Unmanned driving method, system, device and medium in multi-road condition limited space

Abstract

The present application discloses an unmanned driving method, system, device and medium in a multi-road condition restricted space. The method obtains road condition information indicating whether the current driving road condition of the unmanned driving device is a simple road condition or a complex road condition; if the road condition information is a simple road condition, a first unmanned driving strategy corresponding to the simple road condition is selected; a first control output instruction is obtained through the first unmanned driving strategy, and the unmanned driving device is controlled to drive on the simple road condition according to the first control output instruction; if the road condition information changes from a simple road condition to a complex road condition, a second unmanned driving strategy corresponding to the complex road condition is selected; a second control output instruction is obtained through the second unmanned driving strategy, and the unmanned driving device is controlled to drive on the complex road condition according to the second control output instruction. The present application can meet the navigation and control requirements in the mining industry and other multi-road condition restricted space environments.

Description

Technical Field

[0001] The present application relates to the field of unmanned driving technology, and in particular to an unmanned driving method, system, equipment and medium in a multi-road restricted space. Background Art

[0002] With the development of automation and intelligent technology, the application of unmanned driving technology in mines, underground facilities and other confined spaces has become increasingly widespread. In some complex road conditions, unmanned driving technology usually uses traditional simultaneous localization and mapping (SLAM) algorithms, and traditional SLAM algorithms often face challenges due to the complex and changing environment, especially in accurate map construction and real-time path planning. If the method for complex road conditions is used for each road condition, this will increase the computational burden for simple road conditions and increase resource consumption. Therefore, the existing unmanned driving technology cannot meet the navigation and control needs of the mining industry and other multi-road confined space environments. Summary of the invention

[0003] The present application aims to propose an unmanned driving method, system, equipment and medium in a multi-road confined space, which can meet the navigation and control requirements in the mining industry and other multi-road confined space environments.

[0004] In a first aspect, an embodiment of the present application provides an unmanned driving method in a confined space with multiple road conditions, wherein the multiple road conditions include simple road conditions and complex road conditions, and the method includes:

[0005] Obtaining road condition information indicating whether the current road condition of the unmanned driving device is a simple road condition or a complex road condition;

[0006] If the road condition information is a simple road condition, selecting a first unmanned driving strategy corresponding to the simple road condition;

[0007] Obtaining a first control output instruction through the first unmanned driving strategy, and controlling the unmanned driving device to travel on the simple road condition according to the first control output instruction;

[0008] If the road condition information changes from the simple road condition to the complex road condition, selecting a second unmanned driving strategy corresponding to the complex road condition;

[0009] A second control output instruction is obtained through the second unmanned driving strategy, and the unmanned driving device is controlled to travel on the complex road condition according to the second control output instruction.

[0010] Compared with the prior art, the first aspect of the present application has the following beneficial effects:

[0011] This method obtains road condition information indicating whether the current driving condition of the unmanned driving device is a simple road condition or a complex road condition; if the road condition information is a simple road condition, the first unmanned driving strategy corresponding to the simple road condition is selected; the first control output instruction is obtained through the first unmanned driving strategy, and the unmanned driving device is controlled to drive on the simple road condition according to the first control output instruction; if the road condition information changes from a simple road condition to a complex road condition, the second unmanned driving strategy corresponding to the complex road condition is selected; the second control output instruction is obtained through the second unmanned driving strategy, and the unmanned driving device is controlled to drive on the complex road condition according to the second control output instruction. In this way, by comprehensively considering different road conditions and adopting different unmanned driving methods, the performance of the system under various environmental conditions is enhanced, and the navigation and control requirements in the mining industry and other multi-road condition restricted space environments can be met, and the efficiency and reliability of unmanned driving are significantly improved.

[0012] In some implementations, obtaining a first control output instruction through the first unmanned driving strategy includes:

[0013] Performing path planning by using the pre-marked coordinates of the first UWB trajectory point to determine a first target planning path;

[0014] Acquire first UWB position data and first laser radar data of an unmanned driving device traveling on the first target planning path;

[0015] Determine a first current position coordinate of the unmanned driving device according to the first UWB position data and the first laser radar data;

[0016] Determining a polar coordinate solution angle according to the first current position coordinates;

[0017] Determine two first safety boundaries based on the first laser radar data, and determine a first safety boundary correction angle according to the two first safety boundaries;

[0018] Based on the polar coordinate solution angle and the first safety margin correction angle, a target control angle and a target control angular velocity are calculated, and the target control angle and the target control angular velocity are used as a first control output instruction.

[0019] In some implementations, determining the polar coordinate solution angle according to the first current position coordinates includes:

[0020] Acquire the target driving trajectory point coordinates corresponding to the first current position coordinates from the first target planning path;

[0021] Convert the target driving trajectory point coordinates into a polar coordinate system trajectory direction;

[0022] The angle formed by the trajectory direction of the polar coordinate system and the center line of the unmanned driving device is used as the polar coordinate solution angle.

[0023] In some embodiments, determining two first safety boundaries based on the first laser radar data, and determining a first safety boundary correction angle according to the two first safety boundaries includes:

[0024] Inputting the first laser radar data into the SLAM algorithm to obtain two first safety boundaries;

[0025] confirming a first middle line between the two first safety boundaries;

[0026] A first angle between the first middle line and the first center line of the unmanned driving device is used as a first safety margin correction angle.

[0027] In some implementations, the calculating a target control angle and a target control angular velocity based on the polar coordinate solution angle and the first safety margin correction angle includes:

[0028] Calculating a safety margin correction angle difference between the first safety margin correction angle and the polar coordinate solution angle;

[0029] Calculating a target control angle according to the safety margin correction angle difference and the polar coordinate solution angle;

[0030] According to the target control angle, a target control angular velocity is determined.

[0031] In some implementations, obtaining a second control output instruction by using the second unmanned driving strategy includes:

[0032] Perform path planning by using the pre-marked coordinates of the second UWB trajectory point to determine a second target planning path;

[0033] Acquire second laser radar data, IMU posture information, and second UWB position data of the unmanned driving device traveling on the second target planning path;

[0034] Fusing the second laser radar data, the attitude information of the IMU, and the second UWB position data to obtain fused data;

[0035] Inputting the fused data into a SLAM algorithm to obtain a first position of the unmanned driving device, an obstacle position, a spatial change, and two second safety boundaries;

[0036] Processing the second laser radar data, the attitude information of the IMU, and the second UWB position data using a particle filter and a Kalman filter to obtain a second position of the unmanned driving device;

[0037] Obtaining a target driving path according to the second position of the unmanned driving device, the first position of the unmanned driving device, the obstacle position, and the spatial change;

[0038] Determining a second safety boundary correction angle and a position correction angle according to the two second safety boundaries and the target driving path;

[0039] The traveling direction and the steering angle are calculated according to the second safety margin correction angle and the position correction angle, and the traveling direction and the steering angle are used as the second control output instruction.

[0040] In some implementations, determining the second safety boundary correction angle and the position correction angle according to the two second safety boundaries and the target driving path includes:

[0041] confirming a second middle line between the two second safety boundaries;

[0042] Using a second angle between the second middle line and the second middle line of the unmanned driving device as a second safety margin correction angle;

[0043] The target position corresponding to the current position of the unmanned driving device in the target driving path is obtained, and the angle between the current position and the target position is used as the position correction angle.

[0044] In a second aspect, an embodiment of the present application further provides an unmanned driving system in a multi-road condition restricted space, the system comprising:

[0045] A road condition information acquisition unit, used to acquire road condition information indicating whether the current driving road condition of the unmanned driving device is a simple road condition or a complex road condition;

[0046] A first method selection unit, configured to select a first unmanned driving strategy corresponding to the simple road condition if the road condition information is a simple road condition;

[0047] A first method control unit, configured to obtain a first control output instruction through the first unmanned driving strategy, and control the unmanned driving device to travel on the simple road condition according to the first control output instruction;

[0048] A second method selection unit, configured to select a second unmanned driving strategy corresponding to the complex road condition if the road condition information changes from the simple road condition to the complex road condition;

[0049] The second method control unit is used to obtain a second control output instruction through the second unmanned driving strategy, and control the unmanned driving device to travel on the complex road condition according to the second control output instruction.

[0050] In the third aspect, an embodiment of the present application also provides an electronic device, comprising at least one control processor and a memory for communicating with the at least one control processor; the memory stores instructions that can be executed by the at least one control processor, and the instructions are executed by the at least one control processor so that the at least one control processor can execute the unmanned driving method in a multi-road constrained space as described above.

