CN113495281B

CN113495281B - Real-time positioning method and device for movable platform

Info

Publication number: CN113495281B
Application number: CN202110687064.3A
Authority: CN
Inventors: 李昱辰; 钱炜; 杨政; 何晓飞
Original assignee: Hangzhou Fabu Technology Co Ltd
Current assignee: Hangzhou Fabu Technology Co Ltd
Priority date: 2021-06-21
Filing date: 2021-06-21
Publication date: 2023-08-22
Anticipated expiration: 2041-06-21
Also published as: CN113495281A

Abstract

The application provides a real-time positioning method and a device of a movable platform, wherein laser radar equipment, positioning equipment and an inertial measurement unit are arranged on the movable platform, and the method comprises the following steps: acquiring point cloud data acquired by laser radar equipment, position data of a movable platform acquired by positioning equipment and running data acquired by an inertial measurement unit; obtaining estimated absolute pose, actual local pose and actual absolute pose of the movable platform according to the point cloud data, the position data and the driving data; and carrying out calibration processing on the actual local pose according to the estimated absolute pose and the actual absolute pose, and obtaining the calibrated local pose. Compared with the prior art, the position and posture of the movable platform are calibrated in real time through the point cloud data acquired by the laser radar equipment, the position data of the positioning equipment acquisition table and the driving data acquired by the inertial measurement unit, so that a high-precision real-time positioning result is obtained, and the robustness of the positioning method is improved.

Description

Real-time positioning method and device for movable platform

Technical Field

The present application relates to the field of real-time positioning technologies, and in particular, to a real-time positioning method and apparatus for a mobile platform.

Background

The method is characterized in that the position of the vehicle in the world coordinate system is obtained in real time by utilizing sensor information in automatic driving, which is the basis for realizing other functions in unmanned driving, and in an unmanned driving system, a laser radar is an indispensable sensor, and because of the high reliability of data, the high-precision real-time positioning is realized by utilizing the data obtained by the laser radar, which is a common positioning solution for the existing unmanned positioning.

In the existing scheme, a high-precision map of the point cloud drawn in advance and the point cloud obtained on line are utilized for matching to obtain real-time positioning.

However, the existing solutions have the following problems: when the offline map is established and the existing environment is greatly changed, such as surrounding parking vehicles, road construction, surrounding building construction and the like, the information of the offline map can not truly reflect the information of the existing environment, and if the offline map is directly matched through point cloud, the misplaced positioning information can be brought; according to the matching principle, the point cloud on the dynamic object is the point for introducing the online-offline matching error, and the problem that the positioning module is invalid is often caused by too few points which can be used for matching in the online point cloud when the dynamic vehicles around the unmanned vehicle are too many is removed. Therefore, the conventional scheme has a problem that the positioning method of the autonomous vehicle is poor in robustness.

Disclosure of Invention

The embodiment of the application provides a real-time positioning method and device for a movable platform, which are used for solving the problem of poor robustness of a positioning method for an automatic driving vehicle in the prior art.

A first aspect of the present application provides a real-time positioning method of a movable platform on which a laser radar device, a positioning device and an inertial measurement unit are mounted, the method comprising:

acquiring point cloud data acquired by the laser radar equipment, position data of the movable platform acquired by the positioning equipment and running data acquired by the inertial measurement unit;

obtaining estimated absolute pose, actual local pose and actual absolute pose of the movable platform according to the point cloud data, the position data and the driving data;

and carrying out calibration processing on the actual local pose according to the estimated absolute pose and the actual absolute pose to obtain a calibrated local pose.

In an alternative embodiment, the method for obtaining the estimated absolute pose, the actual local pose and the actual absolute pose of the movable platform according to the point cloud data, the position data and the driving data specifically includes:

Processing the driving data by using a dead reckoning method to obtain an estimated absolute pose of the movable platform;

and processing the point cloud data, the estimated absolute pose and the position data to obtain the actual local pose and the actual absolute pose of the movable platform.

In an alternative embodiment, processing the point cloud data, the estimated absolute pose and the position data to obtain an actual local pose of the movable platform specifically includes:

rasterizing the point cloud data according to the estimated absolute pose to obtain online point cloud grid map data of the area around the movable platform;

acquiring offline point cloud grid map data of an area around the movable platform according to the position data and the offline map data;

and obtaining the actual local pose of the movable platform according to the online point cloud grid map data and the offline point cloud grid map data.

In an alternative embodiment, the processing the point cloud data and the position data to obtain the actual absolute pose of the movable platform specifically includes:

Determining a gray map of the area around the movable platform according to the position data and the gray map of the whole area;

and obtaining the actual absolute pose of the movable platform according to the gray map of the area around the movable platform and the point cloud data.

In an alternative embodiment, the driving data includes angular velocity and acceleration, and the driving data is processed by dead reckoning method to obtain an estimated absolute pose of the movable platform, specifically including:

for each moment from the acquisition starting moment to the current moment, carrying out integral processing on the angular speed of each moment, the acceleration of each moment and the noise data of the last moment to obtain pre-integral data of each moment;

processing the pre-integral data from the acquisition starting time to each time of the current time to obtain an integral value of the current time;

and obtaining the estimated absolute pose of the movable platform according to the pre-integral value of the current moment and the gravity direction.

In an alternative embodiment, rasterizing the point cloud data according to the estimated absolute pose to obtain online point cloud raster map data of an area around the movable platform includes:

Performing de-distortion processing on the point cloud data according to the estimated absolute pose to obtain undistorted point cloud data;

constructing a plane model based on a road surface according to the undistorted point cloud data, and obtaining the travelable road surface data of the movable platform according to the centroid, the normal vector and the height threshold value of the plane model;

removing dynamic point cloud data in the undistorted point cloud data according to the drivable pavement data to obtain static point cloud data;

and rasterizing the static point cloud data according to a preset rule to obtain the online point cloud grid map data.

In an optional implementation manner, the undistorted point cloud data is obtained by performing a de-distortion process on the point cloud data according to the estimated absolute pose, and the method specifically includes:

predicting driving data of the laser radar equipment in acquisition time according to the estimated absolute pose, wherein the acquisition time is the time required by the laser radar equipment to rotate for one circle at a preset working frequency;

and removing distortion data caused by self motion in the acquisition time in the point cloud data according to the running data and the linear interpolation method to obtain the undistorted point cloud data.

In an alternative embodiment, obtaining offline point cloud grid map data of an area around the movable platform according to the location data and offline map data includes:

rasterizing the offline map data in the world coordinate system according to the preset rule to obtain the offline point cloud grid map data;

establishing a corresponding relation between the offline point cloud grid map data and the offline point cloud grid map data representation position according to a hash formula;

and determining offline point cloud grid map data of the surrounding area of the movable platform according to the position data and the corresponding relation.

