An unmanned vehicle homing method and system in a complex reconnaissance environment

Abstract

The present application relates to a kind of complex reconnaissance environment under unmanned vehicle homeward journey method and system, belong to unmanned vehicle navigation technical field, method includes real-time monitoring the operating state of unmanned vehicle and environmental information, whether there is homeward journey condition according to monitoring result judges;When there is homeward journey condition, current homeward journey mode is determined based on the existing homeward journey condition;Based on current homeward journey mode, the homeward journey path of unmanned vehicle from current position to homeward journey terminal is planned;Control unmanned vehicle travels along the homeward journey path, and continuously monitor whether homeward journey condition changes in the process of travel, if homeward journey condition changes, then current homeward journey mode is determined again based on the changed homeward journey condition, according to the current homeward journey mode adjusted after being determined again, until unmanned vehicle reaches the homeward journey terminal.The present application solves the problem that existing unmanned vehicle cannot dynamically adjust path according to homeward journey inducement, and cannot be self-adaptively adjusted after path planning.

Description

Technical Field

[0001] This invention relates to the field of unmanned vehicle navigation technology, and in particular to a method and system for unmanned vehicles to return to their home location in complex reconnaissance environments. Background Technology

[0002] With the continuous development of unmanned vehicle technology, its application in fields such as reconnaissance, surveillance, logistics and transportation, and rescue is becoming increasingly widespread. In complex reconnaissance environments, unmanned vehicles need to traverse areas where GPS signals are easily blocked or interfered with, such as dense forests, canyons, urban areas, and ruins. At the same time, communication links may be interrupted due to excessive distance or obstacles. Dynamic obstacles, steep slopes, and ravines in the environment also pose a continuous threat to driving safety. These complex factors place extremely high demands on the autonomous return capability of unmanned vehicles.

[0003] Existing autonomous vehicle return-to-home technologies typically use GPS for positioning and rely on pre-built static maps to plan fixed return paths. For example, after receiving a return-to-home command, the autonomous vehicle uses its current GPS coordinates as the starting point and the mission's starting point as the ending point, searches for the shortest path on the global map, and then travels along that path until it returns. This approach can achieve basic return-to-home functionality under conditions of good signal and a static environment. However, regardless of the reason for triggering the return, the same return-to-home strategy is used, prioritizing the shortest path, and it cannot dynamically adjust the optimization objective of the path planning based on the return-to-home trigger. Furthermore, GPS signals are easily blocked or interfered with in environments such as dense forests and urban areas, leading to positioning drift or even failure, making it impossible to accurately determine the starting and ending points of the return path. Once the return path is generated, it is not adjusted. When sudden obstacles, changes in terrain accessibility, or new dynamic threats appear along the way, the autonomous vehicle cannot adaptively adjust the path. During the return process, there is also a lack of response mechanisms to changes in conditions, making it impossible to sense changes such as communication recovery or the emergence of new dangers and dynamically adjust the strategy. Summary of the Invention

[0004] Based on the above analysis, the embodiments of the present invention aim to provide a method and system for unmanned vehicles to return to their home in complex reconnaissance environments, in order to solve the problems of existing methods that cannot dynamically adjust the path according to the return-to-home cause and cannot adaptively adjust the path after path planning.

[0005] On one hand, embodiments of the present invention provide a method for unmanned vehicles to return to their home position in complex reconnaissance environments, including: Real-time monitoring of the unmanned vehicle's operating status and environmental information; and determination of whether return-to-home conditions exist based on the monitoring results. When return-to-home conditions exist, the current return-to-home mode is determined based on those conditions. Based on the current return mode, plan the return path of the unmanned vehicle from its current location to the return destination; The driverless vehicle is controlled to travel along the return path, and the return conditions are continuously monitored during the journey. If the return conditions change, the current return mode is redefined based on the changed return conditions, and the return path is adjusted according to the redefined current return mode until the driverless vehicle reaches the return destination.

[0006] Furthermore, the return conditions include mission completion conditions, communication interruption conditions, insufficient energy conditions, hazardous environmental conditions, and return command conditions; Determining whether conditions for a return trip exist includes: Obtain task completion progress information from the running status, and determine whether the task completion conditions are met based on the task completion progress information; Obtain communication link status information from the operating state, and determine whether the communication interruption condition is met based on the communication link status information; The remaining energy level information is obtained from the operating status, and the energy consumption required for the return trip is estimated based on the current location. The energy shortage condition is determined based on the comparison between the remaining energy level information and the estimated energy consumption. Risk factors are extracted from the environmental information, and it is determined whether environmental hazard conditions are met based on the risk factors. The system determines whether the conditions for a return-to-home command are met based on the external instructions received by the autonomous vehicle. If any one of the following conditions is met: mission completion, communication interruption, insufficient energy, hazardous environment, or return-to-base instruction, then the return-to-base condition is determined to exist.