[0051] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to enable a computer to execute the unmanned driving method in a multi-road constrained space as described above.

[0052] It can be understood that the beneficial effects of the second to fourth aspects compared with the related art are the same as the beneficial effects of the first aspect compared with the related art. Please refer to the relevant description in the first aspect, and no further details will be given here. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0054] Figure 1 It is a flowchart of an embodiment of an unmanned driving method in a multi-road condition constrained space provided by the present application;

[0055] Figure 2 It is a flowchart of the best embodiment of the unmanned driving method in a multi-road condition restricted space provided by the present application;

[0056] Figure 3 It is a schematic diagram of path planning for simple road conditions and complex road conditions in the best embodiment of the unmanned driving method in a multi-road condition restricted space provided by the present application;

[0057] Figure 4 It is a schematic diagram of polar coordinate posture correction in a local restricted space under simple road conditions in the best embodiment of the unmanned driving method under multiple road conditions and restricted spaces provided by the present application;

[0058] Figure 5 It is a schematic diagram of polar coordinate posture correction in a local restricted space under complex road conditions of the best embodiment of the unmanned driving method under multi-road restricted space provided by the present application;

[0059] Figure 6 It is a structural schematic diagram of an embodiment of an unmanned driving system in a multi-road condition restricted space provided by the present application;

[0060] Description of labels:

[0061] 101. First safety boundary correction angle; 102. Polar coordinate solution angle; 103. Two first safety boundaries; 104. Pre-marked first UWB trajectory point coordinates; 105. First middle line; 106. Current driving direction; 107. Unmanned driving equipment; 108. Polar coordinate system trajectory line; 109. Second safety boundary correction angle; 110. Position correction angle; 111. Two second safety boundaries; 112. Pre-marked second UWB trajectory point coordinates; 113. Second middle line. DETAILED DESCRIPTION

[0062] The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present application, and cannot be understood as limiting the present application.

[0063] In the description of this application, if there is a description of first, second, etc., it is only for the purpose of distinguishing technical features, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features or implicitly indicating the order of the indicated technical features.

[0064] In the description of the present application, it should be understood that descriptions involving orientation, such as orientation or positional relationship indicated as up, down, etc., are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present application.

[0065] In the description of this application, it should be noted that, unless otherwise clearly defined, terms such as setting, installing, connecting, etc. should be understood in a broad sense, and technicians in the relevant technical field can reasonably determine the specific meaning of the above terms in this application based on the specific content of the technical solution.

[0066] With the development of automation and intelligent technology, the application of unmanned driving technology in mines, underground facilities and other confined spaces has become increasingly widespread. In some complex road conditions, unmanned driving technology usually uses traditional SLAM algorithms, which often face challenges due to the complex and changeable environment, especially in accurate map construction and real-time path planning. If the method under complex road conditions is used for each road condition, this will increase the computational burden for simple road conditions and increase resource consumption. Therefore, the existing unmanned driving technology cannot meet the navigation and control needs of the mining industry and other multi-road confined space environments.

[0067] In order to solve the problem that the existing unmanned driving technology cannot meet the navigation and control needs in the mining industry and other multi-road confined space environments, the present application proposes an unmanned driving method, system, equipment and medium in multi-road confined spaces.

[0068] Reference Figure 1 The embodiment of the present application provides an unmanned driving method in a confined space with multiple road conditions, wherein the multiple road conditions include simple road conditions and complex road conditions. The method includes the following steps:

[0069] Step S100, obtaining road condition information indicating whether the current driving condition of the unmanned driving device 107 is a simple road condition or a complex road condition;

[0070] Step S200: If the road condition information is a simple road condition, select a first unmanned driving strategy corresponding to the simple road condition;

[0071] Step S300, obtaining a first control output instruction through a first unmanned driving strategy, and controlling the unmanned driving device 107 to drive on a simple road condition according to the first control output instruction;

[0072] Step S400: If the road condition information changes from a simple road condition to a complex road condition, a second unmanned driving strategy corresponding to the complex road condition is selected;

[0073] Step S500: obtaining a second control output instruction through a second unmanned driving strategy, and controlling the unmanned driving device 107 to drive on complex road conditions according to the second control output instruction.

[0074] In this embodiment, the road condition information indicating that the current driving condition of the unmanned driving device 107 is a simple road condition or a complex road condition is obtained; if the road condition information is a simple road condition, the first unmanned driving strategy corresponding to the simple road condition is selected; the first control output instruction is obtained through the first unmanned driving strategy, and the unmanned driving device 107 is controlled to drive on the simple road condition according to the first control output instruction; if the road condition information changes from a simple road condition to a complex road condition, the second unmanned driving strategy corresponding to the complex road condition is selected; the second control output instruction is obtained through the second unmanned driving strategy, and the unmanned driving device 107 is controlled to drive on the complex road condition according to the second control output instruction. In this way, by comprehensively considering different road conditions and adopting different unmanned driving methods, the performance of the system under various environmental conditions is enhanced, and the navigation and control requirements in the mining industry and other multi-road condition restricted space environments can be met, and the efficiency and reliability of unmanned driving are significantly improved.

[0075] The above-mentioned simple road condition may be a road section artificially set as a simple road condition, and may include a very simple road condition or a relatively simple road condition, that is, a relatively simple road condition is a little more complicated than a very simple road condition.

[0076] The above-mentioned complex road conditions may be artificially set as a section of road being complex road conditions, and may include very complex road conditions or relatively complex road conditions, that is, relatively complex road conditions are simpler than very complex road conditions.

[0077] The first unmanned driving strategy may be an unmanned driving method designed for simple road conditions, or may be a traditional SLAM algorithm for simple road conditions.

[0078] The second unmanned driving strategy may be an unmanned driving method designed for complex road conditions, or may be a traditional SLAM algorithm for complex road conditions.

[0079] In some implementations, obtaining a first control output instruction through a first unmanned driving strategy includes:

[0080] Performing path planning by using the pre-marked first UWB trajectory point coordinates 104 to determine a first target planning path;

[0081] Acquire first UWB position data and first laser radar data of the unmanned driving device 107 traveling on the first target planning path;

[0082] Determine a first current position coordinate of the unmanned driving device 107 according to the first UWB position data and the first laser radar data;

[0083] According to the first current position coordinates, a polar coordinate solution angle 102 is determined;

[0084] Based on the first laser radar data, two first safety boundaries 103 are determined, and according to the two first safety boundaries 103, a first safety boundary correction angle 101 is determined;

[0085] Based on the polar coordinate solution angle 102 and the first safety margin correction angle 101 , a target control angle and a target control angular velocity are calculated, and the target control angle and the target control angular velocity are used as a first control output instruction.

[0086] In this embodiment, path planning is performed by using the pre-marked first UWB trajectory point coordinates 104 to determine the first target planned path, so that the unmanned driving device 107 can travel on the optimal path as much as possible; the first UWB position data and the first laser radar data of the unmanned driving device 107 traveling on the first target planned path are obtained, and the first current position coordinates of the unmanned driving device 107 are determined according to the first UWB position data and the first laser radar data. The current position coordinates of the unmanned driving device 107 are determined according to multiple data, which can improve the accuracy; according to the first current position coordinates, the polar coordinate solution angle 102 is determined, based on the first laser radar data, two first safety boundaries 103 are determined, and based on the two first safety boundaries 103, the first safety boundary correction angle 101 is determined, and based on the polar coordinate solution angle 102 and the first safety boundary correction angle 101, the target control angle and the target control angular velocity are calculated, and the target control angle and the target control angular velocity are used as the first control output instruction, so that the unmanned driving device 107 can travel safely and meet the requirements of efficient operation in narrow underground spaces.

[0087] The pre-marked first UWB trajectory point coordinates 104 may be pre-set UWB markers, and the UWB trajectory point coordinates may be acquired using UWB technology.

[0088] The above-mentioned UWB technology is an ultra-wideband technology, which is a wireless carrier communication technology. It does not use a sinusoidal carrier, but uses nanosecond-level non-sinusoidal narrow pulses to transmit data, so it occupies a wide spectrum range.

[0089] The above-mentioned path planning is performed through the pre-marked first UWB trajectory point coordinates 104 to determine the first target planning path, which can be obtained by using a trajectory point planning algorithm and a route matching algorithm based on the pre-marked first UWB trajectory point coordinates 104. It can also be obtained by using an existing path planning method.

