Autonomous mapping method, system, device and storage medium for unmanned vehicle

Abstract

The present invention discloses an autonomous mapping method, system, device and storage medium for an unmanned vehicle. The method comprises: obtaining laser radar data, and constructing a current real-time grid map based on the laser radar data; extracting a plurality of boundaries from the current real-time grid map, and calculating a cost value of each boundary; storing the boundaries and the cost values ​​in a candidate boundary array; taking the centroid of the boundary with the smallest cost value in the candidate boundary array as an initial frontier target point; based on the initial frontier target point, using a frontier target generation algorithm based on a fixed window to generate a new target point; judging whether the new target point has a collision risk, and if so, using a temporary target point production algorithm considering collision safety to generate a temporary target point, and using the temporary target point as a new target point; if not, using a path planning algorithm to drive the unmanned vehicle to navigate to the new target point; during the operation of the unmanned vehicle, repeating the above steps until there is no boundary in the map, thereby completing the autonomous mapping.

Description

Technical Field

[0001] The present invention relates to the field of autonomous exploration and mapping technology for unmanned vehicles, and in particular to an autonomous mapping method, system, device and storage medium for unmanned vehicles. Background Art

[0002] The statements in this section merely mention background art related to the present invention and do not necessarily constitute prior art.

[0003] In recent years, with the increasing intelligence of unmanned vehicle technology, unmanned vehicles have played an increasingly important role in people's production and life, and the research on unmanned driving has received widespread attention. At present, most unmanned vehicles need to build an environmental map of the application scenario in advance. However, the drawing of environmental maps requires manual collection of environmental information, which is a complex process and a large workload. In addition, in actual applications, there are special environments that are inaccessible or difficult to traverse manually, which brings many difficulties to manual map construction. To solve such problems, the autonomous exploration technology of unmanned vehicles in unknown environments without prior information through autonomous navigation to traverse the entire environment and realize autonomous mapping has become a research focus of scholars at home and abroad.

[0004] Unmanned vehicle autonomous exploration and mapping technology is a technology based on SLAM algorithm and path planning algorithm. It autonomously searches for target points with mapped paths to explore unknown areas of the environment and achieve complete mapping of the environment. It involves multiple technologies such as unmanned vehicle positioning, mapping and path planning.

[0005] In the field of autonomous exploration, Yamauchi proposed a boundary-based autonomous exploration algorithm that defines the boundary between known areas and unknown areas as the frontier, and guides the robot to move continuously to the boundary to continuously increase the exploration area of ​​the environment. However, this method does not consider the optimal path, so when the unknown environment area is relatively large, robot exploration will take a lot of time. Dirk Holz proposed to update and store the path length between the robot and all boundary grids in the grid map through the Dijkstra algorithm, shortening the exploration path of autonomous mapping. However, this method only considers the path information between the robot and the boundary, ignoring the effective information carried by the map itself, so the optimization effect of mapping time is general. Topiwala proposed a detection method based on the WFD algorithm. When looking for the map boundary, the algorithm only scans the free grids and occupied grids of the grid map, without scanning the entire map grid. Although this method reduces the time complexity of the boundary exploration algorithm, it lacks the evaluation of candidate boundaries and is prone to invalid and unreachable boundaries.

[0006] In summary, the existing technologies do not take into account the characteristics of unmanned vehicles based on the Ackerman model, such as the inability to turn around on the spot, the long time it takes to reverse, and the lack of sensors at the rear of the vehicle. Therefore, the existing technologies still have many shortcomings, such as the long time it takes to autonomously explore and build maps, the fixed direction of the front target point, the unreachable front target point, and the risk of collision with the target point. Summary of the invention

[0007] In order to solve the problems existing in the current unmanned vehicle autonomous exploration and mapping process, such as the fixed orientation of the frontier target point, the unreachable frontier target point, the risk of collision of the target point, etc., the present invention provides an unmanned vehicle autonomous mapping method, system, device and storage medium;

[0008] On the one hand, it provides an autonomous mapping method for unmanned vehicles, including:

[0009] Obtain lidar data and build a current real-time grid map based on the lidar data;

[0010] Extract several boundaries from the current real-time grid map and calculate the cost value of each boundary; store the boundaries and their cost values ​​into a candidate boundary array in the order of the cost value from small to large;

[0011] The centroid of the boundary with the smallest cost value in the candidate boundary array is used as the initial frontier target point; based on the initial frontier target point, a frontier target generation algorithm based on a fixed window is used to generate a new target point;

[0012] Determine whether the new target point has a collision risk. If so, use a temporary target point generation algorithm that takes collision safety into consideration to generate a temporary target point, and use the temporary target point as the new target point; use a path planning algorithm to drive the unmanned vehicle to navigate to the new target point; if not, use a path planning algorithm to drive the unmanned vehicle to navigate to the new target point;

[0013] During the operation of the unmanned vehicle, the above steps are executed until there are no boundaries in the map, thus completing autonomous mapping.