[0007] Furthermore, determining the return mode based on existing return conditions includes: When a return-to-home condition exists, a matching return-to-home mode is selected from the preset mode set according to the type of the return-to-home condition; When multiple return-to-home conditions exist, the highest-level return-to-home condition is selected as the dominant return-to-home condition according to a preset priority rule, and a matching return-to-home mode is selected from the preset mode set according to the type of the dominant return-to-home condition. The preset priority rules, from highest to lowest, are: return command condition, insufficient energy condition, communication interruption condition, hazardous environment condition, and mission completion condition. The preset mode set includes at least the instant return mode, low energy consumption priority mode, autonomous navigation priority mode, safety priority mode, and regular return mode.

[0008] Furthermore, based on the current return-to-home mode, the planned return path for the autonomous vehicle from its current location to the return destination includes: Obtain the current location and return destination of the autonomous vehicle; A comprehensive cost function is constructed based on the current location of the unmanned vehicle, the return destination location, and the real-time monitored operating status and environmental information. Based on the current return mode, the weights of each cost term in the comprehensive cost function are adjusted; A path search algorithm is used to minimize the comprehensive cost function after weight adjustment, and the adjusted return path is obtained based on the node sequence with the minimum cumulative cost.

[0009] Furthermore, based on the current return mode, the weights of each cost term in the comprehensive cost function are adjusted, including: Based on the current return mode, determine the adjustment direction and magnitude of the weights for path length cost, environmental risk cost, energy consumption cost, and positioning reliability cost, and generate a weight adjustment vector. Based on the weight adjustment vector, the adjusted weight coefficients of each cost term in the comprehensive cost function are obtained; Normalize all adjusted weight coefficients so that the sum of all weight coefficients equals one, thus obtaining the adjusted comprehensive cost function.

[0010] Furthermore, the adjustment directions and magnitudes of the weights for path length cost, environmental risk cost, energy consumption cost, and location reliability cost are determined, including: When the current return mode is the low energy consumption priority mode, the weight of energy consumption cost is increased by a first preset multiple, while the weight of path length cost is decreased by a second preset multiple, and the weights of environmental risk cost and positioning reliability cost remain unchanged. When the current return mode is the safety priority mode, the weight of environmental risk cost is increased by a third preset multiple, while the weight of path length cost is decreased by a fourth preset multiple, and the weights of energy consumption cost and positioning reliability cost remain unchanged. When the current return mode is the autonomous navigation priority mode, the weight of the positioning reliability cost is increased by a fifth preset multiple, while the weight of the path length cost is decreased by a sixth preset multiple, and the weights of the environmental risk cost and the energy consumption cost remain unchanged. When the current return-to-home mode is instant return-to-home mode or regular return-to-home mode, the weights of path length cost, environmental risk cost, energy consumption cost, and positioning reliability cost remain unchanged.

[0011] Furthermore, the comprehensive cost function is shown in the following equation; ; in, For path length cost, As a cost to environmental risks, As a result of energy consumption, To determine the cost of credibility, The weights are the path length cost. As a weight for environmental risk costs, As a weight for the cost of energy consumption, The weights for determining the credibility cost.

[0012] Furthermore, the path search algorithm is used to minimize the weighted overall cost function, including: The search starts at the current location of the autonomous vehicle and ends at the return point. Starting from the search starting point, obtain all adjacent nodes of the current node; For each neighboring node of the current node, according to the comprehensive cost function, the node with the smallest comprehensive cost function value is selected as the next current node, and its neighboring nodes are obtained until the search target point is reached; from the search target point back to the search starting point, the node sequence with the smallest cumulative cost is obtained, and the return path is obtained based on the node sequence with the smallest cumulative cost.

[0013] Further, extracting risk factors from the environmental information and determining whether environmental hazard conditions are met based on the risk factors includes: Obtain obstacle distribution data within the target area, calculate obstacle density values ​​based on the obstacle distribution data, and use the obstacle density values ​​as the first risk factor; Acquire terrain data within the target area, calculate slope and surface roughness values ​​based on the terrain data, and calculate a second risk factor based on the slope and surface roughness values; The number of dynamic targets within the target area is obtained, and the number of dynamic targets is used as a third risk factor. The first, second, and third risk factors are compared with their corresponding risk thresholds. If the number of risk factors exceeding the corresponding risk threshold is greater than or equal to 2, the environmental hazard condition is deemed to be met; otherwise, the environmental hazard condition is deemed not to be met.

[0014] On the other hand, embodiments of the present invention provide an unmanned vehicle return-to-home system for complex reconnaissance environments, the system comprising: The monitoring module is used to monitor the operating status and environmental information of the unmanned vehicle in real time, and to determine whether there are conditions for returning to base based on the monitoring results; The determination module is used to determine the current return mode based on the existing return conditions when they exist. The planning module is used to plan the return path of the unmanned vehicle from its current location to the return destination based on the current return mode; The adjustment module is used to control the unmanned vehicle to travel along the return path and continuously monitor whether the return conditions change during the journey. If the return conditions change, the current return mode is re-determined based on the changed return conditions, and the return path is adjusted according to the re-determined current return mode until the unmanned vehicle reaches the return destination. The control module is used to determine that the return journey is complete when the unmanned vehicle reaches the return destination and to control the unmanned vehicle to enter the standby state.

