval and a second function for constraining the target error between sensor data and pre-built map data. The first time interval includes any two moments, and the pre-built map data may be pre-built based on practical experience or circumstances. It should be noted that if the pose determination method provided in this application is applied to an odometer system, the pre-built map data is obtained through lidar data within a certain range of the current time sequence. If it is a positioning system, the prior pre-built global map data of the scene needs to be loaded by initializing the pose. It should also be noted that in this application embodiment, the pose can be initialized as needed. If applied to an odometer system, the pose is generally initialized as an identity matrix, or combined with acceleration to initialize roll and pitch so that the xy plane of the odometer coordinate system is horizontal. If it is a positioning system, it needs to be initialized in a highly reliable manner, such as by combining navigation information or manually providing it; no specific limitations are made here.

[0038] Based on this, the first-generation value mentioned above may include the sum of relative pose error and target error. This first-generation value can be used to characterize the error between the estimated object pose and the actual object pose of the target object. It should be noted that the estimated object pose mentioned above may be the object pose estimated by the pose determination method provided in the embodiments of this application, while the actual object pose mentioned above refers to the actual pose of the target object in reality or real environment.

[0039] Specifically, the pose determination device can acquire sensor data of the target object at multiple times and determine the reference object pose of the target object at any two times among the multiple times. Since the reference object pose is determined based on pose variables, the pose determination device can solve the above pose variables based on the first cost function, multiple sensor data and reference object pose, with the first generation value being less than the first threshold as the objective, thereby obtaining the estimated object pose of the target object at the above two times.

[0040] In this embodiment, sensor data of the target object at multiple times can be acquired, and reference object poses of the target object at any two times can be determined. The reference object poses are determined based on pose variables. Then, with the goal that the sum of the relative pose error and the target error included in the first generation value is less than a first threshold, the pose variables are solved based on a first function used to constrain the relative pose error of the target object in the first time period, a second function used to constrain the target error between the sensor data and the pre-built map data, multiple sensor data, and the reference object poses. Since the first generation value is used to characterize the error between the estimated object pose and the actual object pose of the target object, the smaller the first generation value, the closer the estimated object pose is to the actual object pose, thereby effectively improving the accuracy of acquiring the object pose.

[0041] Since in the above embodiments, the first cost function may include a first function for constraining the relative pose error of the target object in the first time period and a second function for constraining the target error between sensor data and pre-built map data, in order to describe the pose determination method provided by the embodiments of this application in detail, in one embodiment, if the first cost function includes a first function for constraining the relative pose error of the target object in the first time period, then the above S130 may specifically include the following steps:

[0042] Based on data from multiple sensors and the pose of a reference object, the relative pose error of the target object in the first time interval is determined.

[0043] With the goal of the relative pose error being less than the second threshold, the pose variables are solved based on the first function and the relative pose error to obtain the estimated pose of the target object at any two time points.

[0044] In some embodiments, the aforementioned relative pose error can characterize the error between the actual pose change of the target object and the reference pose change within a first time interval. The actual pose change can be the specific change of the target object's pose in a real environment from one moment to another, where one moment can be understood as the start of the first time interval and the other moment as the end of the first time interval; no specific limitation is made here. The reference pose change refers to an expected or theoretical pose change obtained based on the reference poses of the target object at any two moments.

[0045] The second threshold can be a pre-set threshold based on practical experience or circumstances, and is not specifically limited here. It should be noted that the first threshold mentioned above is greater than or equal to the second threshold.

[0046] Specifically, when the first cost function includes a first function for constraining the relative pose error of the target object in the first time period, the pose determination device can determine the relative pose error of the target object in the first time period based on multiple sensor data and the reference object pose of the target object at any two of the above-mentioned times. Then, with the goal of the relative pose error being less than a second threshold, the device can solve the pose variables based on the first function and the relative pose error to obtain the estimated object pose of the target object at any two times.

[0047] It should also be noted that the aforementioned relative pose error may include at least one of relative angle error and relative displacement error. Based on this, when the relative pose error includes relative angle error, the first function may include a first sub-function for constraining the relative angle error; when the relative pose error includes relative displacement error, the first function may include a second sub-function for constraining the relative displacement error.

[0048] In this embodiment, the relative pose error of the target object in the first time interval is obtained by acquiring sensor data and the pose of a reference object. With the goal of the relative pose error being less than a second threshold, the pose variables are solved using a first sub-function for constraining the relative angle error and / or a second sub-function for constraining the relative displacement error, along with the aforementioned relative pose error. This yields the estimated pose of the target object at any two time points. Thus, the smaller the relative pose error, the smaller the error between the actual pose change of the target object and the reference pose change in the first time interval, indicating a more accurate estimated pose and effectively improving the accuracy of the acquired pose.

[0049] Based on this, in one embodiment, if the relative pose error includes a relative angle error, the sensor data may include the angular velocity of the target object collected by the inertial measurement unit, and the reference object pose may include a reference attitude angle.

