A mapping method, device, equipment and storage medium

Abstract

The present application provides a mapping method, apparatus, device and storage medium, which are applied to a mobile device, wherein the mobile device is equipped with an observation device, and the observation device includes at least a radar, and the method includes: obtaining a first point cloud frame collected by the radar at the current moment; if the first point cloud frame and the second point cloud frame collected by the radar at the previous moment meet the under-constraint condition, the second point cloud frame is used as the end frame of the first local sub-map of the area to be mapped, and the first point cloud frame is used as the start frame of the second local sub-map of the area to be mapped; wherein different local sub-maps correspond to different sub-areas in the area to be mapped; and a point cloud map of the area to be mapped is generated based on the first local sub-map and the second local sub-map. The present application divides two adjacent point cloud frames at the acquisition moment with the under-constraint problem into two different local sub-maps, so that there is no under-constraint situation inside each local sub-map, thereby ensuring consistency with the real environment of the area to be mapped.

Description

Technical Field

[0001] The present application relates to the field of autonomous driving technology, and in particular to a mapping method, apparatus, device, and storage medium. Background Art

[0002] In the field of autonomous driving, high-precision maps built using Simultaneous Localization and Mapping (SLAM) technology are an indispensable component for achieving autonomous driving capabilities. This is especially true for flying cars, which are one of the future development directions of transportation. However, due to the unknowns of the environment of the mapped area and the positions of mobile devices such as cars and aircraft, areas with few geometric features, such as flat surfaces like lawns and lakes, may be underconstrained. This results in uncertainty in the size, position, and boundaries of the area, resulting in deviations from the actual environment of the mapped area, making full terrain mapping impossible. Summary of the Invention

[0003] To overcome the problems existing in the related art, the present application provides a mapping method, apparatus, device and storage medium.

[0004] According to a first aspect of an embodiment of the present application, a mapping method is provided, which is applied to a mobile device, wherein the mobile device is configured with an observation device, and the observation device includes at least a radar. The method includes:

[0005] Get the first point cloud frame collected by the radar at the current moment;

[0006] If the first point cloud frame and a second point cloud frame acquired by the radar at a moment before the current moment meet an under-constraint condition, the second point cloud frame is used as the end frame of a first local subgraph of the area to be mapped, and the first point cloud frame is used as the start frame of a second local subgraph of the area to be mapped; wherein different local subgraphs correspond to different sub-areas of the area to be mapped;

[0007] A point cloud map of the area to be mapped is generated based on the first local sub-map and the second local sub-map.

[0008] According to a second aspect of an embodiment of the present application, a mapping apparatus is provided, which is applied to a movable device, wherein the movable device is equipped with an observation device, wherein the observation device includes at least a radar, including:

[0009] An acquisition module is used to obtain the first point cloud frame collected by the radar at the current moment;

[0010] a determination module configured to, if the first point cloud frame and a second point cloud frame acquired by the radar at a moment before the current moment satisfy an underconstraint condition, use the second point cloud frame as an end frame of a first local subgraph of the area to be mapped, and use the first point cloud frame as a start frame of a second local subgraph of the area to be mapped; wherein different local subgraphs correspond to different sub-areas in the area to be mapped;

[0011] A generating module is used to generate a point cloud map of the area to be mapped based on the first local sub-map and the second local sub-map.

[0012] According to a third aspect of an embodiment of the present application, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described in the first aspect when executing the program.

[0013] According to a fourth aspect of an embodiment of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and the computer program is used to instruct related hardware to complete the method described in the first aspect above.

[0014] The technical solutions provided by the embodiments of the present application may have the following beneficial effects:

[0015] This application maps the area to be mapped by dividing it into local subgraphs. Underconstraints are detected in real time during the mapping process. When underconstraints occur in the first and second adjacent point cloud frames at the time of acquisition, the first and second point cloud frames are divided into two different local subgraphs. This ensures that underconstraints are eliminated within each local subgraph, and ensures consistency with the real environment of the area to be mapped, even in areas with fewer geometric features.

[0016] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0018] Figure 1A This is a flowchart of a mapping method according to an exemplary embodiment of the present application.

