Logistics distribution method and system for truck and drone collaboration in plateau environment

Abstract

The present invention discloses a logistics distribution method and system for truck-UAV collaboration in plateau environment, which belongs to the field of path planning. By introducing a time-related cost map, the wind and temperature parameters that change in real time are mapped to the energy loss of the UAV. The cost map and the dynamic shortest path algorithm are combined to determine the UAV task execution status, and a truck-UAV collaborative distribution model is established on this basis, so that the UAV can perform more reliable task planning in a dynamically changing weather environment with the assistance of the truck. The present invention is suitable for logistics distribution tasks under changing weather conditions in plateaus.

Description

Technical Field

[0001] The present invention relates to the field of path planning technology, and in particular to a logistics distribution method and system for coordinated truck and drone delivery in a plateau environment. Background Art

[0002] In recent years, the rapid development of drone technology has provided new ideas for solving logistics and distribution problems. Drones are fast, flexible, and not restricted by terrain, making them an ideal choice for logistics and distribution. However, drones have limited endurance and load capacity, which limits their application in long-distance and large-scale logistics and distribution.

[0003] To overcome these challenges, some studies have begun to explore a collaborative logistics distribution model between trucks and drones. This model aims to improve the efficiency and reliability of logistics distribution by combining the large load capacity and long-distance transportation capabilities of trucks with the flexibility and rapid response capabilities of drones. However, most existing studies focus on collaborative models between trucks and drones, and these collaborative models still have some limitations in practical applications. For example, existing models often do not consider the constraints of drones when performing tasks in combination with scene characteristics, resulting in low scene adaptability of the model and higher actual delivery costs. Summary of the invention

[0004] The present invention provides a logistics distribution method and system for the coordination of trucks and drones in a plateau environment. In view of the situation where the drone cannot return smoothly or even cannot reach the delivery location due to the unstable weather in the plateau, the sudden changes and changes in environmental conditions are reflected in a time-related cost map, so as to determine the delivery status of the drone, establish a truck-drone logistics distribution model based on the cost map of the plateau operating environment, dynamically adjust the task allocation of trucks and drones, and use trucks to recover drones that cannot return smoothly and complete the unfinished delivery tasks of the drones.

[0005] The embodiment of the present invention provides a logistics distribution method for a plateau environment using a truck and a drone, comprising the following steps:

[0006] Step S1, considering the impact of plateau environmental changes, establish a time-related distribution cost map, and map the real-time plateau environmental change parameters to the energy loss of the drone;

[0007] Step S2, based on the distribution cost graph, establish a UAV mission feasibility analysis model;

[0008] Step S3, solving the UAV mission feasibility analysis model in combination with a dynamic shortest path algorithm to determine the delivery mission status of the UAV under the influence of plateau environmental changes;

[0009] Step S4, according to the delivery task status of the drone, determine the rules for the collaboration between the drone and the truck, and establish a truck-drone collaborative path planning model to generate a logistics delivery path for the collaboration between the truck and the drone.

[0010] Optionally, in one embodiment of the present invention, in step S1, the plateau environmental changes include changes in wind and temperature. Considering the influence of wind and temperature, a time-dependent distribution cost graph is established. In the distribution cost graph, the weights associated with the edges change dynamically over time. The time-dependent distribution cost is defined as: Gt = (V, E; dt), where V = {V0, V1, ..., V n} refers to the point where there is a wind control unit and temperature detection in the distribution area. Assuming that the customer delivery location of the drone is located at the vertex of the graph, V0 represents the warehouse where the drone starts the delivery mission. Any directed edge e=(V i ,V j )∈E, i≠j means the drone is from the customer point V i To V j The path passed by the drone will change the wind and temperature observed during its movement from each customer point. As time changes, dt represents the cost function of the edge, which indicates the energy consumed by traversing the edge e∈E within time t. Based on the relationship between wind and temperature changes and drone energy consumption, an energy consumption model is established to determine the cost calculation formula: is the energy consumption of the drone passing through edge e with length λ(e), μ Tω (e) represents the specific energy consumption related to temperature and wind.

[0011] Optionally, in one embodiment of the present invention, in step S2, the shortest delivery path is found for drone delivery on the distribution cost graph, a drone mission feasibility analysis model is established, and a weighted graph of cost changes over time is used to describe the distribution area. The goal of the drone mission feasibility analysis model is to determine whether the cargo delivery task from the starting point to the destination can be successfully completed within a specific drone energy budget under given conditions. The drone energy budget represents the energy limit when the drone battery is fully charged.

[0012] Optionally, in one embodiment of the present invention, in step S3, a dynamic shortest path algorithm is used to find the shortest delivery path of the drone on the distribution cost graph, and the distance from the current position V to the current position V is calculated in the graph. i To customer point V jThe shortest path is taken as the constraint of energy budget to solve the task execution status of the drone under each delivery path. O means that the drone gave up the task at the beginning because the estimated energy consumption was greater than the battery capacity; F means that after the task started, the drone failed to reach the customer point for some reason; H means that after the task started, the drone could only reach the delivery destination and could not return to the starting point; Q means that after the task started, the drone successfully provided service to the customer and successfully returned to the warehouse, that is, the actual energy consumption did not exceed the battery capacity.

[0013] Optionally, in one embodiment of the present invention, in step S4, the objective function of establishing the truck-UAV collaborative path planning model is:

[0014]

[0015] Among them, AL is the set of candidate truck stops, C1 is the subset of customers served by drones; a ij ′ means that customer point j is served by the alternative drone takeoff truck stop point i, which is 1, otherwise it is 0; b ju ′ indicates that customer point j is served by the alternative drone landing truck stop u, which is 1, otherwise 0; c drone represents the delivery cost per unit distance of the drone; L i,j,u represents the total flight distance that a drone needs to consume when it takes off from truck stop i and delivers to the jth customer point and lands at truck stop u under the condition of drone flight; c i,truck represents the fixed cost of truck delivery to customer point i; a i It means that the candidate point i is selected as the truck stop point, which is 1, otherwise it is 0;

[0016] The model constraints are:

[0017] This means that each customer point can only be served by one truck stop that releases a drone;

[0018] This means that each customer point can only be served by one truck stop that recovers the drone;

[0019] Indicates that a set of truck stops can serve multiple surrounding customer points, and each drone delivery customer point can only be served by a set of truck stops;

[0020] Indicates that the number of truck stops selected for releasing the drone is equal to the number of truck stops selected for subsequent recovery of the drone;

[0021] It means that when the drone's state task is O, the drone's estimated energy consumption is greater than the battery capacity, so the point is delivered by the truck;

[0022] It means that when the drone mission status is F, the drone fails to reach the customer point, so the point is delivered by the truck;

[0023] L i,j,u =d ij σ1+d ij +v drone t drone σ1, which means the actual range of the drone when it delivers the package, taking into account the climbing, landing, and horizontal flight states of the drone carrying the package;

[0024] L i,j,u ≤S(1-ξ), indicating that the drone needs to meet the set energy consumption reserve coefficient;

[0025] b ST(H) =1, indicating that when the drone mission status is H, the drone can only reach the delivery destination and cannot return to the starting point, so the customer point should be set as a truck stop point;

[0026] a ij ′≤a i ,i∈AL,j∈C1, indicating that only when customer point i is selected as the truck stop for drone release, there may be customer point j that needs drone delivery and is delivered by the drone taking off from this point;

[0027] b ju ′≤b u ,u∈AL,j∈C1, which means that only when customer point u is selected as the truck stop for drone recovery, there may be a drone that has served point j to land;

[0028] a ij ′={0,1},i∈AL,j∈C1, which means that the drone is released from the truck stop at point i and flies to the customer point j, which is 1, otherwise 0;

[0029] a i ={0,1},i∈AL, which means that if i is selected as the truck stop for releasing the drone, it is 1, otherwise it is 0;

[0030] b ju ′={0,1},u∈AL,j∈C1, which means that the drone is 1 when it completes the delivery from customer point j and flies to truck stop u for recovery, otherwise it is 0;

[0031] b u={0,1},u∈AL, which means that u is 1 if it is selected as the truck stop for recovering the drone, and 0 otherwise.

