Autonomous navigation positioning system and method for autonomous vehicle

Abstract

The invention discloses an autonomous navigation positioning system and method for an autonomous vehicle, and belongs to the technical field of automatic driving, and the method specifically comprises the steps: receiving a satellite signal through a GPS receiver, preliminarily determining the position of the vehicle, starting a vehicle sensor to carry out real-time measurement, and preprocessing a result; the preprocessed measurement result is converted into position information in a global coordinate system, and the position and attitude of the vehicle are calculated in combination with a sensor fusion algorithm; generating point cloud data by using a laser radar, matching the point cloud data with a high-precision map to determine the accurate position of a vehicle, and identifying obstacle and road features; a vehicle-mounted camera captures an environment image, extracts visual feature information and fuses the visual feature information with environment perception information; on the basis of fusion information and distance information, the obstacle and road conditions are sensed in real time, the optimal driving route is determined by adopting a path planning algorithm, visual navigation and automatic obstacle avoidance are realized, and the safety and reliability of automatic driving are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of autonomous driving, and specifically relates to an autonomous navigation and positioning system and method for an autonomous driving vehicle. Background Art

[0002] Autonomous driving technology is a technology that uses advanced perception, decision-making, control and vehicle technologies to achieve autonomous driving, automatic obstacle avoidance and automatic path planning. Among them, autonomous navigation and positioning technology refers to the use of various sensors and algorithms to determine the vehicle's position in space and plan the optimal driving route. The development of autonomous navigation and positioning methods for autonomous vehicles has experienced a transition from single-sensor positioning to multi-sensor fusion positioning. In the early days, autonomous vehicles mainly relied on the global satellite navigation system for positioning. However, the positioning accuracy of the global satellite navigation system is limited in scenarios where satellite signals are easily blocked. With the continuous advancement of sensor technology and the reduction of costs, autonomous vehicles have begun to adopt multi-sensor fusion positioning methods. By integrating information from multiple sensors such as GNSS, INS, LiDAR, and cameras, vehicles can achieve more accurate and reliable positioning, but they also face shortcomings and challenges such as increased system complexity, increased data processing and computing burdens, mutual interference between sensors, sensor failures and redundancy issues, environmental adaptability challenges, and cost issues. In order to overcome these shortcomings and challenges, autonomous navigation and positioning algorithms for autonomous vehicles are continuously optimized and innovated. For example, the Kalman filter algorithm is used to fuse information from multiple sensors to improve positioning accuracy, and deep learning algorithms are used to identify and process visual information such as road signs and lane lines to assist vehicle positioning.

[0003] For example, the patent with publication number CN118408558A discloses a vehicle autonomous navigation method, device and computer equipment based on vehicle positioning, including: based on the image acquisition device on the target vehicle, collecting the current driving parameters of the target vehicle during driving, lane line information and obstacle information in the driving environment; determining the target planned path corresponding to the target vehicle and the real-time position information of the target vehicle according to the current driving parameters, lane line information and obstacle information; determining the planned driving parameters of the target vehicle according to the target planned path and the real-time position information of the target vehicle, and the planned driving parameters are used for autonomous navigation of the target vehicle.

[0004] For example, the patent with the authorization announcement number CN107462218B discloses a system and method for relative positioning of a panoramic vision of an autonomous navigation tractor at night, including: three sets of binocular vision systems, an Ethernet switch and an industrial computer; each set of vision systems includes two short-wave infrared cameras fixed in the same horizontal direction, and the three sets of binocular vision systems are arranged in an equilateral triangle to form a panoramic vision unit with a 120° angle between them, and are connected to the industrial computer through an Ethernet switch; the positioning method includes: the three sets of binocular vision systems simultaneously detect the same target in the environment in their respective directions, and infer the vehicle movement based on this, forming displacement vectors in different coordinate systems, and then converting them to the same coordinate system to achieve relative positioning. This technical solution has a reasonable design, a simple structure, and is easy to operate. By detecting the vehicle's motion status from multiple directions and fusing the detection results, the positioning accuracy of the tractor at night is effectively improved, providing data support for navigation.

[0005] The above existing technologies all have the following problems: poor dependence and environmental adaptability; although obstacle avoidance and path planning can be performed through visual information, their obstacle avoidance and path planning capabilities may be limited due to the camera's field of view and resolution, as well as the influence of ambient light and weather conditions. Summary of the invention

[0006] In view of the shortcomings of the prior art, the present invention proposes an autonomous navigation and positioning system and method for an autonomous driving vehicle, which receives satellite signals through a GPS receiver and preliminarily determines the vehicle position, starts the vehicle sensor to perform real-time measurement and preprocesses the results; converts the preprocessed measurement results into position information in a global coordinate system, and calculates the vehicle position and posture in combination with a sensor fusion algorithm; uses a lidar to generate point cloud data, which is matched with a high-precision map to determine the precise position of the vehicle, and identifies obstacles and road features; uses an on-board camera to capture environmental images, extracts visual feature information and fuses it with environmental perception information; based on the fused information and distance information, perceives obstacles and road conditions in real time, and uses a path planning algorithm to determine the optimal driving route, thereby realizing visual navigation and automatic obstacle avoidance, and improving the safety and reliability of autonomous driving.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] An autonomous navigation and positioning method for an autonomous driving vehicle, comprising:

[0009] Step S1: Receive the signal transmitted by the satellite through the GPS receiver, and calculate the distance error from the receiver to the satellite by combining the carrier phase measurement method, so as to preliminarily determine the position of the vehicle, start the vehicle sensor to perform real-time measurement based on the preliminarily determined vehicle position, and pre-process the measurement results;

[0010] Step S2: convert the preprocessed measurement results into position information in the global coordinate system, combine with the preliminary determination of the vehicle's position, and use the fusion algorithm to calculate the position and posture of the vehicle in space;

[0011] Step S3: Based on the position and posture of the vehicle in space, the laser radar emits a laser beam, calculates the relative distance between the vehicle and the surrounding environment according to the reflection time, generates point cloud data, and matches the point cloud data with the high-precision map in the vehicle system. The precise position of the vehicle in the global coordinate system is determined through the matching strategy, and the three-dimensional characteristics of the point cloud data are used to identify obstacles and road features in the environment to generate environmental perception information;

[0012] Step S4: The on-board camera captures the environment image around the vehicle, analyzes the environment image, extracts the visual feature information of road signs, lane lines and obstacles, and fuses the extracted visual feature information with the environment perception information to obtain fused feature information;

[0013] Step S5: Based on the fusion feature information and distance information, a path planning algorithm is used to determine the best driving route, and through visual navigation technology, the vehicle automatically drives along the extracted lane line, automatically takes obstacle avoidance measures when encountering obstacles, and updates its position and posture information.

