An efficient delivery of goods accompanying flight drone path planning system and method

Abstract

This invention discloses an efficient path planning system and method for accompanying unmanned aerial vehicles (UAVs) in cargo delivery. The method includes: Step 1, constructing a cargo delivery path planning model under the condition of UAVs accompanying transport vehicles; Step 2, initializing environmental parameters and performing preliminary delivery path planning to determine the current path planning scheme; Step 3, according to the current path planning scheme, sequentially removing transport vehicle delivery points and calculating the revenue caused by removing each delivery point; Step 4, performing an insertion operation and calculating the cost caused by the insertion to obtain the latest path planning scheme; Step 5, determining whether all transport vehicle delivery points have been removed. If so, proceed to Step 6; otherwise, proceed to Step 3; Step 6, judging the latest path planning scheme. If the stopping condition is met, stop iterating and complete the path planning; otherwise, iterate again and execute Step 3. The system is used to implement the above method.

Description

Technical Field

[0001] This invention relates to a path planning system and method for delivering goods, and more particularly to a path planning system and method for an efficient unmanned aerial vehicle (UAV) accompanying a delivery vehicle. Background Technology

[0002] In recent years, logistics companies have been researching efficient and low-cost delivery methods based on logical networks. Traditional delivery methods optimize solutions to the Traveling Salesman Problem to achieve high efficiency. With the emergence of unmanned technologies, drone delivery, in the low-altitude economy, is becoming a core innovative force in modern logistics networks due to its efficiency, flexibility, and low carbon footprint. Deploying drones for parcel delivery reduces manpower burden, ignores traffic congestion, and offers fast and convenient transportation. However, drones have limited endurance, making it impossible to transport parcels over long distances for extended periods. Combining traditional vehicle delivery with drone delivery, a delivery method where a vehicle accompanies a drone is proposed. The vehicle transports the drone and goods together to the vicinity of the delivery point, where the drone independently completes the delivery task and rejoins the vehicle at the next delivery location for retrieval. This method significantly increases the delivery service radius of the drone, enhancing convenience and flexibility. Simultaneously, the vehicle provides a charging interface for the drone, improving its endurance. When a drone experiences a technical malfunction, the vehicle driver can promptly arrive on-site to resolve the issue.

[0003] like Figure 1 As shown, boxes indicate delivery points unsuitable for drone delivery, while circles indicate delivery points suitable for drone delivery. Figure 1 In the traditional delivery method, the transport vehicles deliver goods one by one according to the order provided by the traveling agent's optimized solution. Figure 2 This is a parallel drone delivery method, where drones and transport vehicles operate independently and in parallel. Due to flight radius constraints, only delivery points 1 and 8 are within the drone delivery range, and delivery point 8 is not suitable for drone delivery. Figure 3 For delivery via accompanying drones, delivery points 2, 3, and 5 are selected as drone delivery points. The arrival times of the drones and transport vehicles at each delivery point under each transportation method are as follows: Figure 4 As shown, the traditional delivery method using transport vehicles takes the longest time, while the delivery method using accompanying drones takes the shortest time. However, this method still lacks an efficient and reliable path planning method for accompanying drones. Summary of the Invention

[0004] Purpose of the invention: The technical problem to be solved by the present invention is to provide an efficient path planning system and method for accompanying unmanned aerial vehicles (UAVs) for cargo delivery, addressing the shortcomings of existing technologies.

[0005] To address the aforementioned technical problems, this invention discloses a path planning system and method for an efficient cargo delivery drone, wherein the system includes:

[0006] The parameter initialization module, revenue calculation module, cost calculation module, and design variable update module obtain the accompanying flight drone path through iterative calculation based on the geographical location information of each delivery point.

[0007] The parameter initialization module is used to construct a cargo delivery route planning model under the condition of drone-accompanied transport vehicle, initialize the environmental parameters of the cargo delivery route planning model, and perform preliminary delivery route planning.

[0008] The revenue calculation module is used to calculate the change in delivery efficiency revenue caused by removing each delivery point in the transport vehicle route;

[0009] The cost calculation module is used to calculate the change in delivery efficiency cost caused by adding delivery nodes of transport vehicles at different locations in each sub-path;

[0010] The design variable update module is used to determine whether an optimized path has been obtained or to continue iteration.