[0050] Based on this, the steps described above for determining the relative pose error of the target object in the first time interval based on multiple sensor data and the pose of a reference object can specifically include the following steps:

[0051] Based on multiple angular velocities, determine the actual angle increment of the target object in the first time interval;

[0052] Based on the first reference attitude angle of the target object at the first moment and the second reference attitude angle of the target object at the second moment, the reference angle increment of the target object in the first time period is determined.

[0053] In some embodiments, the actual angle increment is the product of the duration of the first time and the target angular velocity, wherein the target angular velocity is determined based on multiple angular velocities. For example, the target angular velocity can be any one of the multiple angular velocities, or any one of the angular velocities corresponding to each moment in the first time, or the angular velocity of the target object at the first moment or the second moment, or the average of the angular velocities of the target object at the first moment and the second moment, without any specific limitation here.

[0054] It should be noted that any two moments mentioned above include the first moment and the second moment, with the first moment being earlier than the second moment. Since the first time mentioned above can include any two moments mentioned above, the first time mentioned above can include the first moment and the second moment. The first time can be the start time of the first time, and the second moment can be the end time of the first time. Alternatively, the first moment and the second moment can be any two different moments within the first time, without any specific limitation here.

[0055] In addition, the aforementioned reference angle increment is the product of the transpose of the first reference attitude angle and the second attitude angle.

[0056] Specifically, the pose determination device can determine the actual angle increment of the target object in the first time interval based on multiple angular velocities, and determine the reference angle increment of the target object in the first time interval based on the first reference attitude angle of the target object at the first moment and the second reference attitude angle of the target object at the second moment. This facilitates the subsequent calculation of the relative angle difference based on the actual angle increment and the reference angle increment.

[0057] Based on this, the above steps, which aim to obtain the estimated pose of the target object at any two time points by solving for the pose variables based on the first function and the relative pose error, with the objective of the relative pose error being less than the second threshold, can specifically include the following steps:

[0058] With the objective of ensuring that the relative angle error determined based on the actual angle increment and the reference angle increment is less than the first sub-threshold, the attitude angle variable is solved based on the first sub-function and the relative angle error to obtain the estimated attitude angle of the target object at the first moment and the estimated attitude angle of the target object at the second moment.

[0059] The first sub-threshold mentioned above can be preset based on actual experience or circumstances, and this first sub-threshold can be less than or equal to the second threshold mentioned above, without specific limitations. In addition, the pose variables mentioned above can include attitude angle variables.

[0060] Specifically, after obtaining the actual angle increment of the target object in the first time period and the reference angle increment of the target object in the first time period, the pose determination device can solve the attitude angle variable based on the first sub-function and the relative angle error with the objective that the relative angle error determined based on the actual angle increment and the reference angle increment is less than the first sub-threshold, and obtain the estimated attitude angle of the target object in the first time period and the estimated attitude angle of the target object in the second time period.

[0061] In one example, the first sub-function mentioned above can be represented by the following formula (1):

[0062]

[0063] Among them, R b For the target object at the first time t b The first reference attitude angle, Let R be the transpose matrix of the first reference attitude angle. e For the target object at the second time t e The second reference attitude angle, ω, is the target angular velocity determined based on the angular velocity collected by the inertial measurement unit, (t e -t b ) represents the duration of the first time interval, and f1 is a function used to estimate the error between the actual angle increment and the reference angle increment in the roll, pitch, and yaw directions.

[0064] In this embodiment, the actual angle increment and reference angle increment of the target object in the first time interval can be determined, and the pose variables can be solved with the objective that the relative angle error determined based on the actual angle increment and the reference angle increment is less than a first sub-threshold. This yields the estimated pose angles of the target object at the first and second time intervals. Thus, the smaller the relative angle error, the smaller the error between the actual angle increment and the reference angle increment of the target object in the first time interval, and the more accurate the estimated pose angle, thereby effectively improving the accuracy of obtaining the object's pose.

[0065] In another embodiment, if the relative pose error includes the relative displacement error, the aforementioned sensor data may include the Doppler velocity of the target object acquired by the millimeter-wave radar and the direction vector corresponding to the Doppler velocity, and the aforementioned reference object pose may include a reference attitude angle and a reference position.

[0066] Based on this, the steps described above for determining the relative pose error of the target object in the first time interval based on multiple sensor data and the pose of a reference object can specifically include the following steps:

[0067] Based on multiple Doppler velocities, determine the actual displacement change of the target object in the first time interval;

[0068] Based on the poses of the first and second reference objects, the change in reference displacement of the target object within the first time interval is determined.

[0069] In some embodiments, the actual pose change is the product of the duration of a first time and the target velocity, wherein the target velocity is determined based on a plurality of Doppler velocities and the direction vector corresponding to each Doppler velocity.

[0070] Furthermore, the reference object pose of the target object at any two time points can include a first reference object pose at a first time point and a second reference object pose at a second time point, wherein the first time point is earlier than the second time point. Since the aforementioned reference object pose includes a reference position and a reference attitude angle, the aforementioned first reference object pose can include a first reference position and a first reference attitude angle, and the aforementioned second reference object pose can include a second reference position and a second reference attitude angle. Based on this, the aforementioned change in reference pose is the product of the difference between the second reference position and the first reference position and the first reference attitude angle.