[0014] On the other hand, it provides an autonomous mapping system for unmanned vehicles, including:

[0015] A construction module is configured to: obtain laser radar data, and construct a current real-time grid map based on the laser radar data;

[0016] The extraction module is configured to: extract a number of boundaries from the current real-time grid map, calculate the cost value of each boundary; store the boundaries and the cost values ​​of the boundaries into a candidate boundary array in the order of the cost value from small to large;

[0017] A generation module is configured to: use the centroid of the boundary with the smallest cost value in the candidate boundary array as an initial frontier target point; based on the initial frontier target point, use a frontier target generation algorithm based on a fixed window to generate a new target point;

[0018] The judgment module is configured to: judge whether there is a collision risk at the new target point; if so, use a temporary target point generation algorithm that considers collision safety to generate a temporary target point, and use the temporary target point as a new target point; use a path planning algorithm to drive the unmanned vehicle to navigate to the new target point; if not, use a path planning algorithm to drive the unmanned vehicle to navigate to the new target point;

[0019] The iteration module is configured to iteratively complete the process of building a module, extracting a module, generating a module and judging a module during the operation of the unmanned vehicle until there are no boundaries in the map, thereby completing autonomous mapping.

[0020] On the other hand, there is also provided an electronic device, comprising:

[0021] a memory for non-transitory storage of computer-readable instructions; and

[0022] a processor for executing the computer readable instructions,

[0023] When the computer-readable instructions are executed by the processor, the method described in the first aspect is executed.

[0024] On the other hand, a storage medium is provided, which non-temporarily stores computer-readable instructions, wherein when the non-temporary computer-readable instructions are executed by a computer, the instructions of the method described in the first aspect are executed.

[0025] On the other hand, a computer program product is provided, comprising a computer program, wherein the computer program is used to implement the method described in the first aspect when running on one or more processors.

[0026] The above technical solution has the following advantages or beneficial effects:

[0027] 1. Aiming at the time-consuming characteristic of the unmanned vehicle's reverse turn based on the Ackerman model, the present invention proposes an improved frontier target point cost function. By introducing the angle cost between the target point and the current direction of the unmanned vehicle, the time-consuming behavior of the unmanned vehicle turning around is reduced, thereby speeding up the mapping speed and reducing the mapping time.

[0028] 2. In order to solve the problems that the target points of traditional technologies are unreachable and have fixed directions, the present invention proposes a frontier target point generation method based on a fixed window, which adjusts the position and direction of the initial frontier target point so that the frontier point position is always in an idle grid, ensuring that the frontier target point is reachable. At the same time, the target point direction is always towards the unknown area, thereby obtaining a higher map gain.

[0029] 3. In order to solve the problem of collision risk of target points, the present invention proposes a temporary target point generation method taking collision safety into consideration, performs a breadth-first search around the frontier target points with collision risk, searches for temporary target points with collision safety, avoids collision risk, and thus ensures the safe driving of the unmanned vehicle. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0031] Figure 1 A schematic diagram of the steps of the autonomous mapping method for an unmanned vehicle based on an improved boundary exploration algorithm proposed in the present invention;

[0032] Figure 2 A schematic diagram of a grid map and a boundary area in the present invention;

[0033] Figure 3 Schematic diagram of the boundary search algorithm in the present invention;

[0034] Figure 4 A schematic diagram of the angle between the initial frontier point of the boundary and the line connecting the position of the unmanned vehicle and the direction of the unmanned vehicle in the present invention;

[0035] FIG. 5( a ) and FIG. 5( b ) are simulation experiment effect diagrams comparing the present invention without considering the orientation cost and with considering the orientation cost;

[0036] Figure 6 Schematic diagram of a method for generating frontier target points based on a fixed window in the present invention;