[0019] Figure 1B This is a schematic diagram of factor graph optimization according to an exemplary embodiment of the present application.

[0020] Figure 1CThis is another factor graph optimization schematic diagram shown in the present application according to an exemplary embodiment.

[0021] Figure 2 This is a block diagram of a mapping device according to an exemplary embodiment of the present application.

[0022] Figure 3 This is a hardware structure diagram of an electronic device in which a mapping device is located according to an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0023] Exemplary embodiments will be described in detail herein, with examples illustrated in the accompanying drawings. In the following description, when referring to the drawings, identical numerals in different figures represent identical or similar elements, unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

[0024] The terms used in this application are for the purpose of describing specific embodiments only and are not intended to limit this application. As used in this application and the appended claims, the singular forms "a," "an," "the," and "the" are intended to include the plural forms, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" as used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items.

[0025] It should be understood that although the terms first, second, third, etc. may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from each other. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0026] During the mapping process, for areas with fewer geometric features, such as lawns and lakes, it is impossible to establish a correspondence between the data collected by the observation equipment at two adjacent moments. For example, it is impossible to match the lawn boundary points in two lawn point cloud frames collected by radar at adjacent moments, resulting in uncertainty in the size, boundary, and position of the lawn, an under-constraint problem, and an inaccurate map. To overcome the under-constraint problem existing in the related art, the present application proposes a mapping method that maps the area to be mapped by dividing it into local subgraphs. During the mapping process, possible under-constraint conditions are detected in real time. When under-constraint conditions occur in the first and second point cloud frames adjacent to the acquisition time, the first point cloud frame and the second point cloud frame are divided into two different local subgraphs. This ensures that there is no under-constraint condition within each local subgraph, and even for areas with fewer geometric features, consistency with the actual environment of the area to be mapped can be guaranteed. The present application is applicable to mobile devices with self-positioning and navigation needs, including but not limited to cars, drones, sweeping robots, etc., which are equipped with observation equipment for collecting environmental information of the area to be mapped, including at least a radar.

[0027] Next, the embodiments of the present application are described in detail.

[0028] like Figure 1A As shown, Figure 1A This is a flowchart of a mapping method according to an exemplary embodiment of the present application, which includes the following steps:

[0029] Step 102: Obtain the first point cloud frame collected by the radar at the current moment;

[0030] Step 104: If the first point cloud frame and the second point cloud frame collected by the radar at the previous moment meet the under-constraint condition, the second point cloud frame is used as the end frame of the first local sub-map of the area to be mapped, and the first point cloud frame is used as the start frame of the second local sub-map of the area to be mapped; wherein different local sub-maps correspond to different sub-areas in the area to be mapped;

[0031] Step 106: Generate a point cloud map of the area to be mapped based on the first local sub-map and the second local sub-map.

[0032] Simultaneous localization and mapping builds a map of the area to be mapped based on the mobile device's position and guides the mobile device's movement path based on the created map. Ensuring real-time and accurate mapping is crucial for navigation. Underconstraints can affect mapping accuracy, so underconstraint detection is performed in real time during the mapping process. Each time the radar captures a point cloud frame, it determines whether the first point cloud frame captured at the current moment and the second point cloud frame captured at the previous moment are underconstrained. If underconstraints are present, a local subgraph (the first local subgraph) is constructed based on the second point cloud frame and the previous point cloud frame captured by the radar. A new local subgraph (the second local subgraph) is then constructed using the first point cloud frame as the starting frame. For example, the point cloud frames captured by the radar from the starting time t1 to the current time t3 are p1, p2, and p3, respectively. If p3 and p2 are underconstrained, then p1 and p2 belong to local subgraph a1, and p3 belongs to local subgraph a2. Since the two point cloud frames with under-constraint problems belong to two different local subgraphs, for each local subgraph, there is no under-constraint problem between any two adjacent point cloud frames, and each local subgraph is consistent with the position, size, shape, etc. of the sub-region of the area to be mapped.

