Moon rover driving simulation training system

Abstract

The invention provides a lunar rover driving simulation training system, and relates to the field of analog simulation training. The path trace point intelligent modification module is used for displaying the local cost of a selected point in real time according to the click modification of an operator on the initial driving path, and the local cost is used for representing the feasibility of the selected point; after the operator completes the modification of the initial driving path, checking whether the new driving path meets the safety and feasibility requirements or not; under the condition that the new driving path does not meet the safety and feasibility requirements, an operator is prompted to adjust the new driving path; and under the condition that the new driving path meets the safety and feasibility requirements, the new driving path is sent to the lunar rover for execution. According to the method, an operator is effectively guided to avoid a path infeasible region, so that the operator can quickly adjust the path and obtain an optimal solution, the path planning time is shortened, and the task execution efficiency of the lunar rover is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of simulation training, in particular to a lunar rover driving simulation training system. BACKGROUND

[0002] The lunar rover, also known as a lunar exploration vehicle or a lunar rover, is a mobile robot specially designed for exploration and research on the surface of the moon. Lunar rover driving simulation training is of great significance for verifying the path planning capability of the lunar rover, interactive control capability, and analyzing and evaluating the feasibility and safety of the driving plan.

[0003] Existing lunar rover driving simulation training systems mostly use two-dimensional interfaces or simple three-dimensional virtual environments, and operators cannot intuitively perceive the motion state of the lunar rover and the lunar surface environment feedback, lack natural interaction means, and result in low training efficiency. SUMMARY

[0004] The present application provides a lunar rover driving simulation training system to solve the problem of low training efficiency of the existing lunar rover driving simulation training system and improve training efficiency.

[0005] The present application provides a lunar rover driving simulation training system, which comprises a path point trace intelligent modification module; the path point trace intelligent modification module is used for: According to the point selection modification of the operator on the initial driving path, the local cost of the selected point is displayed in real time, and the local cost is used to represent the feasibility of the selected point; the local cost is generated based on the terrain slope, obstacle distribution and communication coverage range around the selected point; After the operator completes the modification of the initial driving path, it is checked whether the new driving path meets the safety and feasibility requirements; if the new driving path does not meet the safety and feasibility requirements, the operator is prompted to adjust the new driving path; if the new driving path meets the safety and feasibility requirements, the new driving path is sent to the lunar rover for execution.

[0006] In some embodiments, the system further comprises an initial driving path generation module; the initial driving path generation module is used for: Based on the constructed three-dimensional terrain model, a regional terrain layer containing a slope layer, an illumination layer, a road roughness layer and an adoptable area layer is generated; based on the regional terrain layer, a plurality of path planning algorithms are used to automatically generate an optimal path from a set starting point to a target point by comprehensively considering the terrain passability factor, the task safety factor and the energy efficiency factor; the optimal path is used as the initial driving path of the lunar rover, and the key path points and the curvature of the initial driving path are displayed.

[0007] In some embodiments, the plurality of path planning algorithms comprises a weighted A-star algorithm, a hybrid A-star algorithm, and algorithm.

[0008] In some embodiments, the system further comprises: a lunar terrain construction module; the lunar terrain construction module is configured to: acquire digital orthographic image (DOM) data and digital elevation model (DEM) data of the lunar surface by a space probe; construct a digital model of the lunar terrain based on the DOM data and the DEM data using a three-dimensional modeling technique; and generate a virtual lunar terrain by real-time rendering of the digital model using a high-definition pipeline rendering technique.

[0009] In some embodiments, the system further comprises: a model adaptive reconstruction module; the model adaptive reconstruction module is configured to: construct a white model of a UV map, a floor plan, and a wireframe based on an original three-dimensional model of a lunar rover; and render the white model by sequentially assigning material properties, adjusting lighting angles, mapping textures, and calibrating colors.

[0010] In some embodiments, the model adaptive reconstruction module is further configured to: perform LOD face reduction on the original three-dimensional model according to the computing power of a mixed reality device.

[0011] In some embodiments, the system further comprises: a path planning result evaluation module; the path planning result evaluation module is configured to: evaluate a final planned path in terms of path safety, path efficiency, path feasibility, and driving accuracy after completion of a task, and visually display the evaluation results.