[0037] Figure 7 This is an experimental effect diagram of the frontier target point generation method in the present invention;

[0038] FIG8(a) and FIG8(b) are experimental effect diagrams of the temporary target point generation method considering collision safety in the present invention;

[0039] Fig. 9 This is a flow chart of the autonomous mapping method for unmanned vehicles based on the improved boundary exploration algorithm in the present invention. DETAILED DESCRIPTION

[0040] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0041] Embodiment 1

[0042] This embodiment provides an autonomous mapping method for an unmanned vehicle;

[0043] like Figure 1 As shown, the autonomous mapping method for an unmanned vehicle includes:

[0044] S101: Acquire laser radar data, and construct a current real-time grid map based on the laser radar data;

[0045] S102: extracting a number of boundaries from the current real-time grid map, calculating the cost value of each boundary; storing the boundaries and the cost values ​​of the boundaries into a candidate boundary array in the order of the cost value from small to large;

[0046] S103: taking the centroid of the boundary with the smallest cost value in the candidate boundary array as the initial frontier target point; based on the initial frontier target point, using the frontier target generation algorithm based on the fixed window to generate a new target point;

[0047] S104: Determine whether there is a collision risk at the new target point. If so, use a temporary target point generation algorithm that takes collision safety into consideration to generate a temporary target point, and use the temporary target point as a new target point; use a path planning algorithm to drive the unmanned vehicle to navigate to the new target point; if not, use a path planning algorithm to drive the unmanned vehicle to navigate to the new target point;

[0048] S105: During the operation of the unmanned vehicle, S101 to S104 are continuously executed until there are no boundaries in the map, thereby completing autonomous map building.

[0049] Further, S101: construct the current real-time grid map based on the lidar data, which is implemented using the Karto algorithm. The grid map includes: occupied grids, idle grids and unknown grids. The occupied grids are defined as: grids with cost values ​​greater than 253 and less than 255; the idle grids are defined as: grids with cost values ​​less than 128; the unknown grids are defined as: grids with cost values ​​equal to 255; the grids with collision risks are defined as: grids with cost values ​​greater than 128 and less than 253; the boundary grids are defined as: unknown grids connected to the idle grids.

[0050] like Figure 2As shown in the figure, if at least one of the four connected grids above, below, left and right of the unknown grid is an idle grid, the unknown grid connected to the idle grid is defined as a boundary grid, also called a boundary point. The area composed of multiple connected boundary grids is a boundary area, also called a boundary.

[0051] In the path planning of unmanned vehicles, the cost map Costmap provided by the move_base function package is actually used, and its cost value Cost range is [0, 255].

[0052] The move_base package is an important navigation package in the robot operating system ROS, which is used to realize the movement and navigation of the robot in the environment. It provides a high-level interface that enables the robot to plan paths, execute paths, avoid obstacles, etc. to achieve autonomous movement.

[0053] Costmap is an important concept in the move_base package. Costmap is a two-dimensional grid map that represents the feasibility and cost information of the robot's surroundings. It divides the robot's surroundings into different areas and assigns a cost value to each area. The cost value reflects the difficulty or risk of the robot moving in the area.

[0054] Furthermore, the step S102: extracting a plurality of boundaries from the current real-time grid image is implemented by adopting a double breadth-first search algorithm.

[0055] Furthermore, if Figure 3 As shown, the step S102: extracting a plurality of boundaries from the current real-time grid map specifically includes:

[0056] S102-1: Obtain the current position of the unmanned vehicle as the initial position S, add the initial position S to the initial search queue, and mark the initial position S as visited, and start the first level of breadth-first search;

[0057] S102-2: Take out the head position of the initial search queue, search its 8 neighborhoods, and store the idle and unvisited grids in the queue. If the neighborhood grid is an unknown grid, and there is at least one idle grid in the 4 neighborhoods of the current neighborhood grid, mark the current neighborhood grid as a boundary grid f, and add the boundary grid to the boundary search queue, perform a second breadth-first search, obtain a boundary area, and store the boundary area in the candidate boundary array;

[0058] S102-3: Continuously execute S102-2 until the search queue is empty.