[0014] Specifically, the specific steps of step S1 include:

[0015] S1.1: The GPS receiver receives radio frequency signals from multiple GPS satellites, and down-converts and digitizes the radio frequency signals to obtain baseband signals;

[0016] S1.2: Extract carrier phase information from the baseband signal, and calculate the phase difference between the satellite carrier signal received by the receiver and the receiver local oscillator reference signal based on the carrier phase information. in, Indicates the phase of the satellite carrier signal received by the receiver. Represents the phase of the local reference signal generated by the receiver;

[0017] S1.3: Calculate the precise distance from the receiver to the satellite based on the wavelength in the carrier phase information and the phase difference obtained in step S1.2 to obtain the satellite ranging result Among them, d n Indicates the ranging result of the nth satellite, Indicates the phase difference between the nth satellite carrier signal received by the receiver and the receiver local oscillator reference signal, λ n It represents the carrier wavelength of the nth satellite, where n represents the number of satellites.

[0018] Specifically, the specific steps of step S1 also include:

[0019] S1.4: According to the satellite ranging results, correct the initial position, the formula is:

[0020]

[0021] Among them, d n,cor It represents the precise distance from the calibrated receiver to the nth satellite. represents the phase error caused by multipath effect, α dian represents the ionospheric delay correction factor, α dian represents the tropospheric delay correction factor, α wei,t represents the satellite clock correction factor, α jie,t represents the receiver clock error correction factor, and β represents the adjustment coefficient;

[0022] S1.5: Based on the corrected preliminary position information, the vehicle sensor is started, the sensor performs real-time measurement, obtains vehicle dynamic information, and performs preprocessing operations on the vehicle dynamic information measured by the sensor.

[0023] Specifically, the specific steps of step S3 include:

[0024] S3.1: Based on the position and posture of the vehicle in space, the laser radar performs an all-round scan of the vehicle's surrounding environment through array scanning, and uses the difference Δt between the time of sending the laser and the time of receiving the laser and the speed of light c to calculate the relative distance between the vehicle and each point in the surrounding environment.

[0025] S3.2: Obtain the scanning angle information of the laser radar, calculate the position of each scanning point in the three-dimensional space according to the scanning angle information and relative distance data, and present it in the form of point cloud data;

[0026] S3.3: Match the generated point cloud data with the high-precision map in the vehicle system, find the features in the point cloud and the corresponding features in the map, and iteratively adjust the position of the vehicle in the global coordinate system based on the matching results;

[0027] S3.4: According to the position adjustment result in step S3.3, the precise position of the vehicle in the global coordinate system is determined by a matching strategy in combination with the preliminary position information and the vehicle sensor data;

[0028] S3.5: Using the three-dimensional characteristics of point cloud data and the precise position information of the vehicle in the global coordinate system, segment and classify the point cloud data to identify obstacles and road features in the environment;

[0029] S3.6: Calculate the distance and speed of the obstacle relative to the vehicle based on the position and shape of the obstacle, and combine it with the road feature information to generate environmental perception information including the obstacle and road features.

[0030] Specifically, the specific steps of the matching strategy in S3.4 include:

[0031] S3.41: preprocessing the vehicle sensor data, extracting feature points from the preprocessed vehicle sensor data using a feature point detection algorithm, and describing the extracted feature points to generate feature descriptors;

[0032] S3.42: Construct feature points and their descriptors in the high-precision map into a feature database, and use a distance-based matching algorithm to match the feature points collected by the vehicle sensor with the feature points in the feature database to find the most similar matching pairs;

[0033] S3.43: Verify the matching result. If the verification is successful, calculate the initial position of the vehicle in the global coordinate system using the shortest distance formula based on the successfully verified feature point pairs;

[0034] S3.44: Combining the position adjustment result in step S3.3 and the preliminary position information, the preliminary position is optimized using an optimization algorithm, and the optimized vehicle position information is output to the vehicle control system.

[0035] Specifically, the specific steps of step S5 include:

[0036] S5.1: receiving fused feature information and distance information, and integrating the fused feature information and distance information to form vehicle perception data;

[0037] S5.2: Based on the vehicle perception data, the vehicle movement scene is abstracted into an environmental model, and a path planning algorithm is used to search in the known environmental model to determine the global optimal driving route from the current position to the target position;

[0038] S5.3: Perform local path planning based on the global path, the real-time position of the vehicle and the environmental perception information, generate local path planning results, and generate decision information based on the local path planning results and the vehicle status information;

[0039] S5.4: Use a path tracking algorithm to control the vehicle to travel along the global optimal driving route based on the generated decision information;

[0040] S5.5: Based on the global optimal driving route, the visual navigation technology and edge detection algorithm are used to extract lane line information from the road image, and the position and posture of the vehicle relative to the lane line are calculated based on the lane line information. At the same time, steering and speed control instructions are generated so that the vehicle can automatically drive along the extracted lane line;

[0041] S5.6: During driving, the vehicle detects obstacles ahead in real time and determines whether to take obstacle avoidance measures based on the location and speed of the obstacles. If obstacle avoidance is required, the vehicle replans the driving path based on the local path planning algorithm.