[0012] In some embodiments, the path safety is determined based on the minimum distance from the path to obstacles, the length proportion of slope segments on the path that exceed the maximum climbing ability of the lunar rover, and the length proportion of communication unreachable areas on the path; the path efficiency is determined based on the ratio of the actual planned path length to the theoretically shortest feasible path length, the ratio of the estimated driving time to the historical average driving time, and the ratio of the total energy consumption on the path to the theoretically optimal energy consumption; the path feasibility is determined based on the proportion of terrain on the path that meets the driving conditions of the lunar rover, and the number of special operations required on the path; the driving accuracy is determined based on the deviation distance between the actual arrival position of the lunar rover and the target point, and the degree of deviation between the actual driving trajectory of the lunar rover and the planned path.

[0013] In some embodiments, the system further comprises: a path planning comprehensive display module; the path planning comprehensive display module is used for: The motion trajectory of the lunar rover along the selected travel route is displayed in the holographic image, and key points in the travel route are highlighted and marked.

[0014] In some embodiments, the system further comprises: a path planning result storage module; the path planning result storage module is used for: The simulation training result and the final planning path are stored in a local database in XML form synchronously.

[0015] The lunar rover travel simulation training system provided by the application can effectively guide the operator to avoid the infeasible area of the path through convenient path point selection modification, local cost dynamic update, real-time monitoring and intelligent prompting, so that the operator can quickly adjust the path and obtain an optimal solution, the path planning time is reduced, and the efficiency of the lunar rover task execution is improved. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the application or prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0017] Figure 1 is a structural schematic diagram of the lunar rover travel simulation training system provided by the embodiments of the application. DETAILED DESCRIPTION

[0018] In order to make the objects, technical solutions and advantages of the application clearer, the technical solutions in the application will be described clearly and completely in the following with reference to the drawings in the application. Obviously, the described embodiments are some embodiments of the application, but not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the protection scope of the application.

[0019] Figure 1 is a structural schematic diagram of the lunar rover travel simulation training system provided by the embodiments of the application, as Figure 1 shown, the embodiments of the application provide a lunar rover travel simulation training system, which comprises a path point trace intelligent modification module. The path point trace intelligent modification module is used for: According to the point selection modification of the operator to the initial driving path, the local cost of the selected point is displayed in real time, and the local cost is used to represent the feasibility of the selected point; the local cost is generated based on the terrain slope, obstacle distribution and communication coverage range around the selected point; After the operator completes the modification of the initial driving path, it is checked whether the new driving path meets the safety and feasibility requirements; in the case that the new driving path does not meet the safety and feasibility requirements, the operator is prompted to adjust the new driving path; in the case that the new driving path meets the safety and feasibility requirements, the new driving path is sent to the lunar rover for execution.

[0020] Specifically, the lunar surface environment is complex and changeable, and there are challenges such as unknown obstacles, rugged terrain and communication limitations, which makes it difficult to simply rely on the initial driving path generated by the algorithm to meet the actual task requirements. Therefore, the system adopts an intelligent path optimization scheme of human-computer cooperation, and realizes the organic combination of algorithm planning and manual intervention through a path point trace intelligent modification module.

[0021] The module supports the operator to make multi-dimensional adjustments to the initial driving path based on professional experience: the route can be optimized by adding or deleting guide points, the driving direction can be changed by dragging the path curvature, or the position of the lunar rover can be adjusted directly in the three-dimensional model. The system checks the rationality of the modification in real time, and immediately issues a warning if the selected point is located in a communication blind area or a dangerous terrain, to ensure that the adjusted route is safe and feasible.

[0022] In the virtual mixed reality environment, multiple operators can simultaneously view the lunar rover path planning and can modify the route through common discussion to ensure that the final planning result meets the task requirements.

[0023] The modification process adopts an intuitive graphical interaction mode, and the operator can conveniently operate through a mouse or a touch screen, and the system synchronously displays the local cost of the selected point, that is, the cost information of the selected point and its surrounding area, to help the operator understand the feasibility of the selected point. If the local cost of the selected point is larger, the feasibility of the selected point is smaller. After the adjustment is completed, the system automatically verifies the feasibility of the new driving path, and focuses on detecting key indicators such as communication coverage and obstacle avoidance. If the verification is passed, the navigation instruction is updated and executed; if there is a risk, targeted optimization suggestions are provided to guide the operator to further adjust.

[0024] The core basis of path optimization is a high-precision cost map, which is generated through multi-dimensional evaluation of terrain slope, obstacle distribution and communication coverage.

[0025] Terrain slope analysis: obtain the slope data of each area on the lunar surface, and assign a higher cost value to the area with a large slope, because steep slope will increase the risk of rollover, landslide and other risks of the lunar rover, and pose a threat to the driving safety.

[0026] Obstacle Distribution Identification: Accurately identify obstacles on the lunar surface, such as rocks, craters, and rubble piles, and calculate a cost based on their size, density, and distribution. Areas with densely packed, large obstacles carry a higher cost, potentially requiring the rover to frequently detour or adjust its position, increasing both difficulty and risk.