[0032] The second embodiment of the present invention provides a logistics distribution system for trucks and drones in a plateau environment, comprising:

[0033] The mapping module is used to consider the impact of plateau environmental changes to establish a time-related delivery cost map, and map the real-time plateau environmental change parameters to the energy loss of the UAV;

[0034] A construction module is used to build a UAV mission feasibility analysis model based on the distribution cost map;

[0035] An analysis module, used to solve the UAV mission feasibility analysis model in combination with a dynamic shortest path algorithm to determine the delivery mission status of the UAV under the influence of plateau environmental changes;

[0036] The distribution module is used to determine the rules for the collaboration between drones and trucks based on the delivery mission status of the drones, and to establish a truck-drone collaborative path planning model to generate a logistics distribution path for the collaboration between trucks and drones.

[0037] The truck-drone collaborative logistics distribution method and system for plateau environments of the embodiments of the present invention address the impact of weather conditions in plateau areas on the efficiency and safety of drone delivery. By introducing a time-related cost graph, the shortest path between drone customer points is found under the influence of wind and temperature to determine the drone mission execution status, and on this basis, a truck-drone collaborative distribution model is established, so that drones can perform more reliable mission planning in a dynamically changing weather environment with the assistance of trucks.

[0038] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0040] Figure 1 A flowchart of a logistics distribution method for a plateau environment using a truck and a drone in collaboration provided according to an embodiment of the present invention;

[0041] Figure 2 A schematic diagram of the execution process of the logistics distribution method using a truck and a drone in a plateau environment;

[0042] Figure 3 is the unit energy consumption under different wind conditions;

[0043] Figure 4 is the time-varying delivery graph Gt;

[0044] Figure 5 For drone-truck intermodal transport diagram;

[0045] Figure 6 It is a schematic diagram of the combination of the destruction method ds1 and the repair method re1;

[0046] Figure 7 It is a schematic diagram of the combination of the destruction method ds2 and the repair method re2;

[0047] Figure 8 It is a schematic diagram of the combination of the destruction method ds3 and the repair method re3;

[0048] Fig. 9 This is the distribution map of major residential areas in Xining City;

[0049] Fig.10 Select and number a map for distribution points;

[0050] Fig.11 is the approximate Voronoi diagram centered on the monitoring station;

[0051] Fig.12 This is a structural block diagram of a logistics distribution system for use in a plateau environment using a collaboration of trucks and drones according to an embodiment of the present invention. DETAILED DESCRIPTION

[0052] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0053] An embodiment of the present invention provides a method for collaborative logistics distribution of trucks and drones in a plateau environment. The method is used in a truck and drone collaborative logistics distribution scenario in a plateau environment. In this scenario, the distribution process is roughly as follows: Consider plateau environmental factors: temperature and wind, establish an energy consumption model, and plan a flexible distribution path in combination with a time-varying delivery graph. Determine the state of the drone through a dynamic task feasibility algorithm to determine whether the drone can successfully complete the distribution task under given conditions and find the optimal path within a limited energy budget. Finally, introduce a truck-drone distribution model to dynamically adjust the task allocation of trucks and drones to achieve collaborative distribution and improve overall efficiency. Specifically, Figure 1 and Figure 2 As shown, the truck and drone coordinated logistics distribution method suitable for plateau environment includes the following steps:

[0054] Step S1, considering the impact of plateau environmental changes, establish a time-related distribution cost map, and map the real-time plateau environmental change parameters to the energy loss of the drone;

[0055] Step S2, based on the distribution cost graph, establish a UAV mission feasibility analysis model;

[0056] Step S3, solving the UAV mission feasibility analysis model in combination with a dynamic shortest path algorithm to determine the delivery mission status of the UAV under the influence of plateau environmental changes;

[0057] Step S4, according to the delivery task status of the drone, determine the rules for the collaboration between the drone and the truck, and establish a truck-drone collaborative path planning model to generate a logistics delivery path for the collaboration between the truck and the drone.

[0058] Optionally, in one embodiment of the present invention, in step S1, the plateau environmental changes include changes in wind and temperature. Considering the influence of wind and temperature, a time-dependent distribution cost graph is established. In the distribution cost graph, the weights associated with the edges change dynamically over time. The time-dependent distribution cost is defined as: Gt = (V, E; dt), where V = {V0, V1, ..., V n} refers to the point where there is a wind control unit and temperature detection in the distribution area. Assuming that the customer delivery location of the drone is located at the vertex of the graph, V0 represents the warehouse where the drone starts the delivery mission. Any directed edge e=(V i ,V j )∈E, i≠j means the drone is from the customer point V i To V j The path passed by the drone will change the wind and temperature observed during its movement from each customer point. As time changes, dt represents the cost function of the edge, which indicates the energy consumed by traversing the edge e∈E within time t. Based on the relationship between wind and temperature changes and drone energy consumption, an energy consumption model is established to determine the cost calculation formula: is the energy consumption of the drone passing through edge e with length λ(e), μ Tω (e) represents the specific energy consumption related to temperature and wind.

[0059] Optionally, in one embodiment of the present invention, in step S2, the shortest delivery path for drone delivery is found on the delivery cost graph, a drone mission feasibility analysis model is established, and a weighted graph of cost variation over time is used to describe the delivery area. The drone mission feasibility analysis model aims to determine the delivery path under given conditions (for each time point t, wind speed ω s and wind direction dand the temperature τ is known) can successfully complete the cargo delivery mission from the starting point to the destination within a specific UAV energy budget, where the UAV energy budget represents the energy limit when the UAV’s battery is fully charged.

[0060] Optionally, in one embodiment of the present invention, in step S3, a dynamic shortest path algorithm is used to find the shortest delivery path of the drone on the distribution cost graph, and the distance from the current position V to the current position V is calculated in the graph. i To customer point V j The shortest path is taken as the constraint of energy budget to solve the task execution status of the drone under each delivery path. O means that the drone gave up the task at the beginning because the estimated energy consumption was greater than the battery capacity; F means that after the task started, the drone failed to reach the customer point for some reason; H means that after the task started, the drone could only reach the delivery destination and could not return to the starting point; Q means that after the task started, the drone successfully provided service to the customer and successfully returned to the warehouse, that is, the actual energy consumption did not exceed the battery capacity.

