Air-ground cooperative unmanned vehicle path planning method in complex environment

Abstract

The invention discloses an air-ground cooperative unmanned vehicle path planning method in a complex environment, and the method comprises the steps: firstly obtaining distribution point position information, distribution point material demand information, environment information, distribution center position information and unmanned vehicle information; secondly, preliminarily performing global path planning according to the information; then, the unmanned vehicle runs according to a path obtained by global planning, and when the unmanned vehicle senses that the surrounding environment cannot pass through, a built-in unmanned aerial vehicle of the unmanned vehicle lifts off to recollect surrounding environment information to obtain updated environment information; and finally, according to the updated environment data, local path planning is carried out on each path through traversal constraint, re-planning path constraint and energy consumption constraint, and a distribution path of the unmanned vehicle in a complex environment is obtained. Aiming at the requirements of maneuverability, rapidity and continuity of the unmanned vehicle in a complex environment, the unmanned vehicle is not restrained by an existing road, so that the emergency material distribution time is shortened, and the rescue efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of unmanned vehicle path planning for unmanned systems, and in particular to an air-ground collaborative unmanned vehicle path planning method in a complex environment. Background Art

[0002] The frequent occurrence of emergencies warns people that the unpredictability and suddenness of disasters, their high frequency of occurrence and high social harm make it urgent to study emergency rescue work. The path problem of logistics unmanned vehicles that undertake the task of transporting materials is the top priority of the dispatching link. The great destructiveness of disasters often causes varying degrees of damage to local roads. In this case, the advantages and characteristics of unmanned aerial vehicles are fully utilized to pass through complex and changeable off-road environments through the perception of environmental conditions to achieve safe, timely and effective delivery of materials. Under normal conditions, the logistics industry is aimed at maximizing economic benefits, while emergency logistics has the characteristics of weak economy. Its purpose is to deliver emergency rescue materials to the disaster site in a safe and effective way in the shortest possible time.

[0003] There are some significant problems and shortcomings in the existing path planning methods for air-ground collaborative unmanned vehicles. First, most of the current methods mainly consider the terrain elevation and slope information of the area when considering the traversability of unmanned vehicles; however, in complex environments, plants will also have a certain impact on the operation of unmanned vehicles.

[0004] Secondly, there is little research on emergencies that may occur during the operation of unmanned vehicles. Under normal circumstances, unmanned vehicles will operate according to the planned route until all delivery tasks are completed. In complex environments, although the environmental information of the unmanned vehicle operation area is well understood, it is difficult for unmanned vehicles to operate accurately along the planned route. In addition, environmental factors change rapidly, and unmanned vehicles may encounter environmental information that is not mastered during operation, resulting in interruptions in the operation of unmanned vehicles.

[0005] In addition, the current unmanned vehicle path planning methods rarely consider the issue of energy consumption, which will also make these methods difficult to implement. Since the delivery task is carried out in a complex environment, there is no charging equipment such as charging piles in the entire environment. Therefore, it is necessary to carefully control the energy consumption of the entire device so that the entire device can operate within its maximum energy limit.

[0006] This patent expands the problem of emergency logistics path planning by introducing environmental factors, replanning factors, and energy consumption factors. Among environmental factors, the vegetation coverage rate is introduced into the analysis of the traversability of unmanned vehicles; among replanning factors, the detour of unmanned vehicles and potential obstacles are introduced; among energy consumption factors, the energy consumption of drones and unmanned vehicles and the impact of ambient temperature on the energy consumption of unmanned vehicles are considered respectively. An air-ground collaborative unmanned vehicle path planning model is established to solve the problem of emergency logistics distribution path planning. Summary of the invention

[0007] Purpose of the invention: The present invention proposes an air-ground collaborative unmanned vehicle path planning method in a complex environment to solve the path selection and optimization problems in emergency logistics distribution and shorten the emergency rescue time.

[0008] Technical solution: The present invention discloses a method for air-ground coordinated unmanned vehicle path planning in a complex environment, which specifically includes the following steps:

[0009] (1) Obtaining the distribution point location information, distribution point material demand information, environmental information, distribution center location information, and unmanned vehicle information;

[0010] (2) Based on the environmental information, target point information, and distribution center location information obtained in step (1), the entire area is divided into grids and a preliminary global path planning is performed on the area;

[0011] (3) The unmanned vehicle runs according to the path obtained by the global plan. When the unmanned vehicle encounters a detour or there are potential obstacles around the unmanned vehicle that cannot be detected, the drone built into the unmanned vehicle will take off to re-collect the surrounding environmental information and obtain updated environmental information;

[0012] (4) Based on the updated environmental data, a local path optimization model is constructed to perform local path planning by considering regional traversability constraints, re-planning route constraints, and energy consumption constraints; the unmanned vehicle then continues to perform the task according to the re-planned route, repeating steps (3) and (4) until all task points on the local path have been delivered, thus completing the delivery task.

[0013] Furthermore, the area traversability in step (4) includes determining whether the area is traversable and the energy consumption of the unmanned vehicle traveling in the traversable area; wherein the area traversability indicator is determined by the inclination angle of the terrain, and the energy consumption of the unmanned vehicle traveling in the traversable area is determined by the driving distance of the unmanned vehicle and the pitch angle of the unmanned vehicle.

