Lidar and inertial navigation integrating positioning method, and related device

Abstract

A LiDAR and inertial navigation integrating positioning method and system, a storage medium, and an electronic device. The method comprises: according to a first position and orientation state of a vehicle, predicting a second position and orientation state of the vehicle, the first position and orientation state being used for indicating the corrected position and orientation state of the vehicle in a first cycle, the second position and orientation state being used for indicating a predicted position and orientation state of the vehicle in a second cycle, and the first cycle being one cycle ahead of the second cycle (310); observing the vehicle in a local point cloud map and in a vector map to construct observation residuals (320); and, according to the observation residuals, correcting the second position and orientation state to obtain a third position and orientation state (330). The method introduces the observation residual in the local point cloud map and the observation residual in the vector map, corrects the position and orientation state in the current cycle according to the observation residuals and, in the case of the minimum observation residuals, obtains the most accurate corrected position and orientation state in the current cycle, effectively improving the precision of LiDAR and inertial navigation integrating positioning.

Description

Laser radar and inertial navigation fusion positioning method and related equipment Technical Field

[0001] The present application relates to the field of automotive technology, and in particular to a laser radar and inertial navigation fusion positioning method and related equipment. Background Art

[0002] Vehicle posture determination is one of the key technologies in the field of autonomous driving. It mainly refers to the use of sensors such as vehicle-mounted lidar and inertial measurement unit (IMU) to achieve high-precision real-time posture determination of the vehicle.

[0003] Currently, to achieve high-precision, real-time positioning of vehicles, vehicles can be located using lidar odometry or a priori point cloud maps. However, using lidar odometry for vehicle positioning inevitably results in cumulative errors. Using a priori point cloud maps for vehicle positioning requires significant storage and computing resources. Therefore, achieving high-precision, low-resource, real-time positioning of vehicles is an urgent challenge.

[0004] Summary of the Invention

[0005] In view of the above problems, the present application provides a laser radar and inertial navigation fusion positioning method and related equipment, which solves the problem of how to achieve high-precision and low-resource real-time positioning of vehicles.

[0006] To solve the above technical problems, this application proposes the following solutions:

[0007] In the first aspect, the present application provides a laser radar and inertial navigation fusion positioning method, the method comprising: predicting a second posture state of a vehicle based on the first posture state of the vehicle, the first posture state being used to indicate a corrected posture state of the vehicle in a first cycle, and the second posture state being used to indicate a predicted posture state of the vehicle in a second cycle, the first cycle being one cycle prior to the second cycle; observing the vehicle in a local point cloud map and a vector map, and constructing an observation residual; correcting the second posture state based on the observation residual to obtain a third posture state.

[0008] The vehicle's posture state in the previous cycle can be used to preliminarily predict its posture state in the current cycle. However, this initial prediction contains errors and is inaccurate. Therefore, the observation residuals from the local point cloud map and the vector map are introduced to correct the posture state in the current cycle based on these observation residuals. Specifically, when the observation residuals are minimized, the corrected posture state for the current cycle is most accurate, effectively improving the accuracy of fused LiDAR and inertial navigation positioning.

[0009] In combination with the first aspect, in one possible implementation, point clouds within a preset period are accumulated to construct a local point cloud map; the first frame point cloud is matched with the local point cloud map, and the first frame point cloud is used to indicate the second period lidar scanning frame.

[0010] This application performs matching based on the constructed local point cloud map, which can avoid the problem of using the global point cloud map as prior information, which brings about the large storage and computing resources required, and effectively reduces the dependence on storage and computing resources.

[0011] In combination with the first aspect, in another possible implementation method, edge points and plane points of the first frame point cloud are extracted; the first point and the second point are searched in the local point cloud map, and the distance between the first point and the plane point is smaller than the distance between other points and the plane point in the local point cloud map, and the distance between the second point and the edge point is smaller than the distance between other points and the edge point in the local point cloud map.

[0012] In combination with the first aspect, in another possible implementation method, edge points and plane points of the first frame point cloud are extracted, where the edge points include lane line points and rod-shaped object points, and the plane points include ground points; the lane lines, rod-shaped objects and ground in the vector map are discretized to obtain first points to be matched and second points to be matched; the first point is searched in the first points to be matched, and the second point is searched in the second points to be matched.

[0013] This application uses only a small, high-precision vector map as prior information. Observing within the high-precision vector map can further reduce reliance on storage and computing resources. Since the vector map is highly precise, it can also improve global positioning accuracy.

[0014] In combination with the first aspect, in another possible implementation method, the ground is divided into a grid according to a preset length and width, and the center point of the grid is used as the first point to be matched; the lane lines and pole-shaped objects are divided into line segments according to a preset length, and the midpoint of the line segment is used as the second point to be matched.

[0015] According to the above method, lane line points, pole points, and ground points can be discretized into many points, and then more accurate points can be matched in the vector map.

[0016] In combination with the first aspect, in another possible implementation method, the normal vector of the plane point and the direction vector of the edge point are calculated; the plane point residual is calculated based on the normal vector, the coordinates of the plane point in the global coordinate system, and the coordinates of the first point in the global coordinate system; the edge point residual is calculated based on the direction vector, the coordinates of the edge point in the global coordinate system, and the coordinates of the second point in the global coordinate system.

[0017] The residual indicates the degree of match between the LiDAR scan frame and the points it matches in the local point cloud map and vector map. When the residual is less than the threshold, the LiDAR scan frame is essentially consistent with the points it matches in the local point cloud map and vector map. This means the points it matches in the local point cloud map and vector map are the vehicle's actual position. When the residual is greater than the threshold, the LiDAR scan frame is inconsistent with the points it matches in the local point cloud map and vector map, and further correction of the vehicle's position is required.

[0018] In combination with the first aspect, in another possible implementation method, the second posture state covariance of the vehicle is predicted based on the first posture state covariance of the vehicle, the first posture state covariance is used to indicate the corrected posture state covariance of the vehicle in the first period, and the second posture state covariance is used to indicate the predicted posture state covariance of the vehicle in the second period; the second posture state is iteratively extended Kalman filtering based on the second posture state covariance and observation residuals to obtain a third posture state until the third posture state meets the preset convergence conditions.