[0061] Optionally, in one embodiment of the present invention, in step S4, the objective function of establishing the truck-UAV collaborative path planning model is:

[0062]

[0063] Among them, AL is the set of candidate truck stops, C1 is the subset of customers served by drones; a ij ′ means that customer point j is served by the alternative drone takeoff truck stop point i, which is 1, otherwise it is 0; b ju ′ indicates that customer point j is served by the alternative drone landing truck stop u, which is 1, otherwise 0; c drone represents the delivery cost per unit distance of the drone; L i,j,u represents the total flight distance that a drone needs to consume when it takes off from truck stop i and delivers to the jth customer point and lands at truck stop u under the condition of drone flight; c i,truck represents the fixed cost of truck delivery to customer point i; a i It means that the candidate point i is selected as the truck stop point, which is 1, otherwise it is 0;

[0064] The model constraints are:

[0065] This means that each customer point can only be served by one truck stop that releases a drone;

[0066] This means that each customer point can only be served by one truck stop that recovers the drone;

[0067] Indicates that a set of truck stops can serve multiple surrounding customer points, and each drone delivery customer point can only be served by a set of truck stops;

[0068] Indicates that the number of truck stops selected for releasing the drone is equal to the number of truck stops selected for subsequent recovery of the drone;

[0069] It means that when the drone's state task is O, the drone's estimated energy consumption is greater than the battery capacity, so the point is delivered by the truck;

[0070] It means that when the drone mission status is F, the drone fails to reach the customer point, so the point is delivered by the truck;

[0071] L i,j,u =d ij σ1+d ij +v drone t drone σ1, which means the actual range of the drone when it delivers the package, taking into account the climbing, landing, and horizontal flight states of the drone carrying the package;

[0072] L i,j,u ≤S(1-ξ), indicating that the drone needs to meet the set energy consumption reserve coefficient;

[0073] b ST(H) =1, indicating that when the drone mission status is H, the drone can only reach the delivery destination and cannot return to the starting point, so the customer point should be set as a truck stop point;

[0074] a ij ′≤a i ,i∈AL,j∈C1, indicating that only when customer point i is selected as the truck stop for drone release, there may be customer point j that needs drone delivery and is delivered by the drone taking off from this point;

[0075] b ju ′≤b u ,u∈AL,j∈C1, which means that only when customer point u is selected as the truck stop for drone recovery, there may be a drone that has served point j to land;

[0076] a ij ′={0,1},i∈AL,j∈C1, which means that the drone is released from the truck stop at point i and flies to the customer point j, which is 1, otherwise 0;

[0077] a i ={0,1},i∈AL, which means that if i is selected as the truck stop for releasing the drone, it is 1, otherwise it is 0;

[0078] b ju ′={0,1},u∈AL,j∈C1, which means that the drone is 1 when it completes the delivery from customer point j and flies to truck stop u for recovery, otherwise it is 0;

[0079] b u ={0,1},u∈AL, which means that u is 1 if it is selected as the truck stop for recovering the drone, and 0 otherwise.

[0080] The embodiment of the present invention maps the real-time changing wind and temperature parameters to the energy loss of the drone by introducing a time-related cost map. The cost map and the dynamic shortest path algorithm (DSP) are combined to determine the drone task execution status, and a truck-drone collaborative distribution model is established on this basis, so that the drone can perform more reliable task planning in a dynamically changing weather environment with the assistance of the truck. The present invention is suitable for logistics distribution tasks under changing weather conditions on the plateau.

[0081] In this embodiment, the UAV performs logistics delivery services in the plateau area. The unstable weather conditions in the plateau area will affect the endurance of the UAV. Considering the environmental characteristics of the plateau area that have a greater impact on UAV delivery: wind and low temperature, the specific ways in which these environmental characteristics affect UAV logistics delivery are analyzed, and a constraint model is established: (1) Temperature affects the energy consumption of the UAV; (2) The impact of wind on the energy consumption of the UAV.

[0082] 1. Temperature affects drone energy consumption

[0083] The performance of drones will be affected by the ambient temperature during operation. Experimental results show that the battery charging characteristics are almost independent of the charging temperature between 20℃ and 40℃, while the battery charging / discharging performance decreases significantly when the battery temperature is below 20℃. The reduction in battery performance caused by temperature may cause the drone to be unable to return to its launch site or meet service needs. Therefore, it is very important to establish a model of drone endurance and temperature. The performance of lithium-ion batteries is affected by many factors. In this paper, the impact of low temperature on drone endurance in plateau areas is considered. In order to establish a model of drone endurance changing with temperature, the actual battery capacity of lithium-ion batteries at different operating temperatures is used, as shown in Table 1. The functional relationship between the actual capacity of lithium-ion batteries and the ambient temperature is obtained using different fitting methods.

[0084] Table 1 Battery capacity changes with external temperature

[0085]

[0086] The fitted relationship between battery capacity and temperature is:

[0087] δ τ=0.98+6.72×10 -4 τ+3.11×10 -5 τ 2 -8.27×10 -7 τ 3 -20≤τ≤25(1)

[0088] In the formula, τ represents the ambient temperature during the operation of the drone, in °C, and δ τ It indicates the ratio of the actual capacity of the battery to the theoretical capacity when the ambient temperature is τ. Its physical meaning is the ratio of the lithium-ion battery capacity to the standard battery capacity at the current temperature. The horizontal flight speed of the drone is v air , the power during the horizontal flight is P c , within the flight time t, when the temperature is τ, the unit energy consumption of the drone is μ T , expressed as:

[0089]

[0090] When the ambient temperature is τ, the energy E consumed by the drone when crossing edge e T (e) is:

[0091] E T (e) = μ T (e)λ(e)(3).

[0092] 2. Impact of wind on drone energy consumption

[0093] Based on the relative wind force and energy model that measures the energy consumed by the drone when traversing the transportation route, the wind impact on the delivery of packages by logistics drones is estimated. Based on the relative wind direction and speed, the flight energy consumption of the drone under the influence of wind is considered. The effect of wind on the speed and direction of the drone is first considered by calculating the total thrust and total drag. The induced speed is then obtained by estimating the power required for stable flight. Finally, the unit energy consumption is calculated to obtain the unit energy consumption of the drone under the influence of wind (wind speed ω s and wind direction d ) is an energy consumption model when flying along a specified route. Specifically, the model provides a detailed estimate of energy consumption under different wind conditions by analyzing the changes in wind speed and direction of the drone, and estimates the power required for stable flight of the drone under the influence of wind. The calculation formula is:

[0094] P ω =TH[v air sin(θ)+v in ](4)

[0095] Where TH is the total thrust required, i.e., the gravity and drag generated by the forward motion of the drone and the wind, where the drag generated is about v a(e) is a function of a (e) is the speed of the drone relative to the air, which depends on the wind speed ω s and wind direction d .v air is the horizontal flight speed of the drone, θ is the pitch angle, v in It is the induced speed required to give the drone thrust TH.

[0096] In wind ω=(ω s ,ω d ), the unit energy μ of the UAV flying along the route e ω (e) Calculated as power consumption P ω The horizontal flight speed v of the drone air The ratio is as shown in formula (5):

[0097]

[0098] Fixed wind ω=(ω s ,ω d ) under the influence of the UAV passing through the path edge e with a length of λ(e) ω (e) can be expressed as formula (6):

[0099] E ω (e) = μ ω (e)λ(e)(6).

[0100] 3. Consider the influence of wind and temperature at the same time

[0101] The battery capacity will change with temperature. When the temperature is τ, the unit energy consumption of the drone is μ t The relationship between power and temperature is shown in formula (2). When the temperature decreases, the battery capacity becomes smaller, and the energy consumption that can support the drone to fly longer distances is reduced. That is, when the temperature decreases, a certain amount of energy consumption is required to cope with the temperature drop. If the influence of wind is considered on the basis of temperature, the power P in formula (2) can be replaced by the power P under the influence of wind. ω , that is, rewrite (2) into formula (7), so under the influence of temperature and wind, the energy consumption E generated by the drone passing through the edge e with a length of λ(e) is Tω (e) As shown in formula (8).

[0102]

[0103] E Tω (e) = μ Tω (e)λ(e)(8).