In an alternative embodiment, obtaining the actual absolute pose of the movable platform according to the gray scale map of the area around the movable platform and the point cloud data comprises:

registering the point cloud data at the previous moment with the point cloud data at the current moment to obtain a reflection data registration result and a height data registration result; wherein the point cloud data includes reflection data and altitude data;

carrying out weighted average processing on the reflection data registration result and the height data registration result to obtain a height calculation value;

And registering the height calculated value with the height information in the gray map of the area around the movable platform to obtain the actual absolute pose.

In an alternative embodiment, the calibrating the actual local pose according to the estimated absolute pose and the actual absolute pose to obtain the calibrated local pose includes:

constructing a pose diagram of the movable platform according to the estimated absolute pose, the actual local pose and the actual absolute pose;

and optimizing the pose graph through the sliding window optimization method to obtain the calibrated local pose.

A second aspect of the present application provides a real-time positioning apparatus of a movable platform on which a laser radar device, a positioning device, and an inertial measurement unit are mounted, the apparatus comprising:

the acquisition module is used for acquiring point cloud data acquired by the laser radar equipment, position data of the movable platform acquired by the positioning equipment and running data acquired by the inertial measurement unit;

the processing module is used for obtaining estimated absolute pose, actual local pose and actual absolute pose of the movable platform according to the point cloud data, the position data and the driving data;

And the calibration module is used for carrying out calibration processing on the actual local pose according to the estimated absolute pose and the actual absolute pose to obtain the calibrated local pose.

In an alternative embodiment, the processing module is specifically configured to process the driving data using a dead reckoning method to obtain an estimated absolute pose of the movable platform; and processing the point cloud data, the estimated absolute pose and the position data to obtain the actual local pose and the actual absolute pose of the movable platform.

In an optional implementation manner, the processing module is specifically configured to perform rasterization processing on the point cloud data according to the estimated absolute pose, so as to obtain online point cloud grid map data of an area around the movable platform; acquiring offline point cloud grid map data of an area around the movable platform according to the position data and the offline map data; and obtaining the actual local pose of the movable platform according to the online point cloud grid map data and the offline point cloud grid map data.

In an alternative embodiment, the processing module is specifically configured to determine a gray map of an area around the movable platform according to the position data and the gray map of the whole area; and obtaining the actual absolute pose of the movable platform according to the gray map of the area around the movable platform and the point cloud data.

In an optional implementation manner, the driving data includes an angular velocity and an acceleration, and the processing module is specifically configured to integrate, for each time from a collection start time to a current time, the angular velocity at each time, the acceleration at each time, and noise data at a previous time, to obtain pre-integrated data at each time; processing the pre-integral data from the acquisition starting time to each time of the current time to obtain an integral value of the current time; and obtaining the estimated absolute pose of the movable platform according to the pre-integral value of the current moment and the gravity direction.

In an optional implementation manner, the processing module is specifically configured to perform de-distortion processing on the point cloud data according to the estimated absolute pose, so as to obtain undistorted point cloud data; constructing a plane model based on a road surface according to the undistorted point cloud data, and obtaining the travelable road surface data of the movable platform according to the centroid, the normal vector and the height threshold value of the plane model; removing dynamic point cloud data in the undistorted point cloud data according to the drivable pavement data to obtain static point cloud data; and rasterizing the static point cloud data according to a preset rule to obtain the online point cloud grid map data.

In an optional implementation manner, the processing module is specifically configured to predict, according to the estimated absolute pose, driving data of the lidar device within an acquisition time, where the acquisition time is a time required by the lidar device to rotate for one circle at a preset working frequency; and removing distortion data caused by self motion in the acquisition time in the point cloud data according to the running data and the linear interpolation method to obtain the undistorted point cloud data.

In an optional implementation manner, the processing module is specifically configured to rasterize offline map data in a world coordinate system according to the preset rule to obtain the offline point cloud grid map data; establishing a corresponding relation between the offline point cloud grid map data and the offline point cloud grid map data representation position according to a hash formula; and determining offline point cloud grid map data of the surrounding area of the movable platform according to the position data and the corresponding relation.

In an optional implementation manner, the processing module is specifically configured to register the point cloud data at the previous time and the point cloud data at the current time to obtain a reflection data registration result and a height data registration result; wherein the point cloud data includes reflection data and altitude data; carrying out weighted average processing on the reflection data registration result and the height data registration result to obtain a height calculation value; and registering the height calculated value with the height information in the gray map of the area around the movable platform to obtain the actual absolute pose.

In an alternative embodiment, the calibration module is configured to construct a pose map of the movable platform according to the estimated absolute pose, the actual local pose, and the actual absolute pose; and optimizing the pose graph through the sliding window optimization method to obtain the calibrated local pose.

A third aspect of the present application provides an electronic apparatus comprising: a processor and a memory;

the memory is used for storing a computer program;

the processor is configured to invoke and run a computer program stored in the memory for performing the method according to the first aspect.

A fourth aspect of the application provides a computer-readable storage medium storing a computer program for causing a computer to perform the method according to the first aspect.

A fifth aspect of the application provides a computer program product comprising a computer program which, when executed by a processor, implements the method according to the first aspect.

The application provides a real-time positioning method and a device of a movable platform, wherein laser radar equipment, positioning equipment and an inertial measurement unit are arranged on the movable platform, and the method comprises the following steps: acquiring point cloud data acquired by laser radar equipment, position data of a movable platform acquired by positioning equipment and running data acquired by an inertial measurement unit; obtaining estimated absolute pose, actual local pose and actual absolute pose of the movable platform according to the point cloud data, the position data and the driving data; and carrying out calibration processing on the actual local pose according to the estimated absolute pose and the actual absolute pose, and obtaining the calibrated local pose. Compared with the prior art, the method and the device have the advantages that the distortion caused by the movement of the laser radar equipment and the point cloud data of the dynamic object in the point cloud data are removed through the point cloud data acquired by the laser radar equipment and the driving data acquired by the inertial measurement unit, so that the accurate estimated absolute pose is obtained, the offline map obtained according to the position data is maintained according to the point cloud data, the dependence of the positioning technology on the offline map is overcome, the accurate positioning result can be obtained when the environment where the movable platform is located changes, the actual absolute pose of the movable platform in the two-dimensional map is obtained according to the point cloud data, and then the continuity and smoothness of the local poses obtained in different environments are ensured when the estimated absolute pose, the actual local pose and the actual absolute pose are calibrated, so that the robustness of the positioning method is improved.

Drawings

In order to more clearly illustrate the application or the technical solutions of the prior art, the following description of the embodiments or the drawings used in the description of the prior art will be given in brief, it being obvious that the drawings in the description below are some embodiments of the application and that other drawings can be obtained from them without inventive effort for a person skilled in the art.