[0071] Specifically, the pose determination device can determine the actual displacement change of the target object within a first time interval based on multiple Doppler velocities, and can determine the reference displacement change of the target object within the first time interval based on the poses of a first reference object and a second reference object. This allows the relative displacement difference of the target object within the first time interval to be subsequently obtained based on the actual displacement change and the reference displacement change.

[0072] Based on this, the above steps, which aim to obtain the estimated pose of the target object at any two time points by solving for the pose variables based on the first function and the relative pose error, with the objective of the relative pose error being less than the second threshold, can specifically include the following steps:

[0073] With the objective of ensuring that the relative displacement error determined based on the actual displacement change and the reference displacement change is less than the second sub-threshold, the pose variables are solved based on the second sub-function and the relative displacement error to obtain the estimated pose of the target object at the first moment and the estimated position of the target object at the second moment.

[0074] Furthermore, the aforementioned second sub-threshold can be preset based on practical experience or circumstances. This second sub-threshold can be less than or equal to the second threshold, and no specific limitation is made here. In some embodiments, the aforementioned pose variables may include position variables and attitude angle variables. Correspondingly, the aforementioned estimated object pose includes estimated position and estimated attitude angle.

[0075] Specifically, after obtaining the actual displacement change of the target object in the first time period and the reference displacement change of the target object in the first time period, the pose determination device can solve the pose variables based on the second sub-function and the relative displacement error with the objective that the relative displacement error determined based on the actual displacement change and the reference displacement change is less than the second sub-threshold, and obtain the estimated pose of the target object in the first time period and the estimated position of the target object in the second time period.

[0076] In one example, the second sub-function mentioned above can be represented by the following formula (2):

[0077]

[0078] in, T is the transpose matrix of the first reference attitude angle. e As the second reference position, T b As the first reference position, (t) e -t b ) represents the duration of the first moment, and V represents the target speed.

[0079] In this embodiment, by determining the actual displacement change and the reference angular displacement change of the target object within a first time interval, and with the objective that the relative displacement error determined based on the actual displacement change and the reference displacement change is less than a second sub-threshold, the pose variables can be solved to obtain the estimated pose of the target object at the first moment and the estimated position of the target object at the second moment. Thus, the smaller the relative displacement error, the smaller the error between the actual displacement change and the reference displacement change of the target object within the first time interval, and the more accurate the estimated pose, thereby effectively improving the accuracy of obtaining the object pose.

[0080] It should also be noted that, since the target velocity is determined based on multiple Doppler velocities and the direction vector corresponding to each Doppler velocity in the above embodiments, in order to accurately and comprehensively describe the pose determination method provided in the embodiments of this application, in one embodiment, the step of determining the actual displacement change of the target object in the first time interval based on multiple Doppler velocities may specifically include the following steps:

[0081] Determine the object velocity corresponding to each Doppler velocity;

[0082] The velocity cost is obtained by summing the difference between the object velocity and the Doppler velocity corresponding to each Doppler velocity. The velocity variable is solved with the goal of the velocity cost being less than the velocity threshold, and the target velocity is obtained.

[0083] The product of the duration of the first time step and the target velocity is determined as the actual displacement change.

[0084] The aforementioned object velocity is the product of the ratio of the direction vector corresponding to the Doppler velocity to the magnitude of the direction vector, and the velocity variable. Furthermore, the aforementioned velocity threshold can be preset based on practical experience or circumstances, and is not specifically limited here.

[0085] Specifically, the pose determination device can first determine the object velocity corresponding to each Doppler velocity, and then accumulate the difference between the object velocity and the Doppler velocity corresponding to each Doppler velocity to obtain the velocity substitution value. With the velocity substitution value being less than the velocity threshold as the objective, the velocity variable is solved to obtain the target velocity. Thus, after obtaining the target velocity, the product of the duration of the first time and the target velocity can be determined as the actual displacement change.

[0086] In one example, the above-mentioned accumulation of the difference between the object velocity and the Doppler velocity corresponding to each Doppler velocity yields the velocity cost. The velocity variable is solved to obtain the target velocity with the goal of the velocity cost being less than the velocity threshold. This process satisfies the following formula (3):

[0087]

[0088] Among them, J d (V) is the velocity constraint function used to constrain the velocity cost, which represents the difference between the estimated velocity (i.e., the target velocity) and the actual velocity of the target object. Here, V is the velocity variable, and its solved value is the target velocity. di That is, the i-th Doppler velocity acquired by the millimeter-wave radar, m i Let norm(m) be the direction vector of the i-th Doppler velocity. i ) represents the magnitude of the direction vector of the i-th Doppler velocity, and n represents the number of Doppler velocities.

[0089] In this embodiment, a velocity constraint function can be introduced based on sensor data collected by millimeter-wave radar. The target velocity of the target object can be accurately obtained through the velocity constraint function, which facilitates the accurate calculation of the actual displacement change of the target object. This constrains the relative displacement difference between the actual displacement change and the reference displacement change, and accurately obtains the estimated pose of the target object. At the same time, it can effectively overcome the problems of degradation or failure of tunnel state estimation based on pure laser schemes, and instability or failure of state estimation in rainy and foggy weather, and realize true all-weather state estimation.