[0104] In this embodiment, in order to better characterize the dynamic characteristics of energy consumption of drones on the delivery area map, a relevant cost graph is established according to time. In the cost graph, the weights associated with the edges change dynamically over time. Formally, the time-related cost is defined as Gt = (V, E; dt), where V = {V0, V1, ..., V n} refers to the point where there is a wind control unit and temperature detection in the distribution area. To simplify the model, it is assumed that the customer delivery location of the drone is located at the vertex of the graph. V0 represents the warehouse where the drone starts the delivery mission. Any directed edge e=(V i ,V j )∈E, i≠j means the drone is from the customer point V i To V j The path passed, the observed wind and temperature will change as the drone moves from each customer point. So as time changes, dt represents the cost function of the edge, which represents the energy consumed to traverse the edge e∈E in time t.

[0105] During the delivery process of the drone, V and e do not change, only the wind and temperature change with time. Therefore, the energy consumed by the drone to traverse the edge e∈E in time t is shown in formula (9).

[0106]

[0107] like Figure 3 As shown, for each edge in the graph, the cost at different time periods can be pre-calculated, and the energy consumed by the drone to traverse edge e These costs can be given in the graph Gt, which represents the energy cost of traversing the path starting at time t. Although the path cost can be pre-calculated, in practice, the UAV will refresh the edge cost according to the current time-varying cost graph when traversing the path. The actual cost of the path is accumulated gradually according to the actual number of time slots starting from the time the UAV arrives at a vertex. For any two customer points V i To V j There is a path path starting from time t0 ij , the path has the minimum actual cost between the two customer points. In order to calculate the optimal path, the states in different time periods are represented as a time-dependent cost delivery network graph G. The graph may contain path cost information in different time periods to plan the path selection of the drone and minimize the energy cost.

[0108] In this embodiment, based on the above-mentioned impact of wind and low temperature, it is necessary to consider the energy consumption state of the drone in the face of such unstable weather conditions, as well as the drone mission execution in this state, and conduct a drone mission feasibility analysis. Completing the delivery mission means finding a feasible cycle for the drone, that is, it can be completed autonomously using the available energy of the drone. A model can be established to describe the distribution area with a weighted graph of cost changes over time. The goal is to determine whether, under given conditions, the cargo delivery mission from the starting point to the destination can be successfully completed within a specific drone energy budget.

[0109] 1. Task description and constraints:

[0110] (1) Given a delivery area, the topology (vertices and edges) of the time-varying delivery graph Gt is represented by the graph G = (V, E), where V is a set of vertices and E is a set of edges, representing possible paths. It is assumed that the routes that the drone can travel in the delivery area cannot be changed.

[0111] (2) For each time point t, wind speed ω s and wind direction d is known.

[0112] (3) Budget B, which represents the energy limit of the drone when its battery is fully charged.

[0113] (4) The payload of the drone is m p , the speed is v drone .

[0114] (5) You need to select a node V i As a destination for target customers.

[0115] (6) Wrapped in V i The payload becomes 0 when the drone returns.

[0116] (7) The task start time is t, G t Indicates the state of the cost map at this time.

[0117] 2. Problem solving process:

[0118] The goal of the problem is to determine whether there is a feasible cycle Cir, during which the drone needs to start from the starting point V i Start, after time t>0, arrive at customer point V j , and then from the customer point V j Start, return to V within time t′>t i The actual energy consumption cost of the drone running this feasible cycle Cir is shown in formula (10):

[0119]

[0120] If the cost is less than or equal to the energy budget B, the task is feasible, that is, to find the optimal task period Cir to minimize the energy cost. This problem is related to the shortest path problem.

[0121] Since the edge cost changes over time, this increases the complexity of the problem. i ,V j ) depends on the edge cost (energy required), which depends not only on the fixed edge cost, but also on the relative wind conditions and temperature changes caused by environmental factors in the plateau scene. So this is not a simple weighted graph, but a graph that changes over time, so the shortest path calculated previously may no longer be optimal. Even if all the information in the graph G (including wind speed, temperature changes, and edge costs) is known in advance, starting from the starting point V i To the end point V j The minimum energy cost path is not always a simple straight line. The UAV mission execution states that can be generated are shown in Table 2:

[0122] Table 2 UAV mission execution result status

[0123]

[0124] On this basis, a collaborative delivery model of drones and trucks in plateau areas is established. In this model, each customer must receive the service of a drone or a truck driver once. In particular, strong winds and low temperatures in plateau areas may prevent drones from completing their scheduled tasks or returning to base, which requires trucks to provide delivery assistance. Both drones and trucks need to depart from and return to the same warehouse (distribution center). They can operate independently or collaboratively. In collaborative operations, drones are mounted on trucks to save power. Figure 4 As shown in the figure, F represents the node where the drone mission status is determined to be F. Here, the drone cannot complete the package delivery, so the delivery plan is adjusted and this point is set as the truck node. The black line represents the initial optimized route, while the blue line shows the adjustment of the truck and drone delivery plan after the drone status is determined to ensure the smooth completion of the delivery task. represents the energy consumption of the drone traversing edge e at time t.

[0125] Among them, the main variables and parameters involved in the drone and truck joint delivery model can be listed as follows:

[0126] Table 3 Model variables

[0127]

[0128]

[0129] Based on the previous analysis of the feasibility of drone missions, the involvement of trucks is considered according to the node classification results. When the feasibility of drone missions is not enough to meet the delivery needs, trucks can assist in completing the delivery tasks. For nodes with drone states of O and F, drones cannot complete the specified package delivery tasks, so these nodes are set as truck nodes. When the drone state is H, the drone can complete the package delivery, but cannot return to the recycling point from the customer point, so the point is set as a truck stop. For nodes with drone mission states of Q, drones can complete the delivery task and return to the recycling point. The drones in this case are similar to those considered in previous models. This truck and drone collaborative delivery model is established, and the task allocation strategy between drones and trucks is determined to ensure that the delivery needs are met while fully utilizing resources.

[0130] Considering the power consumption of drones, the phased power consumption model of rotary-wing drones is used. The flight path of drones is divided into three parts: vertical take-off, horizontal straight line, and vertical landing. Each drone has a fixed battery capacity. The endurance of each drone (in time units) depends on its speed and payload. Therefore, the power consumption of drones in different flight phases is different. Among them, the power equation of the drone during horizontal cruising can be expressed by equation (11):

[0131]

[0132] Among them, c1, c2, c4, c5 are model coefficients, and their values ​​are 2.8037 (m / kg), 0.3177 (m / kg), 0.0296 (kg / m) and 0.0279 (Ns / m) respectively. air is the horizontal speed of the drone, g is the gravity constant (9.8m / s 2 ), α is the angle of attack, which is generally assumed to be 10°, the weight of the package carried is w, and the weight of the drone is W.