[0090] The above-mentioned acquisition of the first UWB position data and the first laser radar data of the unmanned driving device 107 traveling on the first target planned path can be performed by installing a UWB identification card and a laser radar on the unmanned driving device 107 traveling on the first target planned path, and acquiring the first UWB position data and the first laser radar data of the unmanned driving device 107 through the UWB identification card and the laser radar.

[0091] The first current position coordinates of the unmanned driving device 107 are determined according to the first UWB position data and the first laser radar data by using a SLAM algorithm. Alternatively, the first current position coordinates of the unmanned driving device 107 are determined by first fusing the first UWB position data and the first laser radar data using a Kalman filter method.

[0092] The first laser radar data may include surrounding environment data and more accurate data such as the current position of the unmanned driving device 107 .

[0093] The polar coordinate solution angle 102 is determined according to the first current position coordinates. The traveling target position coordinates are determined according to the first current position coordinates, and then the polar coordinate solution angle 102 can be determined according to the first current position coordinates and the traveling target position coordinates.

[0094] The above-mentioned determination of the two first safety boundaries 103 based on the first laser radar data may be performed by inputting the first laser radar data into the SLAM algorithm, and obtaining the two first safety boundaries 103 through the safety obstacle avoidance calculation in the SLAM algorithm. For example, the current position of the unmanned driving device is obtained through the laser radar data, and then the safety obstacle avoidance in the SLAM algorithm is used to calculate two safety boundary lines that are relatively safe relative to the two sides of the road at the current position, and the two safety boundary lines are at a certain distance from the two sides of the road.

[0095] In some implementations, determining the polar coordinate solution angle 102 according to the first current position coordinates includes:

[0096] Obtaining the target driving trajectory point coordinates corresponding to the first current position coordinates from the first target planning path;

[0097] Convert the coordinates of the target driving trajectory point into the trajectory direction of the polar coordinate system;

[0098] The angle formed by the polar coordinate system trajectory direction and the center line of the unmanned driving device 107 is used as the polar coordinate solution angle 102.

[0099] In this embodiment, the target driving trajectory point coordinates corresponding to the first current position coordinates are obtained from the first target planned path, the target driving trajectory point coordinates are converted into the polar coordinate system trajectory direction, and then the angle formed by the polar coordinate system trajectory direction and the center line of the unmanned driving device 107 is used as the polar coordinate solution angle 102. This can enable the unmanned driving device 107 to continuously approach the target driving trajectory point, allow the unmanned driving device 107 to travel in the first target planned path as much as possible, and improve the efficient operation of the unmanned driving device 107.

[0100] In some embodiments, determining two first safety boundaries 103 based on the first laser radar data, and determining a first safety boundary correction angle 101 according to the two first safety boundaries 103 includes:

[0101] Input the first laser radar data into the SLAM algorithm to obtain two first safety boundaries 103;

[0102] Determine a first middle line 105 between two first safety boundaries 103;

[0103] The first angle between the first center line 105 and the first center line of the unmanned driving device 107 is used as the first safety margin correction angle 101.

[0104] In this embodiment, by inputting the first laser radar data into the SLAM algorithm, two first safety boundaries 103 are obtained, which can ensure the driving safety of the unmanned driving device 107. The first angle between the first center line 105 and the first center line of the unmanned driving device 107 is used as the first safety boundary correction angle 101, which can ensure the driving safety of the unmanned driving device 107.

[0105] In some implementations, the target control angle and the target control angular velocity are calculated based on the polar coordinate solution angle 102 and the first safety margin correction angle 101, including:

[0106] Calculating a safety margin correction angle difference between the first safety margin correction angle 101 and the polar coordinate solution angle 102;

[0107] Calculate the target control angle according to the safety margin correction angle difference and the polar coordinate solution angle 102;

[0108] According to the target control angle, the target control angular velocity is determined.

[0109] In this embodiment, the safety boundary correction angle difference between the first safety boundary correction angle 101 and the polar coordinate solution angle 102 is calculated, and then the target control angle is calculated based on the safety boundary correction angle difference and the polar coordinate solution angle 102. Based on the target control angle, the target control angular velocity is determined. By simplifying the complex control process into intuitive travel direction and steering angle outputs, the unmanned driving device 107 can be driven in the target direction while ensuring the driving safety of the unmanned driving device 107, and the unmanned driving device 107 can be driven in the first target planned path as much as possible, thereby improving the convenience and efficiency of operation.

[0110] In some implementations, obtaining the second control output instruction through the second unmanned driving strategy includes:

[0111] Performing path planning by using the pre-marked second UWB trajectory point coordinates 112 to determine a second target planning path;

[0112] Acquire second laser radar data, IMU posture information, and second UWB position data of the unmanned driving device 107 traveling on the second target planning path;

[0113] Fusing the second laser radar data, the IMU attitude information, and the second UWB position data to obtain fused data;

[0114] The fused data is input into the SLAM algorithm to obtain the first position of the unmanned driving device 107, the obstacle position, the spatial change and the two second safety boundaries 111;

[0115] The second laser radar data, the attitude information of the IMU and the second UWB position data are processed by using a particle filter and a Kalman filter to obtain a second position of the unmanned driving device 107;

[0116] Obtaining a target driving path according to the second position of the unmanned driving device 107, the first position of the unmanned driving device 107, the obstacle position, and the spatial change;

[0117] Determine a second safety boundary correction angle 109 and a position correction angle 110 according to two second safety boundaries 111 and the target driving path;

[0118] The traveling direction and the steering angle are calculated according to the second safety margin correction angle 109 and the position correction angle 110 , and are used as the second control output instruction.

[0119] In this embodiment, path planning is performed by pre-marking the coordinates of the second UWB trajectory point 112 to determine the second target planning path, so that the unmanned driving device 107 can travel on the optimal path as much as possible. The second laser radar data, the posture information of the IMU and the second UWB position data of the unmanned driving device 107 traveling on the second target planning path are obtained, and the second laser radar data, the posture information of the IMU and the second UWB position data are fused to obtain the fused data, and then the fused data is input into the SLAM algorithm to obtain the first position of the unmanned driving device 107, the obstacle position, the spatial change and the two second safety boundaries 111, which can improve the accuracy of the final data. The second laser radar data, the posture information of the IMU and the second UWB position data are processed by particle filtering and Kalman filtering to obtain the second position of the unmanned driving device 107, further improving the accuracy of the second position. According to the second position of the unmanned driving device 107, the first position of the unmanned driving device 107, the obstacle position and the spatial change, the target driving path is obtained, and the second safety boundary correction angle 109 and the position correction angle 110 are determined according to the two second safety boundaries 111 and the target driving path. According to the second safety boundary correction angle 109 and the position correction angle 110, the travel direction and the steering angle are calculated, and the travel direction and the steering angle are used as the second control output instructions. By obtaining the precise target driving path according to the accurate position information, and then performing safety boundary correction and position correction according to the precise target driving path, the unmanned driving device 107 can be driven safely and can also meet the requirements of efficient operation under complex underground road conditions and narrow spaces.

[0120] The second target planning path may be a second target planning path obtained by performing path planning using the least squares method and the A-star algorithm according to the pre-marked second UWB trajectory point coordinates 112 .

[0121] The above-mentioned A-star algorithm is an algorithm existing in the prior art and will not be described in detail in this embodiment.

[0122] The above-mentioned target driving path can be to convert the position coordinates of the second position of the unmanned driving device 107, the first position of the unmanned driving device 107 and the obstacle position in a two-dimensional Cartesian coordinate system into position coordinates in a polar coordinate system, and then use a heuristic algorithm to obtain the optimal path that minimizes the cost function based on the position coordinates and spatial changes after the coordinate conversion, and use the optimal path as the target driving path.

[0123] The above heuristic algorithm may be the A-star algorithm or other algorithms in the prior art.

[0124] In some implementations, determining the second safety boundary correction angle 109 and the position correction angle 110 according to the two second safety boundaries 111 and the target driving path includes:

[0125] Determine a second middle line 113 between the two second safety boundaries 111;

[0126] The second angle between the second middle line 113 and the second middle line of the unmanned driving device 107 is used as the second safety margin correction angle 109;

[0127] The target position corresponding to the current position of the unmanned driving device 107 in the target driving path is obtained, and the angle between the current position and the target position is used as the position correction angle 110 .