[0042] An autonomous navigation and positioning system for an autonomous driving vehicle, comprising: a positioning module, an environment perception module, a path planning module, and a decision control module;

[0043] The positioning module is used to start the sensor to perform real-time measurement according to the initially determined vehicle position, and fuse the measurement results to obtain the precise position and posture of the vehicle in space;

[0044] The environment perception module is used to generate point cloud data using laser radar and match it with a high-precision map to determine the precise location of the vehicle and identify obstacles and road features in the environment;

[0045] The path planning module is used to perceive surrounding obstacles and road conditions in real time based on the fusion feature information and distance information, determine the best driving route, and perform automatic navigation and obstacle avoidance;

[0046] The decision control module is used to make driving decisions and control vehicle execution based on the results of environment perception and path planning.

[0047] Specifically, the positioning module includes: a GPS receiver unit, a carrier phase measurement unit, a data conversion unit, and a position and attitude calculation unit;

[0048] The GPS receiver unit is used to receive signals transmitted by satellites and provide preliminary position information for the vehicle;

[0049] The carrier phase measurement unit is used to calculate the distance error from the receiver to the satellite in combination with the GPS signal to improve the positioning accuracy;

[0050] The data conversion unit is used to convert the preprocessed sensor measurement results into position information in the global coordinate system;

[0051] The position and posture calculation unit is used to calculate the position and posture of the vehicle in space by using a sensor fusion algorithm.

[0052] Specifically, the path planning module includes: a path planning unit, a visual navigation unit, and an obstacle avoidance execution unit;

[0053] The path planning unit is used to determine the best driving route based on the fused feature information and the current state of the vehicle;

[0054] The visual navigation unit is used to navigate using visual information;

[0055] The obstacle avoidance execution unit is used to automatically take obstacle avoidance measures when encountering an obstacle.

[0056] Specifically, the decision control module includes: a decision unit and a control unit;

[0057] The decision-making unit is used to evaluate the current environmental conditions and select a driving mode;

[0058] The control unit is used to generate a control command according to the decision result, and directly control the hardware system of the vehicle to execute according to the control command.

[0059] Compared with the prior art, the present invention has the following beneficial effects:

[0060] 1. The present invention proposes an autonomous navigation and positioning method for an autonomous driving vehicle. By combining GPS signals, sensor measurements and lidar point cloud data, the system can determine the position of the vehicle in the global coordinate system in real time with high precision, while identifying obstacles and road features in the surrounding environment, thereby improving the vehicle's environmental perception capabilities.

[0061] 2. The present invention proposes an autonomous navigation and positioning method for an autonomous driving vehicle. Based on the fused feature information and distance data, the system can intelligently plan the optimal driving route, realize visual navigation and automatic obstacle avoidance. During the entire process, the vehicle continuously updates its position and posture information, realizing precise navigation and intelligent driving, thereby improving the safety and reliability of autonomous driving. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 This is a schematic diagram of an autonomous navigation and positioning method for an autonomous driving vehicle of the present invention;

[0063] Figure 2 This is a principle flow chart of an autonomous navigation and positioning method for an autonomous driving vehicle of the present invention;

[0064] Figure 3 This is an architecture diagram of an autonomous navigation and positioning system for an autonomous driving vehicle of the present invention. DETAILED DESCRIPTION

[0065] Example 1

[0066] See also Figure 1-Figure 2, an embodiment of the present invention provides: an autonomous navigation and positioning method for an autonomous driving vehicle, comprising the following steps:

[0067] Step S1: Receive the signal transmitted by the satellite through the GPS receiver, and calculate the distance error from the receiver to the satellite by combining the carrier phase measurement method, so as to preliminarily determine the position of the vehicle, start the vehicle sensor to perform real-time measurement based on the preliminarily determined vehicle position, and pre-process the measurement results;

[0068] Step S2: convert the preprocessed measurement results into position information in the global coordinate system, combine with the preliminary determination of the vehicle's position, and use the fusion algorithm to calculate the position and posture of the vehicle in space;

[0069] Furthermore, the specific steps of step S2 include:

[0070] S2.1: Determine the origin, direction and unit length of the global coordinate system, and convert the preprocessed measurement results into coordinate information in the global coordinate system through coordinate transformation, rotation and translation operations;

[0071] S2.2: The converted coordinate information and the GPS positioning results are input into the Kalman filter for fusion, and the position and posture of the vehicle in space are obtained according to the output results of the Kalman filter. The Kalman filter is the prior art content in this field and is not the creative solution of this application, so it will not be elaborated here.

[0072] Step S3: Based on the position and posture of the vehicle in space, the laser radar emits a laser beam, calculates the relative distance between the vehicle and the surrounding environment according to the reflection time, generates point cloud data, and matches the point cloud data with the high-precision map in the vehicle system. The precise position of the vehicle in the global coordinate system is determined through the matching strategy, and the three-dimensional characteristics of the point cloud data are used to identify obstacles and road features in the environment to generate environmental perception information;

[0073] Step S4: The on-board camera captures the environment image around the vehicle, uses the computer vision algorithm to analyze the environment image, extracts the visual feature information of the road signs, lane lines and obstacles, and fuses the extracted visual feature information with the environmental perception information to obtain fused feature information;

[0074] Furthermore, the specific steps of step S4 include:

[0075] (1) Use an on-board camera to capture high-definition images of the vehicle's surroundings. Ensure that the camera has high resolution and a wide-angle field of view to capture more environmental details.

[0076] (2) Preprocessing the captured image, including denoising, grayscale conversion, contrast enhancement, etc., using a median filtering method to remove noise in the image and converting the color image into a grayscale image to reduce the amount of calculation, wherein the median filtering method and the grayscale conversion method are prior art contents in the art and are not the inventive solution of the present application, and are not described in detail here;

[0077] (3) Using an edge detection algorithm to extract visual feature information of road signs, lane lines and obstacles, and using a threshold segmentation algorithm to segment the image into different regions, thereby extracting targets such as road signs, lane lines and obstacles. The edge detection algorithm and the threshold segmentation algorithm are prior art contents in this field and are not the inventive solution of the present application, and are not described in detail here.