[0033] Whether there is an underconstraint between two adjacent point cloud frames can be determined based on the state variables of the movable device. These state variables characterize the motion and positional state of the movable device, including at least its pose. Furthermore, state variables may include at least the velocity of the movable device, the gyroscope angular velocity bias, the acceleration bias of the movable device, or the acceleration due to gravity. It should be noted that the state variables of the movable device are all vectors in real space.

[0034] The state variables can be provided by a navigation device installed on the mobile device. The navigation device can provide the posture information of the mobile device, and the state variables of the mobile device at least integrate the posture information provided by the navigation device. Since the navigation device provides global observation information, that is, the posture information in the world coordinate system, the state variables of the mobile device can solve the under-constraint problem on the one hand and the drift problem on the other hand after integrating the data collected by the navigation device. At the same time, the navigation device will also provide a covariance matrix for characterizing the accuracy of the state variables. In one embodiment of the present application, based on the covariance matrix corresponding to the state variables of the mobile device at the current moment, it is determined whether the first point cloud frame collected at the current moment and the second point cloud frame collected at the previous moment meet the under-constraint condition, including the following steps: calculating the eigenvalues ​​of the covariance matrix according to the covariance matrix corresponding to the state variables of the mobile device at the current moment; inputting the calculated eigenvalues ​​into a pre-trained classifier, and determining whether the first point cloud frame and the second point cloud frame meet the under-constraint condition based on the classification results output by the classifier. In addition, other methods can also be used for under-constraint detection, which is not limited by the present application.

[0035] When the state variables of a mobile device incorporate the pose information provided by a navigation device, the pose information collected by the navigation device will have a certain magnitude of error. This error is generally too large for mapping purposes, causing the system state variable error to gradually increase as the mobile device's movement path increases, thereby affecting the accuracy of mapping. To mitigate the error caused by the pose information provided by the navigation device on the state variables of the mobile device, a local subgraph can be reconstructed when the error accumulates to a certain level. Based on the pose information provided by the navigation device at the time of the start frame acquisition of the new local subgraph, the state variables of the mobile device at that time are re-determined. This allows the error caused by the pose information provided by the navigation device to be controlled within a certain range.

[0036] That is to say, it is possible to determine whether to start building a new local subgraph by setting the upper limit of the error brought about by the posture information provided by the navigation device, and to construct a new local subgraph when the upper limit is exceeded at the current moment. The upper limit of the error can be represented by the movement distance of the movable device or by the deviation between the state variable of the movable device at the current moment and the posture information provided by the navigation device at the current moment. When the deviation between the state variable of the movable device at the current moment and the posture information provided by the navigation device at the current moment exceeds a preset threshold, and / or when the cumulative movement distance of the movable device at the current moment is a preset distance threshold, start building a new local subgraph (the second local subgraph), use the first point cloud frame collected at the current moment as the starting frame of the second local subgraph, and use the second point cloud frame collected at the moment before the current moment as the ending frame of the first local subgraph. The deviation threshold between the state variable and the posture information provided by the navigation device and the preset distance threshold of the cumulative movement distance of the navigation device can be set based on the model of the navigation device and usage experience. In addition, it is also possible to determine whether a new local subgraph needs to be constructed based on other conditions that can characterize the upper limit of the error brought about by the posture information provided by the navigation device, and this application does not impose any restrictions on this.

[0037] It is understandable that, based on the aforementioned content, multiple trigger conditions for reconstructing a new local subgraph can be set simultaneously. Regardless of whether an under-constraint problem occurs or the error exceeds the upper limit, a new local subgraph will be reconstructed. This ensures that there is no under-constraint problem within each local subgraph while reducing the error caused by the posture information provided by the navigation device. It should also be noted that the aforementioned state variables for detecting under-constraint conditions and for determining whether the error caused by the posture information provided by the navigation device exceeds the upper limit are the state variables corresponding to the moment when the mobile device collects the point cloud frame by the radar.