[0128] In this embodiment, by using the second angle between the second center line 113 and the second center line of the unmanned driving device 107 as the second safety boundary correction angle 109, and then obtaining the target position corresponding to the current position of the unmanned driving device 107 in the target driving path, and using the angle between the current position and the target position as the position correction angle 110, the unmanned driving device 107 can be driven in the target driving path as much as possible while ensuring the safety of the unmanned driving device 107.

[0129] To facilitate understanding by those skilled in the art, a set of best embodiments is provided below:

[0130] Reference Figure 2 , Figure 2 The device two-dimensional map direction determination refers to the determination of the driving direction obtained by coordinate conversion. This process usually involves comparing the current driving direction of the unmanned driving device 107 and the target direction obtained by polar coordinate conversion. If the two are consistent, the unmanned driving device 107 is in a forward driving state; if they are inconsistent, the unmanned driving device 107 may be in a reverse driving state. This embodiment combines the minimalist SLAM algorithm (i.e., the unmanned driving method used under simple road conditions) and the deep SLAM algorithm (i.e., the unmanned driving method used under complex road conditions) to optimize the navigation and map building capabilities of the unmanned driving device 107 under confined space conditions. This embodiment is particularly suitable for unmanned vehicles that need to perform high-precision navigation and operation in closed, narrow and complex spaces with a variety of road conditions.

[0131] 1. The source of the features of the effective combination of minimalist SLAM and deep SLAM in this embodiment is explained as follows:

[0132] (1) Positioning and composition stage.

[0133] Positioning is the basis of SLAM, which involves the algorithm determining its own position in an unknown or partially known environment. In this embodiment, positioning is initially performed using minimalist SLAM through UWB (ultra-wideband) technology to achieve fast coarse positioning, which provides an initial reference framework for real-time map construction and precise positioning.

[0134] Polar coordinate system: Initial rapid positioning is accomplished through a polar coordinate system, which provides distance and angle information between the device and a reference point (UWB anchor point) for preliminary navigation and obstacle avoidance.

[0135] Transition to Mapping: After completing the initial positioning, the full SLAM algorithm is used to collect lidar sensor data for the construction of real-time high-precision maps. In this process, the position data is converted from polar coordinates to a more refined two-dimensional Cartesian coordinate system, laying the foundation for accurate mapping.

[0136] (2) Transition from polar coordinate system to two-dimensional Cartesian coordinate system in simultaneous positioning and map construction.

[0137] The SLAM process in this embodiment method is a dynamic, continuous update process, including simultaneous positioning and map construction. Polar coordinates provide relative distances and steering angles, while two-dimensional Cartesian coordinates allow for more complex spatial relationship representation and calculation.

[0138] Polar to Cartesian conversion: In SLAM algorithms, data in polar coordinates needs to be converted to Cartesian coordinates to provide a more comprehensive representation of the environment. Mapping is done using Cartesian coordinates, which is the key to accurate map building and can provide data support for advanced path planning and obstacle avoidance.

[0139] (3) Mapping and conversion from point map to surface map.

[0140] Map building is the core part of SLAM, which involves converting various sensor data collected into a map format that can be understood and utilized by machines.

[0141] Conversion from point map to surface map: In the preliminary SLAM process, a point map is usually generated first, that is, a set of discrete points directly marked based on sensor data. These trajectory points represent obstacles and feature points in the environment. Subsequently, these points are used to generate a surface map, that is, by connecting adjacent points and filling the space, a continuous map model that can intuitively describe the environment is formed. The surface map provides more detailed and continuous spatial information for path planning and automatic navigation, enabling unmanned driving equipment to more accurately understand its environment.

[0142] 2. Fusion characteristics of data sources.

[0143] Complementary data usage: In scenarios with high real-time requirements and limited computing resources, data provided by minimalist SLAM is used first, while data from deep SLAM is used in scenarios with higher requirements for high precision.

[0144] Hierarchical processing: Different types of data are used at different operational levels. Basic navigation and obstacle avoidance can use data from minimalist SLAM, while complex mission planning and decision support rely on high-precision maps and lane environment data provided by deep SLAM.

[0145] Dynamic switching and optimization: Dynamically select data to use based on environmental conditions and task requirements, relying on minimalist SLAM when the alley environment is relatively simple, and increasing data input for local nested deep SLAM when the environment is complex or changing.

[0146] In this way, the effective fusion of minimalist SLAM and deep SLAM not only improves the flexibility and adaptability of the system, but also maintains operational efficiency and accuracy, enabling unmanned driving systems to work more reliably in changeable and demanding environments.

[0147] 3. Efficient real-time compression and hierarchical storage of computing data.

[0148] Data storage: Efficient data compression, through advanced data compression algorithms, the system is able to reduce the amount of stored data while maintaining the integrity of key information. Data is allocated to different types of storage media based on access frequency and processing requirements.

[0149] Real-time processing: Multi-threading and parallel processing technologies are used to accelerate data processing and improve the responsiveness of the system, and sensor data is analyzed in real time for rapid decision-making, obstacle avoidance and path planning, trajectory prediction, and environmental modeling.

[0150] 4. Effective constraint features of minimalist SLAM and deep SLAM in safe control of equipment contour and path planning in confined spaces.

[0151] (1) Contour safety control.

[0152] In confined spaces, the management of the physical contours and safety boundaries of unmanned equipment is critical. The main purpose of contour safety control is to ensure that the vehicle remains within a safe operating range at all times and avoids collisions with obstacles in the environment.

[0153] 1) The contour safety control of minimalist SLAM is as follows:

[0154] Real-time feedback and positioning: Take advantage of the fast and real-time position feedback provided by the minimalist SLAM system for basic obstacle avoidance and contour detection. Minimalist SLAM quickly locates the equipment position through simplified sensor inputs (such as UWB) and reacts to upcoming spatial constraints.

[0155] 2) The contour safety control of deep SLAM is as follows:

[0156] Detailed environmental mapping, deep SLAM uses high-precision LiDAR to build a high-precision 3D map of the environment. The algorithm accurately understands the details of the roadway environment, including low-hanging obstacles and narrow passages that pose a threat to the vehicle's outline.

[0157] (2) Path planning

[0158] When it comes to path planning, the system must take into account real-time changing environmental conditions and the physical limitations of the vehicle to determine the safest and most efficient route.

[0159] 1) The path planning of minimalist SLAM is as follows:

[0160] Fast path adjustment: In situations where quick decisions are needed, the basic data provided by Minimalist SLAM is sufficient to perform simple path adjustments and decisions. In emergency obstacle avoidance situations, an alternative path to bypass obstacles can be quickly calculated.

[0161] 2) The path planning of deep SLAM is as follows:

[0162] Precise navigation in complex environments: Deep SLAM provides more complex path planning capabilities through detailed environmental analysis. Using data obtained from detailed maps, you can plan a path that adapts to complex terrain and accurately adjust the route in the changing terrain of the road or adjacent areas.

[0163] Through the effective integration of minimalist SLAM and deep SLAM, the operation of unmanned driving systems in confined spaces is not limited to basic navigation and obstacle avoidance, but can also perform highly complex environmental interactions and decision-making. This system design improves the safety and efficiency of operations, especially in application scenarios that require extremely high accuracy and response speed.

[0164] The unmanned driving method in a multi-road condition restricted space of this embodiment specifically includes the following contents:

[0165] 1. Unmanned driving method used under simple road conditions.

[0166] The unmanned driving method used in this embodiment under simple road conditions specifically includes the following contents:

[0167] 1. Description of the SLAM source in this embodiment.

[0168] (1) SLAM includes real-time positioning.

[0169] This embodiment combines the UWB precise positioning system and the production scheduling system to plan the expected trajectory points of the unmanned equipment and generate a plane minimalist map. The UWB precise positioning system provides real-time position feedback and realizes the instant positioning function within the working range, thereby ensuring the accurate navigation of the unmanned equipment.

[0170] (2) SLAM includes real-time composition and posture correction.

[0171] By using the collaborative work of LiDAR and UWB, this embodiment simplifies the traditional map matching method and uses UWB to instantly locate and correct the position of equipment in a confined space. This process provides real-time position and posture information of the equipment in the confined space without relying on complex map construction, ensuring high-precision environmental perception and rapid response.

[0172] (3) SLAM includes orientation and trajectory planning.

[0173] In response to the safety and collision avoidance requirements in confined spaces, this embodiment converts the posture planning of the plane Cartesian coordinate system into the direction planning described by the polar coordinate system. The vector direction and steering amplitude are accurately calculated through laser radar and UWB data, where the boundary angle of the confined space is used as a passive factor and the driving direction is used as an active factor to ensure that the unmanned driving equipment always drives in the correct target direction.