Fig. 1 is a schematic view of an application scenario of a real-time positioning method of a movable platform according to an embodiment of the present application;

fig. 2 is a schematic structural diagram of a movable platform according to an embodiment of the present application;

fig. 3 is a flow chart of a real-time positioning method of a movable platform according to an embodiment of the present application;

FIG. 4 is a flowchart of another method for positioning a movable platform in real time according to an embodiment of the present application;

FIG. 5 is a flowchart of a real-time positioning method of a movable platform according to an embodiment of the present application;

fig. 6 is a schematic structural diagram of a real-time positioning device for a movable platform according to an embodiment of the present application;

fig. 7 is a schematic structural diagram of an electronic device according to an embodiment of the present application.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the embodiments of the present application more apparent, the technical solutions of the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application, and it is apparent that the described embodiments are some embodiments of the present application, but not all embodiments of the present application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.

The method is characterized in that the position of the vehicle in the world coordinate system is obtained in real time by utilizing sensor information in automatic driving, which is the basis for realizing other functions in unmanned driving, and in an unmanned driving system, a laser radar is an indispensable sensor, and because of the high reliability of data, the high-precision real-time positioning is realized by utilizing the data obtained by the laser radar, which is a common positioning solution for the existing unmanned positioning. In the existing scheme, a high-precision map of the point cloud drawn in advance and the point cloud obtained on line are utilized for matching to obtain real-time positioning.

In order to solve the problems, the application provides a real-time positioning method and device for a movable platform, wherein a laser radar device, a positioning device and an inertial measurement unit are arranged on the movable platform, offline data for positioning, which are determined according to position data acquired by the positioning device, are optimized through point cloud data acquired by the laser radar device, and the positioning data are calibrated according to an optimizing result, so that a positioning result is obtained, dependence on an offline map in a positioning technology is overcome, the problem of inaccurate positioning result caused by environmental change is avoided, and the robustness of the positioning method is further improved.

The application scenario of the present application is described below.

Fig. 1 is a schematic application scenario diagram of a real-time positioning method of a movable platform according to an embodiment of the present application. As shown in fig. 1, by the surrounding environment information of the movable platform 001 and removing the influence of the motion information and the surrounding dynamic information, the pose data of the movable platform 001 is determined, so that the real-time positioning of the movable platform 001 is realized without being influenced by the environment. Fig. 2 is a schematic structural diagram of a movable platform according to an embodiment of the present application, and as shown in fig. 2, the movable platform 001 includes: the laser radar device 002, the positioning device 003 and the inertial measurement unit 004 optimize and calibrate the pose data of the movable platform 001 through the laser radar device 002, the positioning device 003 and the inertial measurement unit 004 to obtain accurate real-time pose data, so that the robustness of the positioning technology is improved.

In the embodiment of the application, the device for realizing the real-time positioning function of the movable platform can be the movable platform, and also can be a device capable of supporting the realization of the function, such as a chip system, and the device can be installed in the movable platform. In the embodiment of the application, the chip system can be composed of chips, and can also comprise chips and other discrete devices.

It should be noted that, the application scenario of the technical solution of the present application may be the scenario in fig. 1, but is not limited thereto, and may be applied to other scenarios where real-time positioning of a movable platform is required.

It can be understood that the above-mentioned real-time positioning method of the movable platform can be implemented by the real-time positioning device of the movable platform provided by the embodiment of the application, and the real-time positioning device of the movable platform can be part or all of a certain device, for example, a chip on the movable platform.

The following takes a real-time positioning device of a movable platform integrated with or installed with related execution codes as an example, and specific embodiments are used for describing the technical scheme of the embodiment of the application in detail. The following embodiments may be combined with each other, and some embodiments may not be repeated for the same or similar concepts or processes.

Fig. 3 is a flow chart of a real-time positioning method of a movable platform according to an embodiment of the present application, where an execution subject of the embodiment is the movable platform, and relates to a specific process of real-time positioning of the movable platform. As shown in fig. 3, the method includes:

s101, acquiring point cloud data acquired by laser radar equipment, position data of a movable platform acquired by positioning equipment and running data acquired by an inertial measurement unit.

Wherein the point cloud data refers to a set of vectors in a three-dimensional coordinate system. The data collected by the lidar device is recorded in the form of points, each of which contains three-dimensional coordinates, possibly color data, altitude data or reflectance data.

The type of the positioning device is not limited in the embodiment of the application, and may be exemplified by a real-time dynamic measurement device (Real Time Kinematic, RTK).

An inertial measurement unit (Inertial Measurement Unit, IMU) is used in embodiments of the application to collect travel data for the movable platform.

Wherein the travel data includes angular velocity and acceleration.

S102, obtaining estimated absolute pose, actual local pose and actual absolute pose of the movable platform according to the point cloud data, the position data and the driving data.

The absolute pose is estimated to be current pose data of the movable platform, which is obtained according to running data acquired by the IMU. The actual local pose is pose data of the movable platform in the laser sensor, which is obtained by performing rasterization processing on the point cloud data, and then matching and updating the point cloud data with an offline grid map determined by the position data. The actual absolute pose is pose data obtained by comparing the height data and the reflection data in the point cloud data with a two-dimensional gray map of the position data and updating gray data in the gray map.

In the embodiment of the application, the distorted point cloud data caused by the movement of the laser radar equipment in the point cloud data is removed through the driving data, and the point cloud information of surrounding dynamic objects is removed according to the point cloud data, so that the error in the point cloud data is removed, and the robustness of the positioning method is improved.

S103, performing calibration processing on the actual local pose according to the estimated absolute pose and the actual absolute pose, and obtaining the calibrated local pose.

According to the estimated absolute pose, the actual absolute pose and the actual local pose, a pose graph of the movable platform is constructed, and the pose graph is calibrated according to a sliding window optimization method to obtain the calibrated local pose.

And when the edge is performed in the sliding window optimization method, constraint relation among pose data is reserved through Jacobian coefficients (Jacobian), so that the obtained pose data cannot be mutated, and continuity and smoothness of a real-time positioning result are ensured.