[0090] In the above embodiments, the first cost function may include, in addition to a first function for constraining the relative pose error of the target object in the first time interval, a second function for constraining the target error between the sensor data and the pre-built map data. Therefore, to accurately and comprehensively describe the pose determination method provided in this application, in one embodiment, the sensor data may include measurement points collected by a lidar, which are the coordinate positions of the target object in the sensor's Cartesian coordinate system. Based on this, when the first cost function includes the second function, the steps described above—aiming for a first-generation value less than a first threshold, solving for pose variables based on the first cost function, multiple sensor data, and the pose of a reference object, to obtain the estimated object pose at any two time intervals—may specifically include the following steps:

[0091] According to the preset time sequence, the interpolation start and end points are respectively taken as the reference object poses of the target object at any two times, and the interpolation is performed according to the preset interpolation ratio to obtain the poses of multiple third reference objects;

[0092] Based on the measurement point, the pose of the third reference object corresponding to the measurement point, and the pre-constructed map data, calculate the distance between the measurement point and the nearest point on the map along the normal direction.

[0093] With the target distance being less than the third threshold as the objective, the pose variables are solved based on the second function and the target distance to obtain the estimated pose of the target object at any two time points.

[0094] The preset time order can be pre-set based on practical experience or circumstances; for example, it can be in order from morning to night, without specific limitations. Similarly, the preset interpolation ratio can also be pre-set based on practical experience or circumstances, without specific limitations. The nearest point on the map is the point in the map data closest to the measurement point, and the normal direction is the direction of the normal at the nearest point on the map.

[0095] Furthermore, the aforementioned target distance is the sum of the distances corresponding to the poses of multiple third reference objects. The aforementioned third threshold can be preset based on practical experience or circumstances; for example, the third threshold can be less than or equal to the aforementioned first threshold, without specific limitations here.

[0096] Specifically, when the first cost function includes the second function, the pose determination device can first interpolate according to a preset time sequence, using the reference object poses of the target object at any two times as the interpolation start point and interpolation end point, and according to a preset interpolation ratio to obtain multiple third reference object poses. Then, based on the measurement point, the third reference object poses corresponding to the measurement point, and the pre-constructed map data, it can calculate the distance between the measurement point and the nearest point on the map along the normal direction. Since multiple measurement points can be obtained in this embodiment, multiple distances can be obtained. Then, based on the target distance obtained from the multiple distances being less than a third threshold as the target, the pose variables can be solved based on the second function and the target distance to obtain the estimated object pose of the target object at any two times.

[0097] In one example, the above preset interpolation ratio can be represented by the following formula (4):

[0098]

[0099] Among them, t b For the first moment, t e For the second time step, t i It is the i-th moment within the time interval from the first moment to the second moment.

[0100] Based on the above preset interpolation ratio, the pose of the i-th third reference object obtained by interpolation can be expressed by the following formula:

[0101]

[0102] T i =(1-s i )T b +s i T e (6)

[0103] Among them, R i Let T be the reference attitude angle of the target object at time i. i Let R be the reference position of the target object at time i. i T i Together they form the pose of the third reference object of the target object at time i.

[0104] Based on this, the second function mentioned above can be represented by the following formula (7):

[0105]

[0106] Where, p i Let q be the i-th measurement point acquired by the lidar. i To search for p in pre-built map data iThe nearest point on the map, N is the nearest point q on the map. i Point normal. It should be noted that the second function mentioned above is constructed based on the cost function from point to surface in the Iterative Closest Point (icp). The difference is that the pose of point p is given by the interpolation of the state at the start and end times. In this way, n cost functions for 12-dimensional pose are constructed with n observation points.

[0107] In this embodiment, the pose of the reference object at multiple times can be obtained by interpolation. Based on the pose of the reference object at multiple times, the measurement points collected by the lidar, and the pre-built map data, the distance between each measurement point and the nearest point on the map along the normal direction can be determined. Then, the target distance can be obtained by accumulating the multiple distances. By constraining the target distance, the estimated pose of the target object at any two times can be obtained. Thus, the smaller the target distance, the more accurate the estimated pose is, thereby effectively improving the accuracy of the estimated pose.

[0108] Furthermore, considering that if the measurement frequency of the lidar is high enough, the pose angle of the target object at the end of the previous time period should be almost the same as the pose of the target object at the beginning of the current time period in two adjacent time periods. Based on this, in order to describe the pose determination method provided in the embodiments of this application more completely and in detail, in one embodiment, the first cost function may further include a third constraint function for constraining the rotational change of the target object; based on this, such as... Figure 2 As shown, the pose determination method provided in this application embodiment may further include the following steps:

[0109] S210, obtain the third attitude angle of the target object at the third time.

[0110] The third time point is earlier than any two times points. It should be noted that the third time point can be the end time of the second time point, and the second time point can be a period of time before the first time point. No specific limitation is made here.