[0038] In addition to the navigation device, an inertial measurement unit (IMU) can also provide state variables for the mobile device. Based on the angular velocity and acceleration collected by the IMU, changes in the speed and position of the mobile device can be determined. In one embodiment of the present application, the observation device also includes an IMU, and the state variables of the mobile device are determined by fusing data collected by the navigation device, radar, and IMU. Whenever the radar or navigation device collects data, iterative updates are performed based on the initial value of the state variable of the mobile device. The initial value of the state variable can be provided by the navigation device or set in advance by the user. The specific fusion process is as follows: Assuming that the initial value of the state variable at initial time t0 is x0, if the radar or navigation device collects data d1 at time t1, the motion data of the mobile device collected by the IMU between time t0 and time t1 and data d1 are fused on the basis of x0 to determine the state variable x1 at time t1; if the radar or navigation device collects data d2 at time t2, the motion data of the mobile device collected by the IMU between time t1 and time t2 and data d2 are fused on the basis of x1 to determine the state variable x2 at time t2, and so on. The data fusion process can be achieved through the Kalman filter.

[0039] In order to further reduce the impact of the error caused by the pose information provided by the navigation device on the accuracy of the state variables of the mobile device, each local subgraph can also be optimized for error. For example, the factor graph optimization method can be used for optimization, introducing two constraints: prior constraints and association constraints. Each local subgraph corresponds to multiple frames of point cloud frames, and each point cloud frame corresponds to the state variables of the mobile device at the time when the frame of the point cloud frame is collected. In other words, each local subgraph corresponds to multiple state variables, and the optimization of the local subgraph is achieved by optimizing multiple state variables. Among them, the prior constraint is the expected value and covariance matrix corresponding to the pose information collected by the navigation device from the time the starting frame of the local subgraph is collected to the time the ending frame of the local subgraph is collected; the association constraint is composed of multiple association factors, each of which is calculated based on the state variables and covariance matrix corresponding to each point cloud frame in any two adjacent point cloud frames corresponding to the local subgraph at the time of collection. Subsequently, the first pose matrix is ​​calculated based on the state variables corresponding to the end frame before optimization, and the second pose matrix is ​​calculated based on the state variables corresponding to the end frame after optimization. The product of the first pose matrix and the second pose matrix is ​​determined as the update matrix; the state variables and local subgraph corresponding to the end frame after optimization are updated based on the update matrix.

[0040] The state variables and the local subgraph corresponding to the optimized end frame are updated based on the update matrix.

[0041] like Figure 1BAs shown, this is a factor graph optimization schematic diagram according to an exemplary embodiment of the present application, which is based on the state variables of the mobile device at different times (time) corresponding to the local subgraph to be optimized. The local subgraph is constructed based on the point cloud frame collected by the radar at time t1-t6, corresponding to the state variables x1-x6 of the mobile device at time t1-t6, and each state variable corresponds to a covariance matrix. The navigation device collects posture information y1, y3 and y6 at time t1, t3 and t6 respectively. The expected value and covariance matrix corresponding to y1 are used as prior factors of the state variable at time t1 to control the adjustment range of the state variable at time t1; the expected value and covariance matrix corresponding to y3 are used as prior factors of the state variable at time t3 to control the adjustment range of the state variable at time t3; the expected value and covariance matrix corresponding to y6 are used as prior factors of the state variable at time t6 to control the adjustment range of the state variable at time t6, and all prior factors are used as prior constraints. Based on the state variable x1 at time t1 and its corresponding covariance matrix z1, and the state variable x2 at time t2 and its corresponding covariance matrix z2, the correlation factor between x1 and x2 is calculated. This is used to control the difference between x1 and x2. Similarly, the correlation factors of two adjacent state variables are obtained, and all correlation factors serve as correlation constraints. The optimized state variables x1'-x6' at each time t1-t6 are then calculated using the probability maximization method.

[0042] It should be noted that due to the inconsistent sampling frequency of the observation equipment, the navigation device may not collect data in the time period corresponding to the local sub-graph (for example, the navigation device did not collect posture information at time t1-t6). In this case, the posture information of the navigation device at each time t1-t6 can be determined by interpolation, and the expected value and covariance matrix corresponding to the posture information at at least one time can be selected as the prior constraint.