[0059] Furthermore, the second breadth-first search in S102-2 specifically includes:

[0060] Take out the head position of the boundary search queue, search its 8-neighborhood, and add the grids that meet the boundary grid conditions in the 8-neighborhood to the boundary search queue, and mark the newly added grids as boundary grids and visited grids, and record the number of boundary grids Count at the same time, and continuously update the shortest distance between the boundary grid and the current unmanned vehicle position as the shortest distance dis of the boundary grid min , accumulate the horizontal coordinates x and vertical coordinates y of the boundary grids respectively until the boundary search queue is empty, then obtain the boundary, and then divide the accumulated horizontal coordinates and vertical coordinates by the number of boundary grids respectively to obtain the centroid of the boundary, and use the centroid of the boundary as the initial frontier point of the boundary. The specific formula is as follows:

[0061]

[0062] Then, the cost value of the boundary is obtained by improving the boundary cost function frontier , the specific formula is as follows:

[0063] Cost frontier =α*dis min *Res-β*Conut*Res+γ*θ orient

[0064] Among them, α, β, γ are weight coefficients, Res is the grid map resolution, and dis min is the shortest distance from the boundary to the position of the unmanned vehicle, Conut is the number of boundary points in the boundary, θ orient is the angle between the line connecting the initial frontier point of the boundary and the position of the unmanned vehicle and the direction of the unmanned vehicle, and its value range is [0,π], such as Figure 4 shown.

[0065] It should be understood that the reason for introducing the orientation cost is that for the unmanned vehicle based on the Ackerman model in the present invention, reversing and turning is a relatively time-consuming behavior. Therefore, the introduction of the orientation cost can allow the unmanned vehicle to give priority to the frontier target point in the front, and prevent the unmanned vehicle from frequently reversing and turning, which is a time-consuming behavior, thereby speeding up the mapping speed and shortening the mapping time. The effect can be shown in Figure 5 (a) and Figure 5 (b). Finally, the boundary area is stored in the candidate boundary array and sorted from small to large according to the cost value.

[0066] Furthermore, if Figure 6 As shown, the S103: taking the centroid of the boundary with the smallest cost value in the candidate boundary array as the initial frontier target point; based on the initial frontier target point, using the frontier target generation algorithm based on the fixed window to generate a new target point; specifically includes:

[0067] S103-1: taking the centroid of the boundary with the smallest cost value in the candidate boundary array as the initial frontier target point;

[0068] S103-2: Figure 6 As shown, taking the initial frontier target point as the center, setting the size of the search window, traversing the search window, and obtaining the centroid of the idle area and the unknown area respectively; the idle area is composed of idle grids, and the unknown area is composed of unknown grids;

[0069] S103-3: The centroid of the idle area is used as a new target point, and the direction of the new target point is defined as a direction pointing to the unknown area.

[0070] It should be understood that the centroid of the idle area is used as the new target point, thereby ensuring the accessibility of the target point. At the same time, the direction of the new target point is defined as the direction pointing to the unknown area, so that a higher map gain can be obtained. The laser radar carried by the unmanned vehicle in the present invention can only detect the range of 180° in front of the front of the vehicle. Therefore, if a higher map gain is to be obtained, the front of the vehicle needs to be facing the unknown area. In addition, by always allowing the front target point to face the unknown area, it can also reduce the more time-consuming behavior of reversing to the target point to a certain extent, and can also shorten the map building time, achieving the effect as follows: Figure 7 shown.

[0071] Further, S104: determining whether the new target point has a collision risk, and if so, using a temporary target point generation algorithm that takes collision safety into consideration to generate a temporary target point, including:

[0072] S104-1: First determine whether the cost of the new target point is less than the set safety threshold Cost safe If it is less than the safety threshold, the new target point is directly used as the navigation target point, otherwise the new target point is defined as a target point with a collision risk, and the process goes to S104-2;

[0073] S104-2: The target point with collision risk is used as the starting point of breadth-first search, stored in the search queue, and marked as visited;

[0074] S104-3: Take out the head position of the search queue, search its 8-neighborhood, and store the grids in the 8-neighborhood that have collision risk and have not been visited into the search queue and mark them as visited;

[0075] S104-4: Repeat S104-3 continuously until an idle grid is found, then end the search, and use the found idle grid as a temporary target point and the actual navigation target point to drive the unmanned vehicle to navigate to the actual navigation target point.