[0078] (4) Fusing the extracted visual feature information with the environmental perception information provided by the vehicle’s lidar to obtain a more accurate and comprehensive environmental perception;

[0079] (5) Output the fused feature information to the vehicle’s autonomous driving system.

[0080] Step S5: Based on the fusion of feature information and distance information, the vehicle perceives the surrounding obstacles and road conditions in real time, uses a path planning algorithm to determine the best driving route, and uses visual navigation technology to automatically drive along the extracted lane lines. When encountering obstacles, it automatically takes obstacle avoidance measures and updates its position and posture information.

[0081] The specific steps of step S1 include:

[0082] S1.1: The GPS receiver receives radio frequency signals from multiple GPS satellites, and down-converts and digitizes the radio frequency signals to obtain baseband signals;

[0083] Furthermore, the specific steps of S1.1 include:

[0084] (1) RF signal reception: The GPS receiving antenna senses various electromagnetic fields and interferences, including all visible GPS signals, and transmits these RF signals to the RF front-end processing module;

[0085] (2) RF front-end processing: The RF front-end processing module filters and amplifies the signals of all visible GPS satellites received by the antenna, and then mixes them with the sine wave local oscillator signal generated by the local oscillator to down-convert them into intermediate frequency signals. Among them, the RF front-end electronic components are generally integrated into an application-specific integrated circuit chip ASIC, usually called a radio frequency integrated circuit RFIC;

[0086] (3) Digital processing: The analog intermediate frequency signal output by the RF front end is converted into a digital intermediate frequency signal through an analog-to-digital converter for subsequent digital signal processing.

[0087] S1.2: Extract carrier phase information from the baseband signal, and calculate the phase difference between the satellite carrier signal received by the receiver and the receiver local oscillator reference signal based on the carrier phase information. in, Indicates the phase of the satellite carrier signal received by the receiver. represents the phase of the local reference signal generated by the receiver, and in the calculation It needs to be converted to radians. If the phase difference is given in degrees, it needs to be converted to radians first.

[0088] It should be noted that when calculating the phase difference, the receiver simultaneously generates a local reference signal with the same frequency as the carrier signal transmitted by the satellite, and obtains the phase difference by comparing the phase of the received satellite carrier signal with the phase of the local reference signal.

[0089] It should be noted that since the carrier signal is a periodic sinusoidal signal, the phase measurement can only measure the part that is less than one wavelength, so there is a problem of integer number uncertainty. In step S1.4 of the present invention, it is necessary to correct the integer ambiguity problem to obtain an accurate phase difference measurement value.

[0090] Furthermore, the specific steps of extracting carrier phase information from the baseband signal include:

[0091] (1) GPS satellites send out modulated signals containing ranging codes based on their onboard clocks. At the same time, the receivers receive these modulated signals containing ranging codes and perform preliminary signal processing.

[0092] (2) Since the carrier signal is modulated on the ranging code, it is necessary to first remove the ranging code and satellite message modulated on the carrier and reacquire the carrier, that is, reconstruct the carrier. The method for reconstructing the carrier may adopt a code correlation method, which is a prior art in the art and is not an inventive solution of the present application, and will not be described in detail herein.

[0093] (3) After reconstructing the carrier, a phase-locked loop circuit is used to achieve phase locking. After phase locking, the phase of the local signal of the receiver is the same as the phase of the GPS carrier signal. At this time, the carrier phase information is extracted. Among them, the phase-locked loop circuit is the existing technology content in this field, and it is not an inventive solution of this application, and it will not be described here.

[0094] S1.3: Calculate the precise distance from the receiver to the satellite based on the wavelength in the carrier phase information and the phase difference obtained in step S1.2 to obtain the satellite ranging result Among them, d n Indicates the ranging result of the nth satellite, Indicates the phase difference between the nth satellite carrier signal received by the receiver and the receiver local oscillator reference signal, λ n represents the carrier wavelength of the nth satellite, where n represents the number of satellites;

[0095] The carrier wavelength is known. For the GPS system, the L1 carrier wavelength is approximately 19.03 cm, and the L2 carrier wavelength is approximately 24.42 cm.

[0096] S1.4: According to the satellite ranging results, correct the initial position, the formula is:

[0097]

[0098] Among them, d n,cor It indicates the precise distance from the receiver to the satellite after correction. represents the phase error caused by multipath effect, α dian Represents the ionospheric delay correction factor, which is used to correct the influence of the ionosphere on signal propagation, α dian represents the tropospheric delay correction factor, which is used to correct the impact of the troposphere on signal propagation, α wei,t Represents the satellite clock correction factor, which is used to correct the satellite clock error, α jie,t It represents the receiver clock error correction factor, which is used to correct the receiver clock error, and β represents the adjustment coefficient;

[0099] It should be noted that the formula in the present invention comprehensively considers various factors that affect the ranging accuracy, including multipath effect, ionospheric delay, tropospheric delay, satellite clock error and receiver clock error, and considers other possible error sources by introducing a correction factor β. This comprehensive consideration makes the ranging result more accurate and reliable, and helps to improve the accuracy of satellite navigation and positioning.

[0100] S1.5: Based on the corrected preliminary position information, the vehicle sensor is started, the sensor performs real-time measurement, obtains vehicle dynamic information, and performs preprocessing operations on the vehicle dynamic information measured by the sensor.

[0101] Among them, vehicle sensors include cameras, lidars, and inertial navigation units; vehicle dynamic information includes distance, speed, acceleration, direction, and images; preprocessing operations include data filtering, denoising, calibration, and time synchronization.