[0015] Compared with the prior art, the present invention can achieve at least one of the following beneficial effects: 1. This invention comprehensively judges whether any of the following conditions exist: mission completion condition, communication interruption condition, insufficient energy condition, environmental hazard condition, and return-to-home instruction condition, by real-time monitoring of operating status and environmental information. When any condition is detected, the return-to-home process is automatically initiated. When a return-to-home condition exists, a matching return-to-home mode is selected from a preset mode set according to the type of the return-to-home condition. When multiple return-to-home conditions exist, the dominant return-to-home condition is selected according to the priority rules of return-to-home instruction condition, insufficient energy condition, communication interruption condition, environmental hazard condition, and mission completion condition, and the corresponding return-to-home mode is matched. This allows the low-energy priority mode, safety priority mode, autonomous navigation priority mode, instant return-to-home mode, and regular return-to-home mode to adaptively match according to the specific reason for triggering the return-to-home process. This overcomes the defect in the prior art where the same strategy is used for all return-to-home scenarios, making it difficult to balance return-to-home efficiency and success rate. It also solves the problem that the existing technology cannot dynamically adjust the path according to the return-to-home trigger.

[0016] 2. This invention constructs a comprehensive cost function based on the current location of the unmanned vehicle, the return destination location, and real-time monitored operating status and environmental information. This function includes path length cost, environmental risk cost, energy consumption cost, and positioning reliability cost. After adjusting the weights of each cost term in the comprehensive cost function according to the current return mode, a path search algorithm is used to minimize the search to obtain the return path. Simultaneously, during the return process, changes in return conditions are continuously monitored, and the return mode is redefined and the return path is adjusted when changes occur. This allows the return path to dynamically adjust and optimize its objectives according to the return mode and respond to new situations that arise during the return journey. This ensures the optimality of the return path under the corresponding return mode, significantly improves the adaptability and robustness of the return process, and solves the problem of the inability to adaptively adjust after path planning.

[0017] In this invention, the above-described technical solutions can be combined with each other to achieve more preferred combinations. Other features and advantages of this invention will be set forth in the following description, and some advantages may become apparent from the description or be learned by practicing the invention. The objects and other advantages of this invention can be realized and obtained from the description and drawings, which are particularly pointed out. Attached Figure Description

[0018] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Throughout the drawings, the same reference numerals denote the same parts. Figure 1 is a schematic flowchart of a method for unmanned vehicles to return to base in a complex reconnaissance environment according to an embodiment of the present invention. Detailed Implementation

[0019] Preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings, which form part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not intended to limit the scope of the present invention.

[0020] Example 1 A specific embodiment of the present invention discloses a method for unmanned vehicles to return to their home location in complex reconnaissance environments, such as... Figure 1 As shown, it includes: S1. Monitor the operating status and environmental information of the unmanned vehicle in real time, and determine whether there are conditions for returning to base based on the monitoring results; The operational status includes mission progress, communication link status, remaining power level, current location coordinates, and positioning system signal quality. Mission progress can include area coverage completion rate, target reconnaissance completion rate, planned path execution completion rate, and effective data acquisition completion rate, specifically determined according to mission requirements. Communication link status includes the signal strength of radio communication links, the signal strength of communication network links, signal-to-noise ratio (SNR), link delay, and packet loss rate. Remaining power level is obtained by reading the remaining battery charge from the battery management system. Current location coordinates are obtained through combined navigation using a GPS receiver and inertial measurement unit. Positioning system (such as GPS, BeiDou, etc.) signal quality includes signal-to-noise ratio and the number of visible satellites.

[0021] Environmental information includes obstacle distribution data, terrain data, and the number of dynamic targets. Obstacle distribution data is derived from the fusion perception results of LiDAR point clouds and visual images; after obstacle detection, the position coordinates of all obstacles within the target area are output. Terrain data comes from pre-stored high-precision maps or a 3D terrain mesh generated in real-time through LiDAR mapping. The number of dynamic targets can be obtained through inter-frame difference analysis of LiDAR point clouds or visual target tracking algorithms. Environmental information is stored in a grid map. The entire environmental space for the autonomous vehicle's task is divided into several square grid cells of fixed size. Each grid cell is called a grid, and each grid is used to store environmental information for that location. The overall operational environment boundary of the autonomous vehicle is obtained, defined by the coordinates of its four vertices. The overall operational environment is divided into uniformly sized grids, each corresponding to an actual 0.5m × 0.5m area. The number of rows and columns of the grid is determined by dividing the side length of the operational environment by the grid size; for example, a 500m × 400m environment corresponds to 1000 × 800 grids. During the mission, the unmanned vehicle continuously fills the corresponding grid with the environmental information collected in real time.