[0019] The goal of revising the second pose state using iterative extended Kalman filtering is to ensure a high degree of match between the points in the LiDAR scan frame and the matching points in the local point cloud map and vector map. By continuously revising the second pose state to minimize the residual error, the vehicle's accurate position and pose in the sub-point cloud map and vector map are determined.

[0020] In combination with the first aspect, in another possible implementation, the second posture state covariance of the vehicle is predicted based on the second posture state covariance, the state transfer matrix, the noise, and the noise matrix.

[0021] In combination with the first aspect, in another possible implementation, the second posture state of the vehicle is predicted based on the first posture state and the angular velocity and acceleration of the inertial navigation corresponding to the second period lidar scanning frame.

[0022] First, the posture state and posture state covariance of the current cycle are predicted based on the posture state and posture state covariance of the vehicle in the previous cycle, and then the predicted posture state of the current cycle is corrected. This can effectively reduce the error caused by directly using the posture state and posture state covariance of the previous cycle to locate the current posture state of the vehicle.

[0023] In combination with the first aspect, in another possible implementation, the second-cycle lidar scanning frame is converted to a global coordinate system, and the second-cycle lidar scanning frame is added to the local point cloud map.

[0024] By adding all the points of the current lidar frame to the local point cloud map constructed above, the local point cloud map can be expanded for use in positioning the vehicle's posture state in the next cycle.

[0025] In combination with the first aspect, in another possible implementation, when the size of the local point cloud map exceeds a first threshold, the third point in the local point cloud map is deleted, and the distance between the third point and the second-cycle lidar scanning frame is greater than a second threshold.

[0026] This application can effectively avoid the problem of local point cloud maps storing too many point cloud locations, which makes the point cloud maps too large and thus relies on storage resources and computing resources.

[0027] In a second aspect, the present application provides a laser radar and inertial navigation fusion positioning device, which includes:

[0028] The prediction module is used to predict the second posture state of the vehicle based on the first posture state of the vehicle. The first posture state is used to indicate the corrected posture state of the vehicle in the first cycle, and the second posture state is used to indicate the predicted posture state of the vehicle in the second cycle. The first cycle is one cycle ahead of the second cycle.

[0029] The observation module is used to observe vehicles in local point cloud maps and vector maps and construct observation residuals.

[0030] The correction module is used to correct the second posture state according to the observation residual to obtain the third posture state.

[0031] In combination with the second aspect, in one possible implementation method, the observation module is specifically used to accumulate point clouds within a preset period and construct a local point cloud map; match the first frame point cloud with the local point cloud map, and the first frame point cloud is used to indicate the second period lidar scanning frame.

[0032] In combination with the second aspect, in another possible implementation method, the observation module is specifically used to extract edge points and plane points of the first frame point cloud; search for the first point and the second point in the local point cloud map, and search for the first point and the second point in the local point cloud map, the distance between the first point and the plane point is smaller than the distance between other points and the plane point in the local point cloud map, and the distance between the second point and the edge point is smaller than the distance between other points and the edge point in the local point cloud map.

[0033] In combination with the second aspect, in another possible implementation method, the observation module is specifically used to extract edge points and plane points of the first frame point cloud, where the edge points include lane line points and rod-shaped object points, and the plane points include ground points; discretize the lane lines, rod-shaped objects and ground in the vector map to obtain first points to be matched and second points to be matched; search for the first point in the first points to be matched, and search for the second point in the second points to be matched.

[0034] In combination with the second aspect, in another possible implementation method, the observation module is specifically used to divide the ground into a grid according to a preset length and width, and use the center point of the grid as the first point to be matched; divide the lane lines and pole-shaped objects into line segments according to a preset length, and use the midpoint of the line segment as the second point to be matched.

[0035] In combination with the second aspect, in another possible implementation method, the observation module is specifically used to calculate the normal vector of the plane point and the direction vector of the edge point; calculate the plane point residual based on the normal vector, the coordinates of the plane point in the global coordinate system, and the coordinates of the first point in the global coordinate system; calculate the edge point residual based on the direction vector, the coordinates of the edge point in the global coordinate system, and the coordinates of the second point in the global coordinate system.

[0036] In combination with the second aspect, in another possible implementation method, the correction module is specifically used to predict the second posture state covariance of the vehicle based on the first posture state covariance of the vehicle, the first posture state covariance is used to indicate the corrected posture state covariance of the vehicle in the first period, and the second posture state covariance is used to indicate the predicted posture state covariance of the vehicle in the second period; iteratively extend the Kalman filter on the second posture state according to the second posture state covariance and the observation residual to obtain the third posture state, until the third posture state meets the preset convergence conditions.

[0037] In combination with the second aspect, in another possible implementation, the prediction module is further used to predict the second posture state covariance of the vehicle based on the second posture state covariance, the state transfer matrix, the noise and the noise matrix.

[0038] In combination with the second aspect, in another possible implementation, the prediction module is specifically used to predict the second posture state of the vehicle based on the first posture state and the angular velocity and acceleration of the inertial navigation corresponding to the second period lidar scanning frame.

[0039] In combination with the second aspect, in another possible implementation, the correction module is further used to convert the second-cycle lidar scanning frame into a global coordinate system and add the second-cycle lidar scanning frame to the local point cloud map.

[0040] In combination with the second aspect, in another possible implementation method, the correction module is also used to delete the third point in the local point cloud map when the size of the local point cloud map exceeds the first threshold, and the distance between the third point and the second cycle lidar scanning frame is greater than the second threshold.

[0041] In order to achieve the above-mentioned purpose, according to the third aspect of the present application, a storage medium is provided, which includes a stored program, wherein when the program is running, the device where the storage medium is located is controlled to execute the laser radar and inertial navigation fusion positioning method of the above-mentioned first aspect.

[0042] In order to achieve the above-mentioned purpose, according to the fourth aspect of the present application, an electronic device is provided, which includes at least one processor, and at least one memory and bus connected to the processor; wherein the processor and the memory communicate with each other through the bus; the processor is used to call the program instructions in the memory to execute the laser radar and inertial navigation fusion positioning method of the above-mentioned first aspect.