[0027] Communication coverage definition: Based on the signal coverage of lunar communication base stations or relay satellites, the communication-reachable and -unreachable areas are marked. The cost of communication-unreachable areas is significantly increased to avoid loss of contact with ground control after the lunar rover enters, which could affect the normal execution of the mission and emergency response.

[0028] The system divides the lunar surface into regular grids, each containing terrain, obstacle, and communication data. This generates a visual cost map, providing a quantitative basis for path planning. Based on this, the system automatically recommends low-risk routes while supporting manual fine-tuning to achieve an optimal balance between safety and efficiency. This collaborative human-machine path optimization mechanism ensures the accuracy of algorithmic planning while retaining the flexibility of human operators to respond to emergencies, significantly improving the lunar rover's mission execution capabilities in complex environments.

[0029] The lunar rover driving simulation training system provided by the embodiment of the present invention effectively guides the operator to avoid infeasible path areas through convenient path point selection and modification, dynamic update of local costs, real-time monitoring and intelligent prompts, enabling the operator to quickly adjust the path and obtain the optimal solution, thereby reducing path planning time and improving the efficiency of lunar rover mission execution.

[0030] In some embodiments, as Figure 1 As shown, the lunar rover driving simulation training system also includes an initial driving path generation module. The initial driving path generation module is used to: Based on the constructed three-dimensional terrain model, a regional terrain layer is generated, including a slope layer, a lighting layer, a rugged road surface layer, and an adoptable area layer. Based on the regional terrain layer, multiple path planning algorithms are used to comprehensively consider terrain accessibility factors, mission safety factors, and energy efficiency factors to automatically generate an optimal path from a set starting point to a target point. The optimal path is used as the initial driving path of the lunar rover, and the key path points and curvature of the initial driving path are displayed.

[0031] Specifically, based on the constructed three-dimensional terrain model, a regional terrain layer containing a slope layer, an illumination layer, a road roughness layer, and an adoptable region layer is generated. Based on the regional terrain layer, the system uses various path planning algorithms to automatically generate an optimal path from the set starting point to the target point, taking into account factors such as terrain accessibility (communication unreachable, terrain restrictions), task safety (driving safety), and energy efficiency (terrain flatness, obstacle distribution, lunar rover energy consumption). The path will serve as the initial driving path for the lunar rover, and the key path points and curvature of the initial driving path will be displayed.

[0032] In some embodiments, the various path planning algorithms include a weighted A-star algorithm, a hybrid A-star algorithm, and a LLM-A-star algorithm. algorithm.

[0033] The weighted A-star algorithm improves the heuristic function by assigning a weight to the heuristic function, allowing the relative size of the actual cost function g and the estimated cost function h to be controlled, thereby balancing search efficiency and path optimality in path planning problems. In certain cases, it can find the optimal solution and reduce the order of magnitude of the search space. The weighted A-star algorithm can shorten the planning time, reduce the algorithm search space, improve the applicability of the algorithm in path planning problems, and improve the efficiency and smoothness of path generation in different environments.

[0034] The hybrid A-star algorithm ( ) combines the advantages of the exploration tree method and the A-star algorithm. The core of this algorithm is the determination of node weights, i.e., the setting of path cost. The system introduces a bidirectional exploration mechanism that combines the heuristic function of the A-star algorithm with path planning considering vehicle kinematic constraints, making the path planning consider both the kinematic characteristics of the vehicle and the distribution of obstacles. The advantage of the hybrid A-star algorithm is that it balances between discrete state space and continuous state space, taking into account both the shortest path and the dynamics of the lunar rover.

[0035] The LLM-A-star algorithm combines the advantages of traditional A-star algorithms and large language models (LLM). The main innovation of the LLM-A-star algorithm is the combination of the global understanding ability of LLM and the precise routing ability of A-star algorithm. The method uses the target state generated by LLM as a waypoint in the path search process, significantly reducing the computational cost and memory usage, By integrating the standard L2 distance heuristic function of A-star with the new heuristic value derived from the waypoint generated by LLM, the granularity problem in the LLM-generated solution is solved, ensuring the effectiveness of the output path while improving the efficiency of the algorithm.

[0036] In some embodiments, the system will use one planning algorithm as the default planning algorithm according to user preference, and provide a quick switching sequence of the default planning algorithm.