[0022] Return-to-base conditions include mission completion conditions, communication interruption conditions, insufficient energy conditions, hazardous environmental conditions, and return-to-base instruction conditions. Determining whether return-to-base conditions exist includes: S11. Obtain task completion progress information from the running state, and determine whether the task completion conditions are met based on the task completion progress information; When a mission is initiated, the autonomous vehicle reads the current mission type identifier from the mission configuration file. Mission types include reconnaissance missions, area coverage missions, and path planning missions. Different mission types have different evaluation criteria for their progress information.

[0023] When the task type is reconnaissance, the task completion condition is met when the number of identified and recorded targets matches the total number of targets specified in the task. Identified targets refer to those successfully detected by the autonomous vehicle using a target detection algorithm, which can employ deep learning-based models such as YOLOv8.

[0024] When the task type is a data acquisition task, the task completion condition is met when the amount of valid data collected reaches the total amount of data required by the task. The autonomous vehicle compares the amount of valid data collected with the total amount of data required by the task. The valid data collected refers to data that meets the quality requirements, including an image sharpness score greater than 0.6 and a point cloud effective echo rate greater than 0.5.

[0025] When the task type is path planning, the task completion condition is met when the completed path mileage reaches the total planned path mileage. The autonomous vehicle compares the completed path mileage with the total planned path mileage, and the completed path mileage is accumulated and calculated by the autonomous vehicle's odometer.

[0026] S12. Obtain communication link status information from the operating state, and determine whether the communication interruption condition is met based on the communication link status information; Communication link status information includes signal strength, signal-to-noise ratio (SNR), and packet loss rate. Corresponding thresholds are set for each of these parameters: signal strength threshold -90dBm (signal strength below -90dBm indicates a weak signal); SNR threshold 10dB (SNR below 10dB indicates poor signal quality); and packet loss rate threshold 30% (packet loss rate above 30% indicates severe data loss). A communication interruption condition is determined when any one of the following conditions is met: signal strength below the signal strength threshold, SNR below the SNR threshold, or packet loss rate above the packet loss rate threshold.

[0027] S13. Obtain the remaining energy level information from the operating status, estimate the energy consumption required for return based on the current position, and determine whether the energy shortage condition is met based on the comparison result between the remaining energy level information and the estimated energy consumption. The autonomous vehicle reads the current remaining battery power. Simultaneously, it estimates the energy required for the return journey based on its current location and the return destination. The return path length is the estimated distance traveled from the vehicle's current location to the return destination, calculated using a path planning algorithm, such as A. The algorithm searches for the shortest path between the current location and the return destination on a grid map and calculates its total length. The energy consumption required for the return trip is obtained by multiplying the return path length by the energy consumption coefficient per unit distance. The energy consumption coefficient per unit distance is obtained through statistical analysis of historical driving data, such as 0.1 watt-hours per meter. If the current remaining battery power is lower than the sum of the energy consumption required for the return trip and the return trip safety redundancy energy, then the energy shortage condition is met. The return trip safety redundancy energy can be set to 15% to 30% of the energy consumption required for the return trip.

[0028] S14. Extract risk factors from the environmental information, and determine whether environmental hazard conditions are met based on the risk factors; Furthermore, determining whether environmental hazard conditions are met based on the aforementioned risk factors includes: S141. Obtain obstacle distribution data within the target area, calculate obstacle density value based on the obstacle distribution data, and use the obstacle density value as the first risk factor; The target area is a circular region with a radius of 50 to 100 meters centered on the current position of the autonomous vehicle. The autonomous vehicle scans its surroundings using a lidar system, receiving reflected echoes to form 3D point cloud data. Simultaneously, it acquires environmental images using a vision camera. A deep learning-based target detection algorithm is used to fuse the point cloud and images, identifying obstacles and outputting the coordinates of each obstacle. The autonomous vehicle counts the total number of obstacles identified within the target area and calculates the area, which is obtained by multiplying pi by the square of the radius. The obstacle density value is obtained by dividing the total number of obstacles by the target area area. The obstacle density value is used as the primary risk factor.

[0029] S142. Obtain terrain data within the target area, calculate slope and surface roughness values ​​based on the terrain data, and calculate a second risk factor based on the slope and surface roughness values. The terrain data comes from pre-stored high-precision maps or 3D terrain meshes generated in real time through LiDAR mapping. The autonomous vehicle calculates slope and surface roughness values ​​based on the terrain data. The slope value is obtained by calculating the angle between the normal vector of each surface element in the terrain mesh and the horizontal plane, while the surface roughness value is obtained by calculating the root mean square of the elevation differences between adjacent nodes in the terrain mesh. The second risk factor is calculated as a weighted sum of the slope and surface roughness values. Both the slope and surface roughness values ​​are normalized, multiplied by their respective weighting coefficients, and then summed. The weighting coefficients are 0.5 for each value.