[0043] By means of the above technical solution, the technical solution provided by this application has at least the following advantages:

[0044] The present application provides a laser radar and inertial navigation fusion positioning method and related equipment. The present application can preliminarily predict the posture state of the vehicle in the current cycle based on the posture state of the vehicle in the previous cycle, but the posture state of the current cycle obtained by the preliminary prediction has errors and is not accurate enough. Therefore, the observation residuals in the local point cloud map and the observation residuals in the vector map are introduced, and the posture state of the current cycle is corrected according to the observation residuals. That is, when the observation residual is the smallest, the accuracy of the posture state of the current cycle obtained by correction is the highest, which effectively improves the accuracy of the laser radar and inertial navigation fusion positioning.

[0045] It can also avoid the problem of using the global point cloud map as prior information, which would require large storage and computing resources, and effectively reduce the dependence on storage and computing resources.

[0046] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Various other advantages and benefits will become apparent to those skilled in the art upon reading the detailed description of the preferred embodiment below. The accompanying drawings are for illustration purposes only and are not to be considered as limiting the present application. The same reference symbols are used throughout the drawings to represent the same components. In the drawings:

[0048] FIG1 shows a schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present application;

[0049] FIG2 shows a schematic diagram of a laser radar and inertial navigation fusion positioning system provided in an embodiment of the present application;

[0050] FIG3 shows a schematic flow chart of a laser radar and inertial navigation fusion positioning method provided in an embodiment of the present application;

[0051] FIG4 shows a schematic diagram of the composition of a laser radar and inertial navigation fusion positioning device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0052] The following describes exemplary embodiments of the present application in more detail with reference to the accompanying drawings. Although exemplary embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present application and to fully convey the scope of the present application to those skilled in the art.

[0053] In the embodiments of the present application, the terms "first" and "second" do not have a logical or temporal dependency, nor do they limit the quantity or execution order. It should also be understood that although the following description uses the terms "first," "second," etc. to describe various elements, these elements should not be limited by the terms. These terms are simply used to distinguish one element from another.

[0054] In the embodiments of the present application, the term "at least one" means one or more, and in the embodiments of the present application, the term "plurality" means two or more.

[0055] It should also be understood that the term "if" may be interpreted as "when" or "upon" or "in response to determining" or "in response to detecting." Similarly, the phrase "if it is determined that..." or "if [stated condition or event] is detected" may be interpreted as "upon determining that..." or "in response to determining that..." or "upon detecting [stated condition or event]" or "in response to detecting [stated condition or event]," depending on the context.

[0056] To facilitate understanding of the solutions of this application, a brief introduction to the relevant concepts is first given as follows.

[0057] Laser Radar: A radar system that uses laser beams to detect the position, speed, and other characteristic quantities of a target. A laser radar consists of a laser transmitter, an optical receiver, a turntable, and an information processing system. Its working principle is that a laser converts electrical pulses into light pulses and transmits them. The optical receiver then converts the light pulses reflected from the target back into electrical pulses. The relative distance of a single point is obtained based on the flight time of the light, and the position (x, y, z) of the target point cloud in the coordinate system is calculated based on the laser radar's basic coordinate system. Laser radar can directly generate point cloud information with coordinate positions. Under the illumination of a multi-line laser radar, it can form multi-point outer contours. Using target classification and recognition technologies such as cluster analysis, semantic segmentation, and deep learning, it builds a digital model of the roads, obstacles, pedestrians, lane lines, vehicles, traffic signs, and other areas surrounding the vehicle.

[0058] In other words, LiDAR measures the distance and orientation of surrounding objects by emitting laser beams, thereby determining the relative position of the vehicle and obstacles. When enough laser beams are emitted, these individual laser points will be combined into a point cloud, outlining the three-dimensional environment information, namely point cloud data.

[0059] A point cloud is a dataset of points in space that can represent three-dimensional shapes or objects, typically acquired by LiDAR. The location of each point in the point cloud is described by a set of Cartesian coordinates (X, Y, Z), some of which may also include color information (R, G, B) or the intensity of the object's reflective surface. Intensity information is obtained by capturing the echo intensity detected by the LiDAR receiver. This intensity information is related to the target's surface material, roughness, angle of incidence, as well as the instrument's emission energy and laser wavelength.

[0060] An inertial measurement unit (IMU), also known as an inertial navigation system, is a device that measures an object's three-axis attitude angles and acceleration. An IMU typically includes a gyroscope, accelerometers, and sometimes a magnetometer. Gyroscopes measure three-axis angles and angular velocities, accelerometers measure three-axis acceleration, and magnetometers provide information on the magnetic field's orientation.

[0061] Plane point: A point on a smooth plane in three-dimensional space, which has a small size difference from the surrounding points, low curvature, and low smoothness.

[0062] Edge point: A point on a sharp edge in three-dimensional space, which has a large size difference from the surrounding points, a high curvature, and a high smoothness.

[0063] The above is an introduction to the technical terms involved in the embodiments of this application, which will not be repeated below.

[0064] As mentioned in the background, LiDAR odometry suffers from unavoidable cumulative errors, making it impossible to achieve high-precision, real-time positioning of vehicles. Furthermore, using a priori point cloud maps to locate vehicles can avoid the errors associated with odometry-based positioning. However, these point cloud maps store a large number of point cloud locations and their attributes, resulting in a large overall size and a high reliance on storage and computing resources.

[0065] In view of this, an embodiment of the present application provides a laser radar and inertial navigation fusion positioning method, the specific method includes: predicting the second posture state of the vehicle based on the first posture state of the vehicle, the first posture state is used to indicate the corrected posture state of the vehicle in the first cycle, and the second posture state is used to indicate the predicted posture state of the vehicle in the second cycle, and the first cycle is one cycle prior to the second cycle; observing the vehicle in the local point cloud map and the vector map, and constructing the observation residual; correcting the second posture state based on the observation residual to obtain the third posture state. The posture state of the vehicle in the current cycle can be preliminarily predicted based on the posture state of the vehicle in the previous cycle, but the posture state of the current cycle obtained by the preliminary prediction has errors and is not accurate enough. Therefore, the observation residuals in the local point cloud map and the observation residuals in the vector map are introduced, and the posture state of the current cycle is corrected based on the observation residuals. That is, when the observation residual is the smallest, the accuracy of the corrected posture state of the current cycle is the highest, which effectively improves the accuracy of the laser radar and inertial navigation fusion positioning.