[0174] 2. The independence and simplicity of data sources.

[0175] (1) LiDAR data: used for preliminary environment perception and position correction.

[0176] (2) UWB data: provides precise location information and minimalist trajectory points for real-time positioning and path constraints, simplifying map creation.

[0177] (3) Polar coordinate data: The position coordinates are extracted from the lidar and UWB data and then converted into polar coordinates for path planning and direction determination.

[0178] The data sources are independent and complementary. By combining UWB data with LiDAR data, efficient and accurate SLAM functions can be achieved. UWB data provides basic positioning, and LiDAR data is used for detail correction to ensure efficient operation of the system in confined spaces.

[0179] 3. Low storage rate characteristics of computing data.

[0180] (1) Data storage optimization: Utilize the minimalist track points provided by UWB to reduce the storage requirements of large-scale map data. Only key track points and necessary location information need to be stored, which greatly reduces the amount of data storage.

[0181] (2) Real-time processing: Simplified data structure and storage method enable the system to achieve real-time processing and rapid response under low storage rate conditions, ensuring efficient operation in complex dynamic environments.

[0182] 4. Constraint characteristics of contour safety control and path planning of unmanned driving equipment in confined spaces.

[0183] (1) Contour safety control: Based on the SLAM output results, the driving direction and position of the equipment are corrected in real time to ensure safe operation in confined spaces. LiDAR data and UWB data are used to accurately perceive environmental boundaries and obstacles, and dynamic obstacle avoidance and path correction are achieved.

[0184] (2) Path planning constraints: In a confined space, local path planning is performed using a polar coordinate system, and effective constraints are imposed by combining the actual contour and boundary information of the equipment. The minimalist SLAM algorithm (i.e., the unmanned driving method used under simple road conditions in this embodiment) gives priority to safety and driving efficiency when planning paths, and ensures efficient operation of the equipment in complex environments by dynamically adjusting the driving path.

[0185] In summary, the minimalist SLAM algorithm for unmanned driving in confined space conditions makes full use of lidar data and UWB data, describes path planning through polar coordinates, and achieves efficient and accurate environmental perception and positioning. This algorithm simplifies the map creation and data storage of traditional SLAM, and provides efficient contour safety control and path planning constraints.

[0186] The unmanned driving method used in the simple road conditions of this embodiment includes the following steps:

[0187] Step S1: Production task path planning.

[0188] When the unmanned driving system is started, the actual equipment operation route is issued through the production scheduling system based on the actual production task. Figure 3 This route (i.e., path) is planned using the UWB pre-marked track point coordinates (i.e., the pre-marked first UWB track point coordinates 104), wherein the track point coordinates are measured as In order to reduce the UWB positioning error, the UWB positioning error model is adopted as , Indicates the error, Indicates the actual distance, which can be the distance actually measured by humans. Represents the measured distance, which is the distance obtained by UWB method. The trajectory point planning algorithm is ,in , The path matching algorithm is based on the weight obtained by minimizing the variance, and the semivariance function is used to evaluate the correlation of spatial data. , ensuring that the task path sequence is unique to the production plan. The task trajectory point coordinates and the UWB tags involved in this embodiment are measured during deployment to provide initial navigation data, establishing a reference point for real-time position monitoring in step S2.

[0189] Step S2: Real-time position feedback.

[0190] The UWB identification card and laser radar installed on the equipment monitor the equipment position in real time, and transmit the position data to the edge computer through the underground wireless network. The Kalman filter method or other existing methods capable of data fusion in the edge computer fuses the UWB data and the laser radar data, and updates the real-time position coordinate data of the unmanned equipment based on the fused data. The updated position coordinate data provides the necessary input for step S3 to ensure the accurate calculation and real-time adjustment of the driving direction.

[0191] Step S3: Determine the driving direction.

[0192] According to the real-time position coordinate data (i.e., the first current position coordinate) provided in step S2, the current required driving direction in the edge calculator involved in this embodiment is obtained by the UWB solution result obtained by the UWB positioning system and edge computing during the equipment's travel, and the two-dimensional Cartesian coordinate system target travel direction (i.e., the target driving trajectory point coordinate) is converted into the polar coordinate system trajectory direction required for the current position of the equipment, and the polar coordinate conversion formula is: , the polar coordinate system trajectory line 108 can be referred to Figure 4 The polar coordinate solution angle 102 is formed by the real-time calculated polar coordinate system trajectory direction and the center line of the unmanned equipment. The trajectory target algorithm can effectively guide the equipment toward the angle target path point of the next trajectory. The trajectory target algorithm is , Indicates the angle corresponding to the current position coordinates (i.e. the current driving direction 106), represents the expected path (i.e., the polar coordinate system trajectory direction corresponding to the coordinates of the target driving trajectory point), Q and R represent weight factors. The driving direction of the unmanned driving device 107 is calculated based on the angle corresponding to the polar coordinate solution angle 102 and the current position coordinates. The driving direction is calculated by the formula: ,in Indicates the angle adjusted according to real-time data (i.e. polar coordinate solution angle 102), Indicates the target driving direction of the unmanned driving device 107.

[0193] Step S4: safety margin correction.

[0194] Reference Figure 4 , this embodiment involves a first safety boundary correction angle 101 formed by a lane center line (i.e., a first middle line 105) formed by the unmanned driving equipment contour safety control lines (i.e., two first safety boundaries 103) and the equipment's actual direction center line. The first safety boundary correction angle 101 is calculated as: ,in, represents the coordinates on the centerline of the lane, Indicates the current position coordinates, Indicates the angle adjustment required for safety margin correction as correction feedback. By comparing the preset trajectory and the real-time calculated trajectory, the polar coordinate solution angle 102 and the first safety margin correction angle 101 in step S3 are fed back to each other to form the angle and angular velocity output by the equipment control. The equipment dynamically adjusts the driving angle and speed to adapt to the lane contour and maintain a safe distance. The specific calculation method of the angle and angular velocity is as follows:

[0195] (1) Combining the polar coordinate solution angle 102 and the first safety margin correction angle 101, the angle of the final control output is calculated by the following formula:

[0196] ,

[0197] in, Indicates the angle of the final control output, represents the safety margin correction factor, It represents the safety margin correction angle difference between the first safety margin correction angle 101 and the polar coordinate solution angle 102 .

[0198] (2) Use the first-order dynamic response model to determine the angular velocity:

[0199] ,

[0200] in, represents the proportional control gain, Represents angular velocity.

[0201] In order to ensure the stability and responsiveness of the system, closed-loop feedback control is used to dynamically adjust and , specifically:

[0202] ,

[0203] in, express The proportional control gain at time, express The proportional control gain at time, and represents the adjustment factor, represents the error between the target position and the actual position, that is , express The safety margin correction factor at the moment, express The safety margin correction factor at the moment.

[0204] The output of step S4 directly affects the generation of control instructions in step S5.

[0205] Step S5: Control output simplification.

[0206] By integrating the analysis results of the above-mentioned steps S1, S2, S3 and S4 of this embodiment to simplify the output of the control instruction, the control output formula may be: ,in, , and They represent proportional gain, integral gain and differential gain respectively, directly converting the unmanned driving control output of confined space equipment into the output of the direction of travel, steering angle and angular velocity under the guidance of UWB, which greatly simplifies the unmanned driving control process of the equipment, thereby improving the simplicity of operation and execution efficiency. The output methods of steering angle and angular velocity include the following:

[0207] (1) Steering angle calculation.

[0208] Define the target direction and the current direction angle , the steering angle difference is : .if If it exceeds 180°, adjust to (corresponding positive adjustment) to ensure the shortest rotation distance, where .

[0209] (2) Angular velocity output.

[0210] Using a proportional-derivative controller (PD controller), the angular velocity is calculated as:

[0211] ,

[0212] in, and They represent proportional gain and differential gain respectively, which are used to adjust response speed and stability. Indicates the rate of change of the angle difference, calculated in real time.

[0213] (3) Angular velocity smoothing and limiting.

[0214] In order to prevent the rapid change of angular velocity from damaging the mechanical system, a limiting function is introduced to smooth the angular velocity:

[0215] ,

[0216] in, Indicates the preset maximum angular velocity.

[0217] (4) Real-time dynamic adjustment.

[0218] Adaptive rules to adjust and Based on the system performance index error and error rate of change :

[0219] ,

[0220] ,

[0221] in, and represents a positive adaptation rate parameter that determines the speed and response of parameter adjustment, Express Seeking guidance, Express Derivative.