[0111] S220, determine the product of the transpose matrix of the third attitude angle and the first reference attitude angle, which is the rotational change of the target object.

[0112] In some embodiments, the first reference attitude angle is the reference attitude angle of the target object at a first moment, wherein the first moment is the earliest of any two moments.

[0113] Specifically, when the first cost function also includes a third constraint function for constraining the rotational change of the target object, the pose determination device can obtain the third attitude angle of the target object at the third time, and then determine the product of the transpose matrix of the determined third attitude angle and the first reference attitude angle as the rotational change of the target object.

[0114] Based on this, the above-mentioned S130 may specifically include the following steps:

[0115] S1301, with the goal of the second-generation value being less than the fourth threshold, the pose variables are solved based on the first cost function, multiple sensor data and the pose of the reference object, to obtain the estimated pose of the target object at any two time points.

[0116] The fourth threshold can be preset based on practical experience or circumstances, and is not specifically limited here. It should be noted that the fourth threshold can be greater than or equal to the first threshold mentioned above.

[0117] In some embodiments, the second-generation value may include the sum of relative pose error, target error, and rotational change. This second-generation value can be used to characterize the error between the estimated object pose and the actual object pose of the target object.

[0118] Specifically, given the rotational change of the target object, the pose determination device can, with the goal that the sum of the relative pose error, the target error, and the rotational change is less than a fourth threshold, solve the pose variables based on the first cost function, multiple sensor data, and the pose of the reference object, and obtain multiple estimated poses of the target object at any two time points.

[0119] In one embodiment, the third function described above can be represented by the following formula (8):

[0120]

[0121] in, The transpose matrix of the target object's third attitude angle at the third time step. f1 represents the first reference attitude angle of the target object at the first moment, and f2 represents the function that converts the rotational change into roll, pitch, and yaw direction errors.

[0122] In this embodiment, based on the use of a first function to constrain the relative pose error of the target object in the first time period and a second function to constrain the target error between sensor data and pre-built map data for pose estimation, considering that the high acquisition frequency of lidar in real life will lead to the pose approximation of two close moments, a function to constrain the rotational change of the target object is introduced. In this way, the estimated pose of the target object can be calculated more accurately and effectively, thus improving the accuracy of the estimated pose acquisition.

[0123] Based on the same inventive concept, embodiments of this application provide an apparatus for a pose determination method, which can be specifically combined with... Figure 3 The pose determination device provided in the embodiments of this application will be described in detail.

[0124] Figure 3 This is a schematic diagram of the structure of a pose determination device provided in an embodiment of this application.

[0125] like Figure 3 As shown, the pose determination device 600 may include:

[0126] The acquisition module 310 is used to acquire sensor data of the target object at multiple times.

[0127] The determination module 320 is used to determine the pose of the reference object at any two times among multiple times, and the pose of the reference object is determined based on the pose variables;

[0128] The solver module 330 is used to solve the pose variables based on the first cost function, multiple sensor data and the pose of the reference object, with the first generation value being less than the first threshold as the objective, and to obtain the estimated pose of the target object at any two time points.

[0129] The first cost function includes a first function for constraining the relative pose error of the target object in the first time period and a second function for constraining the target error between the sensor data and the pre-built map data. The first generation value includes the sum of the relative pose error and the target error, which is used to characterize the error between the estimated object pose and the actual object pose of the target object. The first time period includes any two time periods.

[0130] In one embodiment, the first cost function includes a first function. Based on this, the determination module is further configured to determine the relative pose error of the target object in a first time period based on multiple sensor data and the pose of a reference object. The relative pose error characterizes the error between the actual pose change of the target object and the reference pose change in the first time period.

[0131] The solution module is specifically used to solve the pose variables based on the first function and the relative pose error with the objective of the relative pose error being less than the second threshold, so as to obtain the estimated pose of the target object at any two time points.

[0132] Wherein, when the relative pose error includes the relative angle error, the first function includes a first sub-function for constraining the relative angle error; when the relative pose error includes the relative displacement error, the first function includes a second sub-function for constraining the relative displacement error.

[0133] In one embodiment, the relative pose error includes a relative angle error, the sensor data includes the angular velocity of the target object acquired by the inertial measurement unit, and the reference object pose includes a reference attitude angle; based on this, the aforementioned determining module is specifically used for:

[0134] Based on multiple angular velocities, the actual angle increment of the target object in the first time interval is determined. The actual angle increment is the product of the duration of the first time interval and the target angular velocity. The target angular velocity is determined based on multiple angular velocities.

[0135] Based on the first reference attitude angle of the target object at the first time and the second reference attitude angle of the target object at the second time, the reference angle increment of the target object in the first time is determined. Any two times include the first time and the second time, the first time is earlier than the second time, and the reference angle increment is the product of the transpose matrix of the first reference attitude angle and the second attitude angle.

[0136] The above-mentioned solution module is specifically used for:

[0137] With the objective of ensuring that the relative angle error determined based on the actual angle increment and the reference angle increment is less than the first sub-threshold, the attitude angle variables are solved based on the first sub-function and the relative angle error to obtain the estimated attitude angle of the target object at the first moment and the estimated attitude angle of the target object at the second moment. The pose variables include the attitude angle variables.

