Unmanned autonomous laser mobile measurement method and device

Abstract

The present application relates to an unmanned autonomous laser mobile measurement method and device, wherein the method includes: obtaining environmental information of at least two target robots and determining their viewpoint sets; evaluating the environmental information of each target robot, and merging the environmental information of at least two target robots to generate merged map information based on the evaluation results; calculating the minimum viewpoint set of each target robot in the at least two target robots, and constructing a global roadmap based on the merged map information and the minimum viewpoint set, and using the global roadmap to calculate the communication cost of each preset communication condition under multiple preset communication conditions, so as to control each target robot to perform corresponding travel operations according to the communication cost. Thus, the existing unmanned autonomous laser detection method is solved, and the performance in a highly dynamic and changeable environment is far from meeting the requirements of high autonomy, adaptability, and high efficiency, and is difficult to be applied to the field of automatic exploration and monitoring.

Description

Technical Field

[0001] The present application relates to the technical field of multi-unmanned system collaborative autonomous exploration and 3D perception positioning and mapping, and in particular to an unmanned autonomous laser mobile measurement method and device. Background Art

[0002] With the rapid development of automation and intelligent technology, the need for positioning, mapping and autonomous exploration of multiple unmanned systems in unknown environments has become particularly critical in areas such as robot collaborative operations, post-disaster rescue and environmental monitoring. Traditional mapping and navigation technologies have poor adaptability and low reliability, and face obvious limitations in complex environments.

[0003] Point cloud-based methods rely on geometric feature matching for point cloud registration, and their performance is highly dependent on the robustness of 3D geometric features. Although the SLAM community has developed a variety of methods to improve the accuracy and robustness of sub-map merging, including the use of semantic objects, bag-of-words vectors, full image descriptors, and sequence-based location recognition, in practical applications, problems such as data association errors and transformation matrix errors still need to be overcome.

[0004] Traditional environmental exploration methods mostly use greedy algorithms, focusing only on short-term optimal solutions while ignoring long-term efficiency. This is particularly true in complex scenarios involving multi-robot systems with limited communication. In addition, existing multi-robot collaborative exploration strategies usually prioritize communication over the exploration task itself, which may have a negative impact on the overall task efficiency. Moreover, these strategies usually rely on heuristic methods, do not fully consider the actual impact of information exchange on exploration efficiency, and ignore adaptability and reliability in complex or harsh environments.

[0005] Faced with the limitations of multi-robot sub-map merging, environmental exploration and collaborative exploration strategies in traditional three-dimensional environments, the performance of existing unmanned autonomous laser detection methods in highly dynamic and changeable environments is far from meeting the requirements of high autonomy, adaptability and high efficiency, and is difficult to apply in the field of automatic exploration and monitoring, which urgently needs to be solved. Summary of the invention

[0006] The present application provides an unmanned autonomous laser mobile measurement method and device to solve the problems that the existing unmanned autonomous laser detection methods are far from meeting the requirements of high autonomy, adaptability, and high efficiency in highly dynamic and changeable environments, and are difficult to apply to the field of automatic exploration and monitoring.

[0007] The first aspect of the present application provides an unmanned autonomous laser mobile measurement method, comprising the following steps: acquiring environmental information of at least two target robots and determining a viewpoint set of the at least two target robots; based on a preset factor graph and an AutoMerge framework, evaluating the environmental information of each of the at least two target robots to obtain an evaluation result, and merging the environmental information of the at least two target robots according to a preset merging condition based on the evaluation result to generate merged map information; based on the viewpoint set, calculating the minimum viewpoint set of each of the at least two target robots according to a preset coverage standard and a dynamic viewpoint reward mechanism, constructing a global roadmap based on the merged map information and the minimum viewpoint set, and using the global roadmap to calculate the communication cost of each preset communication condition under multiple preset communication conditions, so as to control each target robot to perform a corresponding travel operation according to the communication cost.

[0008] Optionally, in one embodiment of the present application, based on the preset factor graph and AutoMerge framework, the environmental information of each of the at least two target robots is evaluated to obtain an evaluation result, and the environmental information of the at least two target robots is merged according to the preset merging conditions based on the evaluation result to generate merged map information, including: constructing the factor graph according to the environmental information of each target robot and the internal connection mode between each target robot; obtaining the overlap length between the environmental information of each target robot, and extracting feature descriptors according to the overlap length; based on the AutoMerge framework, the feature descriptors and the factor graph, performing a matching operation on the environmental information of each target robot, generating a matching sequence pair corresponding to each target robot, and performing validity verification on each matching sequence in the matching sequence pair to obtain. The verification result of each matching sequence is determined; whether the verification result meets the preset validity requirement is determined; if the verification result does not meet the preset validity requirement, the corresponding matching sequence is removed; if the verification result meets the preset validity requirement, the inner connection score of the matching sequence pair is calculated, and whether the inner connection score is greater than the preset score threshold is determined; if the inner connection score is greater than the preset score threshold, the environmental information of each target robot is merged to obtain merged sub-map information; if the inner connection score is not greater than the preset score threshold, the merge distance of the matching sequence pair is calculated, and an active merge operation is performed on the matching sequence pair according to the merge distance to obtain an active merge result; based on the active merge result or the merged sub-map information, the environmental information of each robot is traversed until the preset exploration requirement is met to generate the merged map information.

[0009] Optionally, in an embodiment of the present application, calculating the minimum viewpoint set of each of the at least two target robots based on the viewpoint set according to a preset coverage standard and a dynamic viewpoint reward mechanism includes: determining a coverage surface point standard and a local range space of a target sensor; performing a viewpoint reward comparison operation based on the coverage surface point standard and the local range space to obtain a comparison result; selecting an initial viewpoint in the space according to the comparison result, and calculating an uncovered surface of the initial viewpoint in the space, so as to select subsequent viewpoints through the uncovered surface; dynamically adjusting subsequent viewpoint rewards, and repeatedly performing the viewpoint reward comparison operation according to the dynamically adjusted subsequent viewpoint rewards to generate the minimum viewpoint set.

[0010] Optionally, in an embodiment of the present application, constructing a global roadmap based on the merged map information and the minimum viewpoint set includes: constructing a space subdivision model, dividing a target space into a plurality of sub-target spaces according to the space subdivision model, and marking the plurality of sub-target spaces; obtaining candidate viewpoints based on the plurality of marked sub-target spaces, the minimum viewpoint set, a preset resolution, and a preset reuse strategy, and constructing the global roadmap according to the candidate viewpoints and the historical movement trajectories of each target robot.

[0011] Optionally, in an embodiment of the present application, calculating the communication cost of each of a plurality of preset communication conditions by using the global roadmap, so as to control each target robot to perform corresponding traveling operations according to the communication cost includes: determining the environmental understanding information of each target robot, and constructing an information synchronization model according to the environmental understanding information; when target robot i and target robot j are out of a preset communication range, performing non-communication evaluation, semi-communication evaluation, and assumed communication evaluation on target robot i respectively according to the information synchronization model, and calculating the global paths and communication costs corresponding to the non-communication evaluation, the semi-communication evaluation, and the assumed communication evaluation respectively, where i and j are positive integers; calculating the tracking cost of target robot i tracking target robot j according to the global path and the communication cost, and constructing a specified pursuit cost comparison model based on the tracking cost and the communication cost; when M target robots are within the communication range of N robots, using a preset random sampling strategy to select a communication target from N - M target robots, where M and N are positive integers, and M < N; determining the tracking routes of the M target robots tracking the communication target, and obtaining the tracking results of the M target robots, so as to control each target robot to perform corresponding traveling operations based on the tracking routes and tracking results, in combination with a preset pursuit decision and communication strategy.

[0012] The second aspect of the present application provides an unmanned autonomous laser mobile measurement device, including: an acquisition module, used to acquire environmental information of at least two target robots and determine a viewpoint set of the at least two target robots; an evaluation module, used to evaluate the environmental information of each of the at least two target robots based on a preset factor graph and an AutoMerge framework to obtain an evaluation result, and merge the environmental information of the at least two target robots according to a preset merging condition based on the evaluation result to generate merged map information; a control module, used to calculate the minimum viewpoint set of each of the at least two target robots based on the viewpoint set according to a preset coverage standard and a dynamic viewpoint reward mechanism, construct a global roadmap based on the merged map information and the minimum viewpoint set, and use the global roadmap to calculate the communication cost of each preset communication condition under multiple preset communication conditions, so as to control each target robot to perform a corresponding travel operation according to the communication cost.