[0222] Parameter update mechanism and The value of is updated in real time according to the above adaptive method, with the minimum error and optimize the response, update the step size Choose based on system processing capability:

[0223] ,

[0224] ,

[0225] in, express The proportional gain at time, express The proportional gain at time, express The differential gain at time, express The differential gain at time.

[0226] In this embodiment, the unmanned driving method adopted under simple road conditions has the following beneficial effects:

[0227] 1. Efficient and accurate real-time positioning.

[0228] This embodiment provides more accurate real-time positioning capabilities than traditional technologies by combining UWB data obtained by UWB (ultra-wideband) positioning technology with laser radar data. This is particularly critical in complex environments such as confined spaces such as mine tunnels, and can significantly improve the accuracy and reliability of navigation.

[0229] 2. Simplified environmental perception and driving direction determination.

[0230] By using a polar coordinate system to determine the direction of travel, this embodiment simplifies the traditional complex map matching and environment perception process. This method reduces the reliance on high-performance processing hardware, reduces the overall cost of the system, and speeds up the decision-making process.

[0231] 3. Dynamic obstacle avoidance and safety enhancement.

[0232] The method of this embodiment can calculate and update the trajectory in real time to cope with dynamic obstacles and changing environmental conditions. The safety margin correction mechanism further ensures the safety of the device during driving and reduces the risk of collision and accidents.

[0233] 4. Low-cost deployment and high adaptability.

[0234] The use of a minimalist SLAM algorithm (i.e., an unmanned driving method used under simple road conditions) has greatly reduced the complexity and deployment cost of the unmanned driving system. As the cost of UWB equipment gradually decreases, the economic benefits of this unmanned driving system are more significant and suitable for a wide range of industrial scenarios, especially in mining areas with limited resources.

[0235] 5. Enhanced control output simplifies operation.

[0236] By simplifying the complex control process into an intuitive output of travel direction and steering angle, the present embodiment improves the convenience and efficiency of operation. This intuitive control mechanism enables the operator to easily control the unmanned device 107 even in extreme environments.

[0237] 6. Safety and driving stability.

[0238] This embodiment pays special attention to safe driving in confined spaces, and ensures safe operation of the vehicle in narrow environments by real-time monitoring of the vehicle's outline and safety boundaries. Dynamic obstacle avoidance and path correction functions further enhance the safety of the system, and can effectively avoid collisions and other safety accidents. Compared with traditional technologies, this embodiment provides higher safety guarantees.

[0239] In summary, this embodiment significantly improves the navigation and control capabilities of the unmanned driving system in complex and restricted environments through precise UWB positioning, dynamic path planning and obstacle avoidance technology in the minimalist SLAM algorithm. Compared with the existing technology, this embodiment has multiple advantages such as high efficiency, low storage, high precision, safety and reliability, and easy deployment, and has important application value and broad market prospects.

[0240] 1. Unmanned driving methods used under complex road conditions.

[0241] The unmanned driving method used in complex road conditions in this embodiment specifically includes the following contents:

[0242] 1. Description of the source of deep SLAM features in this embodiment.

[0243] Deep SLAM (i.e. the unmanned driving method used in complex road conditions in this embodiment) provides accurate navigation and high-precision map construction in complex environments for unmanned driving systems by combining various advanced technologies and algorithms. The key technical features and data sources of deep SLAM are explained in detail below:

[0244] (1) Localization

[0245] In deep SLAM, positioning does not rely solely on basic sensor inputs, but rather integrates data from multiple sources to enhance positioning accuracy. This includes:

[0246] High-precision LiDAR: Provides the precise location of surrounding objects by scanning the environment and generating highly accurate distance data points.

[0247] Inertial Measurement Unit (IMU): Provides the acceleration and rotation rate of the device, helping to correct any positioning errors derived from the LiDAR source.

[0248] (2) Simultaneous

[0249] The deep SLAM system can simultaneously build maps and locate, and the SLAM algorithm has a high degree of synchronization when processing and updating information. Key technologies include:

[0250] Multi-sensor fusion: Synchronously analyze data from different sensors, such as combining LiDAR SLAM data for more stable and robust positioning and mapping.

[0251] Continuous optimization techniques: Use techniques such as particle filtering or Kalman filtering to continuously optimize the global consistency and position accuracy of the map.

[0252] (3) Mapping

[0253] Mapping is one of the core features in deep SLAM, which involves converting the collected lane data into a high-precision real-time 3D map. Technical features include:

[0254] Point cloud generation and processing: Build high-precision three-dimensional spatial maps by processing point cloud data generated by lidar.

[0255] Surface reconstruction: Utilize point cloud data, process meshing through algorithms, and convert point maps into surface maps, thereby providing more intuitive roadway information.

[0256] Dynamic Object Tracking: Annotate and track the movement of dynamic objects on the map to adjust and update the map in real time.

[0257] (4) Independence of data sources.

[0258] The efficiency of deep SLAM comes from its independence from data sources. The algorithm does not rely on a single data source or sensor type, but can integrate information from multiple independent data sources to ensure that when a sensor fails or the data quality is insufficient, other sensors can supplement the necessary information to ensure the continuous operation and accuracy of the algorithm.

[0259] Through these advanced technical features and multiple data sources, Deep SLAM provides the reliability and accuracy required for unmanned and automated systems in complex and dynamic environments, enabling it to play a key role in modern automated applications.

[0260] The unmanned driving method used in complex road conditions in this embodiment includes the following steps:

[0261] Step S1: Production task route planning.

[0262] When the unmanned driving system is started, it will be operated according to the specific equipment operation route planned by the production scheduling system. Figure 3 Using high-precision UWB positioning technology, the UWB terminal pre-marks and measures the coordinates of the mission trajectory points. Its UWB positioning error model is: ,in, represents the positioning error, Indicates the coordinates of the actual measured trajectory points, which can be the distance obtained by actual measurement. Indicates the preset UWB trajectory point coordinates. The trajectory point coordinate measurement formula is: The accurate matching of the driving route adopts the route matching algorithm, which uses the least square method and the A-star algorithm to ensure the best alignment between the path and the predicted trajectory. High-precision laser radar assistance ensures the correspondence with the physical features in the environment and provides the initial data basis for steps S2 and S3.

[0263] The UWB ranging model is:

[0264] ,

[0265] in, and Indicates the coordinates of the UWB marker point, Represents measurement error.

[0266] The UWB ranging model calculates different target points of a complex path under complex road conditions, such as different locations in a mine. In a path planning and navigation system, this distance calculation can be used to determine the path length required for the unmanned device 107 to move from one point to another, thereby helping to plan the optimal route.

[0267] Step S2: Real-time position feedback and environment perception.

[0268] This embodiment involves unmanned driving equipment using the data obtained by the UWB positioning terminal and the IMU to perform data fusion, and use the high-precision laser radar to scan the environment in real time to collect detailed three-dimensional point cloud data of the complex environment underground. Combined with IMU attitude information , and The data is converted and transmitted to the edge computer through the underground wireless network for real-time calculation and data fusion. The Kalman filter method can be used for data fusion. The UWB data, IMU data and LiDAR data are processed by the deep SLAM algorithm to update the equipment position and extract environmental features in real time, create a real-time updated environmental model, and predict potential lane changes and obstacle positions. Provide real-time environmental perception capabilities for step S3 input.

[0269] The attitude information of the above IMU , and Perform edge computing, specifically:

[0270] 1. Data fusion and coordinate conversion: The attitude angle is usually used to convert the tilt sensor data from the device coordinate system to the global coordinate system. The attitude information of the IMU can help correct the tilt sensor reading error caused by the tilt or rotation of the device, thereby obtaining more accurate position information and movement direction.

[0271] 2. Real-time computing in edge computing: IMU attitude information may be used as an input parameter in the edge calculator to calculate the current state of the device in real time and predict the future position or direction of movement. By combining IMU attitude information, the edge calculator can make more accurate estimates and predictions of the device's motion trajectory.

[0272] The above-mentioned UWB data, IMU data and LiDAR data are processed by the deep SLAM algorithm, specifically:

[0273] 1. Data input: UWB provides location information, IMU provides attitude information, and lidar provides three-dimensional point cloud data of the environment.

[0274] 2. SLAM algorithm: The SLAM (Simultaneous Localization and Mapping) algorithm uses this data to update the device's location in real time and build and update a map model of the surrounding environment. The SLAM algorithm can simultaneously estimate the device's location and the map of the surrounding environment (such as lane structure and obstacle locations).