[0076] Furthermore, a path planning algorithm is used to drive the unmanned vehicle to navigate to a new target point. The path planning algorithm adopts a path planning algorithm that is a fusion of the A* algorithm and the TEB algorithm. The path planning algorithm is a fusion of the global path planning algorithm and the local path planning algorithm. The global path planning algorithm provides the unmanned vehicle with a global optimal path, while the local path planning algorithm is used to achieve real-time obstacle avoidance. The global path planning algorithm adopted in the present invention is the A* algorithm, and the local path planning algorithm is the TEB algorithm.

[0077] FIG8(a) and FIG8(b) are experimental effect diagrams of the temporary target point generation method considering collision safety in the present invention; Fig. 9 This is a flow chart of the autonomous mapping method for unmanned vehicles based on the improved boundary exploration algorithm in the present invention. The present invention proposes an improved frontier target point cost function, which introduces the angle cost between the target point and the current direction of the unmanned vehicle, reduces the time-consuming behavior of the unmanned vehicle turning around, and thus reduces the time consumption of mapping; proposes a frontier target point generation method based on a fixed window, adjusts the position and direction of the initial frontier target point, ensures that the frontier target point is reachable, and at the same time ensures that the direction is always toward the unknown area, thereby obtaining a higher map gain; proposes a temporary target point generation method considering collision safety, searches the periphery of the frontier target point with collision risk, and searches for temporary target points with collision safety, thereby ensuring the safe driving of the unmanned vehicle.

[0078] Embodiment 2

[0079] This embodiment provides an autonomous mapping system for unmanned vehicles;

[0080] Autonomous mapping system for unmanned vehicles, including:

[0081] A construction module is configured to: obtain laser radar data, and construct a current real-time grid map based on the laser radar data;

[0082] The extraction module is configured to: extract a number of boundaries from the current real-time grid map, calculate the cost value of each boundary; store the boundaries and the cost values ​​of the boundaries into a candidate boundary array in the order of the cost value from small to large;

[0083] A generation module is configured to: use the centroid of the boundary with the smallest cost value in the candidate boundary array as an initial frontier target point; based on the initial frontier target point, use a frontier target generation algorithm based on a fixed window to generate a new target point;

[0084] The judgment module is configured to: judge whether there is a collision risk at the new target point; if so, use a temporary target point generation algorithm that considers collision safety to generate a temporary target point, and use the temporary target point as a new target point; use a path planning algorithm to drive the unmanned vehicle to navigate to the new target point; if not, use a path planning algorithm to drive the unmanned vehicle to navigate to the new target point;

[0085] The iteration module is configured to iteratively complete the process of building a module, extracting a module, generating a module and judging a module during the operation of the unmanned vehicle until there are no boundaries in the map, thereby completing autonomous mapping.

[0086] It should be noted that the above-mentioned construction module, extraction module, generation module, judgment module and iteration module correspond to steps S101 to S105 in Embodiment 1, and the examples and application scenarios implemented by the above-mentioned modules and the corresponding steps are the same, but are not limited to the contents disclosed in the above-mentioned Embodiment 1. It should be noted that the above-mentioned modules, as part of the system, can be executed in a computer system such as a set of computer executable instructions.

[0087] The description of each embodiment in the above embodiments has different emphases. For parts not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0088] The proposed system can be implemented in other ways. For example, the system embodiment described above is only illustrative, and the division of the modules is only a logical function division. In actual implementation, there may be other division methods, such as multiple modules can be combined or integrated into another system, or some features can be ignored or not executed.

[0089] Embodiment 3

[0090] This embodiment also provides an electronic device, including: one or more processors, one or more memories, and one or more computer programs; wherein the processor is connected to the memory, and the one or more computer programs are stored in the memory. When the electronic device is running, the processor executes the one or more computer programs stored in the memory so that the electronic device executes the method described in the above embodiment one.

[0091] It should be understood that in this embodiment, the processor may be a central processing unit CPU, and the processor may also be other general-purpose processors, digital signal processors DSP, application-specific integrated circuits ASIC, off-the-shelf programmable gate arrays FPGA or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0092] The memory may include a read-only memory and a random access memory, and provide instructions and data to the processor. A portion of the memory may also include a non-volatile random access memory. For example, the memory may also store information about the device type.

[0093] In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in a processor or an instruction in the form of software.

[0094] The method in the first embodiment can be directly embodied as a hardware processor, or a combination of hardware and software modules in the processor. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware. To avoid repetition, it will not be described in detail here.

[0095] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with this embodiment can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0096] Embodiment 4

[0097] This embodiment further provides a computer-readable storage medium for storing computer instructions. When the computer instructions are executed by a processor, the method described in the first embodiment is completed.