[0014] Furthermore, the process of implementing the regional traversability constraint in step (4) is as follows:

[0015] The regional traversability considers that the weight of the unmanned vehicle loaded with cargo is related to the friction of the unmanned vehicle traveling in the area. When the force generated by the weight of the unmanned vehicle in the direction of the inclined surface is greater than or equal to the friction, the unmanned vehicle cannot pass through the cell; when the force generated by the weight of the unmanned vehicle in the direction of the inclined surface is less than the friction, the unmanned vehicle can pass through the cell; under the specific model:

[0016]

[0017] Where α is the angle of the inclined surface of the region; Δh ii' is the average height difference between the collection point i and the surrounding area i'; d is the horizontal or vertical distance between the cells; μ is the friction coefficient of the area.

[0018] Furthermore, the energy consumption in step (4) includes the energy consumption of the unmanned vehicle and the energy consumption of the unmanned aerial vehicle.

[0019] Furthermore, the energy consumption of the unmanned vehicle includes the running distance, the pitch angle of the unmanned vehicle and the temperature. As for the transport distance, the longer the unmanned vehicle runs, the longer the transport distance is and the higher the energy consumption is. As for the pitch angle, when the unmanned vehicle is driving on a slope, driving in the direction with a smaller pitch angle is conducive to saving energy. When the unmanned vehicle drives diagonally into the slope, the pitch angle of the unmanned vehicle is smaller than the pitch angle when driving vertically into the slope, which is also an important factor affecting energy consumption. The direction of movement of the unmanned vehicle is divided into longitudinal or lateral movement and diagonal movement, as shown below:

[0020]

[0021] Among them, α p is the pitch angle of the unmanned vehicle; after obtaining the relationship between the driving direction of the unmanned vehicle and the pitch angle of the unmanned vehicle through the above formula, the energy consumption of the unmanned vehicle moving from cells ab to different cells is obtained;

[0022] The effect of external temperature on the energy consumption of the unmanned vehicle The energy consumption of the unmanned vehicle is:

[0023]

[0024] Therefore, the energy consumption of the unmanned vehicle is E UGV for:

[0025]

[0026] Among them, E g is the energy consumption per unit distance when the unmanned vehicle is running; h is the height difference of the unmanned vehicle; d is the horizontal or vertical distance between cells; α ρ is the pitch angle of the unmanned vehicle; T is the operating temperature of the unmanned vehicle.

[0027] Furthermore, the energy consumption of the drone in step (4) includes the energy consumption generated by the drone's climbing, hovering, and descending; the energy consumption generated by the drone's climbing and descending is:

[0028] E UAV1 = mgh(E u +E d )

[0029] Energy consumption E generated by the drone during the hovering phase UAV2 for:

[0030]

[0031] Therefore, the energy consumption of the drone is:

[0032]

[0033] Among them, mg is the gravity of the drone; H is the climbing height of the drone; E u The energy consumed per unit height when the drone climbs; E d is the energy consumed per unit height when the drone descends; t is the hovering time of the drone; ρ is the air density; A UAV is the area of ​​the drone propeller.

[0034] Furthermore, the regional traversability constraint in step (4) also takes into account the vegetation coverage, and establishes the following relationship model between vegetation coverage and friction coefficient:

[0035] μ'=(1-f 2 )μ

[0036] Among them, μ' is the friction coefficient after introducing the vegetation coverage correction; f is the vegetation coverage, and two vegetation coverage thresholds f1 and f2 are introduced; the details are as follows:

[0037]

[0038] Furthermore, the re-planning path constraint in step (4) is to consider the re-planning path caused by the detour of the unmanned vehicle. Specifically, the unmanned vehicle will obtain the distance information D between the unmanned vehicle and other target delivery points during the delivery process. ij , when the driverless car uses D ij When the shortest delivery point j is the delivery point for delivery, D ij Should be gradually reduced, and if the unmanned vehicle has to go far due to terrain reasons, then D ij will gradually increase, when the unmanned vehicle takes j as the delivery target D ij Greater than the distance D between the unmanned vehicle and another delivery point j' ij'When , the unmanned vehicle will re-send the drone to plan the route after changing the delivery target, and give priority to delivering to the delivery point j'; at this time, a re-planning point m is set up ab , a m 、b m is the horizontal and vertical coordinates of m; the relevant model is as follows:

[0039]

[0040] When comparing the distance between the unmanned vehicle and other target delivery points, D can be used to reduce the amount of calculation. ij 2 Compare; calculate the actual distance from point i to point j:

[0041]

[0042] Among them, n* represents the number of times the unmanned vehicle moves horizontally, and m* represents the number of times the unmanned vehicle moves longitudinally. If the unmanned vehicle moves diagonally, n* and m* are added once respectively, and h* represents the number of times the height of the unmanned vehicle changes. The changes in the horizontal, longitudinal, and height are accumulated to calculate the overall Euclidean distance.