[0102] The specific steps of S3 include:

[0103] S3.1: Based on the position and posture of the vehicle in space, the laser radar performs an all-round scan of the vehicle's surrounding environment through array scanning, and uses the difference Δt between the time of sending the laser and the time of receiving the laser and the speed of light c to calculate the relative distance between the vehicle and each point in the surrounding environment.

[0104] During the scanning process, the laser radar emits a laser beam to the surrounding environment. After the laser beam hits the target object, part of the laser will be reflected back and received by the receiving module of the laser radar.

[0105] S3.2: Obtain the scanning angle information of the laser radar, calculate the position of each scanning point in the three-dimensional space according to the scanning angle information and relative distance data, and present it in the form of point cloud data;

[0106] Furthermore, the specific steps of S3.2 include:

[0107] (1) Obtain scanning angle information and relative distance data;

[0108] (2) Using the scanning angle information and relative distance data, the coordinates of each scanning point relative to the laser radar in three-dimensional space are calculated through geometric transformation, usually by converting polar coordinates into Cartesian coordinates. The specific conversion process is the prior art content in this field and is not an inventive solution of the present application, so it will not be described in detail here;

[0109] (3) The calculated three-dimensional coordinates are combined into point cloud data and stored in the form of a list or array, where each element represents the three-dimensional coordinate of a scanning point.

[0110] S3.3: Match the generated point cloud data with the high-precision map in the vehicle system, find the features in the point cloud and the corresponding features in the map, and iteratively adjust the position of the vehicle in the global coordinate system based on the matching results;

[0111] Furthermore, the specific steps of S3.3 include:

[0112] (1) Data collection: Use high-precision map collection vehicles to collect road data. During the collection process, ensure that the sensors are working properly and the data is accurate and reliable;

[0113] (2) Data processing: processing the collected point cloud and image data, including point cloud stitching, image enhancement, and noise removal steps, to obtain high-quality point cloud data and image data, wherein the point cloud stitching, image enhancement, and noise removal processes are prior art contents in the art and are not the inventive solutions of the present application, and are not described in detail herein;

[0114] (3) Feature extraction: Extract features from the processed point cloud data. These features can be road elements such as lane lines, street lights, and electric poles, or other objects with obvious markings. At the same time, the corresponding feature information is also extracted from the high-precision map.

[0115] (4) Feature matching: Match the extracted point cloud features with the features in the high-precision map to find the corresponding relationship between them. In the present invention, this step adopts the FPFH feature descriptor matching method, which is the prior art content in this field and is not the inventive solution of this application, and will not be described in detail here;

[0116] (5) Position adjustment: Based on the matching results, the position of the vehicle in the global coordinate system is iteratively adjusted to make the vehicle positioning more accurate.

[0117] S3.4: According to the position adjustment result in step S3.3, the precise position of the vehicle in the global coordinate system is determined by a matching strategy in combination with the preliminary position information and the vehicle sensor data;

[0118] Furthermore, the specific steps of S3.4 include:

[0119] (1) Inputting the position adjustment result in step S3.3, preliminary position information and vehicle sensor data;

[0120] (2) Preprocess and fuse the input data to form unified environmental perception information;

[0121] (3) Extract features from the fused data and select key features for matching;

[0122] (4) Apply matching strategies to match the extracted features with the features in the high-precision map;

[0123] (5) Determine the precise position of the vehicle in the global coordinate system through an optimization algorithm based on the matching results and the sensor data, wherein the optimization algorithm adopts an iterative nearest neighbor algorithm. The iterative nearest neighbor algorithm is a prior art in the art and is not an inventive solution of the present application, and is not described in detail herein;

[0124] (6) Verify the determined location and output the final location information.

[0125] S3.5: Using the three-dimensional characteristics of point cloud data and the precise position information of the vehicle in the global coordinate system, segment and classify the point cloud data to identify obstacles and road features in the environment;

[0126] Furthermore, the specific steps of S3.5 include:

[0127] (1) Input point cloud data collected by vehicle sensors and the precise position information of the vehicle in the global coordinate system;

[0128] (2) Perform preprocessing operations such as denoising and filtering on the point cloud data to improve data quality, and convert the point cloud data into a global coordinate system based on the vehicle position information;

[0129] (3) Using an edge-based segmentation method to segment the point cloud data using the three-dimensional characteristics of the point cloud data, wherein the three-dimensional characteristics of the point cloud data include spatial position and geometric shape. At the same time, the edge-based segmentation method is a prior art in the art and is not an inventive solution of the present application, and is not described in detail herein.

[0130] (4) using a convolutional neural network model to classify the segmented point cloud data and identify different objects or features. The convolutional neural network model is a prior art in the art and is not an inventive solution of the present application, and is not described in detail herein.

[0131] (5) Based on the classification results, obstacles and road features in the environment are identified, where obstacles include vehicles, pedestrians, and trees, and road features include lane lines, shoulders, and traffic signs.

[0132] S3.6: Calculate the distance and speed of the obstacle relative to the vehicle based on the position and shape of the obstacle, and combine it with the road feature information to generate environmental perception information including the obstacle and road features.

[0133] The specific steps of the matching strategy in S3.4 include:

[0134] S3.41: preprocessing the vehicle sensor data, and extracting feature points from the preprocessed vehicle sensor data using a feature point detection algorithm, and describing the extracted feature points to generate feature descriptors, wherein the feature point detection algorithm is a prior art content in the art, is not an inventive solution of the present application, and is not described in detail herein;

[0135] Furthermore, the specific steps of S3.41 include:

[0136] (1) Obtaining raw data from vehicle sensors;

[0137] (2) Filtering, denoising, time alignment, and coordinate conversion of the raw data;

[0138] (3) Detecting feature points in the preprocessed data using a feature point detection algorithm;

[0139] (4) Using SIFT descriptors to describe the detected feature points and generate feature descriptors, wherein a feature descriptor is usually a vector used to represent the local features of the feature points. SIFT descriptors are prior art in this field and are not an inventive solution of the present application, and are not described in detail herein.