[0098] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. The autonomous mapping method for unmanned vehicles is characterized by: include: Obtain lidar data and build a current real-time grid map based on the lidar data; Extract several boundaries from the current real-time grid map and calculate the cost value of each boundary; store the boundaries and their cost values ​​into a candidate boundary array in the order of the cost value from small to large; The centroid of the boundary with the smallest cost value in the candidate boundary array is used as the initial frontier target point; Based on the initial frontier target point, a frontier target generation algorithm based on a fixed window is used to generate a new target point; Determine whether the new target point has a collision risk. If so, use a temporary target point generation algorithm that takes collision safety into consideration to generate a temporary target point, and use the temporary target point as the new target point; use a path planning algorithm to drive the unmanned vehicle to navigate to the new target point; if not, use a path planning algorithm to drive the unmanned vehicle to navigate to the new target point; During the operation of the unmanned vehicle, the above method is executed until there are no boundaries in the map, thus completing autonomous mapping; The cost value of the boundary is obtained by improving the boundary cost function , the specific formula is as follows: in, is the weight coefficient, is the grid map resolution, is the shortest distance from the boundary to the position of the unmanned vehicle, is the number of boundary grids, is the angle between the initial frontier point of the boundary and the line connecting the position of the unmanned vehicle and the direction of the unmanned vehicle, and the value range is ; The centroid of the boundary with the smallest cost value in the candidate boundary array is used as the initial frontier target point; based on the initial frontier target point, a frontier target generation algorithm based on a fixed window is used to generate a new target point; specifically, the method includes: The centroid of the boundary with the smallest cost value in the candidate boundary array is used as the initial frontier target point; Taking the initial frontier target point as the center, setting the size of the search window, traversing the search window, and obtaining the centroid of the idle area and the unknown area respectively; the idle area is composed of idle grids, and the unknown area is composed of unknown grids; The centroid of the idle area is used as the new target point, and the direction of the new target point is defined as the direction pointing to the unknown area; Determine whether the new target point has a collision risk. If so, use a temporary target point generation algorithm that takes collision safety into consideration to generate a temporary target point, including: First, determine whether the cost value of the new target point is less than the set safety threshold If it is less than the safety threshold, the new target point is directly used as the navigation target point, otherwise the new target point is defined as a target point with collision risk and the next step is entered; The target point with collision risk is used as the starting point of breadth-first search, stored in the search queue, and marked as visited; Take out the head position of the search queue, search its 8-neighborhood, and store the grids in the 8-neighborhood that have collision risk and have not been visited into the search queue and mark them as visited; Traverse the search queue until an idle grid is found, end the search, and use it as a temporary target point as the actual navigation target point to drive the unmanned vehicle to navigate to the actual navigation target point; The unmanned vehicle is an unmanned vehicle based on the Ackerman model.

2. The autonomous mapping method for an unmanned vehicle as claimed in claim 1, characterized in that: The current real-time grid map is constructed based on the lidar data and implemented using the Karto algorithm. The grid map includes occupied grids, idle grids and unknown grids. The occupied grids are defined as grids with cost values ​​greater than 253 and less than 255; the idle grids are defined as grids with cost values ​​less than 128; the unknown grids are defined as grids with cost values ​​equal to 255; the grids with collision risks are defined as grids with cost values ​​greater than 128 and less than 253; the boundary grids are defined as unknown grids connected to the idle grids.

3. The autonomous mapping method for an unmanned vehicle as claimed in claim 1, characterized in that: Extracting several boundaries from the current real-time grid map is achieved by using a double breadth-first search algorithm.

4. The autonomous mapping method for an unmanned vehicle as claimed in claim 1, characterized in that: Extract several boundaries from the current real-time grid map, including: Get the current position of the unmanned vehicle as the initial position , and the initial position Join the initial search queue and mark the initial position For visited nodes, start the first level of breadth-first search; Take out the head position of the initial search queue, search its 8 neighborhoods, and store the idle and unvisited grids in the queue. If the neighborhood grid is an unknown grid and there is at least one idle grid in the 4 neighborhoods of the current neighborhood grid, mark the current neighborhood grid as a boundary grid. , and at the same time, add the boundary grid to the boundary search queue, perform a second breadth-first search, obtain a boundary area, and store the boundary area in the candidate boundary array; Repeat the previous step until the search queue is empty.