[0030] S143. Obtain the number of dynamic targets within the target area, and use the number of dynamic targets as a third risk factor; Dynamic target detection is achieved through inter-frame difference analysis of LiDAR point clouds and visual target tracking algorithms. For example, an autonomous vehicle continuously collects two or more frames of LiDAR point cloud data. After spatial matching of point clouds in adjacent frames, it compares whether point clouds exist at the same location at different times. If the location of point clouds in a certain area changes significantly between consecutive frames, and the change exceeds a preset motion judgment threshold (e.g., 0.5 meters), then the area is determined to correspond to a dynamic target. The autonomous vehicle counts the total number of dynamic targets within the target area at the current time and uses the total number of dynamic targets as a third risk factor.

[0031] S144. Compare the first risk factor, the second risk factor, and the third risk factor with their respective risk thresholds. If the number of risk factors exceeding the corresponding risk threshold is greater than or equal to 2, then the environmental hazard condition is satisfied; otherwise, the environmental hazard condition is not satisfied.

[0032] The risk threshold for the first risk factor can be set to 1 per square meter, the second risk threshold to 0.5, and the third risk threshold to 3. Each risk factor is compared to its corresponding risk threshold, and the number of risk factors exceeding the threshold is counted. If the number of risk factors exceeding the corresponding threshold is greater than or equal to 2, the environmental hazard condition is deemed met. Otherwise, the environmental hazard condition is deemed not met.

[0033] S15. Determine whether the conditions for the return command are met based on the external instructions received by the unmanned vehicle; The autonomous vehicle monitors the communication channel between itself and the control unit to detect whether external commands have been received from the control unit. External commands are encapsulated in data frame format, containing a command type field, a timestamp, and a digital signature. The command type field identifies the type of command; for example, a return-to-home command corresponds to a hexadecimal value of 0x01, a stop command to 0x02, and a rerouting command to 0x03, etc. The command type field in the command data frame is parsed. If the value of the command type field is not equal to 0x01 (i.e., it is not a return-to-home command), the command is ignored, and the system continues to monitor subsequent commands. If the value of the command type field is equal to 0x01, the command is determined to be a return-to-home command, satisfying the return-to-home command condition.

[0034] If any one of the following conditions is met: mission completion, communication interruption, insufficient energy, hazardous environment, or return-to-base instruction, then the return-to-base condition is determined to exist.

[0035] S2. When return-to-home conditions exist, determine the current return-to-home mode based on the existing return-to-home conditions; When a return-to-home condition exists, a matching return-to-home mode is selected from the preset mode set according to the type of the return-to-home condition; When multiple return-to-home conditions exist, the highest priority condition is selected as the dominant return-to-home condition according to a preset priority rule. Then, a matching return-to-home mode is selected from the preset mode set according to the type of the dominant return-to-home condition. The preset mode set includes at least the instant return-to-home mode, the low-energy priority mode, the autonomous navigation priority mode, the safety priority mode, and the regular return-to-home mode.

[0036] Each return-to-home condition corresponds to a specific return-to-home mode: mission completion corresponds to the regular return-to-home mode, communication interruption corresponds to the autonomous navigation priority mode, insufficient energy corresponds to the low-energy consumption priority mode, hazardous environment corresponds to the safety priority mode, and valid external command corresponds to the immediate return-to-home mode. When multiple return-to-home conditions exist simultaneously, the autonomous vehicle selects the highest-level condition as the dominant return-to-home condition according to a preset priority rule, and then selects a matching return-to-home mode from a preset mode set based on the type of the dominant return-to-home condition.

[0037] Instant return-to-home mode refers to the mode in which the autonomous vehicle, upon receiving a valid external command, immediately interrupts all currently executing tasks and returns to its destination via the shortest path. Low-energy-consumption-priority mode refers to the mode where, under conditions of insufficient energy, path planning prioritizes minimizing energy consumption. Autonomous navigation-priority mode refers to the mode where, under conditions of communication interruption, path planning prioritizes maximizing positioning reliability and autonomous navigation capabilities. Safety-priority mode refers to the mode where, under hazardous environmental conditions, path planning prioritizes maximizing driving safety. Normal return-to-home mode refers to the default return-to-home mode where, under conditions of normal task completion, path planning maintains a balance among various optimization objectives.

[0038] S3. Based on the current return-to-home mode, plan the return path of the autonomous vehicle from its current location to the return-to-home destination; further, the planning of the return path of the autonomous vehicle from its current location to the return-to-home destination based on the current return-to-home mode includes: S31. Obtain the current location and return destination of the unmanned vehicle; construct a comprehensive cost function based on the current location, return destination, and real-time monitored operating status and environmental information of the unmanned vehicle; The comprehensive cost function is shown in the following equation; ; in, For path length cost, As a cost to environmental risks, As a result of energy consumption, To determine the cost of credibility, The weights are the path length cost. As a weight for environmental risk costs, As a weight for the cost of energy consumption, Weights are assigned to determine the credibility cost. The default value for the weight coefficients is... =0.35、 =0.25、 =0.25、 =0.15.

[0039] Path length cost It is used to indicate the distance of a grid from a target point. It is obtained by calculating the Euclidean distance between the grid and the return destination. The calculated distance value is used as the path length cost of the grid.