[0138] In one embodiment, the relative pose error includes the relative displacement error, the sensor data includes the Doppler velocity of the target object acquired by the millimeter-wave radar and the direction vector corresponding to the Doppler velocity, and the reference object pose includes the reference attitude angle and the reference position.

[0139] Based on this, the aforementioned determining module is specifically used for:

[0140] Based on multiple Doppler velocities, the actual displacement change of the target object in the first time interval is determined. The actual pose change is the product of the duration of the first time interval and the target velocity. The target velocity is determined based on multiple Doppler velocities and the direction vector corresponding to each Doppler velocity.

[0141] Based on the poses of the first and second reference objects, the change in reference displacement of the target object in the first time interval is determined. The poses of the target object at any two time intervals include the first reference object pose of the target object at the first time interval and the second reference object pose of the target object at the second time interval. The first time interval is earlier than the second time interval. The pose of the reference object includes the reference position and the reference attitude angle. The change in reference pose is the product of the difference between the second reference position and the first reference position and the first reference attitude angle.

[0142] The above-mentioned solution module is specifically used for:

[0143] With the objective of ensuring that the relative displacement error determined based on the actual displacement change and the reference displacement change is less than the second sub-threshold, the pose variables are solved based on the second sub-function and the relative displacement error to obtain the estimated pose of the target object at the first moment and the estimated position of the target object at the second moment. The pose variables include position variables and attitude angle variables, and the estimated pose of the target object includes the estimated position and the estimated attitude angle.

[0144] In one embodiment, the aforementioned determining module is specifically used for:

[0145] Determine the object velocity corresponding to each Doppler velocity. The object velocity is the product of the ratio of the direction vector corresponding to the Doppler velocity to the magnitude of the direction vector and the velocity variable.

[0146] The velocity cost is obtained by summing the difference between the object velocity and the Doppler velocity corresponding to each Doppler velocity. The velocity variable is solved with the goal of the velocity cost being less than the velocity threshold, and the target velocity is obtained.

[0147] The product of the duration of the first time step and the target velocity is determined as the actual displacement change.

[0148] In one embodiment, the sensor data includes measurement points acquired by a lidar; the first cost function includes a second function. Based on this, the pose determination device provided in this application embodiment may further include:

[0149] The interpolation module is used to interpolate according to a preset time sequence, taking the poses of the target object at any two times as the interpolation start and end points, and according to a preset interpolation ratio to obtain the poses of multiple third reference objects.

[0150] The calculation module is used to calculate the distance between the measurement point and the nearest point on the map along the normal direction based on the measurement point, the pose of the third reference object corresponding to the measurement point, and the pre-built map data. The nearest point on the map is the point in the map data that is closest to the measurement point, and the normal direction is the direction of the normal at the nearest point on the map.

[0151] The solution module is specifically used to solve for the pose variables based on the second function and the target distance, which is the sum of the distances corresponding to the poses of multiple third reference objects, with the target distance being less than the third threshold. The estimated pose of the target object is obtained at any two time points.

[0152] In one embodiment, the first cost function further includes a third constraint function for constraining the rotational change of the target object; based on this, the above-mentioned acquisition module is also used to acquire the third attitude angle of the target object at the third time moment, the third time moment being earlier than any two time moments.

[0153] The determination module is also used to determine the product of the transpose matrix of the third attitude angle and the first reference attitude angle, which is the rotational change of the target object. The first reference attitude angle is the reference attitude angle of the target object at the first moment, and the first moment is the earliest of any two moments.

[0154] Based on this, the above-mentioned solution module is specifically used for:

[0155] With the goal of the second-generation value being less than the fourth threshold, the pose variables are solved based on the first cost function, multiple sensor data and the pose of the reference object, and the estimated pose of the target object at any two time points is obtained. The second-generation value includes the sum of relative pose error, target error and rotation change.

[0156] In this embodiment, sensor data of the target object at multiple times can be acquired, and reference object poses of the target object at any two times can be determined. The reference object poses are determined based on pose variables. Then, with the goal that the sum of the relative pose error and the target error included in the first generation value is less than a first threshold, the pose variables are solved based on a first function used to constrain the relative pose error of the target object in the first time period, a second function used to constrain the target error between the sensor data and the pre-built map data, multiple sensor data, and the reference object poses. Since the first generation value is used to characterize the error between the estimated object pose and the actual object pose of the target object, the smaller the first generation value, the closer the estimated object pose is to the actual object pose, thereby effectively improving the accuracy of acquiring the object pose.

[0157] The pose determination device provided in this application embodiment can achieve the following: Figure 1 or Figure 2 The method steps of the illustrated embodiment, and the corresponding technical effects they achieve, are not described in detail here.

[0158] Figure 4 A schematic diagram of the hardware structure of the electronic device provided in an embodiment of this application is shown.

[0159] An electronic device may include a processor 401 and a memory 402 storing computer program instructions.