The embodiment of the application provides a real-time positioning method of a movable platform, wherein laser radar equipment, positioning equipment and an inertial measurement unit are arranged on the movable platform, and the method comprises the following steps: acquiring point cloud data acquired by laser radar equipment, position data of a movable platform acquired by positioning equipment and running data acquired by an inertial measurement unit; obtaining estimated absolute pose, actual local pose and actual absolute pose of the movable platform according to the point cloud data, the position data and the driving data; and carrying out calibration processing on the actual local pose according to the estimated absolute pose and the actual absolute pose, and obtaining the calibrated local pose. Compared with the prior art, the method and the device have the advantages that the distortion caused by the movement of the laser radar equipment and the point cloud data of the dynamic object in the point cloud data are removed through the point cloud data acquired by the laser radar equipment and the driving data acquired by the inertial measurement unit, so that the accurate estimated absolute pose is obtained, the offline map obtained according to the position data is maintained according to the point cloud data, the dependence of the positioning technology on the offline map is overcome, the accurate positioning result can be obtained when the environment where the movable platform is located changes, the actual absolute pose of the movable platform in the two-dimensional map is obtained according to the point cloud data, and then the continuity and smoothness of the local poses obtained in different environments are ensured when the estimated absolute pose, the actual local pose and the actual absolute pose are calibrated, so that the robustness of the positioning method is improved.

On the basis of the above embodiments, a method for positioning a movable platform of a first device of a dual system provided in the present application in real time will be further described. Fig. 4 is a flow chart of another real-time positioning method for a movable platform according to an embodiment of the present application, as shown in fig. 4, the method includes:

s201, acquiring point cloud data acquired by laser radar equipment, position data of a movable platform acquired by positioning equipment and running data acquired by an inertial measurement unit.

The technical terms, effects, features, and alternative embodiments of S201 may be understood with reference to S101 shown in fig. 3, and will not be described again here for repeated contents.

S202, obtaining estimated absolute pose, actual local pose and actual absolute pose of the movable platform according to the point cloud data, the position data and the driving data.

Optionally, processing the driving data by using a dead reckoning method to obtain an estimated absolute pose of the movable platform; and processing the point cloud data, the estimated absolute pose and the position data to obtain the actual local pose and the actual absolute pose of the movable platform.

In the embodiment of the application, the running data is processed by using a dead reckoning method, the mode of obtaining the estimated absolute pose of the movable platform is not limited, and for each moment from the acquisition starting moment to the current moment, the angular speed at each moment, the acceleration at each moment and the noise data at the last moment are subjected to integral processing to obtain pre-integral data at each moment; processing the pre-integral data from the acquisition starting time to each time of the current time to obtain an integral value of the current time; and obtaining the estimated absolute pose of the movable platform according to the pre-integral value at the current moment and the gravity direction.

The dead reckoning method is a method for reckoning the position at the next moment by measuring the moving distance and the moving direction under the condition of the position at the current moment.

In the embodiment of the application, the noise data comprise the noise data of the accelerometer and the noise data of the gyroscope.

In the embodiment of the application, the pre-integral data is only relevant to the integral time period and the time length from the starting time to the current time.

And integrating the pre-integrated data to obtain an integrated value, and obtaining an estimated absolute pose according to the gravity direction at the current moment. And estimating absolute pose as pose data of the movable platform at the next moment according to the running data at the current moment. The position data, the speed data, the azimuth data and the like of the pose data packet.

In the embodiment of the application, the processing of the point cloud data, the estimated absolute pose and the position data to obtain the actual local pose and the actual absolute pose of the movable platform comprises the following steps: rasterizing the point cloud data according to the estimated absolute pose to obtain online point cloud raster map data of the surrounding area of the movable platform; acquiring offline point cloud grid map data of an area around the movable platform according to the position data and the offline map data; and obtaining the actual local pose of the movable platform according to the online point cloud grid map data and the offline point cloud grid map data.

Specifically, performing rasterization processing on the point cloud data according to the estimated absolute pose to obtain online point cloud raster map data of an area around the movable platform, including: performing de-distortion processing on the point cloud data according to the estimated absolute pose to obtain undistorted point cloud data; constructing a plane model based on the road surface according to the undistorted point cloud data, and obtaining the travelable road surface data of the movable platform according to the centroid, the normal vector and the height threshold value of the plane model; removing dynamic point cloud data in the undistorted point cloud data according to the drivable pavement data to obtain static point cloud data; and rasterizing the static point cloud data according to a preset rule to obtain online point cloud grid map data.

Performing de-distortion processing on the point cloud data according to the estimated absolute pose to obtain undistorted point cloud data, wherein the de-distortion processing comprises the following steps: predicting the driving data of the laser radar equipment in the acquisition time according to the estimated absolute pose, wherein the acquisition time is the time required by the laser radar equipment to rotate for one circle at the preset working frequency; and removing distortion data of the laser radar equipment caused by self motion in the acquisition time in the point cloud data according to the running data and the linear interpolation method to obtain undistorted point cloud data.

In the embodiment of the present application, the setting of the preset operating frequency is not limited, and may be 10 hz.

The linear interpolation method is to construct a linear motion curve of the laser radar according to the driving data and interpolate the linear motion curve.

In the embodiment of the application, a plane model based on a road surface is constructed according to undistorted point cloud data, and the travelable road surface data of a movable platform is obtained according to the centroid, the normal vector and the height threshold value of the plane model; and removing dynamic point cloud data in the undistorted point cloud data according to the drivable pavement data to obtain static point cloud data.

Specifically, a plane model based on a road surface is constructed according to undistorted point cloud data, the centroid and normal vector of the plane model are obtained through the point cloud data, and the travelable road surface data of the movable platform are obtained according to a height threshold.

The movable platform can drive the road surface data, and the data such as guardrails, trees and the like outside the road surface on which the movable platform can drive are formed by the movable platform.

And whether each point cloud data is the dynamic point cloud data of the dynamic object in the driving road surface data or not can be judged according to the height threshold value, for example, other vehicles, pedestrians and the like, so that the influence of the dynamic point cloud data is removed, and the positioning accuracy is improved.

In the embodiment of the application, the setting of the height threshold is not limited, and can be described according to specific situations, and if the movable platform is an automatic driving vehicle, the height threshold can be set by counting the height information of vehicles on the market.

In the embodiment of the present application, the preset rule is not limited, and an exemplary grid may be set to 50cm by 50 cm.

Further, in the embodiment of the present application, obtaining offline point cloud grid map data of an area around a movable platform according to position data and offline map data includes: gridding the offline map data in the world coordinate system according to a preset rule to obtain offline point cloud grid map data; establishing a corresponding relation between the offline point cloud grid map data and the offline point cloud grid map data representing positions according to a hash formula; and determining offline point cloud grid map data of the surrounding area of the movable platform according to the position data and the corresponding relation.

In the embodiment of the application, the off-line point cloud grid map data is updated through the on-line point cloud grid map data after the actual local pose of the movable platform is obtained according to the on-line point cloud grid map data and the off-line point cloud grid map data.

Further, in the embodiment of the present application, processing the point cloud data and the position data to obtain an actual absolute pose of the movable platform includes: determining a gray map of the area around the movable platform according to the position data and the gray map of the whole area; and obtaining the actual absolute pose of the movable platform according to the gray map and the point cloud data of the area around the movable platform.