[0040] Environmental risk costs This indicates the environmental hazard level of the location of the grid. First, the environmental risk level of the grid is determined, including levels 0, 1, and 2. Level 0 corresponds to a cost value of 0, level 1 corresponds to a cost value of 0.5, and level 2 corresponds to a cost value of 1. When the number of risk factors exceeding the corresponding threshold is greater than or equal to 2, it corresponds to level 2; when the number of risk factors exceeding the corresponding threshold is 1, it corresponds to level 1; and when the number of risk factors exceeding the corresponding threshold is 0, it corresponds to level 0.

[0041] The energy consumption cost represents the amount of energy required for an autonomous vehicle to pass through the grid, as shown in the following formula; ; in, The energy consumption coefficient per unit distance for the pre-calibrated autonomous vehicle represents the energy consumption per unit distance traveled on flat terrain, and its value ranges from 0.05 to 0.15. The second risk factor reflects the ruggedness of the grid terrain. Its value ranges from 0 to 1. The larger the value, the more complex the terrain and the more energy-consuming the passage. The second risk factor is calculated by weighting the slope value and the surface roughness value.

[0042] Location credibility cost This indicates the reliability of the unmanned vehicle's positioning within the grid. It is calculated using the signal quality of the Global Positioning System (GPS). Specifically, it is calculated by subtracting the normalized value of the inertial navigation drift rate from the weighted sum of the current normalized signal-to-noise ratio and the normalized value of the number of visible satellites. The lower the value, the less reliable the positioning in that area.

[0043] Based on the current return mode, the weights of each cost term in the comprehensive cost function are adjusted; The weight adjustment strategy is determined based on the type of the current return mode, with one weight adjustment strategy corresponding to each return mode. Adjusting the weights of each cost term in the comprehensive cost function includes: S321. Based on the current return mode, determine the adjustment direction and magnitude of the weights for path length cost, environmental risk cost, energy consumption cost, and positioning reliability cost, and generate a weight adjustment vector. When the current return mode is low-energy priority mode, the weight of energy consumption cost will be adjusted. Increased to 2.0 times the original size, that is Change the weight from 0.25 to 0.50; simultaneously adjust the weight of the path length cost. Reduced to 0.3 times the original value, that is The weight of environmental risk costs changed from 0.35 to 0.105. =0.25 and the weight of the location credibility cost =0.15 remains unchanged.

[0044] When the current return mode is safety-first, the weight of environmental risk costs will be considered. Increased to 2.2 times the original size, that is Change from 0.25 to 0.55; simultaneously adjust the weight of the path length cost. Reduced to 0.15 times the original value, that is The weighting of energy consumption costs changed from 0.35 to 0.0525. =0.25 and the weight of the location credibility cost =0.15 remains unchanged.

[0045] When the current return-to-home mode is autonomous navigation priority mode, the weight λ4 of the positioning reliability cost is increased to 1.8 times the original value, that is... The value changes from 0.15 to 0.27; simultaneously, the weight λ1 of the path length cost is reduced to 0.65 times its original value. The weight of environmental risk costs changed from 0.35 to 0.2275. =0.25 and the weight of energy consumption cost =0.25 remains unchanged.

[0046] When the current return-to-home mode is either instant return-to-home mode or regular return-to-home mode, all four weighting coefficients remain unchanged at their default values.

[0047] S322. Based on the weight adjustment vector, obtain the adjusted weight coefficients of each cost term in the comprehensive cost function; normalize all the adjusted weight coefficients so that the sum of the weight coefficients equals one, and obtain the adjusted comprehensive cost function.

[0048] The four adjusted weight coefficients are normalized so that their sum equals 1. Normalization is achieved by dividing each adjusted weight coefficient by the sum of the four adjusted weight coefficients. The normalized weight coefficients are then substituted into the comprehensive cost calculation formula, replacing the original default weight coefficients, to obtain the adjusted comprehensive cost value. The calculated comprehensive cost value is then filled into each grid cell to obtain the global cost map.

[0049] S33. A path search algorithm is used to minimize the weighted comprehensive cost function. Based on the node sequence with the minimum cumulative cost, the adjusted return path is obtained, including: S331. Taking the current position of the unmanned vehicle as the search starting point and the return destination position as the search target point, obtain all adjacent nodes of the current node starting from the search starting point; Starting from the starting point, the system explores the surrounding grids layer by layer, selecting the grid with the lowest total cost adjacent to the current grid at each step, until the target point is reached. In the global cost map, each grid already stores its total cost value.

[0050] First, taking the search starting point as the current node, we obtain all its neighboring nodes. A neighboring node is defined as the eight adjacent grid cells surrounding the current node, i.e., the grid cells adjacent in the eight directions: up, down, left, right, upper left, upper right, lower left, and lower right. Excluding the search starting point, the neighboring nodes of other current nodes consist of seven neighboring grid cells, excluding the grid cells of the previous current node.