[0037] To adapt to the lunar rover path planning task, the following algorithm implementation process is adopted: (1) Initialization Input the starting position and target position of the lunar rover, as well as the lunar environment map containing information such as terrain flatness, obstacle distribution (including information such as the location, size, and type of obstacles), communication reachable area range (location and signal coverage range of communication base stations or relay stations), terrain restrictions (such as slope restrictions, non-drivable areas, etc.), and other information.

[0038] Initialize the OpenList and add the starting position node to it, setting its initial cost to 0, and initialize the ClosedList as empty.

[0039] (2) Node expansion and evaluation Select the current node with the smallest comprehensive cost estimate from the OpenList and move it to the ClosedList to indicate that it has been explored.

[0040] Traverse and process the neighbor nodes of the current node: Check terrain restrictions: Determine whether the neighbor node is located within the terrain restriction range (such as non-drivable areas with excessive slope), and if so, skip that node.

[0041] Communication reachability judgment: Detect whether the neighbor node is located in a communication unreachable area, and if so, discard the node if it exceeds the time or resource consumption range allowed by the communication recovery strategy (the communication recovery strategy can be pre-set according to actual conditions, such as considering the deployment of relay devices within a certain range, but the communication-related factors will be considered in the subsequent path planning cost).

[0042] Safety distance and obstacle processing: Calculate the distance between the neighbor node and the surrounding obstacles, and determine whether it meets the minimum safety distance requirement of the lunar rover. If not, evaluate the detour cost according to the nature of the obstacle (such as whether it can be crossed, the distance required for detour, etc.), and perform the corresponding processing, or directly discard the node.

[0043] Terrain flatness evaluation: Obtain the terrain flatness information of the area where the neighbor node is located, and use it as one of the cost factors. Poor flatness will increase the cost of the node, as uneven terrain may increase the energy consumption of the lunar rover and reduce the safety of travel.

[0044] Energy consumption estimation calculation: Based on factors such as terrain flatness, slope, and driving speed, the energy consumption required from the current node to the neighbor node is estimated as a component of the cost function to guide the algorithm to select a path with lower energy consumption.

[0045] For neighbor nodes that meet the conditions, calculate their new generation value (including the actual cost from the starting node to the node and the heuristic estimated cost from the node to the target node), if the node is already in the OpenList or ClosedList and the new generation value is not better than the previously recorded cost, do not update; otherwise, update the relevant information of the node (such as parent node, generation value, etc.), and add the node to the OpenList.

[0046] (3) Target judgment and path backtracking When the target node is added to the ClosedList, it indicates that a feasible path has been found. Then, starting from the target node, backtrack to the starting node along the parent nodes to form a complete driving path.

[0047] If the OpenList is empty and the target node is not found, it indicates that the target position cannot be reached under the current planning conditions. At this time, you can adjust the relevant parameters according to actual needs (such as increasing the flexibility of communication recovery strategy, relaxing some terrain restrictions, etc.) and then re-plan the path.

[0048] (4) Design of heuristic function and cost function Heuristic function: A heuristic function that considers multiple factors is adopted, which can be in the form of: where, is the straight-line distance from node to the target node , represents the terrain flatness of the area where node is located (the value range is [0, 1], 1 represents the most flat), is the distance from node to the nearest communication reachable area, is the corresponding weight coefficient, used to balance the importance of each factor in the heuristic estimation. These weights can be adjusted according to actual task requirements and lunar environment characteristics.

[0049] Cost function: The cost function considers both actual driving cost and estimated cost, which can be in the form of where, is the actual cost from the starting node to node , is the heuristic estimated cost from node to the target node.

[0050] The actual cost includes driving distance cost, energy consumption cost, etc. The energy consumption cost can be calculated and accumulated in detail according to the driving speed, terrain condition, etc. The driving distance cost is calculated according to the actual path length multiplied by the corresponding weight.

[0051] In some embodiments, as shown in FIG. 1, the lunar rover driving simulation training system further includes a lunar surface terrain construction module. Figure 1 The lunar surface terrain construction module is configured to: acquire Digital Elevation Model (DEM) data and Digital Orthophoto Map (DOM) data of the lunar surface through a space probe; construct a digital model of the lunar terrain based on the DEM data and the DOM data by using a three-dimensional modeling technology; and generate a virtual lunar surface terrain by using a high-definition pipeline rendering technology to perform real-time rendering on the digital model.

[0052] Specifically, the lunar surface terrain construction module uses mixed reality technology to construct the lunar surface terrain, and the core technical process includes three key links of data acquisition, three-dimensional modeling, and virtual rendering.