The method comprises the steps of performing de-distortion processing on point cloud data, removing dynamic point cloud data to obtain static point cloud data, and then obtaining height data and reflection data in the static point cloud data.

It is known that the altitude data and the reflection data characterize altitude information in the point cloud data.

Specifically, according to the gray map and the point cloud data of the surrounding area of the movable platform, the actual absolute pose of the movable platform is obtained, which comprises the following steps: registering the point cloud data at the previous moment with the point cloud data at the current moment to obtain a reflection data registration result and a height data registration result; the point cloud data comprises reflection data and height data; carrying out weighted average processing on the reflection data registration result and the height data registration result to obtain a height calculation value; and registering the height calculated value with the height information in the gray map of the surrounding area of the movable platform to obtain the actual absolute pose.

S203, performing calibration processing on the actual local pose according to the estimated absolute pose and the actual absolute pose to obtain the calibrated local pose.

Specifically, a pose diagram of the movable platform is constructed according to the estimated absolute pose, the actual local pose and the actual absolute pose; and optimizing the pose graph by a sliding window optimization method to obtain the calibrated local pose.

On the basis of the above embodiment, fig. 5 is a flow chart of a real-time positioning method of a movable platform according to an embodiment of the present application, as shown in fig. 5, including:

s301, acquiring point cloud data acquired by laser radar equipment, position data of a movable platform acquired by positioning equipment and running data acquired by an inertial measurement unit.

The following describes a case where the inertial measurement unit collects travel data.

According to the embodiment of the application, the angular velocity obtained by the angular velocity meter in the IMU, the acceleration obtained by the accelerometer and the IMU noise data assuming that the movable platform is stationary in the process of acquiring data by the IMU comprise the following steps: the noise data ba of the accelerometer and the noise data bg of the gyroscope are obtained through pre-integration calculation, and the required pre-integration data are respectively the position pre-integration quantity p and differential data of the position pre-integration quantity p on the noise data ba of the accelerometer and the noise data bg of the gyroscope:and->Velocity pre-integral data v and differential amount of the velocity pre-integral data v to noise data ba of accelerometer and noise data bg of gyroscope +.>And->Angle pre-integral data q and its differential to noise data bg of the gyroscope +.>These pre-integration data are only relevant to the time period of integration.

The following describes the case where the laser radar apparatus collects point cloud data.

The laser radar device works at the working frequency of 10 Hz, and the mechanical laser radar device rotates for one circle within 0.1s, so that surrounding geometric information is converted into point cloud data through the principle of light reflection.

The case where the positioning device collects position data of the movable platform will be described below.

Position data of the movable platform in a world coordinate system is obtained through RTK.

S302, processing the driving data by using a dead reckoning method to obtain an estimated absolute pose of the movable platform.

Because the working frequency of the IMU is 100 Hz, the pre-integral data integrated from the first frame is utilized, the time length of integration and the gravity direction are combined, the corresponding estimated absolute pose is generated according to the pre-integral data, and the absolute pose is output to other modules of the movable platform, such as a control module, a sensing module and the like, so that the movable platform can work subsequently. Wherein estimating absolute pose may include: position data P, speed data V, rotation data Q.

Wherein the first frame corresponds to the start time, each frame corresponds to one time, and the time sequence is consistent.

And S303, performing de-distortion processing on the point cloud data according to the estimated absolute pose to obtain undistorted point cloud data.

Specifically, because the laser radar device is moving in 0.1s, the point cloud data are not collected at a static point, each point cloud data contains distortion caused by the movement of the laser radar device, the estimated absolute pose is obtained through the IMU and is used as estimated movement data of the laser radar device, the distortion data of the movement of the laser radar device are eliminated according to the difference value between the time generated by the point cloud data obtained by scanning in each period and the time when the scanning starts by a linear interpolation method, and the undistorted point cloud data are finally obtained.

S304, constructing a plane model based on the road surface according to the undistorted point cloud data, and obtaining the travelable road surface data of the movable platform according to the centroid, the normal vector and the height threshold value of the plane model.

Specifically, point cloud data around a movable platform are possibly from dynamic objects, the point cloud data are against basic static assumption, a plane model based on a road surface is built through the undistorted point cloud data on the basis of undistorted point cloud data, the centroid and normal vector of the plane model are maintained through the point cloud data, after the point cloud data are input, the judgment is carried out according to the point-to-plane distance in the point cloud data, whether the point is a point in the plane model or not is judged, and after the plane is obtained, the travelable road surface data corresponding to the movable platform can be obtained.

And S305, removing dynamic point cloud data in the undistorted point cloud data according to the drivable pavement data to obtain static point cloud data.

The points on the movable platform on the movable road surface data are possible to be dynamic objects, and in order to avoid errors, the points above the movable road surface data are removed as dynamic points to obtain static point cloud data.

S306, determining offline point cloud grid map data of the surrounding area of the movable platform according to the position data and the corresponding relation.

In the embodiment of the application, the off-line point cloud grid map data is established to enable the initialization process of the positioning of the movable platform to be more stable, the establishment process mainly comprises the steps of collecting the point cloud data and the position data of a target area through the movable platform carrying the positioning hardware equipment of the RTK and the laser radar equipment, firstly removing the dynamic point cloud data obtained in the driving process through a manual removing means, then combining with the RTK to obtain the position data, obtaining off-line splice point cloud map data, forming dense point cloud map data of the area, and finally storing the cloud map data into the three-dimensional off-line point cloud grid map data according to the position data. The rasterization processing mode is as follows, the three-dimensional space is divided into grid cubes with the size of 50cm and the size of 50cm, each cube stores relevant point cloud data, wherein each point cloud data corresponds to a corresponding grid according to the position data, the stored relevant point cloud data are the position mean value, the second moment, the feature matrix, the final updating time and the confidence coefficient of the point cloud in the grid, and the final updating time and the confidence coefficient are used for post-matching optimization. The off-line point cloud map data are put into each grid according to the position data, and the corresponding position mean value, second moment, feature matrix, final updating time and confidence coefficient are calculated, wherein a variance matrix can be obtained after the mean value and the second moment are obtained, the distribution condition of the grid point cloud can be obtained after the variance matrix is decomposed through the feature values, and the distribution condition can be specifically divided into: the maximum value Pmax, the intermediate value Pmin and the minimum value Pmin, when Pmin/Pmax is less than 0.05 and Pmin/Pmax is less than 0.05, the grid can be considered as a straight line feature, when Pmin/Pmin is less than 0.2 and Pmin/Pmax is less than 0.2, the grid is considered as a plane feature, when all three feature values are less than 0.05, the grid is considered as a point feature, and finally, a feature matrix is generated according to the obtained feature of the grid so as to store corresponding information. In addition, in order to enable the point cloud data of the determined position data to quickly find out a proper grid, the method and the device utilize the hash table to quickly search according to the position data to obtain the grid of the corresponding position data, and accordingly obtain offline point cloud grid map data which is drawn in advance.