[0051] S332. For each neighboring node of the current node, according to the comprehensive cost function, select the node with the smallest comprehensive cost function value as the next current node, and continue to obtain its neighboring nodes until the search target point is reached; backtrack from the search target point to the search starting point to obtain the node sequence with the smallest cumulative cost, and obtain the return path based on the node sequence with the smallest cumulative cost.

[0052] From all the neighboring nodes of the current node, select the node with the smallest comprehensive cost function as the next current node. Execute S332 on each current node in turn until the target point is reached. Trace back each historical current node in turn and arrange all historical current nodes in the search order to obtain the return path.

[0053] S4. Control the unmanned vehicle to travel along the return path, and continuously monitor whether the return conditions change during the travel. If the return conditions change, redetermine the current return mode based on the changed return conditions, and adjust the return path according to the redetermined current return mode until the unmanned vehicle reaches the return destination. During the autonomous vehicle's operation, S1 is continuously executed, which monitors whether the return-to-home conditions have changed. A change in return-to-home conditions means the emergence of new conditions or the disappearance of a previously monitored condition. For example, if the autonomous vehicle triggers a return-to-home due to a communication interruption, but the communication link is automatically restored during the return journey, the communication interruption condition changes from true to false. Then, S2-S3 are executed again, which involves re-determining the current return-to-home mode based on the changed conditions and replanning the return path based on the re-determined mode. When the autonomous vehicle reaches the return-to-home endpoint, the return is considered complete, and the vehicle is put into standby mode.

[0054] Example 2 A system for returning to home for unmanned vehicles in complex reconnaissance environments, the system comprising: The monitoring module is used to monitor the operating status and environmental information of the unmanned vehicle in real time, and to determine whether there are conditions for returning to base based on the monitoring results; The determination module is used to determine the current return mode based on the existing return conditions when they exist. The planning module is used to plan the return path of the unmanned vehicle from its current location to the return destination based on the current return mode; The adjustment module is used to control the unmanned vehicle to travel along the return path and continuously monitor whether the return conditions change during the journey. If the return conditions change, the current return mode is re-determined based on the changed return conditions, and the return path is adjusted according to the re-determined current return mode until the unmanned vehicle reaches the return destination. The control module is used to determine that the return journey is complete when the unmanned vehicle reaches the return destination and to control the unmanned vehicle to enter the standby state.

[0055] It is understandable that the modules recorded in the unmanned vehicle return-to-home system under this complex reconnaissance environment are similar to those in the reference system. Figure 1 The steps described in the unmanned vehicle return-to-home method for complex reconnaissance environments correspond to those steps in this method. Therefore, the operations, characteristics, and beneficial effects described above for the unmanned vehicle return-to-home method in complex reconnaissance environments also apply to the unmanned vehicle return-to-home system and its constituent modules in complex reconnaissance environments, and will not be repeated here.

[0056] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware, and the program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.

[0057] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for unmanned vehicles to return to base in complex reconnaissance environments, characterized in that, include: Real-time monitoring of the unmanned vehicle's operating status and environmental information; and determination of whether return-to-home conditions exist based on the monitoring results. When return-to-home conditions exist, the current return-to-home mode is determined based on those conditions. Based on the current return mode, plan the return path of the unmanned vehicle from its current location to the return destination; The driverless vehicle is controlled to travel along the return path, and the return conditions are continuously monitored during the journey. If the return conditions change, the current return mode is redefined based on the changed return conditions, and the return path is adjusted according to the redefined current return mode until the driverless vehicle reaches the return destination.

2. The method according to claim 1, characterized in that, The return conditions include mission completion conditions, communication interruption conditions, insufficient energy conditions, hazardous environmental conditions, and return command conditions. Determining whether conditions for a return trip exist includes: Obtain task completion progress information from the running status, and determine whether the task completion conditions are met based on the task completion progress information; Obtain communication link status information from the operating state, and determine whether the communication interruption condition is met based on the communication link status information; The remaining energy level information is obtained from the operating status, and the energy consumption required for the return trip is estimated based on the current location. The energy shortage condition is determined based on the comparison between the remaining energy level information and the estimated energy consumption. Risk factors are extracted from the environmental information, and it is determined whether environmental hazard conditions are met based on the risk factors. The system determines whether the conditions for a return-to-home command are met based on the external instructions received by the autonomous vehicle. If any one of the following conditions is met: mission completion, communication interruption, insufficient energy, hazardous environment, or return-to-base instruction, then the return-to-base condition is determined to exist.

3. The method according to claim 2, characterized in that, Determining the return mode based on existing return conditions includes: When a return-to-home condition exists, a matching return-to-home mode is selected from the preset mode set according to the type of the return-to-home condition; When multiple return-to-home conditions exist, the highest-level return-to-home condition is selected as the dominant return-to-home condition according to a preset priority rule, and a matching return-to-home mode is selected from the preset mode set according to the type of the dominant return-to-home condition. The preset priority rules, from highest to lowest, are: return command condition, insufficient energy condition, communication interruption condition, hazardous environment condition, and mission completion condition. The preset mode set includes at least the instant return mode, low energy consumption priority mode, autonomous navigation priority mode, safety priority mode, and regular return mode.