[0140] S3.42: Construct a feature database from feature points and their descriptors in the high-precision map, and use a distance-based matching algorithm to match feature points collected by the vehicle sensor with feature points in the feature database to find the most similar matching pairs, wherein the distance-based matching algorithm is a prior art in the art and is not an inventive solution of the present application, and is not described in detail herein;

[0141] S3.43: Verify the matching result. If the verification is successful, calculate the initial position of the vehicle in the global coordinate system using the shortest distance formula based on the successfully verified feature point pairs. The shortest distance formula is a prior art in this field and is not an inventive solution of the present application, and is not described in detail here.

[0142] S3.44: Combined with the position adjustment result and preliminary position information in step S3.3, the preliminary position is optimized using an optimization algorithm, and the optimized vehicle position information is output to the vehicle control system. In the present invention, the optimization algorithm adopts the Kalman filter method, and the Kalman filter method is the prior art content in this field, and is not an inventive solution of the present application, and will not be elaborated here.

[0143] The specific steps of step S5 include:

[0144] S5.1: receiving fused feature information and distance information, and integrating the fused feature information and distance information to form vehicle perception data;

[0145] S5.2: Based on the vehicle perception data, the vehicle movement scene is abstracted into an environmental model, and a path planning algorithm is used to search in the known environmental model to determine the global optimal driving route from the current position to the target position;

[0146] Furthermore, the specific steps of S5.2 include:

[0147] (1) Environment modeling and mapping: Creating a software environment for executing the path planning method, based on which the traditional, abstract path problem is mapped into a computer-processable quantitative problem in real space;

[0148] (2) Path search: Find the best path through continuous optimization processing. This path can be evaluated by selecting the shortest time or shortest distance according to actual needs;

[0149] (3) Path optimization: Smooth the searched path to meet the requirements of vehicle driving comfort, stability, etc.

[0150] S5.3: Perform local path planning based on the global path, the real-time position of the vehicle and the environmental perception information, generate local path planning results, and generate decision information based on the local path planning results and the vehicle status information. In the present invention, a dynamic window algorithm is used in local path planning, and the dynamic window algorithm is a prior art content in this field, and is not an inventive solution of the present application, and will not be described in detail here;

[0151] Among them, the decision information is how to drive on the determined path after the path is determined, including turning, accelerating, decelerating, changing lanes, etc.

[0152] S5.4: Using a path tracking algorithm to control the vehicle to travel along the global optimal driving route according to the generated decision information. In the present invention, the path tracking algorithm uses a pure tracking algorithm. The pure tracking algorithm is a prior art in the art and is not an inventive solution of the present application, and will not be described in detail herein.

[0153] S5.5: Based on the global optimal driving route, the visual navigation technology and the edge detection algorithm are used to extract lane line information from the road image, and the position and posture of the vehicle relative to the lane line are calculated according to the lane line information. At the same time, steering and speed control instructions are generated so that the vehicle can automatically drive along the extracted lane line. The visual navigation technology and the edge detection algorithm are the prior art content in this field, and are not the inventive solution of the present application, and are not described in detail here.

[0154] Furthermore, the specific steps of calculating the position and posture of the vehicle relative to the lane line include:

[0155] (1) Determine lane line parameters: According to the lane line tracking results, determine the lane line parameters such as slope and intercept. For curved lane lines, it is necessary to use polynomial fitting and other methods to approximate their shape.

[0156] (2) Calculate vehicle position: Based on the lane line parameters and the vehicle position information in the image, use geometric transformation to convert the vehicle position in the image into the actual road position;

[0157] (3) Calculate vehicle posture: Based on the direction of the lane line and the vehicle's driving direction, calculate the vehicle's heading angle, that is, the angle between the vehicle and the lane line. This can be achieved by comparing the vectors of the vehicle's driving direction and the lane line direction. At the same time, the vehicle's roll angle, pitch angle and other posture information can also be estimated based on information such as the curvature of the lane line and the vehicle's driving speed.

[0158] (4) Fusion of global path information: The calculated vehicle position and posture information is integrated with the global optimal driving route, which helps to assess whether the vehicle has deviated from the planned route and make corresponding adjustments.

[0159] S5.6: During driving, the vehicle detects obstacles ahead in real time and determines whether to take obstacle avoidance measures based on the location and speed of the obstacles. If obstacle avoidance is required, the vehicle replans the driving path based on the local path planning algorithm.

[0160] Furthermore, the specific implementation steps of S5.6 include:

[0161] (1) Obtain the obstacles ahead and their types and locations. After the obstacles are detected, they need to be continuously tracked to obtain their motion status, such as position, speed, and acceleration.

[0162] (2) Based on the location and speed information of the obstacle, as well as the current state and driving target of the vehicle, determine whether obstacle avoidance measures are needed. If obstacle avoidance is required, replan the driving path based on the local path planning algorithm;

[0163] (3) After determining that obstacle avoidance is required, a local path planning algorithm is used to generate a new, safe driving path that avoids known obstacles and is as close as possible to the vehicle's original driving target;

[0164] (4) After the new driving path is generated, the vehicle control system needs to be used to accurately track and control the path, which usually involves precise control of vehicle operations such as steering, braking, and acceleration to ensure that the vehicle can safely travel along the new path.

[0165] Example 2

[0166] See also Figure 3 ,Another embodiment provided by the present invention: An autonomous navigation and positioning system for an autonomous driving vehicle, comprising:;

[0167] Positioning module, environment perception module, path planning module, decision-making and control module;

[0168] The positioning module is used to start the sensor to perform real-time measurement based on the initially determined vehicle position, and fuse the measurement results to obtain the precise position and posture of the vehicle in space;

[0169] The environmental perception module is used to generate point cloud data using LiDAR and match it with high-precision maps to determine the precise location of the vehicle and identify obstacles and road features in the environment;

[0170] The path planning module is used to perceive surrounding obstacles and road conditions in real time based on the fusion of feature information and distance information, determine the best driving route, and perform automatic navigation and obstacle avoidance;

[0171] The decision control module is used to make driving decisions and control vehicle execution based on the results of environmental perception and path planning.