And S307, rasterizing the point cloud data according to the estimated absolute pose to obtain online point cloud raster map data of the surrounding area of the movable platform.

Through static point cloud data, the point cloud data are firstly converted into a local coordinate system of the movable platform from a laser radar coordinate system, wherein the point cloud data marked as ground points are independently formed into large grid plane data to form a plane constraint, and other point cloud data can form corresponding point plane constraint and point line constraint according to the characteristics of grids. In the embodiment of the application, pose data of the frame point cloud data under a local coordinate system is used as a state quantity, and an optimized actual local pose is finally obtained according to the formed constraint relation. In addition, according to the uncertain relation of the point cloud data, sampling is carried out in the ray direction, and the grid with the maximum confidence is added to be optimized.

And S308, obtaining the actual local pose of the movable platform according to the online point cloud grid map data and the offline point cloud grid map data.

Converting the point cloud data into a local coordinate system, adding the point cloud data into a corresponding grid according to the rasterization processing method in the step S306, and updating the information stored in the grid, wherein the updating modes of the position mean value, the second moment and the feature matrix are consistent with those in the step S306, and the confidence coefficient updating modes are as follows: and dividing the characteristic relation corresponding to the grid into a point-surface distance and a point-line distance, converting the distance value into Gaussian distribution with a mean value of 0 and a variance of 2, and accumulating the corresponding probability value to the confidence coefficient of the corresponding grid. And traversing all the point clouds to obtain an updated point cloud grid map and an actual local pose.

S309, according to the gray map and the point cloud data of the surrounding area of the movable platform, the actual absolute pose of the movable platform is obtained.

In the embodiment of the application, the RTK and the laser radar equipment are utilized to draw the target area in advance, wherein the RTK provides accurate world coordinate position data at any moment, and the laser radar equipment provides scene data of a peripheral scene at corresponding time, and the scene data comprises: reflection data and height data. In the embodiment of the application, the point cloud data represents the reflection data and the height data of the corresponding road section or area, and two kinds of data can generate two kinds of matching results when being matched on line, and then the two kinds of results are fused to optimize the final positioning result. Through the two data, various back-end optimization schemes are assisted, and reliable point cloud reflection map data and point cloud height map data are constructed.

In the embodiment of the application, the registration of the point cloud reflection map data and the point cloud height map data is mainly divided into two parts. The first part registers reflection data obtained by scanning the laser radar equipment in real time with point cloud reflection map data constructed in the prior art, and scene surfaces of different materials are presented in different reflectivity values in the scanning result of the laser radar equipment, so that the reflectivity can objectively and comprehensively express the material composition of the scene surfaces, and is a very important positioning basis. In the actual operation process, firstly, acquiring a real-time scanning result through laser radar equipment, extracting a reflectivity value in the real-time scanning result, then finding a corresponding position on point cloud reflection map data with the reflectivity value according to a real-time positioning result on a movable platform at the moment, setting a search area, and searching a position with the highest matching degree with the reflectivity value in the current scanning result in the search area to be used as a registration result; in addition, the second part registers the height data obtained by the laser radar device in real time with the previously constructed point cloud height map data, and the principle is that the point cloud data in the laser scanning result returned by the laser radar device has three-dimensional position information relative to the laser emission points, so that the height data of each laser point can be obtained, and the point cloud height map data can objectively and comprehensively express the height data of the surface in a scene. In theory, the registration result of the second part should have higher robustness than the registration result of the first part, mainly because the height data of surrounding objects are generally unchanged, but the reflection data of the laser point cloud is related to various factors such as the laser working state, the ambient humidity, the distance of the reflecting objects and the like, the robustness is relatively insufficient, and the matching result still has certain referential. Therefore, a reasonable fusion mode is needed for the two matching results to obtain a reasonable output result.

Two registration results are obtained, and during actual operation, it is found that, although theoretically, the two registration results should be completely consistent, in actual situationsBecause of scene information changes and the like, the two registration results are often inconsistent, and a method needs to be found to fuse the two registration results. In the embodiment of the application, the registration result with higher matching degree with the corresponding point cloud data is more reliable, so that an artificial bias term a=1.2 is added on the basis of the covariance-based adaptive fusion result. The covariance item of the matching process is obtained in the matching work of each part, wherein the covariance item of the matching result of the height data is m, the matching covariance item of the reflection data is n, and the fusion weight of the matching result of the final cloud height map data isThe fusion weight of the matching result of the point cloud reflection map data is +.>

And S310, performing calibration processing on the actual local pose according to the estimated absolute pose and the actual absolute pose to obtain the calibrated local pose.

In the embodiment of the application, when the whole state is not initialized, particularly when the state quantity speed data V, IMU noise and the like are not started with zero values, the corresponding values can be aligned in a linear or nonlinear mode, and the initialization processing is performed according to the obtained pre-integral data of the IMU and the pose ordering according to time sequence. Firstly, converting the direction of gravity into the position below a local coordinate system in the embodiment of the application, firstly initializing bg related to angular velocity in IMU noise, and assuming that bg is unchanged in local time, wherein the bg obtained through optimization is a nonlinear alignment result by mathematical expression; after the optimized result bg is obtained, the inter-frame position variation and the speed variation under a fixed coordinate system can be constructed through the well estimated gravity direction, a linear equation is constructed through the construction of the formula, the speed alignment quantity can be obtained through the linear solution of a matrix, and the process is a linear coarse speed alignment process; and then, according to the obtained result, constructing inter-frame speed item constraint and position item constraint in pre-integral constraint by taking the speed data V and the noise item ba of the IMU as state quantity, and finally constructing a nonlinear optimization problem, and solving to obtain nonlinear aligned V and ba.