[0053] First, the DEM data and the DOM data of the lunar surface are acquired through a space probe to form the basic data source of the system. The DOM data provides high-resolution images of the lunar surface, which not only contains texture information of the terrain, but also contains visual features such as color and illumination, which is crucial for constructing a realistic visual model of the lunar surface. The DEM data accurately records the elevation information of the lunar surface, and completely represents the three-dimensional structure and slope characteristics of the lunar terrain.

[0054] In the three-dimensional modeling stage, a double-branch deep convolutional network is used to realize multi-source data fusion. The DOM branch uses a ResNet network architecture to effectively extract multi-scale texture features through its residual connection characteristics. The DEM branch is based on a U-Net architecture, which focuses on extracting key terrain parameters such as elevation gradient and curvature.

[0055] The features of the two branches are adaptively weighted and fused through a channel attention mechanism (SE module), which can dynamically optimize the contribution weight of different data sources and significantly improve the feature expression ability. The fused feature data is converted into a high-precision triangular mesh model through the MarchingCubes algorithm. Through this method, the system can create a three-dimensional terrain model with high detail, ensuring accurate reproduction of macro terrain features (such as impact craters and mountains) and micro details (such as rock distribution).

[0056] In the rendering link, the High Definition Render Pipeline (HDRP) technology is adopted, so that the three-dimensional terrain model can be rendered with extremely high fidelity, providing a solution from offline rendering to real-time rendering. The High Definition Render Pipeline technology realizes an accurate lighting reflection model, real-time shadow generation and multi-level detail rendering through physical level light propagation simulation. This technology makes the virtual lunar surface environment highly consistent with the real environment in terms of optical properties such as diffuse reflection and specular reflection, achieving a realistic degree that is difficult to distinguish visually.

[0057] In some embodiments, as shown in Figure 1 The lunar rover driving simulation training system further comprises a model adaptability reconstruction module. The model adaptability reconstruction module is configured to: Based on the original three-dimensional model of the lunar rover, a white model is built for the UV map, planar graph and wireframe graph required by the business scenario; and the white model is rendered in sequence for material attribute assignment, light angle adjustment, texture mapping and color calibration.

[0058] Specifically, for the adaptability reconstruction and rendering optimization of the three-dimensional model of the lunar rover (or a patrol vehicle or an exploration vehicle), the present application proposes a set of Level of Detail (LOD) dynamic regulation technical solutions based on mixed reality computing power support. Through structure optimization and physical property enhancement of the original three-dimensional model, the balance between high-fidelity visual effect and real-time rendering performance is achieved, and the specific implementation process is as follows: In the model preprocessing stage, based on the basic model library of mechanical structure models such as Chang'e series lunar rovers and Spirit Mars rovers, the physical parameter constraints of the kinematic parameters of each component (mast, wheel) are embedded, and the 6DoF pose data of the Mixed Reality (MR) head-mounted display and the environment perception data of the gaze point tracking information are obtained in real time.

[0059] Based on the original three-dimensional model of the lunar rover, the UV map reconstruction, planar graph topology optimization and wireframe graph structuring processing required by the business scenario are performed to build a white model. The white model is rendered in sequence for material attribute assignment, light angle adjustment, texture mapping and color calibration. By assigning different material properties to the model, the surface effects of different materials such as metal and rock can be simulated. At the same time, the setting of the light angle will affect the light and shadow effect of the model, which is particularly important for simulating the real lighting conditions on the lunar surface. The addition of texture and color further enhances the details and realism of the model, so as to complete the adaptability reconstruction of the three-dimensional model, so that the model can meet the needs of lunar rover driving simulation training.

[0060] In some embodiments, the model adaptability reconstruction module is further configured to: According to the computing power of the mixed reality device, the LOD face reduction processing is performed on the original three-dimensional model.

[0061] Specifically, under the support of the computing power of the mixed reality device, the original three-dimensional model of the lunar rover is subjected to Level of Detail (LOD) reduction processing, allowing the model to be presented at different levels of detail at different distances and angles, thereby optimizing performance and maintaining visual quality, improving display smoothness and rendering speed in scene running. This technology combines physical constraint perception and visual saliency analysis to achieve dynamic LOD control of the model.

[0062] An improved feature-preserving reduction algorithm based on Quadric Error Metrics (QEM) is used to reduce the number of large-scale surfaces of non-critical components (such as the vehicle body shell); at least 70% of the original details of high-dynamic components (such as the end of the robotic arm) are retained, and the simplification threshold is dynamically adjusted through curvature sensitivity analysis: wherein, is the simplification threshold, is the principal curvature, is the motion component weight coefficient.