[0160] Specifically, the processor 401 may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0161] Memory 402 may include mass storage for data or instructions. For example, and not limitingly, memory 402 may include a hard disk drive (HDD), floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, memory 402 may include removable or non-removable (or fixed) media. Where appropriate, memory 402 may be internal or external to an electronic device. In a particular embodiment, memory 402 is a non-volatile solid-state memory.

[0162] Memory may include read-only memory (ROM), random access memory (RAM), disk storage media devices, optical storage media devices, flash memory devices, and electrical, optical, or other physical / tangible memory storage devices. Therefore, typically, memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., memory devices) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the methods according to one aspect of this disclosure.

[0163] The processor 401 reads and executes computer program instructions stored in the memory 402 to implement any of the pose determination methods in the above embodiments.

[0164] In one example, the electronic device may also include a communication interface 403 and a bus 410. For example, Figure 4 As shown, the processor 401, memory 402, and communication interface 403 are connected through bus 410 and complete communication with each other.

[0165] The communication interface 403 is mainly used to realize communication between various modules, devices, units and / or equipment in the embodiments of this application.

[0166] Bus 410 includes hardware, software, or both, that couples components of an electronic device together. For example, and not limitingly, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, bus 410 may include one or more buses. Although specific buses are described and illustrated in embodiments of this application, this application contemplates any suitable bus or interconnect.

[0167] In addition, in conjunction with the pose determination method in the above embodiments, this application embodiment can provide a computer storage medium for implementation. The computer storage medium stores computer program instructions; when these computer program instructions are executed by a processor, they implement the pose determination method provided in this application embodiment.

[0168] This application also provides a computer program product, which includes a computer program that is executed by a processor to implement the pose determination method provided in this application embodiment.

[0169] It should be clarified that this application is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of this application is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of this application.

[0170] It should also be noted that the exemplary embodiments mentioned in this application describe methods or systems based on a series of steps or apparatus. However, this application is not limited to the order of the above steps; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0171] The above are merely specific embodiments of this application. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the protection scope of this application.

Claims

1. A pose determination method, characterized in that, The method includes: Acquire sensor data of the target object at multiple time points; Determine the reference object pose of the target object at any two times during the plurality of times, wherein the reference object pose is determined based on pose variables; With the goal of the first generation value being less than a first threshold, the pose variables are solved based on the first cost function, multiple sensor data, and the pose of the reference object to obtain the estimated object pose of the target object at any two time points. The first cost function includes a first function for constraining the relative pose error of the target object in a first time period and a second function for constraining the target error between the sensor data and the pre-built map data. The first generation value includes the sum of the relative pose error and the target error, and is used to characterize the error between the estimated object pose of the target object and the actual object pose of the target object. The first time period includes any two time periods.

2. The method according to claim 1, characterized in that, When the first cost function includes the first function, the step of solving the pose variable based on the first cost function, multiple sensor data, and the pose of the reference object, with the objective of the first generation value being less than the first threshold, to obtain the estimated object pose of the target object at any two time points, includes: Based on multiple sensor data and the pose of the reference object, the relative pose error of the target object is determined in the first time period. The relative pose error represents the error between the actual pose change of the target object and the reference pose change in the first time period. With the goal of the relative pose error being less than the second threshold, the pose variable is solved based on the first function and the relative pose error to obtain the estimated pose of the target object at any two time points. Wherein, when the relative pose error includes a relative angle error, the first function includes a first sub-function for constraining the relative angle error; when the relative pose error includes a relative displacement error, the first function includes a second sub-function for constraining the relative displacement error.

3. The method according to claim 2, characterized in that, The relative pose error includes the relative angle error, the sensor data includes the angular velocity of the target object collected by the inertial measurement unit, and the reference object pose includes the reference attitude angle; The step of determining the relative pose error of the target object within a first time period based on the multiple sensor data and the pose of the reference object includes: Based on multiple angular velocities, the actual angle increment of the target object within the first time interval is determined. The actual angle increment is the product of the duration of the first time interval and the target angular velocity. The target angular velocity is determined based on the multiple angular velocities. Based on the first reference attitude angle of the target object at a first time and the second reference attitude angle of the target object at a second time, the reference angle increment of the target object in the first time interval is determined. The two times include the first time and the second time, the first time is earlier than the second time, and the reference angle increment is the product of the transpose matrix of the first reference attitude angle and the second reference attitude angle. The step of obtaining the estimated object pose of the target object at any two time points, with the goal of the relative pose error being less than a second threshold, involves solving the pose variable based on a first function and the relative pose error, and then performing the following steps: With the objective of ensuring that the relative angle error determined based on the actual angle increment and the reference angle increment is less than a first sub-threshold, the attitude angle variable is solved based on the first sub-function and the relative angle error to obtain the estimated attitude angle of the target object at the first moment and the estimated attitude angle of the target object at the second moment. The pose variable includes the attitude angle variable.