[0133] Assume that the power of the drone in horizontal flight under no-load condition is P c ′, σ1 is the ratio of the power of the UAV carrying a load when flying horizontally to the power of the UAV flying horizontally without a load, as shown in formula (12). Then the UAV horizontal flight time t under load conditions, corresponding to the standard conditions (UAV without load, operating environment temperature is 25℃) UAV range S′ is shown in (13):

[0134]

[0135] S'=v air tσ1(13)

[0136] In this paper, the objective function is to minimize the distribution cost. The objective function is expressed by formula (14):

[0137]

[0138] The constraints of the model are:

[0139]

[0140]

[0141] L i,j,u =d ij σ1+d ij +v drone t drone σ1(21)

[0142] L i,j,u ≤S(1-ξ)(22)

[0143] b ST(H) =1(23)

[0144] a ij ′≤a i ,i∈AL,j∈C1(24)

[0145] b ju ′≤b u ,u∈AL,j∈C1(25)

[0146] a ij ′={0,1},i∈AL,j∈C1(26)

[0147] a i ={0,1},i∈AL(27)

[0148] b ju ′={0,1},u∈AL,j∈C1(28)

[0149] b u ={0,1},u∈AL(29)

[0150] Formula (15) indicates that each customer point can only be served by one truck stop that releases the drone; Formula (16) indicates that each customer point can only be served by one truck stop that recovers the drone; Formula (17) indicates that a group of truck stops can serve multiple surrounding customer points, and each customer point delivered by the drone can only be served by a group of truck stops; Formula (18) indicates that the number of truck stops selected to release the drone is equal to the number of truck stops that subsequently recover the drone; Formula (19) indicates that when the drone's state task is O, the drone's estimated energy consumption is greater than the battery capacity, so the point is delivered by the truck; Formula (20) indicates that when the drone's mission state is F, the drone fails to reach the customer point, so the point is delivered by the truck; Formula (21) follows Formula (13), indicating that when the drone is delivering packages, the drone takes into account the climbing, landing, and horizontal flight states of the drone carrying the package, and the actual range of the drone is obtained; Formula (22) indicates that the drone needs to meet the set energy consumption pre-storage coefficient; Formula (23) = ... Only when the customer point is selected as the truck stop for drone recovery, there may be a drone that has completed the service at point j and land. Formula (26) indicates that if the drone is released from the truck stop at point i and flies to the customer point j, it is 1, otherwise it is 0. Formula (27) indicates that if i is selected as the truck stop for releasing the drone, it is 1, otherwise it is 0. Formula (28) indicates that if the drone completes the delivery from customer point j and flies to the truck stop u for recovery, it is 1, otherwise it is 0. Formula (29) indicates that if u is selected as the truck stop for recovering the drone, it is 1, otherwise it is 0.

[0151] Considering the influence of wind and temperature in plateau areas, a truck-drone delivery model is established by combining a time-related cost graph and a dynamic shortest path algorithm. The model can dynamically adjust the delivery plan to adapt to sudden environmental changes and ensure the mission status and delivery efficiency of the drone:

[0152] 1. Strong dynamic adaptability: The present invention effectively maps the real-time changing wind and temperature parameters to the energy loss of the drone by establishing a time-related cost map. As shown in the example, through the dynamic shortest path algorithm (DSP) and the adaptive large neighborhood search algorithm, the system can dynamically adjust the flight path of the drone and the delivery plan of the truck according to real-time weather conditions (such as wind speed, wind direction and temperature). This dynamic adaptability enables drones to carry out more reliable mission planning and execution in the complex and changeable weather environment of the plateau area, improving the completion rate and efficiency of the delivery task.

[0153] 2. Optimize resource allocation: By considering the feasibility analysis of drone missions and the truck-drone collaborative delivery model, the present invention can effectively optimize resource allocation. In the example, according to the drone's mission status (O, F, H, Q), the drone and truck tasks are reasonably allocated to ensure that the distribution needs are met while the resources are fully utilized. For example, when the drone cannot complete the delivery task due to bad weather, the truck will intervene to complete the delivery, ensuring the continuity and reliability of the delivery task. This method of optimizing resource allocation not only improves the delivery efficiency, but also ensures the safety of the delivery process.

[0154] 3. Improve delivery efficiency and safety: This invention effectively improves delivery efficiency and safety by introducing a time-related cost graph and a dynamic shortest path algorithm, combined with a collaborative delivery model of drones and trucks. As shown in the example, by dynamically modeling the energy consumption of drones and optimizing delivery plans, drones and trucks can flexibly respond to various delivery challenges under complex and changeable weather conditions in plateau areas. Especially at night when wind and temperature are relatively bad, the system can ensure that drones can safely complete their tasks or be recovered by trucks in a timely manner by adjusting the delivery plan, thereby greatly improving the efficiency and safety of the entire delivery system.

[0155] The following is an introduction through a specific example.

[0156] 1. In the practical application of this embodiment, the feasibility analysis of the UAV mission based on the time-varying cost graph is to find a feasible cycle consisting of two simple paths, one of which leads to the customer point V j , another path returns to the starting point V i . It should be noted that the two paths can intersect at a common node, but the path cannot visit the same node twice in the forward and return process. The purpose of this requirement is to avoid the number of edges in the cycle exceeding 2 (|V|-1), thereby limiting the complexity of the solution. The dynamic shortest path algorithm (DSP) is used to solve the completion of the drone mission, as follows:

[0157] (1) Initialization: Set the initial state CON to undetermined, the time t to 0, the cycle Cir to empty, and the initial position V i V0, the flag variable flag is set to false. The flag variable is used to indicate whether the drone has successfully visited the client point V i At least once, during the initialization phase, the flag is set to false, meaning the drone has not visited any client point yet.

[0158] (2) Main loop: Enter the main loop until the final state (O, F, H, Q) is obtained.

[0159] (3) Calculate the shortest path: Calculate the path from the current position V in each step i To customer point V j The shortest path ij , which is the execution path of the drone in the current time slot.

[0160] (4) Access edge: access path ij And according to formula (10) Update the energy budget B and the current cycle Cir.

[0161] (5) Update status: According to the new cost graph G t Update, including deleting vertex V i and its related edges, while updating the new vertex currently visited.

[0162] (6) Avoiding loops: The algorithm design takes into account the situation where the drone loops endlessly. If the current budget is exhausted during mission execution, the mission is declared a failure.

[0163] (7) Time complexity: Setting the time complexity of the DSP algorithm requires recalculating the new shortest path for each step.

[0164] 2. In the practical application of this embodiment, an adaptive large neighborhood search algorithm is used to solve the drone-truck collaborative delivery model. The algorithm is characterized by multiple constraints and decision variable complexity. Through global search strategies and adaptability, it aims to find the optimal solution under given constraints. The algorithm is applied to the distribution problem described, which involves a complex distribution network of drones and trucks. For the joint delivery problem of drones and trucks, the adaptive large neighborhood search algorithm takes into account multiple constraints, such as the flight distance of drones, load limits, and delivery capacity of trucks. By adaptively selecting different destruction and repair methods, the algorithm is able to search for the optimal solution globally while avoiding falling into local optimality.

[0165] The algorithm reconstructs the solution in the form of a pair of destruction methods and repair methods. Ω- and Ω+ represent the set of destruction and repair methods, respectively. Ω- destroys part of the current solution, while Ω+ reconstructs the destroyed solution. In each iteration, a destruction method ds∈Ω- and a repair method re∈Ω+ are selected to modify the current solution. ALNS assigns a weight to each ds and re method. The ds assignment weight we - , re-assign weights we + , and use roulette wheel for probabilistic selection. First, the weights are initialized with equal probability and iteratively updated according to the reaction factor ρ∈[0,1] and the score ψ. Small reaction factor values ​​ρ may cause the search to be too conservative and converge slowly, while large reaction factor values ​​may cause the search to be too aggressive and may miss potential high-quality solutions.

[0166] we - and we + Represent the weight sets of each ds and re method respectively. - and we + Select the destruction and repair methods, and use the roulette wheel principle to select the method. For the probability pr that the destruction method i is selected i As shown in formula (30):

[0167]

[0168] The weight we is dynamically adjusted according to the score ψ of the method in the iterative process. The score ψ is the evaluation basis of each group of ds and re methods after each iterative search of the algorithm, as shown in Table 4:

[0169] Table 4 Scores of corresponding methods

[0170]

[0171] Set we ij is the weight of method i at the jth iteration. Therefore, after each iteration, the weight of the destruction method we - Update according to formula (31):

[0172]

[0173] In this paper, the same scoring criteria are set for the destruction method and the repair method, β1≥β2≥β3≥β4, β1 is set to 60, β2 is set to 35, β3 is set to 15, and β4 is set to 10.