[0172] The positioning module includes: a GPS receiver unit, a carrier phase measurement unit, a data conversion unit, and a position and attitude calculation unit;

[0173] A GPS receiver unit for receiving signals transmitted by satellites to provide preliminary location information for the vehicle;

[0174] The carrier phase measurement unit is used to combine the GPS signal to calculate the distance error between the receiver and the satellite to improve the positioning accuracy;

[0175] A data conversion unit, used to convert the preprocessed sensor measurement results into position information in a global coordinate system;

[0176] The position and attitude calculation unit is used to calculate the position and attitude of the vehicle in space using a sensor fusion algorithm.

[0177] The environmental perception module includes: LiDAR unit, point cloud data processing unit, map matching unit, and obstacle recognition unit;

[0178] A laser radar unit, which emits a laser beam and calculates the relative distance between the vehicle and the surrounding environment based on the reflection time to generate point cloud data;

[0179] Point cloud data processing unit, used to process point cloud data, such as filtering, noise reduction, and feature extraction;

[0180] A map matching unit, used to match point cloud data with high-precision maps;

[0181] The obstacle recognition unit is used to recognize obstacles and road features using the three-dimensional characteristics of point cloud data.

[0182] The path planning module includes: path planning unit, visual navigation unit, obstacle avoidance execution unit, and update unit;

[0183] The path planning unit is used to determine the best driving route based on the fusion feature information and the current state of the vehicle;

[0184] A visual navigation unit for navigation using visual information, such as lane lines and traffic signs;

[0185] The obstacle avoidance execution unit is used to automatically take obstacle avoidance measures, such as deceleration, steering, and parking, when encountering obstacles;

[0186] The updating unit is used to update the position and posture information of the vehicle according to the real-time driving conditions of the vehicle, so as to provide more accurate input data for subsequent navigation and obstacle avoidance.

[0187] The decision control module includes: a decision unit and a control unit;

[0188] The decision-making unit is used to evaluate the current environmental conditions and select driving modes, such as following a vehicle, changing lanes, and parking;

[0189] The control unit is used to generate control commands based on the decision results, such as acceleration, deceleration, steering, etc., and directly control the vehicle's hardware systems to execute according to the control commands, such as motors, brakes, and steering mechanisms.

[0190] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0191] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented 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 processor, 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 generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0192] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present invention, ordinary technicians in the field can also change, modify, replace and modify the above-mentioned embodiments without departing from the purpose and scope of protection of the present invention, which are all within the protection of the present invention.

Claims

1. An autonomous navigation and positioning method for an autonomous driving vehicle, characterized in that: include: Step S1: Receive the signal transmitted by the satellite through the GPS receiver, and calculate the distance error from the receiver to the satellite by combining the carrier phase measurement method, so as to preliminarily determine the position of the vehicle, start the vehicle sensor to perform real-time measurement based on the preliminarily determined vehicle position, and pre-process the measurement results; Step S2: convert the preprocessed measurement results into position information in the global coordinate system, combine with the preliminary determination of the vehicle's position, and use the fusion algorithm to calculate the position and posture of the vehicle in space; Step S3: Based on the position and posture of the vehicle in space, the laser radar emits a laser beam, calculates the relative distance between the vehicle and the surrounding environment according to the reflection time, generates point cloud data, and matches the point cloud data with the high-precision map in the vehicle system. The precise position of the vehicle in the global coordinate system is determined through the matching strategy, and the three-dimensional characteristics of the point cloud data are used to identify obstacles and road features in the environment to generate environmental perception information; Step S4: The on-board camera captures the environment image around the vehicle, analyzes the environment image, extracts the visual feature information of road signs, lane lines and obstacles, and fuses the extracted visual feature information with the environment perception information to obtain fused feature information; Step S5: Based on the fusion feature information and distance information, a path planning algorithm is used to determine the best driving route, and through visual navigation technology, the vehicle automatically drives along the extracted lane line, automatically takes obstacle avoidance measures when encountering obstacles, and updates its position and posture information.

2. The autonomous navigation and positioning method for an autonomous driving vehicle according to claim 1, characterized in that: The specific steps of step S1 include: S1.1: The GPS receiver receives radio frequency signals from multiple GPS satellites, and down-converts and digitizes the radio frequency signals to obtain baseband signals; S1.2: Extract carrier phase information from the baseband signal, and calculate the phase difference between the satellite carrier signal received by the receiver and the receiver local oscillator reference signal based on the carrier phase information. in, Indicates the phase of the satellite carrier signal received by the receiver. Represents the phase of the local reference signal generated by the receiver; S1.3: Calculate the precise distance from the receiver to the satellite based on the wavelength in the carrier phase information and the phase difference obtained in step S1.2 to obtain the satellite ranging result Among them, d n Indicates the ranging result of the nth satellite, Indicates the phase difference between the nth satellite carrier signal received by the receiver and the receiver local oscillator reference signal, λ n It represents the carrier wavelength of the nth satellite, where n represents the number of satellites.

3. The autonomous navigation and positioning method for an automatic driving vehicle according to claim 2, characterized in that: The specific steps of step S1 also include: S1.4: According to the satellite ranging results, correct the initial position, the formula is: Among them, d n,cor It represents the precise distance from the calibrated receiver to the nth satellite. represents the phase error caused by multipath effect, α dian represents the ionospheric delay correction factor, α dian represents the tropospheric delay correction factor, α wei,t represents the satellite clock correction factor, α jie,t represents the receiver clock error correction factor, and β represents the adjustment coefficient; S1.5: Based on the corrected preliminary position information, the vehicle sensor is started, the sensor performs real-time measurement, obtains vehicle dynamic information, and performs preprocessing operations on the vehicle dynamic information measured by the sensor.