[0043] Then, the first pose matrix T6 is calculated based on the pre-optimization state variable x6 corresponding to the end frame acquisition time t6, and the second pose matrix T6' is calculated based on the post-optimization state variable x6' corresponding to the end frame acquisition time t6, and the product of T6 and T6' is used as the update matrix. The update matrix can be expressed as Where T k is the pose matrix before optimization at time k, is the pose matrix after optimization at time k, Ru is a 3x3 matrix, which corresponds to the rotation part of the update matrix composed of the angular deviation of the pose variables corresponding to the end frame in the local subgraph before and after optimization, and is used to characterize the amplitude of the attitude angle adjustment, which can correct the deviation of the attitude angle; tu is a 3x1 vector, which corresponds to the translation part of the update matrix composed of the position deviation of the pose variables corresponding to the end frame in the local subgraph before and after optimization, which is used to characterize the amplitude of the position adjustment, which can correct the deviation of the position.

[0044] The state variables can be expressed as X = {r, t, V, ba, bg, g}, where r represents the attitude angle of the mobile device, t represents the position of the mobile device, V represents the speed of the mobile device, ba represents the zero bias of the acceleration of the mobile device, bg represents the zero bias of the gyroscope angular velocity on the mobile device, and g represents the acceleration of gravity. The state variables updated by the update matrix are expressed as X = {log(R u exp(r)), R u t+t u , R u V, ba, bg, R u The update of the local sub-image is based on the update matrix to adjust the coordinates of each point in the point cloud frame corresponding to the local sub-image. The local sub-image can be expressed as χ = {p i , i=1,2,...,N}, where p i Represents the coordinates of each point in the point cloud frame, and the updated local subgraph can be expressed as χ u ={R u p i +t u , i=1,2,...,N}.

[0045] After optimizing each local subgraph individually, a global optimization can be performed on the optimized subgraphs, further eliminating modified state variables and reducing errors. The optimization method is similar to the aforementioned optimization method for a single local subgraph, employing factor graph optimization and introducing two constraints: prior constraints and association constraints. The only difference is that the state variables corresponding to a single local subgraph are replaced with state variables corresponding to each local subgraph. The state variables of each local subgraph are the state variables updated by the mobile device at the time of acquisition of the local subgraph's end frame. The difference is that during global optimization, each local subgraph is assigned a prior constraint based on the expected value and covariance matrix corresponding to the pose information collected by the navigation device at the time corresponding to the local subgraph's end frame. The association constraints are composed of multiple association factors, which are calculated based on the updated state variables and covariance matrix corresponding to the end frame of any adjacent first local subgraph and the updated state variables and covariance matrix corresponding to the end frame of a second local subgraph, where the acquisition time of the end frame of the first local subgraph is adjacent to the acquisition time of the start frame of the second local subgraph. Since the pose information collected by the navigation device is introduced as a priori constraint during global optimization, the under-constraint problem between two adjacent local subgraphs can be eliminated.

[0046] like Figure 1C As shown, this is another factor graph optimization schematic diagram shown in accordance with an exemplary embodiment of the present application. Based on the state variable construction at the end frame acquisition time (time) of the constructed multiple local subgraphs, the six local subgraphs g1-g6 are globally optimized. The state variables corresponding to each local subgraph are x1-x6, and each state variable corresponds to a covariance matrix. The navigation device collects pose information y1-y6 at the corresponding time of the end frame of each local subgraph. The expected value and covariance matrix corresponding to y1 are used as the prior factors of the local subgraph g1, and the expected value and covariance matrix corresponding to y2 are used as the prior factors of the local subgraph g2. The prior factors of other local subgraphs are obtained by analogy, and all prior factors are used as prior constraints. Based on the state variable x1 of local subgraph g1 and its corresponding covariance matrix z1, and the state variable x2 of local subgraph g2 and its corresponding covariance matrix z2, the correlation factor between x1 and x2 is calculated. This is used to control the difference between x1 and x2. Similarly, the correlation factors of two adjacent local subgraphs are obtained, and all correlation factors serve as correlation constraints. The optimized state variables x1'-x6' of each local subgraph are then calculated using the probability maximization method.