According to the initialization result, the state quantity initialization quantity is obtained, and in the embodiment of the application, a fixed frame number is finally constructed according to the previous result, and a sliding optimization window mainly based on the mutual constraint of the pose is constructed. The frame unit in the window corresponds to a frame of point cloud data, and the state quantity contained in the frame data comprises position data P, speed data V, rotation data Q, IMU noise data ba and bg, external parameters directly of the laser radar equipment and the IMU, and the external parameters comprise position external parameters Ex.p and azimuth external parameters Ex.q, so that the marginal quantity of window sliding and the priori quantity reserved for global matching results are realized. In the embodiment of the application, the number of windows is kept fixed after operation, but the number can be set to be in the range of 10 frames to 100 frames according to the requirement initially. After the state quantity is determined, carrying out constraint construction according to the results obtained by the modules, and finally constructing a complete optimization problem. The IMU pre-integration quantity can form inter-frame pre-integration constraint in the optimization window, and the state quantity of the IMU pre-integration quantity comprises position data P, speed data V, rotation data Q, IMU noise data ba and bg; in addition, each frame has an obtained actual local pose, so that mutual pose diagram constraint among frames can be formed, and state quantities of the actual local pose are position data P, speed data V and rotation data Q, and the obtained actual local pose provides corresponding state optimization initial values and observation information; in addition, the actual absolute pose is converted to the position below a local coordinate system fixed by the application, a priori constraint of the frame is formed, the corresponding state quantity is position data P and speed data V, and the actual absolute pose obtained by conversion of the registration module is observed; finally, since the number of optimization windows is fixed, when the number of frames participating in optimization is greater than the limit number, two cases are discussed: the first case is that the module of the position difference between the frame of the next to last window and the frame of the next to last window is smaller than 5m and the time difference between the two frames is smaller than 0.1s, the situation directly discards the frame with the next to last window time and the corresponding data, in the other case, the data of the frame with the earliest window time is marginalized and marginalized to the marginalized amount of the frame after the frame, the parameters of the frame sliding outside the window are not optimized due to the fact that after the sliding window is introduced, the frame and the data in the sliding window still have constraint, the frame outside the window is directly discarded to cause constraint information loss, so the frame is packaged into prior information, and the prior information is added into a large nonlinear optimization problem to be used as a part of error, and the step is marginalized; therefore, the marginalizing the linearization state corresponding to a reserved frame when the reserved frame is marginalized includes: and finally, forming an marginalization constraint through the change caused by the change of the linearization point in the subsequent optimization. Finally, according to the result of pose graph optimization, the embodiment of the application can obtain an optimized estimation value of local pose and IMU noise data.

The local pose is used for maintaining the offline point cloud grid map data, the point cloud data grid map of the corresponding frame is used for updating the characteristic information of the offline point cloud grid map data, in addition, the estimation of the IMU noise data is used for updating the result of the state integral recurrence, and the local pose is used for real-time positioning of the movable platform.

Those of ordinary skill in the art will appreciate that: all or part of the steps for implementing the above method embodiments may be implemented by hardware associated with program instructions, where the foregoing program may be stored in a computer readable storage medium, and when executed, the program performs steps including the above method embodiments; and the aforementioned storage medium includes: various media that can store program code, such as ROM, RAM, magnetic or optical disks.

The embodiment of the application also provides a real-time positioning device for a movable platform, and fig. 6 is a schematic structural diagram of the real-time positioning device for a movable platform provided by the embodiment of the application, where the real-time positioning device for a movable platform can be implemented by software, hardware or a combination of the two, so as to execute the real-time positioning method for a movable platform in the embodiment. As shown in fig. 6, a laser radar device, a positioning device, and an inertial measurement unit are mounted on a movable platform, and a real-time positioning apparatus 400 of the movable platform includes: an acquisition module 401, a processing module 402 and a calibration module 403.

The acquisition module 401 is configured to acquire point cloud data acquired by the laser radar device, position data of the movable platform acquired by the positioning device, and running data acquired by the inertial measurement unit;

The processing module 402 is configured to obtain an estimated absolute pose, an actual local pose, and an actual absolute pose of the movable platform according to the point cloud data, the position data, and the driving data;

the calibration module 403 is configured to calibrate the actual local pose according to the estimated absolute pose and the actual absolute pose, and obtain a calibrated local pose.

In an alternative embodiment, the processing module 402 is specifically configured to process the driving data using a dead reckoning method to obtain an estimated absolute pose of the movable platform; and processing the point cloud data, the estimated absolute pose and the position data to obtain the actual local pose and the actual absolute pose of the movable platform.

In an optional implementation manner, the processing module 402 is specifically configured to perform rasterization processing on the point cloud data according to the estimated absolute pose, so as to obtain online point cloud raster map data of an area around the movable platform; acquiring offline point cloud grid map data of an area around the movable platform according to the position data and the offline map data; and obtaining the actual local pose of the movable platform according to the online point cloud grid map data and the offline point cloud grid map data.

In an alternative embodiment, the processing module 402 is specifically configured to determine a gray scale map of an area around the movable platform according to the position data and the gray scale map of the whole area; and obtaining the actual absolute pose of the movable platform according to the gray map and the point cloud data of the area around the movable platform.

In an alternative embodiment, the driving data includes angular velocity and acceleration, and the processing module is specifically configured to integrate, for each time from the acquisition start time to the current time, the angular velocity at each time, the acceleration at each time, and noise data at the previous time, to obtain pre-integrated data at each time; processing the pre-integral data from the acquisition starting time to each time of the current time to obtain an integral value of the current time; and obtaining the estimated absolute pose of the movable platform according to the pre-integral value at the current moment and the gravity direction.

In an alternative embodiment, the processing module 402 is specifically configured to perform de-distortion processing on the point cloud data according to the estimated absolute pose, so as to obtain undistorted point cloud data; constructing a plane model based on the road surface according to the undistorted point cloud data, and obtaining the travelable road surface data of the movable platform according to the centroid, the normal vector and the height threshold value of the plane model; removing dynamic point cloud data in the undistorted point cloud data according to the drivable pavement data to obtain static point cloud data; and rasterizing the static point cloud data according to a preset rule to obtain online point cloud grid map data.

In an alternative embodiment, the processing module 402 is specifically configured to predict, according to the estimated absolute pose, driving data of the lidar device within an acquisition time, where the acquisition time is a time required for the lidar device to rotate one circle at a preset working frequency; and removing distortion data caused by self motion in the point cloud data at the acquisition time according to the running data and the linear interpolation method to obtain undistorted point cloud data.

In an optional implementation manner, the processing module 402 is specifically configured to rasterize the offline map data in the world coordinate system according to a preset rule to obtain offline point cloud grid map data; establishing a corresponding relation between the offline point cloud grid map data and the offline point cloud grid map data representing positions according to a hash formula; and determining offline point cloud grid map data of the surrounding area of the movable platform according to the position data and the corresponding relation.

In an optional implementation manner, the processing module 402 is specifically configured to register the point cloud data at the previous time and the point cloud data at the current time to obtain a reflection data registration result and a height data registration result; the point cloud data comprises reflection data and height data; carrying out weighted average processing on the reflection data registration result and the height data registration result to obtain a height calculation value; and registering the height calculated value with the height information in the gray map of the surrounding area of the movable platform to obtain the actual absolute pose.