[0043] During the transportation, the unmanned vehicle compares the distance between the unmanned vehicle and each delivery point j. When the distance between the initial delivery target point j and the unmanned vehicle is always the smallest, the delivery target does not change and the path length does not change. When there is another delivery point j' whose distance to the unmanned vehicle is smaller than the distance between the initial delivery point j and the unmanned vehicle, a replanning point m is set up, and the delivery target j is changed to j', and the path length also changes. Specifically, L ij ,L im ,L mj' for:

[0044]

[0045] Furthermore, the re-planning path constraint in step (4) is to consider the re-planning path of the drone due to potential obstacles, specifically: define several sets of traversable area P, obstacle area O, and potential obstacle PO, that is, the potential obstacle area is the traversable area when the path is planned for the first time; it is expressed as: When the unmanned vehicle passes a potential obstacle area along the route i It is impassable; at this time, the potential obstacle area is placed in the obstacle area set, PO i Remove O from P and update the set O; launch the drone to replan the path to the new obstacle area O; the specific process is: initial state: P∪O = the entire environment, If an unmanned vehicle passes by the PO i And PO iIf it is not passable, execute: P:P\{PO i},O:=O∪{PO i}; Replan the path to ensure that the new path does not pass through the new set O.

[0046] Furthermore, the implementation process of constructing the local path planning model in step (4) is as follows:

[0047] Design local planning path objective function:

[0048] min Z=αT(i)+βD(i)

[0049]

[0050] Among them, T(i) represents the travel time between the current collection point and the pass area, which is obtained by dividing the distance between the current collection point and the pass area by the running speed of the unmanned vehicle; D(i) represents the distance between the current collection point and the target point. The distance cost D(n) between the current collection point and the target point can be obtained by avoiding the replanning path caused by the detour of the unmanned vehicle. ij express;

[0051] The traversability constraint H(i) is the tangent value of the inclination angle between the collection point and the passage area. H(i) should be less than the corrected friction coefficient of the area to ensure the stable operation of the unmanned vehicle:

[0052]

[0053] Replanning constraint D(i), the unmanned vehicle compares the distance between the unmanned vehicle and each delivery point j during transportation. When the distance between the initial delivery target point j and the unmanned vehicle is always the smallest, the delivery target does not change and the path length does not change; when there is another delivery point j' whose distance to the unmanned vehicle is smaller than the distance between the initial delivery point j and the unmanned vehicle, a replanning point m is set and the delivery target j is changed to j':

[0054]

[0055] Energy consumption constraint E 总 It means that the sum of the energy consumption of the UAV and the unmanned vehicle cannot be greater than the maximum energy of the entire device:

[0056] E UAV +E UGV ≤E 总 .

[0057] Beneficial effects: Compared with the prior art, the beneficial effects of the present invention are: the present invention provides new ideas for emergency logistics rescue work, broadens the innovative thinking, and is a further expansion of the existing emergency logistics unmanned vehicle path problem; in view of the mobility, speed and continuity requirements of unmanned vehicles in complex environments, the unmanned vehicles are not constrained by existing roads, thereby shortening the emergency material distribution time and further improving the rescue efficiency; by providing a scientific path selection decision-making plan for reference, the blindness of the rescue team in path selection is reduced, and the expenditure of manpower, material and financial resources is reduced to a certain extent. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 is a flow chart of the present invention;

[0059] Figure 2 is a two-dimensional global path planning graph;

[0060] Figure 3 It is a two-dimensional optimized local path planning graph;

[0061] Figure 4 It is a two-dimensional local path planning graph without considering replanning;

[0062] Figure 5 It is a three-dimensional optimized local path planning graph. DETAILED DESCRIPTION

[0063] The present invention is further described in detail below based on the accompanying drawings.

[0064] like Figure 1 As shown, the present invention proposes a path planning method for an unmanned vehicle in air-ground collaboration in a complex environment, which specifically includes the following steps:

[0065] Step 1: Obtain the distribution point location information, distribution point material demand information, environmental information, distribution center location information, and unmanned vehicle information.

[0066] Step 2: Based on the environmental information, target point information, and distribution center location information obtained in step 2, the entire area is divided into grids and a preliminary global path planning is performed for the area.

[0067] Step 3: The unmanned vehicle runs along the path obtained by the global plan. When the unmanned vehicle encounters a detour or there are potential obstacles around the unmanned vehicle that cannot be detected, the drone built into the unmanned vehicle will take off to re-collect the surrounding environmental information and obtain updated environmental information;

[0068] Step 4: Based on the updated environmental data, a local path optimization model is constructed to perform local path planning by considering regional traversability constraints, re-route constraints, and energy consumption constraints; the unmanned vehicle then continues the task along the re-planned route, repeating steps 3 and 4 until all task points on the local path are delivered, thus completing the delivery task.

[0069] (1) Regional traversability constraints:

[0070] The study of regional traversability mainly includes judging whether the region is traversable and the energy consumption of the unmanned vehicle in the traversable region. Among them, the regional traversability index is determined by the inclination angle of the terrain, and the energy consumption of the unmanned vehicle in the traversable region is determined by the driving distance of the unmanned vehicle and the pitch angle of the unmanned vehicle.