[0013] Optionally, in one embodiment of the present application, the evaluation module includes: a first construction unit, used to construct the factor graph according to the environmental information of each target robot and the internal connection mode between each target robot; an extraction unit, used to obtain the overlapping length between the environmental information of each target robot, and extract the feature descriptor according to the overlapping length; a matching unit, used to perform a matching operation on the environmental information of each target robot based on the AutoMerge framework, the feature descriptor and the factor graph, generate a matching sequence pair corresponding to each target robot, and perform validity verification on each matching sequence in the matching sequence pair to obtain a verification result of each matching sequence; a first judgment unit, used to judge whether the verification result meets the preset validity requirement, if the verification result does not meet the preset validity requirement requirement, then remove the corresponding matching sequence; a second judging unit, for calculating the inner connection score of the matching sequence pair if the verification result meets the preset validity requirement, and judging whether the inner connection score is greater than a preset score threshold; a first merging unit, for merging the environmental information of each target robot to obtain merged sub-map information if the inner connection score is greater than the preset score threshold; a second merging unit, for calculating the merge distance of the matching sequence pair if the inner connection score is not greater than the preset score threshold, and performing an active merge operation on the matching sequence pair according to the merge distance to obtain an active merge result; a traversal unit, for traversing the environmental information of each robot based on the active merging result or the merged sub-map information until the preset exploration requirement is met, so as to generate the merged map information.

[0014] Optionally, in an embodiment of the present application, the control module includes: a determination unit configured to determine the coverage surface point standard and the local range space of the target sensor; an execution unit configured to perform a viewpoint reward comparison operation based on the coverage surface point standard and the local range space to obtain a comparison result; a first calculation unit configured to select an initial viewpoint of the space according to the comparison result and calculate the uncovered surface of the initial viewpoint of the space, so as to select subsequent viewpoints through the uncovered surface; an adjustment unit configured to dynamically adjust the subsequent viewpoint reward and repeatedly perform the viewpoint reward comparison operation according to the dynamically adjusted subsequent viewpoint reward to generate the minimum viewpoint set of viewpoints.

[0015] Optionally, in an embodiment of the present application, the control module further includes: a modeling unit configured to construct a space subdivision model, divide the target space into a plurality of sub-target spaces according to the space subdivision model, and label the plurality of sub-target spaces; a second construction unit configured to obtain candidate viewpoints based on the plurality of labeled sub-target spaces, the minimum viewpoint set, a preset resolution, and a preset reuse strategy, and construct the global roadmap according to the candidate viewpoints and the historical movement trajectories of each target robot.

[0016] Optionally, in an embodiment of the present application, the control module further includes: a third construction unit configured to determine the environmental understanding information of each target robot and construct an information synchronization model according to the environmental understanding information; a second calculation unit configured to perform non-communication evaluation, semi-communication evaluation, and assumed communication evaluation on the target robot i respectively according to the information synchronization model when the target robot i and the target robot j are out of the preset communication range, and calculate the global paths and communication costs corresponding to the non-communication evaluation, the semi-communication evaluation, and the assumed communication evaluation respectively, where i and j are positive integers; a third calculation unit configured to calculate the tracking cost of the target robot i tracking the target robot j according to the global path and the communication cost, and construct a specified pursuit cost comparison model based on the tracking cost and the communication cost; a selection unit configured to, when M target robots are within the communication range of N robots, select a communication target from N - M target robots by using a preset random sampling strategy, where M and N are positive integers and M < N; a tracking unit configured to determine the tracking routes of the M target robots tracking the communication target and obtain the tracking results of the M target robots, so as to control each target robot to perform corresponding traveling operations based on the tracking routes and the tracking results, in combination with a preset pursuit decision and a communication strategy.

[0017] The third aspect of the present application provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the unmanned autonomous laser mobile measurement method as described in the above embodiment.

[0018] The fourth aspect of the present application provides a computer-readable storage medium, which stores a computer program. When the program is executed by a processor, it implements the unmanned autonomous laser mobile measurement method as described above.

[0019] The fifth aspect of the present application provides a computer program product, including a computer program, which is executed to implement the above-mentioned unmanned autonomous laser mobile measurement method.

[0020] Therefore, the embodiments of the present application have the following beneficial effects:

[0021] The embodiment of the present application can obtain the environmental information of at least two target robots and determine the viewpoint set of at least two target robots; based on the preset factor graph and AutoMerge framework, evaluate the environmental information of each target robot in the at least two target robots, obtain the evaluation result, and merge the environmental information of at least two target robots according to the preset merging conditions according to the evaluation result to generate merged map information; calculate the minimum viewpoint set of each target robot in the at least two target robots according to the preset coverage standard and dynamic viewpoint reward mechanism, build a global roadmap based on the merged map information and the minimum viewpoint set, and use the global roadmap to calculate the communication cost of each preset communication condition under multiple preset communication conditions, so as to control each target robot to perform the corresponding travel operation according to the communication cost. The present application integrates the advantages of laser radars on multiple vehicles, overcomes the limitations of each single robot and traditional methods through efficient map merging and communication navigation algorithms, and provides map construction and navigation solutions with both high precision and high reliability in a variety of complex environments. Thus, the existing unmanned autonomous laser detection methods are solved, which are far from meeting the requirements of high autonomy, adaptability, and high efficiency in highly dynamic and changeable environments, and are difficult to be applied in the field of automatic exploration and monitoring.

[0022] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0024] Figure 1A flowchart of an unmanned autonomous laser mobile measurement method provided according to an embodiment of the present application;

[0025] Figure 2 A schematic diagram of execution logic of an unmanned autonomous laser mobile measurement method provided for one embodiment of the present application;

[0026] Figure 3 A schematic diagram of an inner join calculation and adaptive merging process provided for an embodiment of the present application;

[0027] Figure 4 A schematic diagram of a viewpoint sampling local planning process provided for an embodiment of the present application;

[0028] Figure 5 A schematic diagram of a spatial subdivision global planning process provided for an embodiment of the present application;

[0029] Figure 6 A schematic diagram of a pursuit strategy information synchronization cost analysis process provided for an embodiment of the present application;

[0030] Figure 7 A schematic diagram of the execution logic of an unmanned autonomous laser mobile measurement system provided for one embodiment of the present application;

[0031] Figure 8 This is an example diagram of an unmanned autonomous laser mobile measurement device according to an embodiment of the present application;

[0032] Fig. 9 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0033] Among them, 10 is an unmanned autonomous laser mobile measurement device; 100 is an acquisition module, 200 is an evaluation module, 300 is a control module; 901 is a memory, 902 is a processor, and 903 is a communication interface. DETAILED DESCRIPTION

[0034] 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 intended to be used to explain the present application, and should not be construed as limiting the present application.

[0035] The following describes the unmanned autonomous laser mobile measurement method and device of the embodiment of the present application with reference to the accompanying drawings. In view of the problems mentioned in the above background technology, the present application provides an unmanned autonomous laser mobile measurement method, in which the environmental information of at least two target robots is obtained and the viewpoint set of at least two target robots is determined; based on the preset factor graph and AutoMerge framework, the environmental information of each target robot in the at least two target robots is evaluated to obtain an evaluation result, and the environmental information of at least two target robots is merged according to the preset merging conditions according to the evaluation result to generate merged map information; the minimum viewpoint set of each target robot in the at least two target robots is calculated according to the preset coverage standard and dynamic viewpoint reward mechanism, and a global route map is constructed based on the merged map information and the minimum viewpoint set, and the communication cost of each preset communication condition under multiple preset communication conditions is calculated using the global route map, so as to control each target robot to perform the corresponding travel operation according to the communication cost. The present application integrates the advantages of laser radars on multiple vehicles, overcomes the limitations of each single robot and traditional methods through efficient map merging and communication navigation algorithms, and provides both high-precision and high-reliability map construction and navigation solutions in a variety of complex environments. This solves the problem that the existing unmanned autonomous laser detection methods are far from meeting the requirements of high autonomy, adaptability, and high efficiency in highly dynamic and changeable environments, and are difficult to apply in the field of automatic exploration and monitoring.

[0036] Specifically, Figure 1 A flow chart of an unmanned autonomous laser mobile measurement method provided in an embodiment of the present application.