4. The method according to claim 2 or 3, characterized in that, The relative pose error includes the relative displacement error, the sensor data includes the Doppler velocity of the target object acquired by the millimeter-wave radar and the direction vector corresponding to the Doppler velocity, and the reference object pose includes the reference attitude angle and the reference position. The step of determining the relative pose error of the target object within a first time period based on multiple sensor data and the pose of the reference object includes: Based on multiple Doppler velocities, the actual displacement change of the target object in the first time interval is determined. The actual pose change is the product of the duration of the first time interval and the target velocity. The target velocity is determined based on the multiple Doppler velocities and the direction vector corresponding to each Doppler velocity. Based on the poses of the first and second reference objects, the reference displacement change of the target object in the first time interval is determined. The reference object poses of the target object at any two times include the first reference object pose of the target object at the first time and the second reference object pose of the target object at the second time. The first time is earlier than the second time. The reference object pose includes a reference position and a reference attitude angle. The reference pose change is the product of the difference between the second reference position and the first reference position and the first reference attitude angle. The step of obtaining the estimated object pose of the target object at any two time points, with the goal of the relative pose error being less than a second threshold, involves solving the pose variable based on a first function and the relative pose error, and then performing the following steps: With the objective of ensuring that the relative displacement error determined based on the actual displacement change and the reference displacement change is less than a second sub-threshold, the pose variables are solved based on the second sub-function and the relative displacement error to obtain the estimated pose of the target object at the first time and the estimated position of the target object at the second time. The pose variables include position variables and attitude angle variables, and the estimated pose of the target object includes the estimated position and the estimated attitude angle.

5. The method according to claim 4, characterized in that, The determination of the actual displacement change of the target object within the first time interval based on multiple Doppler velocities includes: Determine the object velocity corresponding to each Doppler velocity, wherein the object velocity is the product of the ratio of the direction vector corresponding to the Doppler velocity to the magnitude of the direction vector and the velocity variable; The velocity value is obtained by summing the difference between the object velocity and the Doppler velocity corresponding to each Doppler velocity. The velocity variable is solved with the goal of the velocity value being less than the velocity threshold to obtain the target velocity. The product of the duration of the first time and the target velocity is determined as the actual displacement change.

6. The method according to claim 1, characterized in that, The sensor data includes measurement points acquired by the lidar; when the first cost function includes a second function, the step of solving the pose variables based on the first cost function, multiple sensor data, and the pose of the reference object, with the objective of the first generation value being less than a first threshold, to obtain the estimated object pose of the target object at any two time points, includes: According to a preset time sequence, the interpolation start point and interpolation end point are respectively taken as the reference object poses of the target object at any two times, and interpolation is performed according to a preset interpolation ratio to obtain the poses of multiple third reference objects; Based on the measurement point, the pose of the third reference object corresponding to the measurement point, and the pre-constructed map data, the distance between the measurement point and the nearest point on the map along the normal direction is calculated. The nearest point on the map is the point in the map data that is closest to the measurement point, and the normal direction is the direction of the normal at the nearest point on the map. With the target distance being less than a third threshold as the objective, the target distance being the sum of the distances corresponding to the poses of the plurality of third reference objects, the pose variables are solved based on the second function and the target distance to obtain the estimated object poses of the target object at any two time points.

7. The method according to claim 1, characterized in that, The first cost function further includes a third constraint function for constraining the rotational change of the target object; the method further includes: Obtain the third attitude angle of the target object at the third time point, wherein the third time point is earlier than any two of the time points; The product of the transpose matrix of the third attitude angle and the first reference attitude angle is determined to be the rotation change of the target object. The first reference attitude angle is the reference attitude angle of the target object at the first moment, and the first moment is the earliest of any two moments. The step of obtaining the estimated object pose of the target object at any two time points, with the first generation value being less than a first threshold as the objective, involves solving the pose variables based on a first cost function, multiple sensor data, and the pose of the reference object, and then including: With the goal of the second-generation value being less than the fourth threshold, the pose variables are solved based on the first cost function, multiple sensor data, and the pose of the reference object to obtain the estimated pose of the target object at any two time points. The second-generation value includes the sum of the relative pose error, the target error, and the rotation change.

8. A pose determination device, characterized in that, The device includes: The acquisition module is used to acquire sensor data of the target object at multiple times. The determination module is used to determine the reference object pose of the target object at any two times among the plurality of times, wherein the reference object pose is determined based on pose variables; The solution module is used to solve the pose variables based on a first cost function, multiple sensor data and the pose of the reference object, with the first generation value being less than a first threshold as the objective, to obtain the estimated object pose of the target object at any two time points. The first cost function includes a first function for constraining the relative pose error of the target object in a first time period and a second function for constraining the target error between the sensor data and the pre-built map data. The first generation value includes the sum of the relative pose error and the target error, and is used to characterize the error between the estimated object pose of the target object and the actual object pose of the target object. The first time period includes any two time periods.

9. An electronic device, characterized in that, The device includes: a processor and a memory storing computer program instructions; The processor reads and executes the computer program instructions to implement the pose determination method as described in any one of claims 1-7.

10. A computer storage medium, characterized in that, The computer storage medium stores computer program instructions, which, when executed by a processor, implement the pose determination method as described in any one of claims 1-7.

Images