In an alternative embodiment, the calibration module 403 is configured to construct a pose map of the movable platform according to the estimated absolute pose, the actual local pose, and the actual absolute pose; and optimizing the pose graph by a sliding window optimization method to obtain the calibrated local pose.

It should be noted that, the real-time positioning device for a movable platform provided by the embodiment of the present application may be used to execute the method provided by any of the foregoing embodiments, and specific implementation manners and technical effects are similar, and are not repeated here.

Fig. 7 is a schematic structural diagram of an electronic device according to an embodiment of the present application. As shown in fig. 7, the first device 500 may include: at least one processor 501 and a memory 502. Fig. 7 shows a first device, for example a processor.

A memory 502 for storing a program. In particular, the program may include program code including computer-operating instructions.

The memory 502 may comprise high-speed RAM memory or may further comprise non-volatile memory (non-volatile memory), such as at least one disk memory.

The processor 501 is configured to execute computer-executable instructions stored in the memory 502 to implement the above-mentioned real-time positioning method of the mobile platform;

The processor 501 may be a central processing unit (Central Processing Unit, abbreviated as CPU), or an application specific integrated circuit (Application Specific Integrated Circuit, abbreviated as ASIC), or one or more integrated circuits configured to implement embodiments of the present application.

Alternatively, in a specific implementation, if the communication interface, the memory 502, and the processor 501 are implemented independently, the communication interface, the memory 502, and the processor 501 may be connected to each other through a bus and perform communication with each other. The bus may be an industry standard architecture (Industry Standard Architecture, abbreviated ISA) bus, an external device interconnect (Peripheral Component, abbreviated PCI) bus, or an extended industry standard architecture (Extended Industry Standard Architecture, abbreviated EISA) bus, among others. Buses may be divided into address buses, data buses, real-time positioning buses of a movable platform, etc., but do not represent only one bus or one type of bus.

Alternatively, in a specific implementation, if the communication interface, the memory 502, and the processor 501 are integrated on a chip, the communication interface, the memory 502, and the processor 501 may complete communication through an internal interface.

The present application also provides a computer-readable storage medium, which may include: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, etc., in which program codes can be stored, and specifically, the computer-readable storage medium stores program information for the real-time positioning method of the removable platform.

The embodiment of the application also provides a program which is used for executing the real-time positioning method of the movable platform provided by the embodiment of the method when being executed by a processor.

Embodiments of the present application also provide a program product, such as a computer readable storage medium, having instructions stored therein, which when run on a computer, cause the computer to perform the method for real-time positioning of a movable platform provided by the above method embodiments.

In the above embodiments, it may be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions in accordance with embodiments of the present application are produced in whole or in part. The computer may be a general purpose computer, a special purpose computer, a computer network, or other programmable apparatus. The computer instructions may be stored in or transmitted from one computer-readable storage medium to another, for example, by wired (e.g., coaxial cable, fiber optic, digital Subscriber Line (DSL)), or wireless (e.g., infrared, wireless, microwave, etc.) means from one website, computer, server, or data center. Computer readable storage media can be any available media that can be accessed by a computer or data storage devices, such as servers, data centers, etc., that contain an integration of one or more available media. The usable medium may be a magnetic medium (e.g., floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid State Disk (SSD)), etc.

Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some or all of the technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the invention.

Claims

1. A method of real-time positioning of a movable platform on which a lidar device, a positioning device and an inertial measurement unit are mounted, the method comprising:

performing calibration processing on the actual local pose according to the estimated absolute pose and the actual absolute pose to obtain a calibrated local pose;

Obtaining estimated absolute pose, actual local pose and actual absolute pose of the movable platform according to the point cloud data, the position data and the driving data, wherein the method specifically comprises the following steps:

processing the point cloud data, the estimated absolute pose and the position data to obtain an actual local pose and an actual absolute pose of the movable platform;

processing the point cloud data, the estimated absolute pose and the position data to obtain an actual local pose of the movable platform, wherein the processing comprises the following steps:

obtaining the actual local pose of the movable platform according to the online point cloud grid map data and the offline point cloud grid map data;

processing the point cloud data and the position data to obtain the actual absolute pose of the movable platform, wherein the processing comprises the following steps:

2. The method according to claim 1, wherein the travel data comprises angular velocity and acceleration, and wherein the travel data is processed using dead reckoning to obtain an estimated absolute pose of the movable platform, comprising:

3. The method of claim 1, wherein rasterizing the point cloud data according to the estimated absolute pose to obtain online point cloud raster map data for an area surrounding the movable platform comprises:

4. A method according to claim 3, wherein the performing a de-distortion process on the point cloud data according to the estimated absolute pose to obtain undistorted point cloud data, comprises:

and removing distortion data of the laser radar equipment, which is caused by self motion, in the point cloud data according to the running data and the linear interpolation method, so as to obtain the undistorted point cloud data.

5. The method of any of claims 1-4, wherein obtaining offline point cloud grid map data of an area surrounding the movable platform from the location data and offline map data comprises:

gridding the offline map data in the world coordinate system according to a preset rule to obtain the offline point cloud grid map data;

6. The method of claim 1, wherein obtaining an actual absolute pose of the movable platform from the gray scale map of the area surrounding the movable platform and the point cloud data comprises:

7. The method of claim 1, wherein calibrating the actual local pose based on the estimated absolute pose and the actual absolute pose to obtain a calibrated local pose comprises:

and optimizing the pose graph through a sliding window optimization method to obtain the calibrated local pose.

8. A real-time positioning apparatus of a movable platform on which a laser radar device, a positioning device, and an inertial measurement unit are mounted, the apparatus comprising:

The calibration module is used for carrying out calibration processing on the actual local pose according to the estimated absolute pose and the actual absolute pose to obtain a calibrated local pose;

the processing module is specifically used for processing the driving data by using a dead reckoning method to obtain an estimated absolute pose of the movable platform; processing the point cloud data, the estimated absolute pose and the position data to obtain an actual local pose and an actual absolute pose of the movable platform;

the processing module is specifically configured to perform rasterization processing on the point cloud data according to the estimated absolute pose, so as to obtain online point cloud grid map data of an area around the movable platform; acquiring offline point cloud grid map data of an area around the movable platform according to the position data and the offline map data; obtaining the actual local pose of the movable platform according to the online point cloud grid map data and the offline point cloud grid map data;

the processing module is specifically configured to determine a gray map of an area around the movable platform according to the position data and the gray map of the whole area; and obtaining the actual absolute pose of the movable platform according to the gray map of the area around the movable platform and the point cloud data.

9. An electronic device, comprising: a processor and a memory;

the memory is used for storing a computer program;

the processor is configured to invoke and run a computer program stored in the memory to perform the method of any of claims 1-7.

10. A computer readable storage medium storing a computer program for causing a computer to perform the method of any one of claims 1-7.