[0037] like Figure 1 As shown, the unmanned autonomous laser mobile measurement method includes the following steps:

[0038] In step S101, environmental information of at least two target robots is acquired, and viewpoint sets of at least two target robots are determined.

[0039] The embodiments of the present application can first obtain the robot's environmental information (i.e., map fragments), such as lidar data, point cloud data, and system IMU direction information; in the actual execution process, the embodiments of the present application can use lidar sensors assembled on multiple machines to collect point cloud status, motion status information, and distance information in various environments. The sensors of each vehicle work together to maintain time synchronization and spatial alignment, thereby obtaining the collection of multiple indoor and outdoor scene data.

[0040] Secondly, the embodiment of the present application can construct a viewpoint set model, such as Figure 2 As shown, the specific definition of the viewpoint set model is as follows:

[0041] In this viewpoint set model, Defined as the workspace to be explored, is a traversable subspace; the viewpoint v∈SE(3) describes the posture of the vehicle-mounted sensor v=[p v ,q v ], where p v ∈W trav and q v ∈SO(3) represent the position and direction respectively; L={v1,…,v n :v i ∈SE(3)} is represented as a set of n viewpoints along the vehicle’s past trajectory; is the surface perceived by the sensor at v, and the union of the surfaces perceived at the viewpoint along L is make represents the subset of the sensed surface that has been covered so far during the exploration process; the sensed but not yet covered surface is denoted by

[0042] Afterwards, the embodiment of the present application further needs to define a workspace and a traversable subspace, so as to describe the posture and perception surface of the sensor through a set of viewpoints and determine the coverage path.

[0043] In step S102, based on the preset factor graph and AutoMerge framework, the environmental information of each of the at least two target robots is evaluated to obtain an evaluation result, and the environmental information of the at least two target robots is merged according to the preset merging conditions based on the evaluation result to generate merged map information.

[0044] Furthermore, the embodiment of the present application also needs to maintain a factor graph for representing the internal connections between robots, such as Figure 3 As shown in the figure, the AutoMerge framework is used to evaluate the quality score of the current feature association, and the data association is verified and the overlap length is extended by planning the path. The sub-maps are merged when specific threshold conditions are met, and the relative positions of the merged robots are determined, so as to dynamically adjust and merge the sub-maps generated by each robot to obtain the final merged map information.

[0045] Optionally, in one embodiment of the present application, based on a preset factor graph and an AutoMerge framework, the environmental information of each target robot in at least two target robots is evaluated to obtain an evaluation result, and the environmental information of at least two target robots is merged according to a preset merging condition based on the evaluation result to generate merged map information, including: constructing a factor graph based on the environmental information of each target robot and the internal connection method between each target robot; obtaining the overlap length between the environmental information of each target robot, and extracting a feature descriptor based on the overlap length; based on the AutoMerge framework, the feature descriptor and the factor graph, performing a matching operation on the environmental information of each target robot, generating a matching sequence pair corresponding to each target robot, and verifying the validity of each matching sequence in the matching sequence pair. , to obtain the verification result of each matching sequence; determine whether the verification result meets the preset validity requirement, if the verification result does not meet the preset validity requirement, remove the corresponding matching sequence; if the verification result meets the preset validity requirement, calculate the inner connection score of the matching sequence pair, and determine whether the inner connection score is greater than the preset score threshold; if the inner connection score is greater than the preset score threshold, merge the environmental information of each target robot to obtain merged sub-map information; if the inner connection score is not greater than the preset score threshold, calculate the merge distance of the matching sequence pair, and perform an active merge operation on the matching sequence pair according to the merge distance to obtain an active merge result; based on the active merge result or the merged sub-map information, traverse the environmental information of each robot until the preset exploration requirement is met to generate merged map information.

[0046] It should be noted that the specific process of evaluating and merging data collected by different robots through the factor graph and the AutoMerge framework in the embodiment of the present application, and merging map fragments when specific conditions are met, is as follows:

[0047] Step 1. In the specific implementation process, the embodiment of the present application can maintain a factor graph G = {V, E} through MUI-TARE to represent the internal connection between robots, the node V represents the map fragment obtained by each robot, and the edge E = {ωi, j} represents the connection between these fragments, where ωi, j represents the internal connection between robots vi and vj, and the calculation method is shown in Formula 1. In AutoMerge, for location recognition, the internal connection between two fragments is determined by the overlap length (ie, the number of matching frames) and the difference in location recognition descriptors. The internal connection has a higher score, and the longer the overlap length, the more stable the internal connection.

[0048]

[0049] Among them, F i ,F j represents the concatenation of feature descriptors extracted from overlapping parts, Li,j Indicates v i ,v j The overlap length between w ,σ is a hyperparameter, c w Adjust ω i,j The dependence on the overlap length, ∈, is a constant to avoid division by zero, and only the inner join fraction ω i,j The fragment v that is larger than the set threshold i ,v j The relative positions of the robots after merging can be determined through the transformation matrix.

[0050] Step 2: Stable inner connections usually require a longer overlap distance, but it is difficult to achieve when the robot explores independently. Therefore, a shorter matching frame sequence is needed to detect overlaps, but data association errors and large errors in the transformation matrix must also be avoided. Therefore, the embodiment of the present application can use an adaptive merging strategy and an active method to enable the robot to verify data association and extend the overlap length by planning a path. Once a new inner connection is detected during the exploration process, a robot that is closer to the overlapping area will be used for adaptive merging. The robot will verify and increase the inner connection by actively exploring the potential overlapping area. The merging process will continue until the inner connection is proven to be a data association error or the submaps are successfully merged.

[0051] Step 3: MUI-TARE periodically calls AutoMerge to match the point cloud from the robot and returns the matching sequence pairs. The embodiment of the present application can traverse all matching sequence pairs, verify the inner connection, calculate the transformation matrix, and estimate the merge distance A. diat , the calculation method is shown in Formula 2; if the inner join verification fails, the fault detection mechanism is started and a fallback strategy is adopted until the sub-map merging condition is met;

[0052]

[0053] Among them, ω i,j Indicates the current inner connection, ω thresh represents the inner join threshold guiding map merging, F i and F j is the expected feature descriptor when the overlap length reaches the threshold, A dist is the estimated distance the agent needs to travel to establish a stable inner connection;

[0054] Step 4: The complete exploration path is planned by combining the local planner and the global planner; the local planner first samples the viewpoints covering the surface of the subspace and plans the shortest path by solving the traveling salesman problem; the global planner calculates a set of global coarse paths for the N robots in the submap m In order to travel in the “under exploration” subspace and try to minimize the longest travel distance among all paths, it is defined as follows:

[0055]

[0056] Then, based on the A calculated in step 3 dist , the embodiment of the present application can actively merge the planner to plan a path so that the merge agent accesses the A of the overlapping agent dst Frame viewpoint, and use a greedy strategy to navigate the agent to the nearest uncovered frame viewpoint, and solve the problem of erroneous data association that may occur during active merging through a fault detection mechanism; at the same time, the embodiment of the present application defines a verification gain, and its calculation method is shown in Formula 3, which evaluates the relationship between the increase in internal connections and the duration of active merging.

[0057]

[0058] in, is the current inner connection value, The inner connection value before the proxy starts to actively increase the overlap, C t is the hyperparameter that controls how G(i,j,t) changes over time;

[0059] In the active merging process of the embodiment of the present application, if an erroneous data association is encountered, the overlapping area will not be increased by active merging, so the verification gain decreases as the duration of active merging increases. When the verification gain drops below a certain threshold, the server notifies the robot to exit the active merging and marks the internal connection as invalid. Otherwise, the agent continues to actively merge until the sub-map merging requirements are met. Once a reliable internal connection between sub-maps is established, the sub-maps will be merged into a single sub-map, that is, the final merged map information, thereby providing a reliable data basis and theoretical support for the subsequent construction of the global roadmap.

[0060] In step S103, the minimum viewpoint set of each target robot among at least two target robots is calculated according to the preset coverage standard and the dynamic viewpoint reward mechanism, and a global roadmap is constructed based on the merged map information and the minimum viewpoint set. The global roadmap is used to calculate the communication cost of each preset communication condition under multiple preset communication conditions, so as to control each target robot to perform corresponding movement operations according to the communication cost.