[0047] Corresponding to the above-mentioned method embodiment, the present application also provides an embodiment of a mapping device and a terminal used therein. Figure 2 As shown, Figure 2 2 is a block diagram of a mapping device 200 according to an exemplary embodiment of the present application, the device comprising:

[0048] An acquisition module 210 is configured to acquire a first point cloud frame collected by the radar at a current moment;

[0049] Determining module 220 is configured to, if the first point cloud frame and a second point cloud frame acquired by the radar at a moment before the current moment satisfy an underconstraint condition, use the second point cloud frame as an end frame of a first local subgraph of the area to be mapped, and use the first point cloud frame as a start frame of a second local subgraph of the area to be mapped; wherein different local subgraphs correspond to different sub-areas in the area to be mapped;

[0050] The generating module 230 is configured to generate a point cloud map of the area to be mapped based on the first local sub-map and the second local sub-map.

[0051] The implementation process of the functions and effects of each module in the above-mentioned device is specifically described in the implementation process of the corresponding steps in the above-mentioned method, and will not be repeated here.

[0052] For the device embodiment, since it basically corresponds to the method embodiment, the relevant parts can be referred to the partial description of the method embodiment. The device embodiment described above is merely illustrative, wherein the modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules, that is, they may be located in one place, or they may be distributed on multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the present application scheme. Those of ordinary skill in the art can understand and implement it without paying any creative work.

[0053] The embodiment of the mapping device in this application document can be installed on an electronic device. The device embodiment can be implemented by software, or by hardware or a combination of software and hardware. Taking software implementation as an example, as a device in a logical sense, it is formed by the processor reading the corresponding computer program instructions in the non-volatile memory into the memory and running them. From the hardware level, such as Figure 3 FIG. 1 is a hardware structure diagram of the electronic device 300 in which the mapping device is located in an embodiment of the present application. Figure 3 In addition to the processor 310, memory 330, network interface 320, and non-volatile memory 340 shown, the electronic device where the device 331 is located in the embodiment may also include other hardware according to the actual function of the electronic device, which will not be described in detail.

[0054] Accordingly, the present application further provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method described in any of the aforementioned method embodiments is implemented. The present application further provides a computer-readable storage medium storing a computer program, wherein the computer program is used to instruct related hardware to implement the method described in any of the aforementioned method embodiments.

[0055] The foregoing description describes specific embodiments of the present application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in an order different from that described in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order shown or the sequential order to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0056] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention claimed herein. This application is intended to cover any variations, uses, or adaptations of the present application that follow the general principles of this application and include common knowledge or customary techniques in the art not claimed herein. The description and examples are to be considered merely as exemplary, and the true scope and spirit of the present application are indicated by the claims.

[0057] It should be understood that the present application is not limited to the exact structures described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

[0058] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

Claims

1. A mapping method, characterized in that: The method is applied to a mobile device, wherein the mobile device is equipped with an observation device, the observation device includes at least a radar and a navigation device, and the state variables of the mobile device at least integrate the posture information in the world coordinate system provided by the navigation device, and correspond to the covariance matrix provided by the navigation device that characterizes the accuracy of the state variables; the method includes: Get the first point cloud frame collected by the radar at the current moment; If the first point cloud frame and a second point cloud frame acquired by the radar at a moment before the current moment meet an under-constraint condition, the second point cloud frame is used as the end frame of a first local subgraph of the area to be mapped, and the first point cloud frame is used as the start frame of a second local subgraph of the area to be mapped; wherein different local subgraphs correspond to different sub-areas of the area to be mapped; Performing global optimization on the first local subgraph and the second local subgraph; during the global optimization, each local subgraph corresponds to an expected value and a covariance matrix corresponding to the pose information collected by the navigation device at a time corresponding to the end frame of the local subgraph as a first prior constraint; A point cloud map of the area to be mapped is generated based on the globally optimized first local sub-graph and the second local sub-graph.

2. The method according to claim 1, characterized in that Determine whether the first point cloud frame and the second point cloud frame meet the under-constraint condition based on the following steps: Calculating the eigenvalues ​​of the covariance matrix according to the covariance matrix corresponding to the state variable of the movable device at the current moment; The feature value is input into a pre-trained classifier, and based on a classification result output by the classifier, it is determined whether the first point cloud frame and the second point cloud frame meet an under-constraint condition.