[0063] Based on visual saliency detection (using U-Net to segment the gaze area), the focus area is kept at 2048x2048 texture, and the peripheral area is downgraded to 512x512 intelligent texture compression.

[0064] In some embodiments, as shown in Figure 1 the lunar rover driving simulation training system further includes a path planning comprehensive display module. The path planning comprehensive display module is configured to: superimpose and display the motion trajectory of the lunar rover along the selected travel route in the holographic image, and highlight and mark the key points in the travel route.

[0065] Specifically, the motion trajectory of the lunar rover along the selected travel route is superimposed and displayed in the holographic image, and the key points in the travel route are highlighted and marked, which can be scientific target points, navigation nodes or dangerous areas, etc., facilitating the operator to study and modify the presented travel route.

[0066] In some embodiments, as shown in Figure 1 the lunar rover driving simulation training system further includes a path planning result storage module. The path planning result storage module is configured to: synchronously store the simulation training results and the final planned path in the form of XML in a local database.

[0067] Specifically, the simulation training result is data generated by the system in the simulation environment, and the final planned path is the final path planning scheme determined by the system after calculation or optimization. The simulation training result and the final planned path are stored in the local database in synchronization. The storage form is extensible markup language (XML) format, which has the characteristics of strong readability, easy parsing and exchange.

[0068] In some embodiments, as shown in Figure 1 The lunar rover driving simulation training system further includes a path planning result evaluation module. The path planning result evaluation module is configured to: After the task is completed, the final planned path is evaluated in terms of path safety, path efficiency, path feasibility and driving accuracy, and the path evaluation result is visualized and displayed.

[0069] Specifically, an evaluation index system including path safety, path efficiency, path feasibility and driving accuracy is constructed. After the lunar rover path planning task is completed, the final planned path is evaluated in terms of path safety, path efficiency, path feasibility and driving accuracy, and the path evaluation result is visualized and displayed, so as to provide improvement measures for the planned path of the lunar rover in the future and provide further reference for further training, teaching and discussion in the future.

[0070] In some embodiments, the path safety is determined based on the minimum distance from the path to the obstacles, the length proportion of the slope segment on the path that exceeds the maximum climbing ability of the lunar rover, and the length proportion of the communication unreachable area on the path; The path efficiency is determined based on the ratio of the actual planned path length to the theoretically shortest feasible path length, the ratio of the estimated driving time to the historical average driving time, and the ratio of the total energy consumption on the path to the theoretically optimal energy consumption; The path feasibility is determined based on the proportion of the terrain on the path that meets the driving conditions of the lunar rover and the number of special operations required on the path; The driving accuracy is determined based on the deviation distance between the actual arrival position of the lunar rover and the target point, and the deviation degree between the actual driving trajectory of the lunar rover and the planned path.

[0071] Specifically, (1) path safety 1) Obstacle avoidance degree: the minimum distance from the path to the obstacles (such as rocks, meteor craters, etc.) is counted. The smaller the distance, the worse the avoidance effect. If there is a collision with the obstacles on the path, it is directly determined as an unsafe path. For example, set the safe minimum distance threshold d. When the distance between a certain point on the path and the obstacle is less than d, deduct a point for each occurrence. If a collision occurs, the score for this item is 0.

[0072] 2) Slope adaptability: analyze the slope conditions of the area the path passes through, calculate the length proportion of the slope section on the path that exceeds the maximum climbing ability of the lunar rover. The higher the proportion, the lower the safety of the path. If the proportion exceeds a certain proportion, this score is greatly reduced, linearly deducted according to the specific proportion, until 0.

[0073] 3) Communication coverage guarantee: statistics the length proportion of the path in the communication unreachable area. The smaller the proportion, the better. When the proportion exceeds a certain threshold, it is also deducted in proportion until it is deducted.

[0074] (B) Path efficiency 1) Path length: compare the actual planned path length with the theoretically shortest feasible path length between two points, and calculate the ratio of the two. The closer the ratio is to 1, the higher the path efficiency. Set a reasonable maximum ratio limit, and deduct the part that exceeds the limit in proportion.

[0075] 2) Travel time: estimate the time required to travel along the planned path, and compare it with the average travel time of paths generated by historical similar tasks or other planning algorithms. Calculate the time ratio and set an upper limit for the time ratio. Deduct the part that exceeds the limit.

[0076] 3) Energy consumption estimation: consider factors such as terrain and travel speed to estimate the total energy consumption on the path. Compare it with the theoretically optimal energy consumption and calculate the energy consumption ratio. The smaller the ratio, the better. Deduct the part that exceeds the set ratio limit in proportion.