4. The method according to claim 3, characterized in that, Based on the current return-to-home mode, the planned return path for the autonomous vehicle from its current location to the return destination includes: Obtain the current location and return destination of the autonomous vehicle; A comprehensive cost function is constructed based on the current location of the unmanned vehicle, the return destination location, and the real-time monitored operating status and environmental information. Based on the current return mode, the weights of each cost term in the comprehensive cost function are adjusted; A path search algorithm is used to minimize the comprehensive cost function after weight adjustment, and the adjusted return path is obtained based on the node sequence with the minimum cumulative cost.

5. The method according to claim 4, characterized in that, Based on the current return mode, the weights of each cost term in the comprehensive cost function are adjusted, including: Based on the current return mode, determine the adjustment direction and magnitude of the weights for path length cost, environmental risk cost, energy consumption cost, and positioning reliability cost, and generate a weight adjustment vector. Based on the weight adjustment vector, the adjusted weight coefficients of each cost term in the comprehensive cost function are obtained; Normalize all adjusted weight coefficients so that the sum of all weight coefficients equals one, thus obtaining the adjusted comprehensive cost function.

6. The method according to claim 5, characterized in that, The weighting adjustment directions and adjustment magnitudes for path length cost, environmental risk cost, energy consumption cost, and location reliability cost are determined separately, including: When the current return mode is the low energy consumption priority mode, the weight of energy consumption cost is increased by a first preset multiple, while the weight of path length cost is decreased by a second preset multiple, and the weights of environmental risk cost and positioning reliability cost remain unchanged. When the current return mode is the safety priority mode, the weight of environmental risk cost is increased by a third preset multiple, while the weight of path length cost is decreased by a fourth preset multiple, and the weights of energy consumption cost and positioning reliability cost remain unchanged. When the current return mode is the autonomous navigation priority mode, the weight of the positioning reliability cost is increased by a fifth preset multiple, while the weight of the path length cost is decreased by a sixth preset multiple, and the weights of the environmental risk cost and the energy consumption cost remain unchanged. When the current return-to-home mode is instant return-to-home mode or regular return-to-home mode, the weights of path length cost, environmental risk cost, energy consumption cost, and positioning reliability cost remain unchanged.

7. The method according to claim 6, characterized in that, The comprehensive cost function is shown in the following equation; ； in, For path length cost, As a cost to environmental risks, As a result of energy consumption, To determine the cost of credibility, The weights are the path length cost. As a weight for environmental risk costs, As a weight for the cost of energy consumption, The weights for determining the credibility cost.

8. The method according to claim 7, characterized in that, The path search algorithm is used to minimize the weighted overall cost function, including: The search starts at the current location of the autonomous vehicle and ends at the return point. Starting from the search starting point, obtain all adjacent nodes of the current node; For each neighboring node of the current node, according to the comprehensive cost function, select the node with the smallest comprehensive cost function value as the next current node, and repeatedly obtain the neighboring nodes of the next current node until the search target point is reached; By tracing back from the search target point to the search starting point, the node sequence with the minimum cumulative cost is obtained, and the return path is obtained based on the node sequence with the minimum cumulative cost.

9. The method according to claim 2, characterized in that, Extracting risk factors from the environmental information and determining whether environmental hazard conditions are met based on the risk factors includes: Obtain obstacle distribution data within the target area, calculate obstacle density values ​​based on the obstacle distribution data, and use the obstacle density values ​​as the first risk factor; Acquire terrain data within the target area, calculate slope and surface roughness values ​​based on the terrain data, and calculate a second risk factor based on the slope and surface roughness values; The number of dynamic targets within the target area is obtained, and the number of dynamic targets is used as a third risk factor. The first, second, and third risk factors are compared with their corresponding risk thresholds. If the number of risk factors exceeding the corresponding risk threshold is greater than or equal to 2, the environmental hazard condition is deemed to be met; otherwise, the environmental hazard condition is deemed not to be met.

10. A return-to-home system for unmanned vehicles in complex reconnaissance environments, characterized in that: The system includes: The monitoring module is used to monitor the operating status and environmental information of the unmanned vehicle in real time, and to determine whether there are conditions for returning to base based on the monitoring results; The determination module is used to determine the current return mode based on the existing return conditions when they exist. The planning module is used to plan the return path of the unmanned vehicle from its current location to the return destination based on the current return mode; The adjustment module is used to control the unmanned vehicle to travel along the return path and continuously monitor whether the return conditions change during the journey. If the return conditions change, the current return mode is re-determined based on the changed return conditions, and the return path is adjusted according to the re-determined current return mode until the unmanned vehicle reaches the return destination. The control module is used to determine that the return journey is complete when the unmanned vehicle reaches the return destination and to control the unmanned vehicle to enter the standby state.

Images