[0061] Furthermore, the embodiments of the present application also need to utilize the coverage standards of the sensors and the dynamic viewpoint reward mechanism to calculate the minimum set of viewpoints to cover all unperceived surfaces to optimize the level of detail of the map; then, the embodiments of the present application can divide the space through known trajectories and create a global roadmap to ensure that the paths are collision-free and connected, while using efficient algorithms to optimize path selection, synchronize the states of multiple robots and solve path planning problems under limited communication conditions; thereafter, the embodiments of the present application can compare costs under different communication assumptions to determine whether to execute a pursuit strategy to optimize information sharing. If the pursuit is unsuccessful, a preset rendezvous strategy is adopted to ensure the efficiency of information exchange.

[0062] Therefore, the embodiments of the present application utilize adaptive fusion and advanced collaborative strategies to integrate and construct a dual-resolution global-local map of three-dimensional space, thereby improving the accuracy of the map; in addition, the embodiments of the present application optimize information sharing and improve communication efficiency through a multi-aircraft formation pursuit strategy, ultimately reducing the algorithm running time and improving the efficiency of environmental exploration.

[0063] Optionally, in one embodiment of the present application, the minimum viewpoint set of each target robot in at least two target robots is calculated according to a preset coverage standard and a dynamic viewpoint reward mechanism, including: determining the coverage surface point standard and the local range space of the target sensor; based on the coverage surface point standard and the local range space, performing a viewpoint reward comparison operation to obtain a comparison result; selecting an initial viewpoint of the space according to the comparison result, and calculating the uncovered surface of the initial viewpoint of the space to select subsequent viewpoints through the uncovered surface; dynamically adjusting subsequent viewpoint rewards, and repeatedly performing the viewpoint reward comparison operation according to the dynamically adjusted subsequent viewpoint rewards to generate a minimum viewpoint set.

[0064] In the specific implementation process, the embodiment of the present application constructs a local map planning path based on viewpoint sampling, such as Figure 4 As shown in the figure, the standard of sensor coverage surface and the configuration space within the local range are specified. The viewpoint reward is dynamically adjusted according to the sub-model of viewpoint sampling to solve the set coverage problem, so as to optimize the selection of additional viewpoints to maximize the total coverage area. The specific process is described as follows:

[0065] Step 1: Define the criteria for sensor coverage of surface points:

[0066] |p s -p v |≤D

[0067]

[0068] Among them, p s To explore the surface patches in space, n sis the normal pointing to the free space side, the viewpoint v is the center point on the surface patch, D and T are two constants, in the specific implementation process, D is set to be shorter than the physical sensor range; it is used to limit the relative distance and direction of the surface patch relative to the viewpoint, and this standard ensures that the surface is well perceived; in addition, the embodiment of the present application sets the local planning range make To consider collision and connectivity, determine the traversable subspace. To consider the corresponding configuration space of rotation and translation;

[0069] Step 2: As more viewpoints are selected, the reward for selecting additional viewpoints decreases. This is mainly because nearby viewpoints have overlapping fields of view and the same surface can be perceived from multiple viewpoints. Therefore, the reward for a viewpoint depends on the previously selected viewpoint. Let For the i-th selected viewpoint, v i Uncovered surface Adjusted to:

[0070]

[0071] Among them, the embodiment of the present application uses the set covering problem from Finding coverage of unperceived surfaces The minimum viewpoint set of Indicates from The uncovered surface perceived in the image represents the sub-modularity of the uncovered area. In the embodiment of the present application, the initial viewpoint set is selected. Dynamically adjust the viewpoint's contribution to uncovered areas i Rewards The surface area A v , to optimize the selection of additional viewpoints to maximize the total coverage area.

[0072] Therefore, the embodiments of the present application calculate the minimum set of viewpoints to cover all unperceived surfaces based on the coverage standard of the sensor and the dynamic viewpoint reward mechanism, thereby optimizing the level of detail of the map.

[0073] Optionally, in one embodiment of the present application, a global roadmap is constructed based on the merged map information and the minimum viewpoint set, including: constructing a spatial subdivision model, and dividing the target space into multiple sub-target spaces according to the spatial subdivision model, and marking the multiple sub-target spaces; based on the marked multiple sub-target spaces, the minimum viewpoint set, the preset resolution and the preset reuse strategy, obtaining candidate viewpoints, and constructing a global roadmap based on the candidate viewpoints and the historical motion trajectory of each target robot.

[0074] It should be noted that the embodiments of the present application can construct a global map planning path based on spatial subdivision, such as Figure 5 As shown, firstly, the embodiment of the present application may specify a spatial subdivision model to construct a global route map based on past trajectories, obtain collision-free viewpoints that can connect unknown areas through viewpoint reuse, and optimize global path selection with a probabilistic integrity algorithm; in addition, the same initial state is adopted under the condition of limited multi-machine communication, and multi-hop communication is used to share new information to solve the VRP. The specific process is as follows:

[0075] Step 1. Specify the spatial subdivision model: divide the space outside H into equal rectangular subspaces. Each subspace stores the covered and uncovered surfaces formed during the exploration process. The data is only used for storage and is retained in the subspace, while the data in H is updated as the exploration proceeds. Each subspace stores the states of "unexplored", "exploring" and "explored". If the subspace does not contain any covered or uncovered surfaces, the state is unexplored; if the subspace only contains covered surfaces, the state is explored; if the subspace contains any uncovered surfaces, the state is explored; only the exploration subspace is considered in the global planning. h∈ represents the exploration subspace, Represented as the set of explored subspaces;

[0076] Step 2: Global roadmap construction: Nodes represent physical traversable locations in the environment. If there is a traversable path between a pair of nodes, they are connected by edges. In order to avoid redundant calculations of collision checks and path planning, the viewpoints obtained by solving the above viewpoint sampling problem are reused; since the viewpoint candidates are in In other words, each viewpoint candidate is collision-free and consistent with Any other viewpoint candidates with traversable paths in the global roadmap are connected, and a subset of viewpoint candidates are reused as nodes in the global roadmap to avoid redundant collisions and connectivity checks. To ensure the sparsity of the roadmap, viewpoint candidates are randomly sampled at a fixed resolution to maintain a sufficiently connected roadmap without incurring high computational costs. The embodiments of the present application may use other data structures or route planning algorithms instead of the global roadmap as long as they have probabilistic completeness when calculating the shortest path between two locations;

[0077] Therefore, it can be understood that in order to find a path through the current viewpoint v current and The global path of the centroid of each subspace in The embodiment of the present application can construct a global roadmap in the traversable space along the past trajectory of the vehicle. The traversable nodes are connected by edges. The redundant calculation of collision check and path planning is avoided by viewpoint reuse to obtain viewpoints without collision and that can connect unknown areas. Finally, fixed-resolution random sampling is adopted for viewpoint candidates to ensure roadmap connection and acceptable computational cost, and A* Algorithms or other probabilistic integrity algorithms are used to optimize the selection of the global path and reduce redundant calculations for collision detection and path planning;

[0078] Step 3: Define a multi-robot exploration model under limited communication: All robots are initialized in the same coordinate system and know the starting positions of other robots in advance. They can communicate when they are physically within a certain distance range. Additionally, robots can communicate through multi-hop, where information is relayed by other intermediate robots. All robots share the current positions of the robots, the states of the explored and unexplored subspaces, and their traversability. Only new information is shared each time to avoid redundant transmissions. At this time, the calculation of the optimal solution for solving the VRP is as follows:

[0079]

[0080] where is a set of global paths for N robots. The lowest-cost solution, i.e., the optimal solution, is shared by all robots and used as the initial guess for further optimization.

[0081] Optionally, in an embodiment of the present application, the communication costs for each preset communication condition under multiple preset communication conditions are calculated using a global roadmap, so as to control each target robot to perform corresponding traveling operations according to the communication costs, including: determining the environmental understanding information of each target robot and constructing an information synchronization model based on the environmental understanding information; when target robot i and target robot j are outside the preset communication range, perform no-communication evaluation, semi-communication evaluation, and assumed communication evaluation on target robot i respectively according to the information synchronization model, and calculate the global paths and communication costs corresponding to the no-communication evaluation, semi-communication evaluation, and assumed communication evaluation respectively, where i and j are positive integers; calculating the tracking cost of target robot i tracking target robot j based on the global path and the communication cost, and constructing a specified pursuit cost comparison model based on the tracking cost and the communication cost; when M target robots are within the communication range of N robots, select communication targets from N - M target robots using a preset random sampling strategy, where M and N are positive integers and M < N; determining the tracking routes of M target robots tracking the communication targets and obtaining the tracking results of M target robots, so as to control each target robot to perform corresponding traveling operations based on the tracking routes and the tracking results, in combination with a preset pursuit decision and communication strategy.