[0275] 3. Environmental feature extraction and prediction: SLAM can not only extract features in the environment in real time, but also predict the location of potential lane changes and obstacles. This is achieved by analyzing historical and current data and predicting the future movement path of the equipment.

[0276] 4. Safe obstacle avoidance: The safe obstacle avoidance module in the SLAM algorithm evaluates the device's motion path and updates the device's driving path to avoid these obstacles when potential dangers or obstacles are detected.

[0277] Therefore, inputting UWB, IMU and laser data into the SLAM algorithm and combining it with the safety obstacle avoidance module of the SLAM algorithm can realize device position update, environmental feature extraction, potential lane changes and obstacle position prediction.

[0278] Step S3: Determine driving direction and optimize route.

[0279] The deep SLAM algorithm combines the data obtained from the lidar, IMU and UWB for depth mapping and feature extraction in step S3. The sensor data is processed using particle filter and Kalman filter technology. The processed data is mainly filtered state estimation, which usually includes the location information of the device (such as (x, y) coordinates) and other related state variables (such as speed and acceleration). These data are the basis for further coordinate transformation and path optimization. At the same time, combined with the UWB data and the results of edge computing, the direction of travel of the two-dimensional Cartesian coordinate system is converted to the trajectory direction of the polar coordinate system required by the equipment, see Figure 5 , and its trajectory conversion algorithm is: Then, according to the coordinates in the converted polar coordinate system, the optimized path selection strategy is determined, the three-dimensional environment mapping is converted into a practical two-dimensional path planning, and the target driving path is obtained by optimizing the path selection strategy. In this process, the dynamic obstacle position is avoided, and the driving path of the unmanned driving equipment is adjusted in real time to ensure the correct driving direction and the safety boundary correction in step S4.

[0280] The particle filter described above can be a nonlinear, non-Gaussian filtering method suitable for processing state estimation in high-noise or nonlinear systems. It approximates state distribution through sampling and weight updating, and is particularly suitable for processing more complex scenarios or sensor data with large uncertainty.

[0281] The above Kalman filter performs well in estimating the state of systems with linear and Gaussian noise. It updates the state estimate and covariance matrix recursively and is a classic linear filtering method.

[0282] The above two methods of using particle filter and Kalman filter to process sensor data together can process different types of data at the same time and improve the accuracy of state estimation. Particle filter can handle highly nonlinear scenarios, while Kalman filter is more efficient and accurate when processing linear parts.

[0283] The above optimization path selection strategy can be the following method:

[0284] These data after particle filtering and Kalman filtering provide the current state of the system (for example, position and speed). After processing by Kalman filtering and particle filtering, the system can obtain more accurate position information and state estimation.

[0285] Then, in the coordinate system transformation, the (x, y) coordinates in the 2D Cartesian coordinate form are converted to the polar coordinate form (r, θ). This transformation is usually done to simplify path planning or analysis, and to express the distance and direction of the device relative to the target point in a more intuitive way.

[0286] Then optimize the path selection. The path optimization strategy is usually based on multiple factors, such as the shortest distance of the path, obstacle avoidance, and steering angle. Specific methods include:

[0287] Select a path with the smallest cost function, which may include factors such as path length, energy consumption, and obstacle avoidance. Then combine historical data and current estimates to select an optimal path. For example, the A-star algorithm is used to evaluate the cost of each path node to find the optimal path.

[0288] The above comprehensively considers the precise location information obtained by particle filtering and Kalman filtering, as well as the data obtained through UWB and edge computing, and ultimately determines the optimal driving path (i.e., the target driving path).

[0289] Step S4: safety boundary correction and closed-loop detection.

[0290] Using the dynamic object tracking and surface reconstruction of the deep SLAM algorithm, the system dynamically adjusts the second safety boundary correction angle 109, and the safety boundary dynamic adjustment calculation is: ,in, To adjust the angle difference, The current driving direction 106 in the polar coordinates of step S3 is used to ensure the safe driving of the equipment in the complex environment of the underground tunnel. Figure 5 At the same time, loop closure detection plays an important role in step S4. When the system detects that the equipment has returned to a known area, When the target driving path is obtained in step S3, position correction is performed to reduce the cumulative error and optimize the navigation accuracy. is the current location, The known target position in the target driving path corresponding to the current position. Through continuous safety margin adjustment, the equipment driving safety and efficiency are ensured.

[0291] exist When the current position of the device is close to the known position, it means that the current position of the device is close to the known position, but this does not mean that the unmanned driving device 107 does not need correction. On the contrary, the system will perform position correction at this time to further reduce the cumulative error and optimize the navigation accuracy. This correction is to eliminate small errors caused by measurement noise or dynamic environmental changes, thereby ensuring that the device always travels along the optimal path and maintains high-precision navigation in complex environments.

[0292] S5: Simplified control output and real-time adjustment.

[0293] This invention combines the advantages of deep SLAM and minimalist SLAM algorithms to output accurate travel direction and steering angle instructions: ,in and They are proportional control gain and differential control gain, respectively, which are used to adjust the control output for quick response. The system dynamically adjusts the control instructions according to real-time environmental data and prediction models, simplifies the control process, and improves the system's response speed and intuitive operation. Through data fusion processing and real-time feedback mechanism, the accuracy and reliability of the control output are ensured.

[0294] Through the above technical solutions, the method of this embodiment not only enhances the environmental perception and positioning capabilities of the unmanned driving system, but also significantly improves the accuracy of navigation and the safety of the system through precise path planning and dynamic adjustment. This integrated method makes the unmanned driving system more adaptable to complex and changing production environments.

[0295] Compared with the prior art, this embodiment has the following beneficial effects:

[0296] This embodiment uses unmanned equipment to detect and obtain the real-time contour of the lane according to the laser radar, effectively integrates minimalist SLAM and deep SLAM, and efficiently plans the dynamic path. The vehicle's steering angle and posture are adjusted by the on-board edge calculator and the forward lane contour, which brings many beneficial effects and significantly improves the performance and reliability of the unmanned driving system:

[0297] 1. Improve positioning accuracy and environmental adaptability.

[0298] By combining the high-precision environmental mapping of the deep SLAM algorithm with the rapid response capability of the minimalist SLAM, this embodiment significantly improves the positioning accuracy of the unmanned driving system in complex environments. This enables the system to more accurately identify and adapt to changes in the surrounding environment, thereby improving the safety and reliability of navigation.

[0299] 2. Optimize resource utilization and computing efficiency.

[0300] This embodiment reduces the computational burden by using a minimalist SLAM algorithm in scenarios with low requirements, and introduces a deep SLAM algorithm in situations with high precision requirements. This dynamic switching strategy optimizes the use of computing resources, improves overall system efficiency, and reduces energy consumption.

[0301] 3. Enhance the ability to handle dynamic obstacles.

[0302] Combining the detailed environmental perception capability of deep SLAM and the fast data processing capability of minimalist SLAM, this embodiment can handle dynamic obstacles more effectively. The system can update and adjust the path planning in real time to adapt to the dynamically changing environment and reduce the risk of collision.

[0303] 4. Improve the reliability and robustness of the system.

[0304] This embodiment enhances the performance of the system under various environmental conditions by combining multi-sensor data fusion and two SLAM technologies. Even when some sensors are limited or fail, the system can still maintain a high level of operational performance.

[0305] This embodiment proposes an unmanned driving method suitable for complex and narrow environments such as underground mines. This method is particularly suitable for applications under confined space conditions, such as various types of articulated equipment such as articulated transport vehicles, and provides an efficient dynamic path planning solution. In these application scenarios, the environment is harsh and the space is limited, and the accuracy and reliability of the navigation system are extremely high. Through the effective fusion of minimalist SLAM and deep SLAM, this system can accurately perceive the environment and update the navigation in real time, ensuring that unmanned equipment can perform tasks safely and efficiently.

[0306] Reference Figure 6The embodiment of the present application also provides an unmanned driving system in a multi-road condition restricted space, the system comprising a road condition information acquisition unit 100, a first method selection unit 200, a first method control unit 300, a second method selection unit 400 and a second method control unit 500, wherein:

[0307] The road condition information acquisition unit 100 is used to acquire road condition information indicating whether the current driving road condition of the unmanned driving device 107 is a simple road condition or a complex road condition;

[0308] The first method selection unit 200 is used to select a first unmanned driving strategy corresponding to the simple road condition if the road condition information is a simple road condition;

[0309] The first method control unit 300 is used to obtain a first control output instruction through a first unmanned driving strategy, and control the unmanned driving device 107 to drive on a simple road condition according to the first control output instruction;

[0310] The second method selection unit 400 is used to select a second unmanned driving strategy corresponding to the complex road condition if the road condition information changes from a simple road condition to a complex road condition;

[0311] The second method control unit 500 is used to obtain a second control output instruction through a second unmanned driving strategy, and control the unmanned driving device 107 to travel on complex road conditions according to the second control output instruction.