[0077] (Three) Path feasibility 1) Terrain adaptability: based on the performance parameters of the lunar rover, evaluate the proportion of different terrain types (such as plains, hills, sand, etc.) on the planned path, and calculate the proportion of terrain that meets the driving conditions of the lunar rover. The higher the proportion, the higher the feasibility, and the lower the proportion, the lower the score.

[0078] 2) Operation complexity: count the number of special operations (such as turning, climbing slopes, etc.) required for the lunar rover on the path. The fewer the times, the lower the operation complexity, the easier the path to execute, and the score is deducted according to the comparison between the actual number and the preset maximum allowed number.

[0079] (Four) Travel accuracy 1) Target point accuracy: calculate the deviation distance between the actual arrival position of the lunar rover and the target point. The smaller the deviation distance, the higher the accuracy. If the deviation exceeds a certain threshold (such as the set accuracy range), deduct the part in proportion.

[0080] 2) Path tracking accuracy: compare the deviation between the actual travel trajectory of the lunar rover and the planned path, calculate the average deviation distance and the maximum deviation distance. The smaller the average deviation distance and the maximum deviation distance does not exceed the allowed range, the higher the score. Deduct the score according to the deviation.

[0081] The evaluation process is as follows: 1) Data Collection: After the mission is completed, the lunar rover's mission execution data will be collected, including detailed data such as driving trajectory, speed, energy consumption records, communication status logs, information on obstacles encountered, and terrain slope of the passed area.

[0082] 2) Indicator calculation: Based on the above evaluation indicator system, the collected data is used to calculate the score of each evaluation indicator. The score of each sub-indicator is determined by a preset scoring rule (such as a linear deduction model or a step-by-step deduction model).

[0083] 3) Comprehensive score: The scores of each indicator are weighted and summed to obtain the comprehensive score of the path planning result. The weight distribution is set according to the importance of each indicator.

[0084] 4) Visual display: 3D path display: On a 3D lunar terrain model, the planned path (e.g., a solid green line indicates a safe and efficient path segment, and a dashed red line indicates a risky path segment), the actual driving path (a dotted blue line), and key nodes (such as the starting point, end point, and obstacle locations, marked with specific icons) are intuitively presented using different colors and line types, allowing evaluators to clearly compare the planned and actual driving conditions.

[0085] Indicator Radar Chart: Draws a radar chart of each evaluation indicator (path safety, path efficiency, path feasibility, and driving plan accuracy) to intuitively display the relative strengths and weaknesses of each indicator and quickly identify the advantages and disadvantages of path planning.

[0086] Data statistics charts: Use bar charts and line charts to display specific evaluation data, such as the distribution of obstacle avoidance distances on the path, slope change curves, and energy consumption changes over time, providing detailed data support for in-depth analysis.

[0087] Through the path planning result evaluation scheme of the present invention, the advantages and disadvantages of the lunar rover path planning can be comprehensively and objectively evaluated, providing strong support for the continuous improvement of path planning technology and optimization of driving plans, and continuously promoting the efficient and safe execution of lunar rover missions.

[0088] The use of the lunar rover driving simulation training system provided by the present invention is as follows: The experimental team employed a hierarchical permissions management system, comprised of an administrator and several commanders. System permissions were divided into administrator privileges (the main control end) and commander privileges (the collaborative interaction end). Administrators had core control functions, such as scene activation, supporting scene selection through various methods, including eye tracking gaze interaction, gesture interaction, and controller interaction, and managing the display and hiding of layers. Commanders had path editing capabilities, supporting operations such as adding guide points and dragging curvatures using eye tracking combined with controller or gesture interaction.

[0089] The user needs to wear an MR helmet when logging in, and the MR helmet is installed with the lunar rover driving simulation training system provided by the application. Identity authentication is completed through account password or iris recognition, the system automatically matches the role permission and displays it in real time on the MR interface. Before the scene starts, all users need to pass through the double verification of eye tracking and voice recognition, and the administrator also needs to additionally complete the scene selection and start operation.

[0090] After entering the scene, the system will generate the initial driving path of the lunar rover according to the lunar space and lunar terrain data, combined with the planning algorithm, and project it to the actual physical space, while superimposing the information of communication unreachable area, terrain restricted area, etc. The administrator can adjust the layer display through point selection or voice control, and the team members can discuss and research in different perspectives, optimize the driving path by using voice, gesture, handle or artificial path editing (such as guide point adjustment, curvature optimization) and other ways.

[0091] After the lunar rover executes the final path to reach the end point, the system automatically generates a visual evaluation report, including path planning effect, task execution data and other analysis results. Team members can conduct in-depth research based on these data, optimize subsequent path planning strategies and improve task success rate.