[0174] 1. Neighborhood Structure

[0175] According to this model, four destruction methods and four repair algorithms are designed, as shown in Table 5:

[0176] Table 5 Method settings

[0177]

[0178] For the above destruction methods and repair methods, the methods are combined to form the following neighborhood:

[0179] (1) Destruction method ds1 and repair method re1

[0180] Remove a point from the current path and reinsert it into a random section. If the path of the drone changes after using this destruction and repair method, it is necessary to check whether the drone path involved in this operation meets the load and endurance constraints, and adjust the solution if it does not meet the constraints. Figure 5 As shown in the figure, if the drone mission status of delivery point 3 is determined to be F, the point is selected and reinserted into the drone section 2-4, the truck path will change and delivery point 3 will become a truck service point. The solid line in the figure represents the truck route and the dotted line represents the drone path. Figure 6 and Figure 7 The same is true.

[0181] (2) Destruction method ds2 and repair method re2

[0182] Remove two service points and swap their positions before reinserting them, involving one point in the drone’s path, to check the drone’s load capacity and endurance constraints. Figure 6 As shown, this is a schematic diagram of the combination of the destruction method ds2 and the repair method re2.

[0183] (3) Destruction method ds3 and repair method re3

[0184] Removing a take-off point (or landing point) from a drone’s path and setting it to a new truck delivery point involves a change in the drone’s path, and requires checking the load and endurance constraints of the drone’s path. Figure 7 As shown, the drone mission status of customer point 6 is determined to be H, the drone cannot return to customer point 6, the landing point 6 in the drone path 5-7-6 is removed, and then inserted at the location of the truck delivery point 8.

[0185] 2. Specific implementation process of adaptive large neighborhood search algorithm

[0186] (1) Initialize the optimal solution cost_best and the current solution cost to the initial solution cost0. In order to improve the diversity and global exploration ability of the search process, ALNS introduces the temperature setting in the simulated annealing algorithm, sets the initial temperature T0, the final temperature T_min, the temperature cooling rate γ, and the maximum number of iterations it_max, and sets the weights of the destruction method and the repair method and Initialize to the same weight, the number of iterations it is initialized to 1, and the current temperature T is initialized to the initial temperature T0;

[0187] (2) Based on method weight and Select the destruction and repair methods ds and re from the neighborhood sets Ω- and Ω+ to generate a new solution cost_new

[0188] (3) Compare the new solution cost_new with the current solution cost. If the new solution is better than the current solution, update the current solution cost to the new solution cost_new. Otherwise, submit the new solution to the Metropolis acceptance criterion for judgment (Metropolis is an acceptance criterion used in simulated annealing algorithms). Then compare the updated current solution cost with the optimal solution cost_new. If the current solution cost is better than the optimal solution cost_new, update the optimal solution cost_new. Otherwise, keep it unchanged.

[0189] (4) Record the scores of each method in each generation. At the end of an iteration, adjust the weights of each method according to formula (31).

[0190] (5) Repeat (2) to (4) until the number of iterations it reaches the maximum number of iterations it_max or the temperature T reaches the minimum temperature T_min, and the global optimal solution cost_new is obtained.

[0191] Specific examples of this embodiment in numerical experiments: The case studied in this paper is the joint delivery of trucks and drones in Xining City. Considering the distribution needs, this paper focuses on the four densely populated urban areas (Chengbei District, Chengdong District, Chengxi District, Chengzhong District), and selects appropriate distribution points based on the locations of industrial parks, commercial centers and residential areas surveyed. Considering the plateau environment, it is necessary to select drones with a higher maximum altitude, longer range and longer endurance. The main model of drone used in this paper is Matrice 300RTK, which has the characteristics of high-altitude flight, strong stability, waterproof and dustproof, and is equipped with a real-time positioning system, which can provide more accurate positioning information. Its parameter data is shown in Table 6:

[0192] Table 6 UAV performance

[0193]

[0194] like Figure 8 As shown in the figure, the larger logistics companies in Xining are mainly concentrated in the center of Xining City (the intersection of four urban areas), so the distribution area of ​​this article is limited to this area of ​​Xining City, and several relatively densely populated residential areas are selected as distribution locations, and the warehouse location is set at Debang Logistics Company. Fig. 9As shown in the figure, the red asterisks represent the distribution points, and the black squares represent the logistics companies. Their coordinate positions are shown in Table 7.

[0195] Table 7 Distribution point locations

[0196]

[0197] Obtain meteorological observation sites. The data source is the site data in the China Meteorological Element Site Observation Daily Dataset of the Resource and Environmental Science and Data Center of the Chinese Academy of Sciences. Download the original file from the website, unzip it and open the shp file through QGIS. Use python to generate csv and json formats, and reverse geocode the longitude and latitude to obtain the csv file. The required information of each site is shown in Table 8.

[0198] Table 8 Location of each testing station

[0199]

[0200] In this embodiment, the detected wind and temperature data are obtained. Taking the selected monitoring station Datong Station as an example, Tables 9 and 10 show the wind and temperature sample data of Datong Station for one day, and the data is updated every 60 minutes. As can be seen from the table, the wind and temperature at night are relatively bad, so the wind and temperature data at 3 am are selected in the subsequent calculation. If you want to view the wind and temperature impact at other times, just change the data information at this time.

[0201] Table 9 Wind data of Datong Station on a certain day in October

[0202]

[0203]

[0204] Table 10 Temperature data of Datong Station on a certain day in October

[0205]

[0206]

[0207] In this embodiment, a VG diagram is constructed to determine the drone mission status. Based on the data in Tables 7 and 8, any point is mapped to the delivery area covered by a specific monitoring station. To solve this problem, a Voronoi diagram based on monitoring stations is used to divide the delivery area. Then, the customer points in the delivery area are mapped to the nearest monitoring station, forming multiple Voronoi cells centered on different monitoring stations, such as Fig.11 shown.

[0208] First, a V-oronoi tessellation method called monitoring station construction is used to divide the delivery area into different parts in detail. Customer points in each delivery area are mapped to the nearest monitoring station, so that multiple Voronio cells are formed, each centered on a different monitoring station. Then it is necessary to plan the routes of the drone between these delivery points. In order to better understand the relationship between the delivery points, a VG graph (V-oronoi-graph) is established. This graph divides the delivery space into multiple areas, each of which belongs to the nearest monitoring station. Such an area is centered on the detection station and forms a polygon consisting of this point and its other nearest points. With this method, the connection between the delivery area and the delivery point can be seen more clearly, which helps to optimize the route planning of the drone, improve the efficiency of the entire delivery process, and enable the drone to deliver quickly and accurately between different areas.

[0209] In this embodiment, in order to plan the drone path, an interconnection strategy is used to connect the delivery points to form a VG (Vertex Graph). V are the vertices of each delivery point and V crown cells (V-oronoi), and the edge set E V will consist of a set of undirected edges. These undirected edges connect each cell's vertex to its central delivery point, as well as all edges that divide the cell, and then split each edge into two edges in opposite directions, making VG directed. To determine the single energy cost function in (32), each edge e is associated with the global wind and temperature observed at time t in the V array where it is located, which is reflected by the energy cost function in the formula. Fig.10 As shown, all edges are affected by at most 2 wind and temperature factors, some of which belong entirely to a single cell, while others span two cells, and their costs are weighted averaged based on the wind and temperature data of the two cells.