3. The method according to claim 2, characterized in that The state variable includes at least a posture, and the method further includes: If a deviation between the state variable of the movable device at the current moment and the pose information provided by the navigation device at the current moment exceeds a preset threshold, the second point cloud frame is used as the end frame of the first local sub-map of the area to be mapped, and the first point cloud frame is used as the start frame of the second local sub-map of the area to be mapped; and / or, If the cumulative moving distance of the movable device at the current moment is a preset distance threshold, the second point cloud frame is used as the end frame of the first local sub-map of the area to be mapped, and the first point cloud frame is used as the starting frame of the second local sub-map of the area to be mapped.

4. The method according to claim 2, characterized in that The observation device also includes an inertial measurement unit, and the state variables of the movable device are determined by fusing data collected by the navigation device, the radar and the inertial measurement unit, wherein the navigation device can provide position information of the movable device in the world coordinate system.

5. The method according to claim 4, characterized in that The fusion of the data collected by the navigation device, the radar and the inertial measurement unit is achieved through a Kalman filter.

6. The method according to claim 2, characterized in that Each of the local subgraphs corresponds to multiple point cloud frames, each point cloud frame corresponds to the state variable of the movable device at the time when the point cloud frame is collected, and the method further includes: Optimizing a plurality of state variables corresponding to the local subgraph based on a second priori constraint and a second association constraint; The second priori constraint is the expected value and covariance matrix corresponding to the pose information collected by the navigation device from the time when the start frame of the local sub-graph is collected to the time when the end frame of the local sub-graph is collected; The second association constraint is composed of a plurality of second association factors, and the second association factors are calculated based on the state variables and covariance matrices corresponding to each point cloud frame in any two adjacent point cloud frames corresponding to the local subgraph at acquisition moments; Calculating a first pose matrix based on the state variables corresponding to the end frame before optimization, calculating a second pose matrix based on the state variables corresponding to the end frame after optimization, and determining the product of the first pose matrix and the second pose matrix as an update matrix; The state variables and the local subgraph corresponding to the optimized end frame are updated based on the update matrix.

7. The method according to claim 6, characterized in that The method further comprises: Optimizing the point cloud map of the area to be mapped based on the first prior constraint and the first association constraint; The first association constraint is composed of multiple first association factors, which are calculated based on the updated state variables and covariance matrix corresponding to the end frame of any adjacent first local sub-graph and the updated state variables and covariance matrix corresponding to the end frame of the second local sub-graph. The acquisition time of the end frame of the first local sub-graph is adjacent to the acquisition time of the start frame of the second local sub-graph.

8. The method according to claim 2, characterized in that The state variables also include at least one of the following: Velocity, gyroscope angular velocity bias, acceleration bias, or acceleration due to gravity.

9. A mapping device, characterized in that: Applied to a mobile device, the mobile device is configured with an observation device, the observation device includes at least a radar and a navigation device, the state variables of the mobile device at least integrate the posture information in the world coordinate system provided by the navigation device, and correspond to the covariance matrix provided by the navigation device that characterizes the accuracy of the state variables; the device includes: An acquisition module is used to obtain the first point cloud frame collected by the radar at the current moment; a determination module configured to, if the first point cloud frame and a second point cloud frame acquired by the radar at a moment before the current moment satisfy an underconstraint condition, use the second point cloud frame as an end frame of a first local subgraph of the area to be mapped, and use the first point cloud frame as a start frame of a second local subgraph of the area to be mapped; wherein different local subgraphs correspond to different sub-areas in the area to be mapped; A generation module is used to perform global optimization on the first local subgraph and the second local subgraph; during global optimization, each local subgraph corresponds to an expected value and covariance matrix corresponding to the posture information collected by the navigation device at the time corresponding to the end frame of the local subgraph as a first prior constraint; based on the globally optimized first local subgraph and the second local subgraph, a point cloud map of the area to be mapped is generated.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the method according to any one of claims 1 to 8 is implemented.

11. A computer-readable storage medium storing a computer program, wherein the computer program is used to instruct related hardware to perform the method according to any one of claims 1 to 8.

Images