[0092] The whole process supports multi-modal interaction methods such as eye tracking, gesture recognition, voice control and handle operation, and realizes accurate mapping of virtual information and physical space, ensuring the efficiency and collaboration of task execution.

[0093] Finally, it should be noted that: the above examples are only used to illustrate the technical solutions of the application, and not to limit them; although the application has been described in detail with reference to the foregoing examples, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing examples, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the application.

Claims

1. A lunar rover driving simulation training system, characterized in that: include: Path point trace intelligent modification module; the path point trace intelligent modification module is used to: Based on the operator's point selection and modification of the initial driving path, the local cost of the selected point is displayed in real time, and the local cost is used to represent the feasibility of the selected point; the local cost is generated based on the terrain slope, obstacle distribution and communication coverage around the selected point; After the operator completes the modification of the initial driving path, the new driving path is checked to see whether it meets the safety and feasibility requirements; if the new driving path does not meet the safety and feasibility requirements, the operator is prompted to adjust the new driving path; if the new driving path meets the safety and feasibility requirements, the new driving path is sent to the lunar rover for execution.

2. The lunar rover driving simulation training system according to claim 1, characterized in that: The system further includes an initial driving path generation module; the initial driving path generation module is configured to: Based on the constructed three-dimensional terrain model, a regional terrain layer is generated, including a slope layer, a lighting layer, a rugged road surface layer, and an adoptable area layer. Based on the regional terrain layer, multiple path planning algorithms are used to comprehensively consider terrain accessibility factors, mission safety factors, and energy efficiency factors to automatically generate an optimal path from a set starting point to a target point. The optimal path is used as the initial driving path of the lunar rover, and the key path points and curvature of the initial driving path are displayed.

3. The lunar rover driving simulation training system according to claim 2, characterized in that: The multiple path planning algorithms include weighted A-star algorithm, hybrid A-star algorithm and algorithm.

4. The lunar rover driving simulation training system according to claim 1, characterized in that: The system further includes a lunar surface topography construction module; the lunar surface topography construction module is configured to: The method involves collecting digital orthophoto DOM data and digital elevation model DEM data of the lunar surface through space probes; constructing a digital model of the lunar terrain based on the DOM data and the DEM data using three-dimensional modeling technology; and rendering the digital model in real time using high-definition pipeline rendering technology to generate a virtual lunar terrain.

5. The lunar rover driving simulation training system according to claim 1, characterized in that: The system further includes: a model adaptive reconstruction module; the model adaptive reconstruction module is used to: Based on the original 3D model of the lunar rover, a white model of the UV map, plane map and wireframe map required for the business scenario is constructed; the white model is rendered by assigning material properties, adjusting the lighting angle, texture mapping and color calibration.

6. The lunar rover driving simulation training system according to claim 5, characterized in that: The model adaptive reconstruction module is also used to: According to the computing power of the mixed reality device, LOD surface reduction processing is performed on the original three-dimensional model.

7. The lunar rover driving simulation training system according to claim 1, characterized in that: The system further includes a path planning result evaluation module; the path planning result evaluation module is used to: After the task is completed, the final planned path is evaluated based on path safety, path efficiency, path feasibility, and driving accuracy, and the path evaluation results are visualized.

8. The lunar rover driving simulation training system according to claim 7, characterized in that: The path safety is determined based on the minimum distance to obstacles on the path, the proportion of slopes on the path that exceed the maximum climbing capability of the lunar rover, and the proportion of communication-inaccessible areas on the path. The path efficiency is determined based on the ratio of the actual planned path length to the theoretical shortest feasible path length, the ratio of the estimated travel time to the historical average travel time, and the ratio of the total energy consumption on the path to the theoretical optimal energy consumption; The feasibility of the path is determined based on the percentage of terrain on the path that meets the driving conditions of the lunar rover and the number of special operations required on the path; The driving accuracy is determined based on the deviation distance between the actual arrival position of the lunar rover and the target point, and the degree of deviation between the actual driving trajectory of the lunar rover and the planned path.

9. The lunar rover driving simulation training system according to claim 1, characterized in that: The system further includes a path planning integrated display module; the path planning integrated display module is used to: The motion trajectory of the lunar rover along the selected route is superimposed on the holographic image, and key points in the route are highlighted.

10. The lunar rover driving simulation training system according to claim 1, characterized in that: The system further includes a path planning result storage module; the path planning result storage module is used to: The simulation training results and the final planned path are stored in a localized database in XML format simultaneously.

Images