[0066] The present application also provides a laser radar and inertial navigation fusion positioning device, which can be used to perform the above-mentioned laser radar and inertial navigation fusion positioning method. Optionally, the laser radar and inertial navigation fusion positioning device can be an electronic device with data processing capabilities, or a functional module within such an electronic device, without limitation.

[0067] For example, the electronic device may be a server, which may be a single server, or a server cluster composed of multiple servers. For another example, the electronic device may be a mobile phone, a tablet computer, a desktop, a laptop, a handheld computer, a notebook computer, an ultra-mobile personal computer (UMPC), a netbook, as well as a cellular phone, a personal digital assistant (PDA), an augmented reality (AR), a virtual reality (VR) device and other terminal devices. For another example, the electronic device may also be a recording device, a video surveillance device and other devices. This application does not impose any special restrictions on the specific form of the electronic device.

[0068] The following takes the laser radar and inertial navigation fusion positioning device as an example of an electronic device, as shown in Figure 1, which is a hardware structure of an electronic device 100 provided in this application.

[0069] As shown in FIG. 1 , the electronic device 100 includes a processor 110 , a communication circuit 120 , and a communication interface 130 .

[0070] Optionally, the electronic device 100 may further include a memory 140 , wherein the processor 110 , the memory 140 and the communication interface 130 may be connected via a communication line 120 .

[0071] The processor 110 may be a central processing unit (CPU), a general-purpose processor, a network processor (NP), a digital signal processor (DSP), a microprocessor, a microcontroller, a programmable logic device (PLD), or any combination thereof. The processor 101 may also be any other device with processing capabilities, such as a circuit, a device, or a software module, without limitation.

[0072] In one example, the processor 110 may include one or more CPUs, such as CPU0 and CPU1 in FIG. 1 .

[0073] As an optional implementation, the electronic device 100 includes multiple processors. For example, in addition to the processor 110, it may also include a processor 170. The communication line 120 is used to transmit information between the various components included in the electronic device 100.

[0074] Communication interface 130 is used to communicate with other devices or other communication networks. Such other communication networks may be Ethernet, Radio Access Network (RAN), Wireless Local Area Network (WLAN), etc. Communication interface 130 may be a module, circuit, transceiver, or any other device capable of communication.

[0075] The memory 140 is used to store instructions, where the instructions may be computer programs.

[0076] The memory 140 may be a read-only memory (ROM) or other type of static storage device that can store static information and / or instructions, or a random access memory (RAM) or other type of dynamic storage device that can store information and / or instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compact disc, laser disc, optical disc, digital versatile disc, Blu-ray disc, etc.), magnetic disk storage medium or other magnetic storage device, etc., without limitation.

[0077] It should be noted that the memory 140 can exist independently of the processor 110 or can be integrated with the processor 110. The memory 140 can be used to store instructions, program codes, or some data. The memory 140 can be located inside the electronic device 100 or outside the electronic device 100, without limitation.

[0078] The processor 110 is configured to execute instructions stored in the memory 140 to implement the communication method provided in the following embodiments of the present application. For example, when the electronic device 100 is a terminal or a chip in a terminal, the processor 110 may execute instructions stored in the memory 140 to implement the steps performed by the transmitting end in the following embodiments of the present application.

[0079] As an optional implementation, the electronic device 100 further includes an output device 150 and an input device 160. The output device 150 may be a device such as a display screen or a speaker that can output data from the electronic device 100 to a user. The input device 160 may be a device such as a keyboard, a mouse, a microphone, or a joystick that can input data to the electronic device 100.

[0080] It should be noted that the structure shown in FIG1 does not constitute a limitation on the computing device. In addition to the components shown in FIG1 , the computing device may include more or fewer components than shown, or a combination of certain components, or a different arrangement of components.

[0081] The laser radar and inertial navigation fusion positioning device and application scenarios described in the embodiments of this application are intended to more clearly illustrate the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided in the embodiments of this application. Ordinary technicians in this field can know that with the evolution of laser radar and inertial navigation fusion positioning devices and the emergence of new business scenarios, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.

[0082] Figure 2 is a schematic diagram of a laser radar and inertial navigation system fusion positioning system provided in an embodiment of the present application. As shown in Figure 2, the system includes a computing device 210, a laser radar 220, and an inertial navigation system 230. In this embodiment of the present application, computing device 210 may be electronic device 100 in Figure 1. Computing device 210 is in communication with laser radar 220 and inertial navigation system 230, and laser radar 220 and inertial navigation system 230 are mounted on the vehicle whose position is to be determined.

[0083] It should be noted that during vehicle travel, LiDAR 220 can collect point cloud data at preset time intervals. Each time a frame of point cloud data is collected, LiDAR 220 sends it to computing device 210. Point cloud data refers to a collection of data consisting of multiple laser points. The data for each laser point includes the corresponding three-dimensional position coordinates and reflection intensity.

[0084] The inertial navigation system 230 is used to measure the linear acceleration and angular velocity of the vehicle at certain time intervals and send the measured linear acceleration and angular velocity to the computing device 210 .

[0085] The computing device 210 receives the point cloud data sent by the laser radar 220 and the data sent by the inertial navigation 230, and whenever the computing device 210 receives a frame of point cloud data, the computing device determines the current posture of the vehicle according to the laser radar and inertial navigation fusion positioning method provided in the embodiment of the present application. After determining the current posture of the vehicle, the computing device can send the determined current posture to the control system of the vehicle so that the control system of the vehicle makes driving decisions based on the determined current posture. Among them, driving decisions include adjusting the vehicle's route, adjusting the vehicle's speed, etc., which are not limited in the embodiment of the present application. Among them, the vehicle posture includes the position and posture of the vehicle.

[0086] It should be noted that in the embodiments of the present application, the computing device 210 may be a terminal device such as an in-vehicle terminal, a smart phone, a tablet computer, or a personal computer. Alternatively, the computing device 210 may also be a computing device of a cloud service platform, such as a server or server cluster of a cloud service platform, and the embodiments of the present application do not limit this. In addition, in the embodiments of the present application, the vehicle whose position is to be determined may be an unmanned vehicle or an intelligent vehicle with an autonomous driving function, and the embodiments of the present application do not limit this either.