[0082] During the actual execution process, such as Figure 6As shown, the embodiment of the present application can set a multi-robot information synchronization model, calculate the global path planning as the cost without communication and assuming communication, and further calculate the additional cost caused by the pursuit if the cost is lower under the assumed communication. If the additional cost is still lower than the global cost without communication, the pursuit decision is taken. In particular, when the pursuit fails, the communication strategy based on the meeting is adopted to ensure effective information exchange. The specific process is as follows:

[0083] Step 1: Information synchronization model: The robot's understanding of the environment is defined as the set of global subspaces that have been explored and are being explored, using Indicates that each robot keeps track of its own knowledge and the knowledge of other robots; to represent robot i’s knowledge of robot j’s knowledge of the environment, the symbol Where i and j are integers between 1 and N, and the symbol Represents robot i’s own knowledge;

[0084] Before exploring, The prior knowledge of robot j is initialized with robot i, for example, a single global subspace containing the starting position of robot j. When the two robots i and j can communicate with each other, they will synchronize their knowledge with the updated knowledge denoted by superscript t. The synchronization is represented by the following equation:

[0085]

[0086] At the same time, the knowledge understanding of other robots is updated, which is expressed as:

[0087]

[0088] Where k is an integer between 1 and N, and k is not equal to i and j. Here, Represents the information combination of the two robots. When updating the state of the global subspace, the explored state takes precedence over the exploring state.

[0089] Step 2, communication cost comparison model: When robot i and robot j are out of communication range, robot i must decide whether to pursue robot j for communication by evaluating three options (no communication, semi-communication, and hypothetical communication); represents the global subspace currently known only to robot i, such that Without communication, robot i plans the global path The cost is expressed as:

[0090]

[0091] in, No access Under the assumed communication, robot j knows So that:

[0092]

[0093] Robot i can plan the global path as The cost is expressed as:

[0094]

[0095] In the case of semi-communication, Perform spatial division and randomly select half of the regions or sampling points to obtain Make Ensure that robot i can effectively plan the global path based on incomplete data The half communication cost is expressed as:

[0096]

[0097] The preferred embodiment of this application is compared and c, if If is less than c, it can be considered that the communication between robot i and robot j is beneficial, otherwise Compare with c, if c + If it is less than c, it is beneficial for robot i to communicate with robot j, and robot j can share the access workload; however, the above steps require robot i to deviate from its current exploration path, thereby incurring additional costs; thus, the embodiment of the present application can examine the costs incurred by robot i when chasing robot j, and evaluate the situation in which the costs incurred can be justified by potential benefits;

[0098] Step 3: Chasing cost comparison model: estimate the cost of robot i chasing robot j to share relevant The cost of information generated by assuming that robot j follows the global path Explore In the subspace in the image, robot i can adopt a tracking strategy, that is, visit the same subspace to locate robot j; by estimating the time when robot j arrives at each subspace, robot i can plan a path to maximize the possibility of encountering robot j while reducing the travel time. The embodiment of the present application can be approximated by solving the TSP problem constrained by the time window, which calculates the number of subspaces visited by robot i from the current position And finally return the global path of the same position. The path followed by robot i and robot j is expressed as Taking this tracking into account, the total cost is given by:

[0099]

[0100] Assuming that the information exchange between the two robots is successful, if is less than the current cost c, robot i should interrupt its exploration and follow Chase robot j; otherwise, continue to explore; after completing information synchronization, robot i calculates the difference ΔI between the information obtained through pursuit and the currently known information, including the increase in the exploration area ΔG exp , the reduction of unexplored area ΔG unexp and the refinement of the explored area information ΔG redifine ;

[0101] Based on the newly acquired information ΔI and the current resource status R, decide whether to move forward to a new exploration area. This decision is guided by the ratio of information gain to resource consumption ΔI / R to maximize the efficiency of information collection when resources allow. If the information gain exceeds the set threshold and the robot's energy is sufficient to support further exploration tasks, the system will decide to continue exploring, otherwise the robot will return.

[0102] After completing the exploration of a known area, robots should chase other robots to obtain more information about the environment or to pass on information about previously explored spaces. The process of determining whether to chase a robot and where to chase it is the same as described above, with the main goal of minimizing the overall exploration time. However, in this case, the focus of information exchange shifts to the possible discovery of new exploration areas or avoiding repeated visits to already explored spaces, rather than sharing the workload of exploring a large space as in the typical case.

[0103] Step 4: Chasing Correction Model: When there are a total of When a robot is within the communication range of a total of N robots, a random sampling method is used to iteratively select the communication target from the remaining NM robots;

[0104] Specifically, there are 2 (N-M) possible target robot combinations to communicate, distributing the selection of the target robot in planning cycles, where each planning cycle only selects from a total of 2 (N-M) Selecting a smaller number of combinations from the combinations without replacement, targeting a smaller number of robots (one or two) is sufficient to improve the non-communication strategy, so the biased selection is to prioritize fewer targets, where the probability of selecting a combination with fewer targets is higher than selecting a combination with more targets; In order to determine the route for M robots to pursue the selected target robot, solve the VRP with time window constraints, and the VRP solution for pursuit is jointly optimized and shared by all M robots. If the pursuit attempt fails, a strategy similar to rendezvous is used as a backup, in which the robots return to a predetermined location for rendezvous to ensure effective information exchange and enhance exploration efficiency;

[0105] Understandably, in the worst case, the pursuit strategy degenerates into a rendezvous-based strategy, but due to the symmetry of the robot reasoning, this situation is rare, and a target robot that finds enough new information will also try to pursue, which often meets the pursuit robot halfway, otherwise, the target robots will not stray far from their original exploration path, and the pursuit robot can easily find them.

[0106] Therefore, the embodiment of the present application utilizes adaptive fusion and advanced collaborative strategies to fuse and construct a dual-resolution global-local map of three-dimensional space, thereby improving the accuracy of the map and the efficiency of exploration time, reducing the computing running time, and optimizing information sharing and improving the overall efficiency of the system through a multi-machine formation pursuit strategy. For example, in environments including but not limited to large-scale spaces, undulating terrain, complex topological structures, cluttered obstacles and unstructured environments, the adaptive merging and verification technology under the MUI-TARE structure is used to reduce the exploration time of the embodiment of the present application by 7% to 56%, and speed up the planning time by 13% to 52%, compared with existing methods: Passive-Explorer, SMMR-Explorer, etc. It can effectively fuse and enhance the lidar data on multiple vehicles, reduce the errors of overlapping parts and ensure the efficient merging of sub-maps; in addition, through the dual-resolution map construction technology, compared with the existing NBVP, GB P, MBP method, the running time of the embodiment of the present application is reduced by 50% and always maintained below 1 second, the exploration time efficiency is improved by 80%, and the location environment can be explored faster and more comprehensively; through multi-machine system chasing technology, compared with the traditional strategy based on meeting, the average exploration efficiency of the embodiment of the present application is improved by 39% in the model operation of more than one thousand times; in summary, the embodiment of the present application integrates the advantages of lidars on multiple vehicles, and through efficient map merging and communication navigation algorithms, it not only overcomes the limitations of each single robot and traditional methods, but also provides high-precision and high-reliability map construction and navigation solutions in a variety of complex environments.

[0107] In addition, the present application constructs a corresponding unmanned autonomous laser mobile measurement system based on the unmanned autonomous laser mobile measurement method. The unmanned autonomous laser mobile measurement system mainly includes: sensor module, data transmission module, motion control module, data processing module and data processor and other components.

[0108] Among them, the sensor module is used to obtain lidar data for mapping and navigation, including a multi-machine multi-lidar device;

[0109] The data transmission module is used to connect each sensor with the data processing module;

[0110] The motion control module is used to realize the planned path, including a four-wheel drive ROS robot device;

[0111] A data processing module is used to process the data collected by the sensor module and perform unmanned autonomous laser mobile measurement;

[0112] The data processor includes: sub-map adaptive fusion module, dual-resolution map construction module, pursuit strategy cost analysis module, etc.