[0312] It should be noted that since the unmanned driving system in a constrained space with multiple road conditions in this embodiment and the unmanned driving method in a constrained space with multiple road conditions mentioned above are based on the same inventive concept, the corresponding contents in the method embodiment are also applicable to the system embodiment and will not be described in detail here.

[0313] An embodiment of the present application also provides an electronic device, comprising: at least one control processor and a memory for communicating with the at least one control processor.

[0314] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0315] The non-transient software program and instructions required to implement the above-mentioned embodiment of the unmanned driving method in a multi-road condition restricted space are stored in the memory. When executed by the processor, the above-mentioned embodiment of the unmanned driving method in a multi-road condition restricted space is executed, for example, the above-mentioned Figure 1 The method comprises steps S100 to S500.

[0316] The system embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0317] The present application also provides a computer-readable storage medium storing computer-executable instructions, which are executed by one or more control processors to enable the one or more control processors to execute an unmanned driving method in a multi-road condition restricted space in the above method embodiment, for example, to execute the above described Figure 1 The functions of method steps S100 to S500 in the method.

[0318] It will be appreciated by those skilled in the art that all or some of the steps and systems in the methods disclosed above may be implemented as software, firmware, hardware, and appropriate combinations thereof. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or transient medium). As known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disk (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and can be accessed by a computer. Furthermore, it is well known to those skilled in the art that communication media typically embodies computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0319] The above is a specific description of the preferred implementation of the embodiments of the present application, but the embodiments of the present application are not limited to the above-mentioned implementation methods. Technical personnel familiar with the field can also make various equivalent modifications or substitutions without violating the spirit of the embodiments of the present application. These equivalent modifications or substitutions are all included in the scope defined by the claims of the embodiments of the present application.

Claims

1. An unmanned driving method in a multi-road restricted space, characterized in that: The multiple road conditions include simple road conditions and complex road conditions, and the method includes: Obtaining road condition information indicating whether the current road condition of the unmanned driving device is a simple road condition or a complex road condition; If the road condition information is a simple road condition, selecting a first unmanned driving strategy corresponding to the simple road condition; A first control output instruction is obtained through the first unmanned driving strategy, and the unmanned driving device is controlled to travel on the simple road condition according to the first control output instruction; wherein: The obtaining of the first control output instruction by using the first unmanned driving strategy includes: Performing path planning by using the pre-marked coordinates of the first UWB trajectory point to determine a first target planning path; Acquire first UWB position data and first laser radar data of an unmanned driving device traveling on the first target planning path; Determine a first current position coordinate of the unmanned driving device according to the first UWB position data and the first laser radar data; Determining a polar coordinate solution angle according to the first current position coordinates; Determine two first safety boundaries based on the first laser radar data, and determine a first safety boundary correction angle according to the two first safety boundaries; Calculating a target control angle and a target control angular velocity based on the polar coordinate solution angle and the first safety margin correction angle, and using the target control angle and the target control angular velocity as a first control output instruction; If the road condition information changes from the simple road condition to the complex road condition, selecting a second unmanned driving strategy corresponding to the complex road condition; A second control output instruction is obtained through the second unmanned driving strategy, and the unmanned driving device is controlled to travel on the complex road condition according to the second control output instruction.

2. The unmanned driving method in a multi-road confined space according to claim 1, characterized in that: The step of determining the polar coordinate solution angle according to the first current position coordinates includes: Acquire the target driving trajectory point coordinates corresponding to the first current position coordinates from the first target planning path; Convert the target driving trajectory point coordinates into a polar coordinate system trajectory direction; The angle formed by the trajectory direction of the polar coordinate system and the center line of the unmanned driving device is used as the polar coordinate solution angle.

3. The unmanned driving method in a multi-road restricted space according to claim 1, characterized in that: The determining of two first safety boundaries based on the first laser radar data, and determining a first safety boundary correction angle according to the two first safety boundaries, includes: Inputting the first laser radar data into a SLAM algorithm to obtain two first safety boundaries; confirming a first middle line between the two first safety boundaries; A first angle between the first middle line and the first center line of the unmanned driving device is used as a first safety margin correction angle.

4. The unmanned driving method in a multi-road restricted space according to claim 1, characterized in that: The calculating a target control angle and a target control angular velocity based on the polar coordinate solution angle and the first safety margin correction angle includes: Calculating a safety margin correction angle difference between the first safety margin correction angle and the polar coordinate solution angle; Calculating a target control angle according to the safety margin correction angle difference and the polar coordinate solution angle; According to the target control angle, a target control angular velocity is determined.

5. The unmanned driving method in a multi-road restricted space according to claim 1, characterized in that: The obtaining of the second control output instruction by the second unmanned driving strategy includes: Perform path planning by using the pre-marked coordinates of the second UWB trajectory point to determine a second target planning path; Acquire second laser radar data, IMU posture information, and second UWB position data of the unmanned driving device traveling on the second target planning path; Fusing the second laser radar data, the attitude information of the IMU, and the second UWB position data to obtain fused data; Inputting the fused data into a SLAM algorithm to obtain a first position of the unmanned driving device, an obstacle position, a spatial change, and two second safety boundaries; Processing the second laser radar data, the attitude information of the IMU, and the second UWB position data using a particle filter and a Kalman filter to obtain a second position of the unmanned driving device; Obtaining a target driving path according to the second position of the unmanned driving device, the first position of the unmanned driving device, the obstacle position, and the spatial change; Determining a second safety boundary correction angle and a position correction angle according to the two second safety boundaries and the target driving path; The traveling direction and the steering angle are calculated according to the second safety margin correction angle and the position correction angle, and the traveling direction and the steering angle are used as the second control output instruction.

6. The unmanned driving method in a multi-road restricted space according to claim 5, characterized in that: The determining of the second safety boundary correction angle and the position correction angle according to the two second safety boundaries and the target driving path includes: confirming a second middle line between the two second safety boundaries; Using a second angle between the second middle line and the second middle line of the unmanned driving device as a second safety margin correction angle; The target position corresponding to the current position of the unmanned driving device in the target driving path is obtained, and the angle between the current position and the target position is used as the position correction angle.

7. An unmanned driving system in a multi-road restricted space, characterized in that: The system comprises: A road condition information acquisition unit, used to acquire road condition information indicating whether the current driving road condition of the unmanned driving device is a simple road condition or a complex road condition; A first method selection unit, configured to select a first unmanned driving strategy corresponding to the simple road condition if the road condition information is a simple road condition; The first method control unit is used to obtain a first control output instruction through the first unmanned driving strategy, and control the unmanned driving device to travel on the simple road condition according to the first control output instruction; wherein: The obtaining of the first control output instruction by using the first unmanned driving strategy includes: Performing path planning by using the pre-marked coordinates of the first UWB trajectory point to determine a first target planning path; Acquire first UWB position data and first laser radar data of an unmanned driving device traveling on the first target planning path; Determine a first current position coordinate of the unmanned driving device according to the first UWB position data and the first laser radar data; Determining a polar coordinate solution angle according to the first current position coordinates; Determine two first safety boundaries based on the first laser radar data, and determine a first safety boundary correction angle according to the two first safety boundaries; Calculating a target control angle and a target control angular velocity based on the polar coordinate solution angle and the first safety margin correction angle, and using the target control angle and the target control angular velocity as a first control output instruction; A second method selection unit, configured to select a second unmanned driving strategy corresponding to the complex road condition if the road condition information changes from the simple road condition to the complex road condition; The second method control unit is used to obtain a second control output instruction through the second unmanned driving strategy, and control the unmanned driving device to travel on the complex road condition according to the second control output instruction.

8. An electronic device, characterized in that: It includes at least one control processor and a memory for communicating with the at least one control processor; the memory stores instructions that can be executed by the at least one control processor, and the instructions are executed by the at least one control processor so that the at least one control processor can execute the unmanned driving method in a multi-road constrained space as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to enable a computer to execute the unmanned driving method in a multi-road constrained space as described in any one of claims 1 to 6.

Images