[0087] Next, the laser radar and inertial navigation fusion positioning method is described in detail with reference to the accompanying drawings. Figure 3 is a flow chart of a laser radar and inertial navigation fusion positioning method provided by this application. This method is applied to the laser radar and inertial navigation fusion positioning device having the hardware structure shown in Figure 1, and specifically includes the following steps:

[0088] Step 310: Predict a second posture state of the vehicle based on the first posture state of the vehicle.

[0089] Obtain the vehicle's first corrected pose state in the first cycle, its first pose state covariance, the acceleration and angular velocity of the inertial navigation corresponding to the second cycle's lidar scan frame, and the point cloud of the second cycle's lidar scan frame in the lidar coordinate system. In one embodiment, the first cycle represents the previous cycle, and the second cycle represents the current cycle.

[0090] Taking the inertial navigation as the center, the vehicle's posture state can be expressed as:

[0091] in, is the attitude of the inertial navigation in the global coordinate system, is the position of the inertial navigation in the global coordinate system, is the velocity of the inertial navigation in the global coordinate system, is the zero bias of the inertial navigation gyroscope, is the zero bias of the inertial navigation accelerometer, is the acceleration due to gravity.

[0092] Furthermore, the second posture state of the vehicle is predicted based on the first posture state and the angular velocity and acceleration of the inertial navigation corresponding to the second period lidar scanning frame.

[0093] Specifically, in, is the first posture state, u i Including the angular velocity of the inertial navigation gyroscope and the acceleration of the inertial navigation accelerometer.

[0094] Because the frequency of the INS is much greater than that of the LiDAR, one LiDAR cycle is equal to multiple INS cycles. Therefore, the vehicle's second posture state (i.e., the INS posture state) is the posture state of the last cycle of the multiple INS cycles corresponding to one LiDAR cycle.

[0095] For example, one cycle of the laser radar is equal to three cycles of the inertial navigation, and the first position state of the vehicle after the first cycle correction is obtained. First posture state Equal to the posture state of the first cycle of the inertial navigation, that is, According to the above formula, the second periodic attitude state of the inertial navigation is The third cycle attitude state of inertial navigation This is the second posture state.

[0096] In one embodiment, the second posture state covariance of the vehicle is predicted based on the second posture state covariance, the state transfer matrix, the noise, and the noise matrix.

[0097] Specifically, in, is the state transfer matrix, Q is the noise, F W is the noise matrix.

[0098] First, the posture state and posture state covariance of the current cycle are predicted based on the posture state and posture state covariance of the vehicle in the previous cycle, and then the predicted posture state of the current cycle is corrected. This can effectively reduce the error caused by directly using the posture state and posture state covariance of the previous cycle to locate the current posture state of the vehicle.

[0099] Step 320: Observe the vehicle in the local point cloud map and the vector map, and construct an observation residual.

[0100] HD maps contain all the key static features of the road / environment required for autonomous driving (for example, roads, buildings, traffic lights, and road markings), including objects that cannot be detected by sensors due to occlusion. HD maps for autonomous driving are known for their high accuracy and rich geometric and semantic information. They are closely linked to the vehicle positioning function and continuously interact with different sensors, including lidar, radar, and cameras, to build the perception module of the autonomous driving system. This interaction ultimately supports the mission and motion planning of autonomous vehicles.

[0101] Therefore, vehicles are observed based on high-precision maps to construct observation residuals. High-precision maps are divided into high-precision vector maps and high-precision point cloud maps. High-precision point cloud maps are 3D reconstructions of real scenes and can rely on real-time lidar observations for high-precision positioning. High-precision vector maps contain semantic and geometric information such as lane lines, road markings, and traffic regulations. They can provide navigation information such as lane center lines and road markings, providing prior knowledge for the vehicle's lane keeping function.

[0102] In one implementation, a local point cloud map is constructed by accumulating each frame's point cloud. The current frame's point cloud is then matched against the local point cloud map to construct an observation residual. Matching against the constructed local point cloud map avoids the storage and computational overhead associated with using the global point cloud map as prior information, effectively reducing reliance on these resources.

[0103] First, in an embodiment of the present application, the laser radar collects point cloud data at a certain time interval and accumulates each frame of point cloud to construct a local point cloud map. Whenever the laser radar collects a frame of point cloud data, the collected frame of point cloud data is sent to the computing device. Based on this, after the laser radar collects the first point cloud data at the current moment, the first point cloud data is sent to the computing device. Accordingly, the computing device receives the first point cloud data sent by the laser radar and obtains at least one frame of point cloud data before the current moment, thereby obtaining multiple frames of point cloud data. Among them, a frame of point cloud data refers to a collection of data including multiple laser points. The data of each laser point includes the three-dimensional position coordinates and reflection intensity of the corresponding laser point.

[0104] It should be noted that since the point cloud data collected by the lidar is data in the radar coordinate system, and the radar coordinate system is constantly moving as the vehicle travels, in order to subsequently fuse the first point cloud data with the point cloud data collected by the lidar before the current moment to determine the vehicle's posture, the computing device can convert the received first point cloud data and the point cloud data before the current moment into the same coordinate system.

[0105] Secondly, all plane points and edge points of the current frame point cloud are extracted through plane and edge extraction algorithms.

[0106] In one embodiment, the plane smoothness is calculated based on the curvature of the point as an indicator for extracting the feature information of the current frame. This extraction method is to determine whether a point is a plane point or an edge point by calculating the relationship between all points in a set (the current frame point cloud). After obtaining the indicator of plane smoothness, the feature points can be divided into two categories: plane points and edge points. Plane points: points on a smooth plane in three-dimensional space, which have a small size difference from the surrounding points, low curvature, and low smoothness. Edge points: points on a sharp edge in three-dimensional space, which have a large size difference from the surrounding points, high curvature, and high smoothness. The points in the set are sorted according to the plane smoothness, and the points with a plane smoothness less than the threshold are regarded as plane points, and the points with a plane smoothness greater than the threshold are regarded as edge points.