[0071] The research on regional traversability mainly considers the weight of the unmanned vehicle loaded with cargo and the friction of the unmanned vehicle in the area. When the force generated by the weight of the unmanned vehicle in the direction of the inclined surface is greater than or equal to the friction, the unmanned vehicle cannot pass through the cell. When the force generated by the weight of the unmanned vehicle in the direction of the inclined surface is less than the friction, the unmanned vehicle can pass through the cell. The specific model is as follows:

[0072]

[0073] Where α is the angle of the inclined surface of the region; Δh ii' is the average height difference between the collection point i and the surrounding area i'; d is the horizontal or vertical distance between the cells; μ is the friction coefficient of the area.

[0074] Secondly, the vegetation coverage and soil moisture content in the area also have different degrees of influence on the passage of unmanned vehicles. The greater the vegetation coverage in the area, the higher the moisture content in the soil, which will reduce the friction coefficient when the unmanned vehicle passes through the area. Therefore, when conducting traversability analysis, the vegetation coverage of the area is also one of the important factors to be considered. The following relationship model between vegetation coverage and friction coefficient is established.

[0075] μ'=(1-f 2 )μ (2)

[0076] Among them, μ' is the friction coefficient after introducing the vegetation coverage correction; f is the vegetation coverage, and two vegetation coverage thresholds f1 and f2 are introduced here. When the vegetation coverage in the area is less than f1, it will not affect the current friction coefficient; when the vegetation coverage in the area is greater than f1 and less than f2, the friction coefficient of the area will change according to formula (2); when the vegetation coverage in the area is greater than f2, it is considered that the friction coefficient of the area is too small and it is judged that the area cannot be crossed. The specific situation is expressed by formula (3):

[0077]

[0078] (2) Research on re-planning routes:

[0079] Under normal circumstances, when unmanned vehicles are transporting in complex environments, they will first transport according to the path obtained after the regional traversability study, and then carry out the delivery task according to the order of users in the path. However, in reality, unmanned vehicles cannot smoothly complete the delivery task according to the planned path and the order of task points. This study mainly considers the situation where unmanned vehicles detour and drones re-plan routes due to potential obstacles.

[0080] Avoid re-routing caused by driverless vehicles:

[0081] The unmanned vehicle performs delivery according to the constraints in the crossing analysis during transportation. At this time, due to terrain problems, the unmanned vehicle may deviate from the target delivery point. In this regard, this method proposes a method to change the delivery target in time. Specifically, the unmanned vehicle will obtain the distance information D between the unmanned vehicle and other target delivery points during delivery. ij When the driverless car uses D ij When the shortest delivery point j is the delivery point for delivery, D ij Should be gradually reduced, and if the unmanned vehicle has to go far due to terrain reasons, then D ij will gradually increase, when the unmanned vehicle takes j as the delivery target D ij Greater than the distance D between the unmanned vehicle and another delivery point j' ij' When , the unmanned vehicle will re-send the drone to plan the route after changing the delivery target, and give priority to delivering to the delivery point j'. At this time, a re-planning point m is set up ab , a m 、b m is the horizontal and vertical coordinates of m. The relevant model is as follows:

[0082]

[0083] Formula (4) is the distance calculation formula between the unmanned vehicle and other target delivery points, which is represented by the Euclidean distance between the unmanned vehicle and the delivery point. When comparing the distance between the unmanned vehicle and other target delivery points, D can be used to reduce the amount of calculation. jk 2 Make a comparison.

[0084]

[0085] Formula (5) is the actual distance formula from point i to point j, where n* represents the number of times the unmanned vehicle moves horizontally, and m* represents the number of times the unmanned vehicle moves longitudinally. If the unmanned vehicle moves diagonally, n* and m* are added once respectively, and h* represents the number of times the height of the unmanned vehicle changes. The changes in the horizontal, longitudinal, and height are accumulated to calculate the overall Euclidean distance.

[0086] During the transportation, the unmanned vehicle compares the distance between the unmanned vehicle and each delivery point j. When the distance between the initial delivery target point j and the unmanned vehicle is always the smallest, the delivery target does not change and the path length does not change. When there is another delivery point j' whose distance to the unmanned vehicle is smaller than the distance between the initial delivery point j and the unmanned vehicle, a replanning point m is set up, and the delivery target j is changed to j', and the path length also changes. ij ,L im ,L mj' , obtained from formula (6).

[0087]

[0088] Rerouting due to potential obstacles:

[0089] When the drone senses the environment and plans the path in the air, there may be obstacles in the blind spot due to the perspective problem, which makes the drone unable to run along the predetermined path. In view of this situation, several sets of passable areas (P), obstacle areas (O), and potential obstacles (PO) are defined here. That is, when the path is planned for the first time, the potential obstacle area is the passable area. It is manifested as: When the unmanned vehicle passes a potential obstacle area (PO i ) is not passable. At this time, the potential obstacle area is placed in the obstacle area set (PO i Remove O from P, update the set O), and launch the drone to replan the path to the new obstacle area O. The specific process can be expressed as:

[0090] Initial state: P∪O = the entire environment,

[0091] If an unmanned vehicle passes by the PO i And PO i If it is not passable, execute: P:P\{PO i},O:=O∪{PO i}

[0092] Replan the path to ensure that the new path does not pass through the new set O.