[0113] Specifically, the above submap adaptive fusion module is used to merge only the segments whose inner connection scores are greater than the set threshold. Stable inner connection requires a longer overlap distance. The robot verifies data association and extends the overlap length by planning the path. During the exploration process, MUI-TARE periodically calls AutoMerge to match the point cloud from the robot and returns the matching sequence pairs. It traverses all matching sequence pairs, verifies the inner connection, calculates the transformation matrix and estimates the merge distance A. dist ; If the inner connection verification fails, the fault detection mechanism is started and a fallback strategy is adopted until the sub-map merging conditions are met; the merging distance A dist The calculation is as follows:

[0114]

[0115] Among them, ω i,j Indicates the current inner connection, ω thresh is the inner join threshold for guiding map merging, F i and F j is the expected feature descriptor when the overlap length reaches the threshold, A dist is the estimated distance the agent needs to travel to establish a stable inner connection. To evaluate the relationship between the increase in inner connections and the duration of active merging, the gain needs to be verified:

[0116]

[0117] in, is the current inner connection value, The inner connection value before the proxy starts to actively increase the overlap, C t is a hyperparameter that controls how G(i,j,t) changes over time.

[0118] In the submap adaptive fusion module, if wrong data association is encountered, the overlapping area will not be increased by active merging, and the verification gain decreases with the increase of active merging duration. When the verification gain drops below the set threshold, the server notifies the robot to exit the active merging and marks the inner connection as invalid, otherwise the agent continues to actively merge until the submap merging requirements are met. Once reliable inner connections between submaps are established, the submaps will be merged into a single submap;

[0119] Dual-resolution map building module that specifies local viewpoint sampling criteria to ensure that local surfaces are well perceived. Viewpoints are defined and selected through points in space. and viewpoint p v The distance between them and the normal n s The conditions for viewpoint v to cover the surface patch are as follows:

[0120] |p s -p v |≤D

[0121]

[0122] Where D and T are two constants that limit the distance and field of view of the surface patch relative to the viewpoint, respectively. Seeking to cover the surface The viewpoint sampling problem exhibits submodularity. The reward of the viewpoint depends on the previously selected viewpoint and needs to be dynamically adjusted as follows:

[0123]

[0124] in, Represents the viewpoint v i The perceived uncovered surface, while is the total uncovered surface. Global path planning uses A * The algorithm, combined with the viewpoint reuse technology, reduces the redundancy of collision checking and path calculation, and improves the efficiency of path planning. In addition, in a multi-robot system, the path planning is optimized by the vehicle routing problem, and effective coordination and information sharing between robots are achieved. Especially in a communication-restricted environment, the collaborative exploration capability of the system is enhanced through multi-hop communication and dynamic information synchronization.

[0125] Chasing strategy cost analysis module, in which a cost-benefit analysis model for chasing decisions is constructed. This model is used to decide whether a robot should chase another robot to exchange information. This requires evaluating the potential benefits and costs of chasing. The following formula is used to compare the costs before and after communication:

[0126]

[0127] Assuming communication, the cost is calculated as:

[0128]

[0129] In the case of semi-communication, the cost is calculated as:

[0130]

[0131] in, The global path planned for N robots, Global path planning for robots to acquire knowledge of other robots under the assumption of communication, For the global path planning of the robot after acquiring half of the other robots' knowledge, if Less than c, indicating that the benefits of pursuit and communication exceed the costs, otherwise continue to compare with c;

[0132] In addition, in the information synchronization model, robot information synchronization is performed in the following ways:

[0133]

[0134] in, The merging operation of the representation information ensures that after each communication, the global subspace knowledge of the participating robots is up-to-date and complete.

[0135] Afterwards, the embodiment of the present application can calculate the path required for robot i to track robot j through the tracking calculation model after determining that tracking is beneficial, and use the TSP problem with time window constraints to approximate:

[0136]

[0137] in, Trace the desired path for robot j for robot i.

[0138] In addition, it should be noted that Figure 7 As shown, the MID-360 multi-line laser radar in the embodiment of the present application is designed for precise environmental scanning, with an effective working range of up to 70 meters at 80% reflectivity and a minimum blind area of ​​0.1 meters. The 360° horizontal field of view and the -7° to +52° vertical field of view ensure comprehensive spatial coverage. The laser radar outputs a point cloud at a high resolution of 200,000 points per second, with a typical frame rate of 10Hz, connected via 100BASE-TX Ethernet, and synchronized via IEEE 1588-2008 (PTPv2) or GPS. Advanced features include anti-interference capabilities and extremely low false alarm rates in strong sunlight. The integrated ICM40609 IMU supports improved positioning accuracy, which is critical for dynamic environments.

[0139] The HEPBURN P27J data transmission module provides a frequency range of 1.40GHz to 1.46GHz and supports Mesh networking of up to 32 nodes. Its RF bandwidth can be selected from 4 / 8 / 10 / 14 / 20MHz, supports up to 90Mbps stream transmission and 9-hop multi-hop transmission. The module has a single-hop delay of less than 7ms and a network access time of less than 1s. It can operate stably at a line-of-sight distance of 30km between air and ground and a line-of-sight distance of 1-2km between ground and ground. It weighs less than 175g and provides a variety of interfaces, suitable for long-distance, high-bandwidth data transmission needs.

[0140] The features of the four-wheel drive motion control module are as follows: it provides a transmission ratio of 1:27 and a maximum speed of 1.82m / s, can carry a maximum weight of 12kg, and has a deadweight of 7.5kg. The dimensions are 381mm x 466mm x 152mm, and the turning radius is 0m. It is equipped with a 22.2V and 5000mAh battery, supports a maximum running time of 6.5 hours, and a charging time of 2 hours. It is equipped with an S20F 20kgf load sensor, a solid rubber tire with a wheelbase of 125mm, and is driven by an MD36N 35W high-torque motor. It supports wireless data transmission within a range of 500 meters, integrates CAN, serial port and APP remote control functions, is compatible with ROS, has an OLED display, a light and sound prompt system, safety protection measures, provides real-time monitoring and remote operation capabilities, and is suitable for efficient operation in changing environments.

[0141] The data processing module is located in the center of the drone equipment and consists of an embedded high-performance CPU processor and other related processors, including a CPU with 2 cores and 4 threads, a basic operating frequency of 3.5GHz, and a turbo acceleration of up to 4.0GHz. Core TM The i7-7567U CPU, with integrated Intel Iris Plus Graphics 650, runs between 300 and 1100MHz, is suitable for high-performance environments that focus on energy efficiency, and can provide stable performance for general computing and graphics tasks (CPU benchmark), as well as a 32GB LPDDR5 memory module. The module has a built-in Ubuntu operating system and is equipped with an unmanned autonomous laser mobile measurement method that can accept data received by the data acquisition module for real-time processing.

[0142] Therefore, the unmanned autonomous laser mobile measurement system of the present application can also adopt adaptive fusion and advanced collaborative strategies to integrate and construct a dual-resolution global-local map of three-dimensional space, thereby improving the accuracy of the map and the efficiency of exploration time. The system optimizes information sharing through a multi-machine formation pursuit strategy, evaluates and merges data collected by different robots through factor graphs and the AutoMerge framework, and optimizes global path selection through a dynamic viewpoint reward mechanism and spatial subdivision technology, so that it can be applied to fields such as robot collaborative operations, post-disaster rescue, and environmental monitoring. It has significant advantages in improving the accuracy, adaptability, and efficiency of autonomous mobile measurement systems, and is particularly suitable for automatic exploration and monitoring tasks in highly dynamic and changeable environments.

[0143] According to the unmanned autonomous laser mobile measurement method proposed in the embodiment of the present application, the environmental information of at least two target robots is obtained, and the viewpoint set of at least two target robots is determined; based on the preset factor graph and AutoMerge framework, the environmental information of each of the at least two target robots is evaluated to obtain an evaluation result, and the environmental information of at least two target robots is merged according to the preset merging conditions based on the evaluation result to generate merged map information; the minimum viewpoint set of each of the at least two target robots is calculated according to the preset coverage standard and dynamic viewpoint reward mechanism, and a global roadmap is constructed based on the merged map information and the minimum viewpoint set, and the global roadmap is used to calculate the communication cost of each preset communication condition under multiple preset communication conditions, so as to control each target robot to perform corresponding travel operations according to the communication cost. The present application integrates the advantages of laser radars on multiple vehicles, overcomes the limitations of each single robot and traditional methods through efficient map merging and communication navigation algorithms, and provides map construction and navigation solutions with both high accuracy and high reliability in a variety of complex environments.