[0107] Finally, find the first point closest to the above plane point and the second point closest to the above edge point in the local point cloud map of the current frame. And calculate the normal vector u of the plane point j , the direction vector v of the edge point j The observation residual is constructed based on the normal vector of the plane point and the direction vector of the edge point. The observation residual includes the plane point residual and the edge point residual.

[0108] The plane point residual can be expressed as:

[0109] The edge point residual can be expressed as:

[0110] in, represents the coordinates of point j in the global coordinate system, Represents the coordinates of the first or second point closest to point j in the local point cloud map of the current frame in the global coordinate system.

[0111] In another implementation, lane lines, poles, and the ground detected in the current frame's point cloud are matched against a high-precision vector map to construct observation residuals. This application uses only a relatively small high-precision vector map as prior information. Observation within the high-precision vector map further reduces reliance on storage and computing resources. Because the vector map is highly precise, it also improves global positioning accuracy.

[0112] Specifically, the lane line points, pole points and ground points in the current frame point cloud are extracted through semantic segmentation. In the embodiment of the present application, the lane line points and pole points are regarded as edge points, and the plane points are regarded as plane points. The lane lines, poles and ground in the vector map are discretized to obtain the first point to be matched and the second point to be matched. That is, the ground is divided into a grid according to the preset length and width, and the center point of the grid is used as the first point to be matched; the lane lines and poles are divided into line segments according to the preset length, and the midpoint of the line segment is used as the second point to be matched. The first point closest to the ground point is searched in the first point to be matched, and the second point closest to the lane line point or pole point is searched in the second point to be matched, and the normal vector of the ground point, the direction vector of the lane line point and the pole point are calculated. The observation residual is constructed based on the normal vector of the ground point and the direction vector of the lane line point and the pole point.

[0113] The specific steps for finding the closest points to lane points, pole points, and ground points in the high-precision vector map and constructing observation residuals based on the normal vectors of the ground points and the direction vectors of the lane points and pole points are the same as those for matching and constructing observation residuals in the local point cloud map, so they are not repeated here.

[0114] The observation residual indicates the degree of match between the LiDAR scan frame and the points it matches in the local point cloud map and vector map. When the residual is less than the threshold, the LiDAR scan frame is essentially consistent with the points it matches in the local point cloud map and vector map. This means the points it matches in the local point cloud map and vector map are the vehicle's actual position. When the residual is greater than the threshold, the LiDAR scan frame is inconsistent with the points it matches in the local point cloud map and vector map, and further correction of the vehicle's position is required.

[0115] Step 330: Correct the second posture state according to the observation residual to obtain a third posture state.

[0116] After calculating the observation residuals of the local point cloud map and the vector map, the second pose state is iteratively extended with the Kalman filter according to the covariance of the second pose state and the observation residual to obtain the third pose state until the third pose state meets the preset convergence conditions to obtain an accurate third pose state.

[0117] Specifically, set the number of iterations of the iterative extended Kalman filter γ = 0, and calculate the third posture state: Where J represents the Jacobian matrix; H represents the observation matrix containing all matching points: y j represents the observation noise of the lidar, Represents the error state of the second posture state; K represents the Kalman gain: K = PH T (HPH T +R) -1 , P represents the predicted covariance, that is, the second pose state covariance, and R represents the observation noise of all matching points.

[0118] Furthermore, the convergence condition is set as When the calculated posture state meets the above convergence conditions, This is the third posture state. It is the third pose state covariance.

[0119] The goal of revising the second pose state using iterative extended Kalman filtering is to ensure a high degree of match between the points in the LiDAR scan frame and the matching points in the local point cloud map and vector map. By continuously revising the second pose state to minimize residual errors, the vehicle's accurate position and attitude in the sub-point cloud map and vector map are determined, improving vehicle positioning accuracy.

[0120] For example, when the residual is less than a threshold, the points in the LiDAR scan frame are essentially consistent with the points matched in the local point cloud map and vector map. This means the points matched in the local point cloud map and vector map are the actual vehicle position. When the residual is greater than the threshold, the LiDAR scan frame is inconsistent with the points matched in the local point cloud map and vector map, and further correction of the vehicle's posture is required.

[0121] In another embodiment, after obtaining the accurate third posture state, all points of the current lidar frame are converted to the global coordinate system, and then all points of the current lidar frame are added to the local point cloud map constructed above.

[0122] By adding all the points of the current lidar frame to the local point cloud map constructed above, the local point cloud map can be expanded for use in positioning the vehicle's posture state in the next cycle.

[0123] Additionally, the size of the constructed local point cloud map is checked. When the size of the local point cloud map exceeds a first threshold, points in the local point cloud map that are far from the points in the current period's LiDAR scan frame are deleted. This effectively prevents the local point cloud map from storing too many point cloud locations, which can lead to excessive point cloud size and reliance on storage and computing resources.

[0124] It is understood that in order to implement the functions in the above embodiments, the computer device includes hardware structures and / or software modules corresponding to the execution of each function. It should be readily apparent to those skilled in the art that, in combination with the units and method steps of each example described in the embodiments disclosed in this application, the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed in a hardware-driven manner or in a computer software-driven hardware manner depends on the specific application scenario and design constraints of the technical solution.

[0125] Furthermore, as an implementation of the method embodiment shown in FIG3 above, the present application embodiment provides a laser radar and inertial navigation fusion positioning device. The embodiment of this device corresponds to the aforementioned method embodiment. For ease of reading, this embodiment will not further elaborate on the details of the aforementioned method embodiment, but it should be clear that the device in this embodiment can implement all the contents of the aforementioned method embodiment. Specifically, as shown in FIG4, the laser radar and inertial navigation fusion positioning device 400 includes:

[0126] The prediction module 410 is used to predict the second posture state of the vehicle based on the first posture state of the vehicle. The first posture state is used to indicate the corrected posture state of the vehicle in the first cycle, and the second posture state is used to indicate the predicted posture state of the vehicle in the second cycle. The first cycle is one cycle ahead of the second cycle.

[0127] The observation module 420 is used to observe the vehicle in the local point cloud map and the vector map and construct the observation residual.