[0210] The length of the shortest path to each cell vertex in the VG graph is calculated, as shown in Table 11. The path lengths shown in these graphs are very close to the Euclidean distance between the same delivery points.

[0211] Table 11 The shortest path under VG graph

[0212]

[0213]

[0214] In this embodiment, the truck-drone collaborative delivery model is solved, the temperature and wind information at a certain moment are selected, the energy consumption of the drone is dynamically modeled through the cost map, and the cost map is taken into consideration for the feasibility of the task. If the drone's state is judged to be H or F, the drone's launch position i and customer service point j are both truck nodes. Because if the state is F, this constraint condition is that the truck can recover the drone that can complete the delivery task but cannot return; if the drone's state is judged to be F, the truck can recover the drone that has set off but cannot reach the delivery point due to uncertain factors (sudden strong wind, rapid temperature drop) and complete the delivery task of the customer point. If the drone's state is O, the drone's launch position i and customer service point j are both truck nodes, and the drone does not launch. If the drone's state is judged to be Q, the drone recovery point u is a point in the point set that is different from the drone's launch point i and customer service point j, and this point is also the node where the truck stops.

[0215] The model is solved using the adaptive large neighborhood search algorithm. The mission status of the drone is not considered for the time being. That is, the drone can serve any customer point within the energy limit, and then the truck and the drone jointly complete the delivery task. The initialized delivery plan is shown in Table 12, including truck stops, drone delivery points, etc.

[0216] Table 12 Initialized delivery plan

[0217]

[0218] According to the established model, the model is solved by considering the constraints, and the final optimized distribution plan is shown in Table 13:

[0219] Table 13 Optimized distribution plan

[0220]

[0221] It can be seen from Table 13 that the task allocation of drones and trucks needs to be adjusted because the impact of temperature and wind on the drone mission status is taken into account. Since the impact of temperature and wind may cause the drone status to deteriorate, the optimized delivery results have relatively more truck stops. Because if the drone status is F, H, or O, the truck must not only recover the drone, but also consider the delivery service that the drone has not completed. For example, the drone mission status at point 4 is H. The drone can complete the delivery task at point 4, but it cannot return. Point 4 needs to be set as a truck stop to recover the drone. This solution can ensure the integrity of the entire delivery task. If the delivery plan in Table 10 is followed, the delivery task of some customer points cannot be completed or the drone cannot be recovered.

[0222] Based on the challenges of logistics and distribution in plateau areas, this embodiment proposes a new truck-drone collaborative logistics and distribution model to address the impact of weather conditions in plateau areas on the efficiency and safety of drone delivery, aiming to solve the logistics and distribution problems under the ever-changing weather conditions in plateau areas. Previous truck-drone collaborative models mostly only set energy limits, but did not consider whether the changing weather conditions would affect the completion of drone delivery tasks.

[0223] This embodiment takes into account the influence of wind and temperature in the plateau area and constructs a truck-drone distribution model that combines a cost map and a dynamic shortest path algorithm. The model can dynamically adjust the distribution plan to adapt to sudden environmental changes and ensure the mission status and distribution efficiency of the drone. The impact of real-time wind and temperature changes on the drone is reflected in the time-related cost map, and the drone mission execution status of each distribution point is determined by combining the cost map and the dynamic shortest path algorithm. On this basis, the distribution rules are set to establish a joint distribution model of trucks and drones. If the mission status of the drone is not considered for the time being, that is, the influence of the changing wind and temperature is ignored, an initial truck and drone distribution plan is obtained. Then consider the drone mission execution status, according to the established model, consider the constraints to solve the model, and finally obtain the optimized distribution plan. Since the influence of temperature and wind on the drone mission status is taken into account, the task allocation of drones and trucks is readjusted. In order to ensure the completion rate of the distribution task, the truck stops in the optimized distribution result should be relatively more. The model established in this article enables drones to perform more reliable mission planning in a dynamically changing weather environment with the assistance of trucks.

[0224] Next, a logistics distribution system for plateau environments using the collaboration of trucks and drones proposed in accordance with an embodiment of the present application will be described with reference to the accompanying drawings.

[0225] Fig.12 This is a structural block diagram of a logistics distribution system for use in a plateau environment using a collaboration of trucks and drones according to an embodiment of the present invention.

[0226] like Figure 2 As shown, the truck and drone collaborative logistics distribution system 10 for plateau environments includes: a mapping module 100, a construction module 200, an analysis module 300 and a distribution module 400.

[0227] The mapping module 100 is used to establish a time-related distribution cost map by considering the impact of plateau environmental changes, and to map real-time plateau environmental change parameters to the energy loss of the drone.

[0228] The construction module 200 is used to establish a drone mission feasibility analysis model based on the distribution cost map.

[0229] The analysis module 300 is used to solve the UAV mission feasibility analysis model in combination with a dynamic shortest path algorithm to determine the delivery mission status of the UAV under the influence of plateau environmental changes.

[0230] The distribution module 400 is used to determine the rules for the collaboration between drones and trucks based on the delivery mission status of the drones, and to establish a truck-drone collaborative path planning model to generate a logistics distribution path for the collaboration between trucks and drones.

[0231] It should be noted that the above explanation of the embodiment of the logistics distribution method for the cooperation of trucks and drones in plateau environments is also applicable to the logistics distribution system for the cooperation of trucks and drones in plateau environments of this embodiment, and will not be repeated here.

[0232] According to the truck and drone collaborative logistics distribution system for plateau environments proposed in the embodiment of the present application, in view of the impact of weather conditions in plateau areas on the efficiency and safety of drone delivery, a time-related cost graph is introduced to find the shortest path between the drone customer points under the influence of wind and temperature to determine the drone mission execution status, and on this basis, a truck-drone collaborative delivery model is established, so that drones can perform more reliable mission planning in a dynamically changing weather environment with the assistance of trucks.

[0233] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0234] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "N" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0235] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present invention belong.