[0093] If the UAV does not pass through any potential obstacle area or the potential obstacle area it passes through is still passable, there is no need to update the obstacle area set O or re-plan the path.

[0094] (3) Energy confinement research:

[0095] Since the delivery task is carried out in a complex environment, there is no charging equipment such as charging piles in the entire environment. Therefore, it is necessary to carefully control the energy consumption of the entire device. This patent takes into account that the drone can be charged inside the unmanned vehicle, so the energy consumption of the unmanned vehicle and the drone are combined and do not exceed the energy limit of the entire device. The following considers the effects of altitude, air density, and the area of ​​the drone propeller on the energy consumption of the drone, and the effects of distance, pitch angle of the unmanned vehicle, and temperature on the energy consumption of the unmanned vehicle, and establishes and compares the energy consumption formula of the drone and the energy consumption formula of the unmanned vehicle.

[0096] Drone energy consumption E UAV :

[0097] With the advancement of technology, the application of drones in transportation systems has become increasingly feasible. At present, the main disadvantage of drones is their limited endurance, and the energy consumption of drones must be carefully managed. Among them, the energy consumption of drones can be subdivided into: climbing, hovering, descending and horizontal flight stages, and the consumption in the horizontal flight stage accounts for 87.63% of the total consumption. In addition, in this study, drones are mainly used for environmental perception and route replanning, and drones are not required to perform transportation work, so only the energy consumption generated by the climbing, hovering and descending of drones needs to be considered. The energy consumption generated by the climbing and descending of drones can be expressed by formula (7):

[0098] E UAV1 = mgh(E u +E d )(7)

[0099] In the hovering stage, as the altitude increases, the air density decreases. The smaller the density, the less energy can be obtained. The decrease in air density will increase the energy consumption of the drone. Therefore, the energy consumption E generated by the drone in the hovering stage is UAV2 It can be expressed by formula (8):

[0100]

[0101] Finally, by combining the energy consumption generated by the UAV's climbing and descending stages with the energy consumption generated by the hovering stage, the energy consumption formula of the UAV during the stage of environmental perception and re-route planning is obtained, which is expressed by formula (9):

[0102]

[0103] Among them, mg is the gravity of the drone; H is the climbing height of the drone; E u The energy consumed per unit height when the drone climbs; E d is the energy consumed per unit height when the drone descends; t is the hovering time of the drone; ρ is the air density; A UAV is the area of ​​the drone propeller.

[0104] Energy consumption of unmanned vehicles E UGV :

[0105] During the operation of the unmanned vehicle, many factors affect its energy consumption, including the operating distance, the pitch angle of the unmanned vehicle, and the temperature. The first is the transportation distance. Generally speaking, the longer the operation of the unmanned vehicle, the longer its transportation distance, resulting in higher energy consumption. Secondly, the pitch angle is also an important factor affecting energy consumption. Road surfaces with large pitch angles will increase the energy required for the unmanned vehicle to overcome gravity, especially on uphill sections. At the same time, although energy consumption is reduced when going downhill, the energy recovery efficiency of the braking system will also affect the overall energy consumption.

[0106] When the unmanned vehicle is driving along the contour line, the pitch angle is 0. In this case, the climbing energy is minimal, which prevents soil erosion to a large extent. Therefore, we link the energy consumption of the unmanned vehicle in the traversable area with the pitch angle of the unmanned vehicle chassis. Simply put, when the unmanned vehicle is driving on a slope, driving in the direction with a smaller pitch angle is beneficial to saving energy. In addition, when the unmanned vehicle drives diagonally into the slope, the pitch angle of the unmanned vehicle is smaller than the pitch angle when it drives vertically into the slope. In order to better show the effect of the pitch angle on energy consumption, the direction of the unmanned vehicle's movement is divided into longitudinal or lateral movement and diagonal movement. As expressed in formula (10):

[0107]

[0108] Among them, α p is the pitch angle of the unmanned vehicle. After obtaining the relationship between the driving direction of the unmanned vehicle and the pitch angle of the unmanned vehicle through formula (10), the energy consumption of the unmanned vehicle moving from the ab cell to the a'b' cell can be obtained.

[0109] Finally, the outside temperature has a significant impact on the energy consumption of the unmanned vehicle, which can be expressed by formula (11):

[0110]

[0111] The temperature range of unmanned vehicles during operation is usually between -40℃ and 85℃. This temperature range is determined based on the environmental conditions of the unmanned vehicle and the design of the unmanned vehicle's own heat dissipation and cooling system. Battery performance is greatly affected by temperature changes. In low temperature environments, the capacity of the power battery will decrease, and the charging and discharging efficiency will decrease, thereby increasing energy consumption. The changes in different temperatures cause the available capacity of lithium batteries to decay to varying degrees: the available capacity is 70% at -10℃, 85% at 0℃, and 100% at 25℃. High temperatures may accelerate battery aging and reduce its efficiency. After the operating temperature exceeds 40℃, the battery cycle life is halved for every 10℃ increase. The temperature difference inside the battery pack causes different decay rates of the single cells, reducing performance. According to the above research, the energy consumption E of the unmanned vehicle UGV It can be expressed by the following formula:

[0112]

[0113] Among them, E g is the energy consumption per unit distance when the unmanned vehicle is running; h is the height difference of the unmanned vehicle; d is the horizontal or vertical distance between cells; α ρ is the pitch angle of the unmanned vehicle; T is the operating temperature of the unmanned vehicle.