5. The autonomous mapping method for an unmanned vehicle as claimed in claim 4, characterized in that: The second level of breadth-first search includes: Take out the head position of the boundary search queue, search its 8-neighborhood, and add the grids that meet the boundary grid conditions in the 8-neighborhood to the boundary search queue, mark the newly added grids as boundary grids and visited grids, and record the number of boundary grids at the same time , continuously update the shortest distance between the boundary grid and the current unmanned vehicle position as the shortest distance of the boundary grid , the horizontal coordinate of the boundary grid and the vertical coordinate Accumulate them separately until the boundary search queue is empty, then get the boundary, and then divide the accumulated horizontal and vertical coordinates by the number of boundary grids to get the center of mass of the boundary. , taking the centroid of the boundary as the initial frontier point of the boundary, the specific formula is as follows: Then, the cost value of the boundary is obtained by improving the boundary cost function .

6. The autonomous mapping system for unmanned vehicles is characterized by including: A construction module is configured to: obtain laser radar data, and construct a current real-time grid map based on the laser radar data; The extraction module is configured to: extract a number of boundaries from the current real-time grid map, calculate the cost value of each boundary; store the boundaries and the cost values ​​of the boundaries into a candidate boundary array in the order of the cost value from small to large; A generation module is configured to: use the centroid of the boundary with the smallest cost value in the candidate boundary array as the initial frontier target point; Based on the initial frontier target point, a frontier target generation algorithm based on a fixed window is used to generate a new target point; The judgment module is configured to: judge whether there is a collision risk at the new target point; if so, use a temporary target point generation algorithm that considers collision safety to generate a temporary target point, and use the temporary target point as a new target point; use a path planning algorithm to drive the unmanned vehicle to navigate to the new target point; if not, use a path planning algorithm to drive the unmanned vehicle to navigate to the new target point; An iteration module is configured to: during the operation of the unmanned vehicle, iteratively complete the process of building a module, extracting a module, generating a module, and judging a module until there are no boundaries in the map, thereby completing autonomous map building; The cost value of the boundary is obtained by improving the boundary cost function , the specific formula is as follows: in, is the weight coefficient, is the grid map resolution, is the shortest distance from the boundary to the position of the unmanned vehicle, is the number of boundary grids, is the angle between the initial frontier point of the boundary and the line connecting the position of the unmanned vehicle and the direction of the unmanned vehicle, and the value range is ; The centroid of the boundary with the smallest cost value in the candidate boundary array is used as the initial frontier target point; based on the initial frontier target point, a frontier target generation algorithm based on a fixed window is used to generate a new target point; specifically, the method includes: The centroid of the boundary with the smallest cost value in the candidate boundary array is used as the initial frontier target point; Taking the initial frontier target point as the center, setting the size of the search window, traversing the search window, and obtaining the centroid of the idle area and the unknown area respectively; the idle area is composed of idle grids, and the unknown area is composed of unknown grids; The centroid of the idle area is used as the new target point, and the direction of the new target point is defined as the direction pointing to the unknown area; Determine whether the new target point has a collision risk. If so, use a temporary target point generation algorithm that takes collision safety into consideration to generate a temporary target point, including: First, determine whether the cost value of the new target point is less than the set safety threshold If it is less than the safety threshold, the new target point is directly used as the navigation target point, otherwise the new target point is defined as a target point with collision risk and the next step is entered; The target point with collision risk is used as the starting point of breadth-first search, stored in the search queue, and marked as visited; Take out the head position of the search queue, search its 8-neighborhood, and store the grids in the 8-neighborhood that have collision risk and have not been visited into the search queue and mark them as visited; Traverse the search queue until an idle grid is found, end the search, and use it as a temporary target point as the actual navigation target point to drive the unmanned vehicle to navigate to the actual navigation target point; The unmanned vehicle is an unmanned vehicle based on the Ackerman model.

7. An electronic device, comprising: a memory for non-transitory storage of computer readable instructions; as well as a processor for executing the computer readable instructions, When the computer-readable instructions are executed by the processor, the method described in any one of claims 1 to 5 is executed.

8. A storage medium characterized by being non-temporary Storing computer-readable instructions, wherein when the non-transitory stored computer-readable instructions are executed by a computer, the method according to any one of claims 1 to 5 is performed.

Images