[0144] Next, the unmanned autonomous laser mobile measurement device proposed according to the embodiment of the present application is described with reference to the accompanying drawings.

[0145] Figure 8 It is a block diagram of an unmanned autonomous laser mobile measurement device according to an embodiment of the present application.

[0146] like Figure 8 As shown, the unmanned autonomous laser mobile measurement device 10 includes: an acquisition module 100 , an evaluation module 200 and a control module 300 .

[0147] The acquisition module 100 is used to acquire environmental information of at least two target robots and determine a viewpoint set of the at least two target robots.

[0148] The evaluation module 200 is used to evaluate the environmental information of each of the at least two target robots based on a preset factor graph and an AutoMerge framework to obtain an evaluation result, and merge the environmental information of the at least two target robots according to a preset merging condition based on the evaluation result to generate merged map information.

[0149] The control module 300 is used to calculate the minimum viewpoint set of each of the at least two target robots based on the viewpoint set according to a preset coverage standard and a dynamic viewpoint reward mechanism, construct a global roadmap based on the merged map information and the minimum viewpoint set, and use the global roadmap to calculate the communication cost of each preset communication condition under multiple preset communication conditions, so as to control each target robot to perform a corresponding travel operation according to the communication cost.

[0150] Optionally, in one embodiment of the present application, the evaluation module 200 includes: a first construction unit, an extraction unit, a matching unit, a first judgment unit, a second judgment unit, a first merging unit, a second merging unit and a traversal unit.

[0151] Wherein, the first construction unit is used to construct the factor graph according to the environmental information of each target robot and the internal connection mode between each target robot.

[0152] The extraction unit is used to obtain the overlap length between each target robot environment information and extract the feature descriptor according to the overlap length.

[0153] A matching unit is used to perform a matching operation on the environmental information of each target robot based on the AutoMerge framework, the feature descriptor and the factor graph, generate a matching sequence pair corresponding to each target robot, and verify the validity of each matching sequence in the matching sequence pair to obtain a verification result for each matching sequence.

[0154] The first judgment unit is used to judge whether the verification result meets the preset validity requirement, and if the verification result does not meet the preset validity requirement, remove the corresponding matching sequence.

[0155] The second judgment unit is configured to calculate the inner connection score of the matching sequence pair if the verification result meets the preset validity requirement, and judge whether the inner connection score is greater than a preset score threshold.

[0156] The first merging unit is configured to merge the environment information of each target robot to obtain merged sub-map information if the inner connection score is greater than the preset score threshold.

[0157] The second merging unit is configured to calculate a merging distance of the matching sequence pair if the inner connection score is not greater than the preset score threshold, and perform an active merging operation on the matching sequence pair according to the merging distance to obtain an active merging result.

[0158] A traversal unit is used to traverse the environmental information of each robot based on the active merging result or the merged sub-map information until a preset exploration requirement is met to generate the merged map information.

[0159] Optionally, in one embodiment of the present application, the control module 300 includes: a determination unit, an execution unit, a first calculation unit and an adjustment unit.

[0160] Wherein, the determination unit is used to determine the coverage surface point standard and the local range space of the target sensor.

[0161] An execution unit is used to perform a viewpoint reward comparison operation based on the coverage surface point standard and the local range space to obtain a comparison result.

[0162] The first calculation unit is used to select an initial viewpoint of the space according to the comparison result, and calculate an uncovered surface of the initial viewpoint of the space, so as to select a subsequent viewpoint through the uncovered surface.

[0163] The adjusting unit is used to dynamically adjust the subsequent viewpoint rewards and repeatedly perform the viewpoint reward comparison operation according to the dynamically adjusted subsequent viewpoint rewards to generate the minimum viewpoint set.

[0164] Optionally, in one embodiment of the present application, the control module 300 further includes: a modeling unit and a second building unit.

[0165] The modeling unit is used to construct a space segmentation model, divide the target space into a plurality of sub-target spaces according to the space segmentation model, and mark the plurality of sub-target spaces.

[0166] The second construction unit is used to obtain candidate viewpoints based on the marked multiple sub-target spaces, the minimum viewpoint set, the preset resolution and the preset reuse strategy, and to construct the global roadmap according to the candidate viewpoints and the historical motion trajectory of each target robot.

[0167] Optionally, in one embodiment of the present application, the control module 300 further includes: a third construction unit, a second calculation unit, a third calculation unit, a selection unit and a tracking unit.

[0168] The third construction unit is used to determine the environmental understanding information of each target robot and construct an information synchronization model according to the environmental understanding information.

[0169] The second calculation unit is used to perform a no-communication evaluation, a semi-communication evaluation and a hypothetical communication evaluation on the target robot i according to the information synchronization model when the target robot i and the target robot j are beyond the preset communication range, and calculate the global path and communication cost corresponding to the no-communication evaluation, the semi-communication evaluation and the hypothetical communication evaluation, respectively, where i and j are positive integers.

[0170] The third calculation unit is used to calculate the tracking cost of the target robot i tracking the target robot j according to the global path and the communication cost, and to construct a prescribed chasing cost comparison model based on the tracking cost and the communication cost.

[0171] A selection unit is used to select communication targets from NM target robots using a preset random sampling strategy when M target robots are within the communication range of N robots, where M and N are positive integers and M <N。

[0172] A tracking unit is used to determine the tracking routes of the M target robots for tracking the communication target, and obtain the tracking results of the M target robots, so as to control each target robot to perform corresponding movement operations based on the tracking routes and tracking results and in combination with preset chasing decisions and communication strategies.

[0173] It should be noted that the above explanation of the unmanned autonomous laser mobile measurement method embodiment is also applicable to the unmanned autonomous laser mobile measurement device of this embodiment, and will not be repeated here.

[0174] The unmanned autonomous laser mobile measurement device proposed in the embodiment of the present application includes an acquisition module for acquiring environmental information of at least two target robots and determining a viewpoint set of at least two target robots; an evaluation module for evaluating the environmental information of each of the at least two target robots based on a preset factor graph and an AutoMerge framework to obtain an evaluation result, and merging the environmental information of the at least two target robots according to the preset merging conditions based on the evaluation result to generate merged map information; a control module for calculating the minimum viewpoint set of each of the at least two target robots according to a preset coverage standard and a dynamic viewpoint reward mechanism, building a global route map based on the merged map information and the minimum viewpoint set, and using the global route map to calculate the communication cost of each preset communication condition under multiple preset communication conditions, so as to control each target robot to perform a corresponding travel operation according to the communication cost. The present application integrates the advantages of laser radars on multiple vehicles, overcomes the limitations of each single robot and traditional methods through efficient map merging and communication navigation algorithms, and provides a map construction and navigation solution with both high accuracy and high reliability in a variety of complex environments.

[0175] Fig. 9 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device may include:

[0176] A memory 901 , a processor 902 , and a computer program stored in the memory 901 and executable on the processor 902 .

[0177] When the processor 902 executes the program, the unmanned autonomous laser mobile measurement method provided in the above embodiment is implemented.

[0178] Furthermore, the electronic device further comprises:

[0179] The communication interface 903 is used for communication between the memory 901 and the processor 902 .

[0180] The memory 901 is used to store computer programs that can be executed on the processor 902 .

[0181] The memory 901 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0182] If the memory 901, the processor 902 and the communication interface 903 are implemented independently, the communication interface 903, the memory 901 and the processor 902 can be connected to each other through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Fig. 9 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0183] Optionally, in a specific implementation, if the memory 901, the processor 902 and the communication interface 903 are integrated on a chip, the memory 901, the processor 902 and the communication interface 903 can communicate with each other through an internal interface.

[0184] The processor 902 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0185] An embodiment of the present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the unmanned autonomous laser mobile measurement method as described above.

[0186] An embodiment of the present application also provides a computer program product, including a computer program, which, when executed, is used to implement the above-mentioned unmanned autonomous laser mobile measurement method.

[0187] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0188] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0189] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.

[0190] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or N wirings (electronic devices), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways as necessary and then storing it in a computer memory.