[0114] (4) Constructing a local path planning model:

[0115] i is the collection point set, i=1, 2, ..., n; j is the target point set, i=1, 2, ..., j; v is the running speed of the unmanned vehicle; Δh is the height difference between the current collection point and the passage area or the height difference between the current collection point and the target point; d is the horizontal or vertical distance between the cells; T(i) is the travel time between the current collection point and the passage area; D(i) is the distance between the current collection point and the target point; α, β are the corresponding weight coefficients; H(i) is the tangent value of the inclination angle between the collection point and the passage area; μ' is the corrected friction coefficient; D ij is the distance between collection point i and target delivery point j; D ij' is the distance between collection point i and other target delivery points j'; x i is a 0-1 variable. When x i =1, indicating that at the collection point i, the drone needs to perceive the environment in the air and re-plan the path, otherwise, x i =0; E UAV is the energy consumption of the drone; Mg is the gravity of the drone; H is the climbing height of the drone; E u The energy consumed per unit height when the drone climbs; E d is the energy consumed per unit height when the drone descends; t is the hovering time of the drone; ρ is the air density; A UAVis the area of ​​the drone propeller; E UGV is the energy consumption of the unmanned vehicle; E g is the energy consumption per unit distance when the unmanned vehicle is running; α ρ is the pitch angle of the unmanned vehicle; T is the operating temperature of the unmanned vehicle.

[0116] According to the guidance of the global path points, the global path is corrected in the local environment to plan the local optimal path.

[0117] min Z=αT(i)+βD(i) (13)

[0118] The local planning path objective function (13) is designed to comprehensively consider the travel time and distance during the path planning process, so as to minimize the comprehensive objective of the final path and meet the transportation requirements of unmanned vehicles in complex environments.

[0119]

[0120] Formula (14) represents the travel time between the current collection point and the pass area, which is obtained by dividing the distance between the current collection point and the pass area by the operating speed of the unmanned vehicle.

[0121]

[0122] (15) represents the distance between the current collection point and the target point. The distance cost D(n) between the current collection point and the target point can be obtained by avoiding the re-planning path caused by the detour of the unmanned vehicle. jk express.

[0123]

[0124] Traversability constraint (16). Through traversability research, it can be obtained that the tangent value H(i) of the inclination angle between the collection point and the passage area should be smaller than the corrected friction coefficient of the area to ensure the stable operation of the unmanned vehicle.

[0125]

[0126] The replanning constraint (17) indicates that the unmanned vehicle compares the distance between the unmanned vehicle and each delivery point j during transportation. When the distance between the initial delivery target point j and the unmanned vehicle is always the smallest, the delivery target does not change and the path length does not change. When there is another delivery point j' whose distance to the unmanned vehicle is smaller than the distance between the initial delivery point j and the unmanned vehicle, a replanning point m is set and the delivery target j is changed to j'.

[0127] E UAV +E UGV ≤E 总 (18)

[0128] Energy consumption constraint (18): The energy consumption of the entire device is divided into the energy consumption of the drone and the energy consumption of the unmanned vehicle. This constraint means that the sum of the energy consumption of the drone and the unmanned vehicle cannot be greater than the maximum energy of the entire device.

[0129]

[0130] Formulas (19) and (20) are the energy consumption calculation formulas for drones and unmanned vehicles, respectively. In formula (19), the energy consumption generated by the drone's climbing and descending stages is combined with the energy consumption generated by the hovering stage to obtain the energy consumption formula for the drone during the stage of environmental perception and route replanning. In formula (20), the battery performance of unmanned vehicles is greatly affected by temperature changes. Changes in temperature cause the available capacity of lithium batteries to decay to varying degrees: the available capacity is 70% at -10°C, 85% at 0°C, and 100% at 25°C. High temperatures may accelerate battery aging and reduce its efficiency. After the operating temperature exceeds 40°C, the battery cycle life is halved for every 10°C increase.

[0131] The present invention considers a map range of 10km*10km, a grid size of 100m*100m, and a distribution center location of (1,1) according to the grid order. The location information of the four target task points are (37,22), (33,43), (76,80), and (93,23), respectively. The device can detect a range of 1 grid, the drone detection range is 5 grids, and the vegetation coverage rate range is (0.05-0.3). And when any of the following three situations occurs, the unmanned vehicle stops running; 1) the device energy reaches the upper limit; 2) all target points have been completed; 3) it is impossible to pass to any target point.