[0128] The correction module 430 is used to correct the second posture state according to the observation residual to obtain a third posture state.

[0129] Furthermore, as shown in FIG4 , the observation module 420 is specifically used to accumulate point clouds within a preset period and construct a local point cloud map; match the first frame point cloud with the local point cloud map, and the first frame point cloud is used to indicate the second period lidar scanning frame.

[0130] Furthermore, as shown in FIG4 , the observation module 420 is specifically used to extract edge points and plane points of the first frame point cloud; search for the first point and the second point in the local point cloud map, and search for the first point and the second point in the local point cloud map, wherein the distance between the first point and the plane point is smaller than the distance between other points in the local point cloud map and the plane point, and the distance between the second point and the edge point is smaller than the distance between other points in the local point cloud map and the edge point.

[0131] Furthermore, as shown in FIG4 , the observation module 420 is specifically used to extract edge points and plane points of the first frame point cloud, where the edge points include lane line points and rod-shaped object points, and the plane points include ground points; perform discrete processing on the lane lines, rod-shaped objects, and ground in the vector map to obtain first points to be matched and second points to be matched; search for the first point in the first points to be matched, and search for the second point in the second points to be matched.

[0132] Furthermore, as shown in FIG4 , the observation module 420 is specifically configured to divide the ground into a grid according to a preset length and width, and use the center point of the grid as the first point to be matched; divide the lane lines and pole-shaped objects into line segments according to a preset length, and use the midpoint of the line segment as the second point to be matched.

[0133] Furthermore, as shown in Figure 4, the observation module 420 is specifically used to calculate the normal vector of the plane point and the direction vector of the edge point; calculate the plane point residual based on the normal vector, the coordinates of the plane point in the global coordinate system, and the coordinates of the first point in the global coordinate system; calculate the edge point residual based on the direction vector, the coordinates of the edge point in the global coordinate system, and the coordinates of the second point in the global coordinate system.

[0134] Furthermore, as shown in Figure 4, the correction module 430 is specifically used to predict the second posture state covariance of the vehicle based on the first posture state covariance of the vehicle, the first posture state covariance is used to indicate the corrected posture state covariance of the vehicle in the first period, and the second posture state covariance is used to indicate the predicted posture state covariance of the vehicle in the second period; iteratively extend the Kalman filter on the second posture state according to the second posture state covariance and the observation residual to obtain the third posture state, until the third posture state meets the preset convergence condition.

[0135] Furthermore, as shown in FIG4 , the prediction module 410 is further configured to predict the second posture state covariance of the vehicle based on the second posture state covariance, the state transfer matrix, the noise, and the noise matrix.

[0136] Furthermore, as shown in FIG4 , the prediction module 410 is specifically configured to predict the second posture state of the vehicle based on the first posture state and the angular velocity and acceleration of the inertial navigation corresponding to the second period lidar scanning frame.

[0137] Furthermore, as shown in FIG4 , the correction module 430 is further configured to convert the second-cycle laser radar scanning frame into a global coordinate system and add the second-cycle laser radar scanning frame to the local point cloud map.

[0138] Furthermore, as shown in FIG4 , the correction module 430 is also used to delete the third point in the local point cloud map when the size of the local point cloud map exceeds a first threshold, and the distance between the third point and the second-cycle lidar scanning frame is greater than a second threshold.

[0139] Furthermore, an embodiment of the present application provides an electronic device comprising a processor and a memory. The prediction module 410, observation module 420, and correction module 430 are stored in the memory as program units, and the processor executes the program units stored in the memory to implement corresponding functions. The processor includes a kernel, which retrieves the corresponding program units from the memory.

[0140] An embodiment of the present application provides a storage medium on which a program is stored. When the program is executed by a processor, the laser radar and inertial navigation fusion positioning method is implemented.

[0141] An embodiment of the present application provides a processor, which is used to run a program, wherein the laser radar and inertial navigation fusion positioning method is executed when the program is run.

[0142] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing an initialization program having the following method steps: predicting a second posture state of a vehicle based on a first posture state of the vehicle, the first posture state being used to indicate a corrected posture state of the vehicle in a first cycle, and the second posture state being used to indicate a predicted posture state of the vehicle in a second cycle, the first cycle being one cycle prior to the second cycle; observing the vehicle in a local point cloud map and a vector map, and constructing an observation residual; and correcting the second posture state based on the observation residual to obtain a third posture state.

[0143] Furthermore, the point clouds within a preset period are accumulated to construct a local point cloud map; the first frame point cloud is matched with the local point cloud map, and the first frame point cloud is used to indicate the second period laser radar scanning frame.

[0144] Furthermore, edge points and plane points of the first frame point cloud are extracted; the first point and the second point are searched in the local point cloud map, the distance between the first point and the plane point is smaller than the distance between other points and the plane point in the local point cloud map, and the distance between the second point and the edge point is smaller than the distance between other points and the edge point in the local point cloud map.

[0145] Furthermore, edge points and plane points of the first frame point cloud are extracted, where the edge points include lane line points and rod-shaped object points, and the plane points include ground points; the lane lines, rod-shaped objects and ground in the vector map are discretized to obtain first points to be matched and second points to be matched; the first point is searched in the first points to be matched, and the second point is searched in the second points to be matched.

[0146] Furthermore, the ground is divided into grids according to a preset length and width, and the center point of the grid is used as the first point to be matched; the lane lines and pole-shaped objects are divided into line segments according to a preset length, and the midpoint of the line segment is used as the second point to be matched.

[0147] Furthermore, the normal vector of the plane point and the direction vector of the edge point are calculated; the plane point residual is calculated based on the normal vector, the coordinates of the plane point in the global coordinate system, and the coordinates of the first point in the global coordinate system; the edge point residual is calculated based on the direction vector, the coordinates of the edge point in the global coordinate system, and the coordinates of the second point in the global coordinate system.

[0148] Furthermore, the second posture state covariance of the vehicle is predicted based on the first posture state covariance of the vehicle. The first posture state covariance is used to indicate the corrected posture state covariance of the vehicle in the first period, and the second posture state covariance is used to indicate the predicted posture state covariance of the vehicle in the second period. The second posture state is iteratively extended by Kalman filtering based on the second posture state covariance and the observation residual to obtain the third posture state until the third posture state meets the preset convergence condition.