Claims

1. A logistics distribution method using trucks and drones in a plateau environment, characterized in that: The following steps are involved: Step S1, considering the influence of plateau environmental changes, establish a time-related distribution cost map, and map the real-time plateau environmental change parameters to the energy loss of the drone; plateau environmental changes include changes in wind and temperature, and considering the influence of wind and temperature, establish a time-related distribution cost map. In the distribution cost map, the weights associated with the edges change dynamically over time. The time-related distribution cost is defined as: Gt = (V, E; dt), where V = {V0, V1, ..., V n } refers to the point where there is a wind control unit and temperature detection in the distribution area. Assuming that the customer delivery location of the drone is located at the vertex of the graph, V0 represents the warehouse where the drone starts the delivery mission. Any directed edge e=(V i , V j )∈E, i≠j means the drone is from the customer point V i To V j The path passed by the drone will change the wind and temperature observed during its movement from each customer point. As time changes, dt represents the cost function of the edge, which indicates the energy consumed by traversing the edge e∈E within time t. Based on the relationship between wind and temperature changes and drone energy consumption, an energy consumption model is established to determine the cost calculation formula: is the energy consumption of the drone passing through edge e with length λ(e), μ Tω (e) represents the specific energy consumption related to temperature and wind; Step S2, based on the distribution cost graph, establish a UAV mission feasibility analysis model; Step S3, solving the UAV mission feasibility analysis model in combination with a dynamic shortest path algorithm to determine the delivery mission status of the UAV under the influence of plateau environmental changes; Step S4, according to the delivery task status of the drone, determine the rules for the collaboration between the drone and the truck, and establish a truck-drone collaborative path planning model to generate a logistics delivery path for the collaboration between the truck and the drone; In step S2, the shortest delivery path for drone delivery is found on the delivery cost graph, and a drone mission feasibility analysis model is established. A weighted graph of cost changes over time is used to describe the delivery area. The goal of the drone mission feasibility analysis model is to determine whether the cargo delivery task from the starting point to the destination can be successfully completed within a specific drone energy budget under given conditions. The drone energy budget represents the energy limit of the drone when the battery is fully charged. In step S3, the dynamic shortest path algorithm is used to find the shortest delivery path of the drone on the distribution cost graph. The distance from the current position V to the current position V is calculated in the graph. i To customer point V j The shortest path of the delivery path is solved with the energy budget as the constraint. O means that the drone gave up the task at the beginning because the estimated energy consumption was greater than the battery capacity; F means that after the task started, the drone failed to reach the customer point for some reason; H means that after the task started, the drone could only reach the delivery destination and could not return to the starting point; Q means that after the task started, the drone successfully provided service to the customer and successfully returned to the warehouse, that is, the actual energy consumption did not exceed the battery capacity; In step S4, the rules for the cooperation between the drone and the truck are: and Among them, a ij ′ is 1 if customer point j is served by the alternative drone takeoff truck stop point i, otherwise 0, b ju ′ is 1 if customer point j is served by the alternative UAV landing truck stop u, otherwise it is 0. ST(O) and ST(F) are nodes with UAV states O and F respectively. The first formula indicates that when the UAV state task is O, the estimated energy consumption of the UAV is greater than the battery capacity, and the point is delivered by the truck; the second formula indicates that when the UAV mission state is F, the UAV fails to reach the customer point, and the point is delivered by the truck.

2. The method according to claim 1, characterized in that In step S4, the objective function of establishing the truck-UAV collaborative path planning model is: Among them, AL is the set of candidate truck stops, C1 is the subset of customers served by the drone; c drone represents the delivery cost per unit distance of the drone; L i,j,u represents the total flight distance that a drone needs to consume when it takes off from truck stop i and delivers to the jth customer point and lands at truck stop u under the condition of drone flight; c i,truck represents the fixed cost of truck delivery to customer point i; a i It means that the candidate point i is selected as the truck stop point, which is 1, otherwise it is 0; The model constraints are: This means that each customer point can only be served by one truck stop that releases a drone; This means that each customer point can only be served by one truck stop that recovers the drone; Indicates that a set of truck stops can serve multiple surrounding customer points, and each drone delivery customer point can only be served by a set of truck stops; Indicates that the number of truck stops selected for releasing the drone is equal to the number of truck stops selected for subsequent recovery of the drone; It means that when the drone's state task is O, the drone's estimated energy consumption is greater than the battery capacity, so the point is delivered by the truck; It means that when the drone mission status is F, the drone fails to reach the customer point, so the point is delivered by the truck; L i,j,u =d ij σ1+d ij +v drone t drone σ1, which means the actual range of the drone when it delivers the package, taking into account the climbing, landing, and horizontal flight states of the drone carrying the package; L i,j,u ≤S(1-ξ), indicating that the drone needs to meet the set energy consumption reserve coefficient; b ST(H) =1, indicating that when the drone mission status is H, the drone can only reach the delivery destination and cannot return to the starting point, so the customer point should be set as a truck stop point; a ij ′≤a i ,i∈AL,j∈C1, indicating that only when customer point i is selected as the truck stop for drone release, there may be customer point j that needs drone delivery and is delivered by the drone taking off from this point; b ju ′≤b u ,u∈AL,j∈C1, which means that only when customer point u is selected as the truck stop for drone recovery, there may be a drone that has served point j to land; a ij ′={0,1},i∈AL,j∈C1, which means that the drone is released from the truck stop at point i and flies to the customer point j, which is 1, otherwise 0; a i ={0,1},i∈AL, which means that if i is selected as the truck stop for releasing the drone, it is 1, otherwise it is 0; b ju ′={0,1},u∈AL,j∈C1, which means that the drone is 1 when it completes the delivery from customer point j and flies to truck stop u for recovery, otherwise it is 0; b u ={0,1},u∈AL, which means that u is 1 if it is selected as the truck stop for recovering the drone, and 0 otherwise.

3. A logistics distribution system for trucks and drones in plateau environments, characterized in that: include: The mapping module is used to establish a time-related distribution cost map considering the impact of plateau environmental changes, and map the real-time plateau environmental change parameters to the energy loss of the drone; the plateau environmental changes include changes in wind and temperature, and the time-related distribution cost map is established considering the impact of wind and temperature. In the distribution cost map, the weights associated with the edges change dynamically over time, and the time-related distribution cost is defined as: Gt = (V, E; dt), where V = {V0, V1, ..., V n } refers to the point where there is a wind control unit and temperature detection in the distribution area. Assuming that the customer delivery location of the drone is located at the vertex of the graph, V0 represents the warehouse where the drone starts the delivery mission. Any directed edge e=(V i , V j )∈E, i≠j means the drone is from the customer point V i To V j The path passed by the drone will change the wind and temperature observed during its movement from each customer point. As time changes, dt represents the cost function of the edge, which indicates the energy consumed by traversing the edge e∈E within time t. Based on the relationship between wind and temperature changes and drone energy consumption, an energy consumption model is established to determine the cost calculation formula: is the energy consumption of the drone passing through edge e with length λ(e), μ Tω (e) represents the specific energy consumption related to temperature and wind; A construction module is used to build a UAV mission feasibility analysis model based on the distribution cost map; An analysis module, used to solve the UAV mission feasibility analysis model in combination with a dynamic shortest path algorithm to determine the delivery mission status of the UAV under the influence of plateau environmental changes; The distribution module is used to determine the rules for the collaboration between drones and trucks based on the delivery mission status of drones, establish a truck-drone collaborative path planning model, and generate a logistics distribution path for the collaboration between trucks and drones; In the building module, the shortest delivery path for drone delivery is found on the delivery cost graph, and a drone mission feasibility analysis model is established. A weighted graph of cost changes over time is used to describe the delivery area. The goal of the drone mission feasibility analysis model is to determine whether the cargo delivery task from the starting point to the destination can be successfully completed within a specific drone energy budget under given conditions. The drone energy budget represents the energy limit of the drone when the battery is fully charged. In the analysis module, a dynamic shortest path algorithm is used to find the shortest delivery path of the drone on the distribution cost graph, and the distance from the current position V to the current position V is calculated in the graph. i To customer point V j The shortest path of the delivery path is solved with the energy budget as the constraint. O means that the drone gave up the task at the beginning because the estimated energy consumption was greater than the battery capacity; F means that after the task started, the drone failed to reach the customer point for some reason; H means that after the task started, the drone could only reach the delivery destination and could not return to the starting point; Q means that after the task started, the drone successfully provided service to the customer and successfully returned to the warehouse, that is, the actual energy consumption did not exceed the battery capacity; In the delivery module, the rules for the cooperation between drones and trucks are: and Among them, a ij ′ is 1 if customer point j is served by the alternative drone takeoff truck stop point i, otherwise 0, b ju ′ is 1 if customer point j is served by the alternative UAV landing truck stop u, otherwise it is 0. ST(O) and ST(F) are nodes with UAV states O and F respectively. The first formula indicates that when the UAV state task is O, the estimated energy consumption of the UAV is greater than the battery capacity, and the point is delivered by the truck; the second formula indicates that when the UAV mission state is F, the UAV fails to reach the customer point, and the point is delivered by the truck.

Images