[0132] Obtain the location information, obstacle information, terrain information and vegetation coverage information of each disaster point. And carry out global path planning based on the information, such as Figure 2 As shown in the figure, the transport path is 0-1-2-3-4. Then, according to the potential obstacles and detours of the unmanned vehicle during operation, the drone needs to detect the environment and replan the path. After that, the locally optimized path is obtained, as shown in Figure 3 As shown in the figure, the transport path is 0-2-1-4-3, the total distance is 27503m, and the number of drone explorations is 28. Figure 4 As shown, the present invention also aims at not considering the re-planning situation, the device is transported according to the globally planned path 0-1-2-3-4, the total distance is 30350m, and the number of drone explorations is 30 times.

[0133] By comparing the two results, the distribution distance of emergency logistics is shortened when the re-routing is considered. It can be seen that considering the re-routing in a complex environment shortens the total distance of emergency logistics transportation and improves the rescue efficiency. Figure 5 shown.

[0134] The above implementation modes are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A path planning method for an air-ground collaborative unmanned vehicle in a complex environment, characterized in that: The following steps are involved: (1) Obtaining the distribution point location information, distribution point material demand information, environmental information, distribution center location information, and unmanned vehicle information; (2) Based on the environmental information, target point information, and distribution center location information obtained in step (1), the entire area is divided into grids and a preliminary global path planning is performed on the area; (3) The unmanned vehicle runs according to the path obtained by the global plan. When the unmanned vehicle encounters a detour or there are potential obstacles around the unmanned vehicle that cannot be detected, the drone built into the unmanned vehicle will take off to re-collect the surrounding environmental information and obtain updated environmental information; (4) Based on the updated environmental data, a local path optimization model is constructed to perform local path planning by considering regional traversability constraints, re-planning route constraints, and energy consumption constraints; the unmanned vehicle then continues to perform the task according to the re-planned route, repeating steps (3) and (4) until all task points on the local path have been delivered, thus completing the delivery task.

2. According to the method for path planning of an air-ground coordinated unmanned vehicle in a complex environment in claim 1, it is characterized in that: The area traversability in step (4) includes determining whether the area is traversable and the energy consumption of the unmanned vehicle traveling in the traversable area; wherein the area traversability index is determined by the inclination angle of the terrain, and the energy consumption of the unmanned vehicle traveling in the traversable area is determined by the driving distance of the unmanned vehicle and the pitch angle of the unmanned vehicle.

3. According to the method for path planning of an air-ground coordinated unmanned vehicle in a complex environment in claim 1, it is characterized in that: The process of implementing the regional traversability constraint in step (4) is as follows: The regional traversability considers that the weight of the unmanned vehicle loaded with cargo is related to the friction of the unmanned vehicle traveling in the area. When the force generated by the weight of the unmanned vehicle in the direction of the inclined surface is greater than or equal to the friction, the unmanned vehicle cannot pass through the cell; when the force generated by the weight of the unmanned vehicle in the direction of the inclined surface is less than the friction, the unmanned vehicle can pass through the cell; under the specific model: Where α is the angle of the inclined surface of the region; Δh ii' is the average height difference between the collection point i and the surrounding area i'; d is the horizontal or vertical distance between the cells; μ is the friction coefficient of the area.

4. According to the method for path planning of an air-ground coordinated unmanned vehicle in a complex environment in claim 1, it is characterized in that: The energy consumption in step (4) includes the energy consumption of the unmanned vehicle and the energy consumption of the unmanned aerial vehicle.

5. The method for path planning of an air-ground coordinated unmanned vehicle in a complex environment according to claim 4 is characterized in that: The energy consumption of the unmanned vehicle includes the running distance, the pitch angle of the unmanned vehicle and the temperature. As for the transportation distance, the longer the unmanned vehicle is running, the longer the transportation distance is, and the higher the energy consumption is. As for the pitch angle, when the unmanned vehicle is driving on a slope, driving in the direction with a smaller pitch angle is conducive to saving energy. When the unmanned vehicle enters the slope obliquely, the pitch angle of the unmanned vehicle is smaller than the pitch angle when entering the slope vertically, which is also an important factor affecting energy consumption. The direction of movement of the unmanned vehicle is divided into longitudinal or lateral movement and diagonal movement, as shown below: Among them, α p is the pitch angle of the unmanned vehicle; after obtaining the relationship between the driving direction of the unmanned vehicle and the pitch angle of the unmanned vehicle through the above formula, the energy consumption of the unmanned vehicle moving from cells ab to different cells is obtained; The effect of external temperature on the energy consumption of the unmanned vehicle The energy consumption of the unmanned vehicle is: Therefore, the energy consumption of the unmanned vehicle is E UGV for: Among them, E g is the energy consumption per unit distance when the unmanned vehicle is running; h is the height difference of the unmanned vehicle; d is the horizontal or vertical distance between cells; α ρ is the pitch angle of the unmanned vehicle; T is the operating temperature of the unmanned vehicle.