4. The autonomous navigation and positioning method for an automatic driving vehicle according to claim 3, characterized in that: The specific steps of step S3 include: S3.1: Based on the position and posture of the vehicle in space, the laser radar performs an all-round scan of the vehicle's surrounding environment through array scanning, and uses the difference Δt between the time of sending the laser and the time of receiving the laser and the speed of light c to calculate the relative distance between the vehicle and each point in the surrounding environment. S3.2: Obtain the scanning angle information of the laser radar, calculate the position of each scanning point in the three-dimensional space according to the scanning angle information and relative distance data, and present it in the form of point cloud data; S3.3: Match the generated point cloud data with the high-precision map in the vehicle system, find the features in the point cloud and the corresponding features in the map, and iteratively adjust the position of the vehicle in the global coordinate system based on the matching results; S3.4: According to the position adjustment result in step S3.3, the precise position of the vehicle in the global coordinate system is determined by a matching strategy in combination with the preliminary position information and the vehicle sensor data; S3.5: Using the three-dimensional characteristics of point cloud data and the precise position information of the vehicle in the global coordinate system, segment and classify the point cloud data to identify obstacles and road features in the environment; S3.6: Calculate the distance and speed of the obstacle relative to the vehicle based on the position and shape of the obstacle, and combine it with the road feature information to generate environmental perception information including the obstacle and road features.

5. The autonomous navigation and positioning method for an automatic driving vehicle according to claim 4, characterized in that: The specific steps of the matching strategy in S3.4 include: S3.41: preprocessing the vehicle sensor data, and extracting feature points from the preprocessed vehicle sensor data using a feature point detection algorithm, and describing the extracted feature points to generate feature descriptors; S3.42: Construct feature points and their descriptors in the high-precision map into a feature database, and use a distance-based matching algorithm to match the feature points collected by the vehicle sensor with the feature points in the feature database to find the most similar matching pairs; S3.43: Verify the matching result. If the verification is successful, calculate the initial position of the vehicle in the global coordinate system using the shortest distance formula based on the successfully verified feature point pairs; S3.44: Combining the position adjustment result in step S3.3 and the preliminary position information, the preliminary position is optimized using an optimization algorithm, and the optimized vehicle position information is output to the vehicle control system.

6. The autonomous navigation and positioning method for an automatic driving vehicle according to claim 5, characterized in that: The specific steps of step S5 include: S5.1: receiving fused feature information and distance information, and integrating the fused feature information and distance information to form vehicle perception data; S5.2: Based on the vehicle perception data, the vehicle movement scene is abstracted into an environmental model, and a path planning algorithm is used to search in the known environmental model to determine the global optimal driving route from the current position to the target position; S5.3: Perform local path planning based on the global path, the real-time position of the vehicle and the environmental perception information, generate local path planning results, and generate decision information based on the local path planning results and the vehicle status information; S5.4: Use a path tracking algorithm to control the vehicle to travel along the global optimal driving route based on the generated decision information; S5.5: Based on the global optimal driving route, the visual navigation technology and edge detection algorithm are used to extract lane line information from the road image, and the position and posture of the vehicle relative to the lane line are calculated based on the lane line information. At the same time, steering and speed control instructions are generated so that the vehicle can automatically drive along the extracted lane line; S5.6: During driving, the vehicle detects obstacles ahead in real time and determines whether to take obstacle avoidance measures based on the location and speed of the obstacles. If obstacle avoidance is required, the vehicle replans the driving path based on the local path planning algorithm.

7. An autonomous navigation and positioning system for an autonomous driving vehicle, which is used to implement an autonomous navigation and positioning method for an autonomous driving vehicle according to any one of claims 1 to 6, characterized in that: include: Positioning module, environment perception module, path planning module, decision-making and control module; The positioning module is used to start the sensor to perform real-time measurement according to the initially determined vehicle position, and fuse the measurement results to obtain the precise position and posture of the vehicle in space; The environment perception module is used to generate point cloud data using laser radar and match it with a high-precision map to determine the precise location of the vehicle and identify obstacles and road features in the environment; The path planning module is used to perceive surrounding obstacles and road conditions in real time based on the fusion feature information and distance information, determine the best driving route, and perform automatic navigation and obstacle avoidance; The decision control module is used to make driving decisions and control vehicle execution based on the results of environment perception and path planning.

8. The autonomous navigation and positioning system for an autonomous driving vehicle according to claim 7, characterized in that: The positioning module includes: a GPS receiver unit, a carrier phase measurement unit, a data conversion unit, and a position and attitude calculation unit; The GPS receiver unit is used to receive signals transmitted by satellites and provide preliminary position information for the vehicle; The carrier phase measurement unit is used to calculate the distance error from the receiver to the satellite in combination with the GPS signal to improve the positioning accuracy; The data conversion unit is used to convert the preprocessed sensor measurement results into position information in the global coordinate system; The position and posture calculation unit is used to calculate the position and posture of the vehicle in space by using a sensor fusion algorithm.

9. The autonomous navigation and positioning system for an automatic driving vehicle according to claim 8, characterized in that: The path planning module includes: a path planning unit, a visual navigation unit, and an obstacle avoidance execution unit; The path planning unit is used to determine the best driving route based on the fused feature information and the current state of the vehicle; The visual navigation unit is used to navigate using visual information; The obstacle avoidance execution unit is used to automatically take obstacle avoidance measures when encountering an obstacle.

10. The autonomous navigation and positioning system for an automatic driving vehicle according to claim 9, characterized in that: The decision control module includes: a decision unit and a control unit; The decision-making unit is used to evaluate the current environmental conditions and select a driving mode; The control unit is used to generate a control command according to the decision result, and directly control the hardware system of the vehicle to execute according to the control command.

Images