[0149] Furthermore, the second posture state covariance of the vehicle is predicted based on the second posture state covariance, the state transfer matrix, the noise and the noise matrix.

[0150] Furthermore, the second posture state of the vehicle is predicted based on the first posture state and the angular velocity and acceleration of the inertial navigation corresponding to the second period lidar scanning frame.

[0151] Furthermore, the second-cycle lidar scanning frame is converted to the global coordinate system, and the second-cycle lidar scanning frame is added to the local point cloud map.

[0152] Furthermore, when the size of the local point cloud map exceeds a first threshold, the third point in the local point cloud map is deleted, and the distance between the third point and the second-cycle laser radar scanning frame is greater than a second threshold.

[0153] The present application is described with reference to the flow chart and / or block diagram of the method, device (system), and computer program product according to the embodiment of the present application. It should be understood that each flow process and / or box in the flow chart and / or block diagram and the combination of the flow process and / or box in the flow chart and / or block diagram can be realized by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processing machine or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device for realizing the function specified in one flow chart flow or multiple flows and / or one box or multiple boxes of the block diagram.

[0154] In a typical configuration, the device includes one or more processors (CPUs), memory, and a bus. The device may also include input / output interfaces, network interfaces, and the like.

[0155] Memory may include non-permanent memory in a computer-readable medium, random access memory (RAM), and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory includes at least one memory chip. Memory is an example of a computer-readable medium.

[0156] Computer-readable media includes permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0157] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0158] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0159] The above are merely embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should all be included within the scope of the claims of the present application.

Claims

1. A laser radar and inertial navigation fusion positioning method, characterized in that: The method comprises: Predicting a second posture state of the vehicle according to a first posture state of the vehicle, wherein the first posture state is used to indicate a corrected posture state of the vehicle in a first cycle, and the second posture state is used to indicate a predicted posture state of the vehicle in a second cycle, and the first cycle precedes the second cycle by one cycle; Observe the vehicle in the local point cloud map and the vector map to construct an observation residual; The second posture state is corrected according to the observation residual to obtain a third posture state.

2. The method according to claim 1, characterized in that Observing the vehicle in a local point cloud map includes: Accumulating point clouds within a preset period to construct the local point cloud map; The first frame point cloud is matched with the local point cloud map, and the first frame point cloud is used to indicate the second period laser radar scanning frame.

3. The method according to claim 2, characterized in that Matching the first frame point cloud with the local point cloud map includes: Extracting edge points and plane points of the first frame point cloud; Search for a first point and a second point in the local point cloud map, wherein a distance between the first point and the plane point is smaller than a distance between other points in the local point cloud map and the plane point, and a distance between the second point and the edge point is smaller than a distance between other points in the local point cloud map and the edge point.

4. The method according to claim 3, characterized in that Observing the vehicle in a vector map includes: Extracting edge points and plane points of the first frame point cloud, wherein the edge points include lane line points and rod-shaped object points, and the plane points include ground points; The lane lines, poles and ground in the vector map are discretized to obtain the first The point to be matched and the second point to be matched; The first point is searched among the first points to be matched, and the second point is searched among the second points to be matched.

5. The method according to claim 4, characterized in that The lane lines, poles and ground in the vector map are discretized, including: Divide the ground into grids according to preset length and width, and use the center point of the grid as the first point to be matched; The lane line and the rod-shaped object are divided into line segments according to a preset length, and the midpoint of the line segment is used as the second point to be matched.

6. The method according to claim 3 or 4, characterized in that: Construct observation residuals, including: Calculate the normal vector of the plane point and the direction vector of the edge point; Calculating a plane point residual according to the normal vector, the coordinates of the plane point in the global coordinate system, and the coordinates of the first point in the global coordinate system; The edge point residual is calculated according to the direction vector, the coordinates of the edge point in the global coordinate system, and the coordinates of the second point in the global coordinate system.

7. The method according to claim 1, characterized in that Correcting the second posture state according to the observation residual includes: Predicting a second posture state covariance of the vehicle according to a first posture state covariance of the vehicle, wherein the first posture state covariance is used to indicate a corrected posture state covariance of the vehicle in a first period, and the second posture state covariance is used to indicate a predicted posture state covariance of the vehicle in a second period; The second posture state is iteratively extended by Kalman filtering according to the covariance of the second posture state and the observation residual to obtain the third posture state until the third posture state meets a preset convergence condition.

8. The method according to claim 1, characterized in that Predicting a second posture state of the vehicle according to the first posture state of the vehicle includes: The second posture state of the vehicle is predicted based on the first posture state and the angular velocity and acceleration of the inertial navigation corresponding to the second periodic laser radar scanning frame.

9. The method according to claim 1, characterized in that: After correcting the second posture state according to the observation residual to obtain a third posture state, the method further includes: The second-cycle laser radar scanning frame is converted into a global coordinate system, and the second-cycle laser radar scanning frame is added to the local point cloud map.

10. The method according to claim 9, characterized in that The method further comprises: When the size of the local point cloud map exceeds a first threshold, a third point in the local point cloud map is deleted, and a distance between the third point and the second period laser radar scanning frame is greater than a second threshold.

11. A laser radar and inertial navigation fusion positioning system, characterized in that: The system includes: a processor, a laser radar and an inertial navigation system; The laser radar is used to scan the vehicle, obtain a first scanning frame, and send the first scanning frame to the processor; The inertial navigation system is used to scan the vehicle, obtain a second scanning frame, and send the second scanning frame to the processor; The processor is used to execute the laser radar and inertial navigation fusion positioning method as described in any one of claims 1-10.

12. A storage medium, characterized in that: The storage medium includes a stored program, wherein when the program is running, the device where the storage medium is located is controlled to execute the laser radar and inertial navigation fusion positioning method as described in any one of claims 1-10.

13. An electronic device, characterized in that: The device includes at least one processor, and at least one memory and a bus connected to the processor; wherein the processor and the memory communicate with each other through the bus; the processor is used to call the program instructions in the memory to execute the laser radar and inertial navigation fusion positioning method as described in any one of claims 1-10.

Images