6. The method for path planning of an air-ground coordinated unmanned vehicle in a complex environment according to claim 4 is characterized in that: The energy consumption of the drone in step (4) includes the energy consumption generated by the drone's climbing, hovering, and descending; the energy consumption generated by the drone's climbing and descending is: AND UAV1 =mgh(E u +E d ) Energy consumption E generated by the drone during the hovering phase UAV2 for: Therefore, the energy consumption of the drone is: Among them, mg is the gravity of the drone; H is the climbing height of the drone; E u The energy consumed per unit height when the drone climbs; E d is the energy consumed per unit height when the drone descends; t is the hovering time of the drone; ρ is the air density; A UAV is the area of ​​the drone propeller.

7. The method for path planning of an air-ground coordinated unmanned vehicle in a complex environment according to claim 1, characterized in that: The regional traversability constraint in step (4) also takes into account the vegetation coverage rate, and establishes the following relationship model between vegetation coverage rate and friction coefficient: μ'=(1-f 2 )m Among them, μ' is the friction coefficient after introducing the vegetation coverage correction; f is the vegetation coverage, and two vegetation coverage thresholds f1 and f2 are introduced; the details are as follows:

8. The method for path planning of an air-ground coordinated unmanned vehicle in a complex environment according to claim 1 is characterized in that: The re-planning path constraint in step (4) is to consider the re-planning path caused by the detour of the unmanned vehicle. Specifically, the unmanned vehicle will obtain the distance information D between the unmanned vehicle and other target delivery points during the delivery process. ij , when the driverless car uses D ij When the shortest delivery point j is the delivery point for delivery, D ij Should be gradually reduced, and if the unmanned vehicle has to go far due to terrain reasons, then D ij will gradually increase, when the unmanned vehicle takes j as the delivery target D ij Greater than the distance D between the unmanned vehicle and another delivery point j' ij' When , the unmanned vehicle will re-send the drone to plan the route after changing the delivery target, and give priority to delivering to the delivery point j'; at this time, a re-planning point m is set up ab , a m 、b m is the horizontal and vertical coordinates of m; the relevant model is as follows: When comparing the distance between the unmanned vehicle and other target delivery points, D can be used to reduce the amount of calculation. ij 2 Compare; calculate the actual distance from point i to point j: Among them, n* represents the number of times the unmanned vehicle moves horizontally, and m* represents the number of times the unmanned vehicle moves longitudinally. If the unmanned vehicle moves diagonally, n* and m* are added once respectively, and h* represents the number of times the height of the unmanned vehicle changes. The changes in the horizontal, longitudinal, and height are accumulated to calculate the overall Euclidean distance. During the transportation, the unmanned vehicle compares the distance between the unmanned vehicle and each delivery point j. When the distance between the initial delivery target point j and the unmanned vehicle is always the smallest, the delivery target does not change and the path length does not change. When there is another delivery point j' whose distance to the unmanned vehicle is smaller than the distance between the initial delivery point j and the unmanned vehicle, a replanning point m is set up, and the delivery target j is changed to j', and the path length also changes. Specifically, L ij ,L im ,L mj' for:

9. The method for path planning of an air-ground coordinated unmanned vehicle in a complex environment according to claim 1, characterized in that: The re-planning path constraint in step (4) is to consider the re-planning path of the drone due to potential obstacles, specifically: define several sets of traversable area P, obstacle area O, and potential obstacle PO, that is, the potential obstacle area is the traversable area when the path is planned for the first time; it is expressed as: When the unmanned vehicle passes a potential obstacle area along the route i It is impassable; at this time, the potential obstacle area is placed in the obstacle area set, PO i Remove O from P and update the set O; launch the drone to re-plan the path to the new obstacle area O; the specific process is: initial state: P∪O = the entire environment, If an unmanned vehicle passes by the PO i And PO i If it is not passable, execute: Replan the path to ensure that the new path does not pass through the new set O.

10. The method for path planning of an air-ground coordinated unmanned vehicle in a complex environment according to claim 1, characterized in that: The implementation process of building the local path planning model in step (4) is as follows: Design local planning path objective function: min Z=αT(i)+βD(i) Among them, T(i) represents the travel time between the current collection point and the pass area, which is obtained by dividing the distance between the current collection point and the pass area by the running speed of the unmanned vehicle; D(i) represents the distance between the current collection point and the target point. The distance cost D(n) between the current collection point and the target point can be obtained by avoiding the replanning path caused by the detour of the unmanned vehicle. ij express; The traversability constraint H(i) is the tangent value of the inclination angle between the collection point and the passage area. H(i) should be less than the corrected friction coefficient of the area to ensure the stable operation of the unmanned vehicle: Replanning constraint D(i), the unmanned vehicle compares the distance between the unmanned vehicle and each delivery point j during transportation. When the distance between the initial delivery target point j and the unmanned vehicle is always the smallest, the delivery target does not change and the path length does not change; when there is another delivery point j' whose distance to the unmanned vehicle is smaller than the distance between the initial delivery point j and the unmanned vehicle, a replanning point m is set and the delivery target j is changed to j': Energy consumption constraint E 总 It means that the sum of the energy consumption of the UAV and the unmanned vehicle cannot be greater than the maximum energy of the entire device: AND UAV +E UGV ≤E 总 。

Images