[0191] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above embodiment, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0192] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0193] In addition, each functional unit in each embodiment of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0194] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present application. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. An unmanned autonomous laser mobile measurement method, characterized in that: The following steps are involved: Acquire environmental information of at least two target robots, and determine a viewpoint set of the at least two target robots; Based on a preset factor graph and an AutoMerge framework, evaluating the environmental information of each of the at least two target robots to obtain an evaluation result, and merging the environmental information of the at least two target robots according to a preset merging condition based on the evaluation result to generate merged map information; Based on the viewpoint set, a minimum viewpoint set of each of the at least two target robots is calculated according to a preset coverage standard and a dynamic viewpoint reward mechanism, a global roadmap is constructed based on the merged map information and the minimum viewpoint set, and a communication cost of each preset communication condition under multiple preset communication conditions is calculated using the global roadmap, so as to control each target robot to perform a corresponding travel operation according to the communication cost; The step of evaluating the environmental information of each of the at least two target robots based on the preset factor graph and the AutoMerge framework to obtain an evaluation result, and merging the environmental information of the at least two target robots according to the evaluation result and a preset merging condition to generate merged map information includes: Constructing the factor graph according to the environmental information of each target robot and the internal connection mode between each target robot; Obtaining the overlapping length between each target robot environment information, and extracting a feature descriptor according to the overlapping length; Based on the AutoMerge framework, the feature descriptor and the factor graph, a matching operation is performed on the environmental information of each target robot to generate a matching sequence pair corresponding to each target robot, and each matching sequence in the matching sequence pair is validated to obtain a verification result of each matching sequence; Determining whether the verification result meets the preset validity requirement, and if the verification result does not meet the preset validity requirement, removing the corresponding matching sequence; If the verification result meets the preset validity requirement, then calculating the inner connection score of the matching sequence pair, and determining whether the inner connection score is greater than a preset score threshold; If the inner connection score is greater than the preset score threshold, merging the environment information of each target robot to obtain merged sub-map information; If the inner connection score is not greater than the preset score threshold, calculating the merging distance of the matching sequence pair, and performing an active merging operation on the matching sequence pair according to the merging distance to obtain an active merging result; Based on the active merging result or the merged sub-map information, traverse the environmental information of each target robot until a preset exploration requirement is met to generate the merged map information; The step of calculating the minimum viewpoint set of each of the at least two target robots based on the viewpoint set according to a preset coverage standard and a dynamic viewpoint reward mechanism includes: Determine the coverage surface point criteria and local range space of the target sensor; Based on the coverage surface point standard and the local range space, performing a viewpoint reward comparison operation to obtain a comparison result; Selecting a spatial initial viewpoint according to the comparison result, and calculating an uncovered surface of the spatial initial viewpoint, so as to select subsequent viewpoints through the uncovered surface; Dynamically adjusting subsequent viewpoint rewards, and repeatedly performing the viewpoint reward comparison operation according to the dynamically adjusted subsequent viewpoint rewards to generate the minimum viewpoint set; The mathematical expression of the coverage surface point standard is: |p s -p v |≤D Among them, p s To explore surface patches in space; p v Indicates the position corresponding to the viewpoint of the vehicle-mounted sensor; n s is the normal pointing to the free space side; the viewpoint v is the center point on the surface patch, and D and T are two constants.

2. The method according to claim 1, characterized in that The constructing a global roadmap based on the merged map information and the minimum viewpoint set includes: Constructing a space segmentation model, dividing the target space into a plurality of sub-target spaces according to the space segmentation model, and marking the plurality of sub-target spaces; Based on the marked multiple sub-target spaces, the minimum viewpoint set, the preset resolution and the preset reuse strategy, candidate viewpoints are obtained, and the global roadmap is constructed according to the candidate viewpoints and the historical motion trajectories of each target robot.

3. The method according to claim 2, characterized in that The method of calculating the communication cost of each preset communication condition under multiple preset communication conditions by using the global roadmap, so as to control each target robot to perform a corresponding moving operation according to the communication cost, includes: Determining the environmental understanding information of each target robot, and building an information synchronization model according to the environmental understanding information; When the target robot i and the target robot j are beyond the preset communication range, the target robot i is evaluated without communication, evaluated semi-communication and evaluated as if it were communication according to the information synchronization model, and the global path and communication cost corresponding to the evaluation without communication, the evaluation semi-communication and the evaluation as if it were communication are calculated, respectively, where i and j are positive integers; Calculating the tracking cost of the target robot i tracking the target robot j according to the global path and the communication cost, and constructing a prescribed chasing cost comparison model based on the tracking cost and the communication cost; When M target robots are within the communication range of N robots, a preset random sampling strategy is used to select communication targets from NM target robots, where M and N are positive integers and M <N; Determine the tracking routes of the M target robots to track the communication target, and obtain the tracking results of the M target robots, so as to control each target robot to perform corresponding travel operations based on the tracking routes and tracking results and in combination with preset chasing decisions and communication strategies.

4. An unmanned autonomous laser mobile measurement device, characterized in that: include: An acquisition module, used to acquire environmental information of at least two target robots and determine a viewpoint set of the at least two target robots; An evaluation module, configured to evaluate the environmental information of each of the at least two target robots based on a preset factor graph and an AutoMerge framework to obtain an evaluation result, and merge the environmental information of the at least two target robots according to a preset merging condition based on the evaluation result to generate merged map information; A control module, configured to calculate a minimum viewpoint set of each of the at least two target robots based on the viewpoint set according to a preset coverage standard and a dynamic viewpoint reward mechanism, construct a global roadmap based on the merged map information and the minimum viewpoint set, and calculate a communication cost of each preset communication condition under multiple preset communication conditions using the global roadmap, so as to control each target robot to perform a corresponding travel operation according to the communication cost; Wherein, the evaluation module comprises: A first construction unit, configured to construct the factor graph according to the environmental information of each target robot and the internal connection mode between each target robot; An extraction unit, used for obtaining the overlap length between the environment information of each target robot, and extracting a feature descriptor according to the overlap length; A matching unit, configured to perform a matching operation on the environmental information of each target robot based on the AutoMerge framework, the feature descriptor and the factor graph, generate a matching sequence pair corresponding to each target robot, and perform validity verification on each matching sequence in the matching sequence pair to obtain a verification result for each matching sequence; A first judging unit, configured to judge whether the verification result meets a preset validity requirement, and if the verification result does not meet the preset validity requirement, remove the corresponding matching sequence; A second judgment unit, configured to calculate an inner connection score of the matching sequence pair if the verification result meets the preset validity requirement, and judge whether the inner connection score is greater than a preset score threshold; A first merging unit, configured to merge the environment information of each target robot to obtain merged sub-map information if the inner connection score is greater than the preset score threshold; a second merging unit, configured to calculate a merging distance of the matching sequence pair if the inner connection score is not greater than the preset score threshold, and perform an active merging operation on the matching sequence pair according to the merging distance to obtain an active merging result; A traversal unit, configured to traverse the environment information of each target robot based on the active merging result or the merged sub-map information until a preset exploration requirement is met, so as to generate the merged map information; The control module comprises: a determination unit for determining a coverage surface point standard and a local range space of a target sensor; an execution unit, configured to execute a viewpoint reward comparison operation based on the coverage surface point standard and the local range space to obtain a comparison result; A first calculation unit, configured to select an initial viewpoint of the space according to the comparison result, and calculate an uncovered surface of the initial viewpoint of the space, so as to select a subsequent viewpoint through the uncovered surface; an adjusting unit, configured to dynamically adjust subsequent viewpoint rewards, and repeatedly perform the viewpoint reward comparison operation according to the dynamically adjusted subsequent viewpoint rewards, so as to generate the minimum viewpoint set; The mathematical expression of the coverage surface point standard is: |p s -p v |≤D Among them, p s To explore surface patches in space; p v Indicates the position corresponding to the viewpoint of the vehicle-mounted sensor; n s is the normal pointing to the free space side; the viewpoint v is the center point on the surface patch, and D and T are two constants.

5. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the unmanned autonomous laser mobile measurement method as described in any one of claims 1 to 3.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the unmanned autonomous laser mobile measurement method as described in any one of claims 1 to 3.

7. A computer program product, comprising a computer program, characterized in that The computer program is executed to implement the unmanned autonomous laser mobile measurement method as described in any one of claims 1 to 3.

Images