[0011] This invention also proposes a method for efficient path planning of accompanying drones for cargo delivery, comprising:

[0012] Step 1: Construct a cargo delivery route planning model under the condition of drone-accompanied transport vehicle;

[0013] Step 2: Initialize the environmental parameters of the goods delivery route planning model and perform preliminary delivery route planning to determine the current route planning scheme;

[0014] Step 3: Based on the current route planning scheme, remove the delivery points of the transport vehicles in sequence and calculate the revenue generated by removing the delivery point;

[0015] Step 4: Based on the delivery points removed in Step 3, perform the insertion operation and calculate the cost caused by the insertion to obtain the latest route planning scheme.

[0016] Step 5: Determine whether all delivery points of the transport vehicles have been removed. If yes, proceed to step 6; otherwise, proceed to step 3.

[0017] Step 6: Judge the latest path planning scheme. If the stopping condition is met, stop the iteration and complete the path planning. Otherwise, take the path planning scheme as the current path planning scheme and repeat step 3.

[0018] Beneficial effects:

[0019] (1) This invention combines traditional vehicle delivery with drone delivery. The optimized path planned extends the delivery radius of the drone, improves delivery efficiency, and reduces manpower burden.

[0020] (2) The accompanying flight drone path planning method designed in this invention can automatically generate path information based on delivery point data, providing transport vehicle drivers with an efficient delivery solution.

[0021] (3) The computational method proposed in this invention is for large-scale problem solving and does not require the calculation of all feasible solutions. It is a simple but efficient heuristic solution method. Attached Figure Description

[0022] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments, and the advantages of the present invention in the above and / or other aspects will become clearer.

[0023] Figure 1 This is a schematic diagram of the route for traditional delivery methods using transport vehicles.

[0024] Figure 2 This is a schematic diagram of the route for parallel drone delivery.

[0025] Figure 3 This is a schematic diagram of the delivery route using drones that accompany the drones.

[0026] Figure 4 This is a diagram showing the time taken for various delivery methods.

[0027] Figure 5 This is a schematic diagram of the path planning process for an accompanying drone.

[0028] Figure 6 This is a schematic diagram of the Traveling Salesman's Path results.

[0029] Figure 7 This is a diagram illustrating the update of design variables.

[0030] Figure 8 A schematic diagram showing the optimized path results for accompanying drones. Detailed Implementation

[0031] This invention addresses the major trend of "low-altitude economy" in cities, and the complex coupling relationship between customer groups, transport vehicles, and drones, which makes it difficult to solve the optimal delivery solution by traversing all variables. It proposes an efficient path planning system and method for accompanying drones for goods delivery. The system includes a parameter initialization module, a revenue calculation module, a cost calculation module, and a design variable update module. Based on the geographical location information of each delivery point, the optimal path of the accompanying drone is obtained after multiple iterations of calculation and processing by each module.

[0032] The parameter initialization module is used to solve the traveling salesman problem and supports the initialization of transport vehicle routes and design variables.

[0033] The revenue calculation module is used to calculate the change in delivery efficiency revenue when each delivery point is removed from the transport vehicle's route;

[0034] The cost calculation module is used to calculate the change in delivery efficiency cost caused by adding delivery nodes of transport vehicles at different locations in each sub-path;

[0035] The design variable update module is used to determine whether an optimized path has been obtained or to continue iteration.

[0036] The accompanying drone path planning method proposed in this invention has the following general idea: The drone, mounted on a transport vehicle, takes off and delivers goods at a suitable location, then meets up with the transport vehicle at another delivery point, returning to the transport vehicle to recharge. The transport vehicle continues its delivery service, taking the drone to the next suitable location, where it takes off again, delivers goods, and returns to the transport vehicle. This cycle forms the optimal delivery path. This invention is suitable for scenarios where most delivery points are outside the drone's range, and the drone's takeoff from the transport vehicle allows it to deliver to farther locations.

[0037] For example, the drone delivery point starts at 0. After removing delivery points from the delivery route, the removed points become the starting drone delivery points, forming the drone delivery path with any two adjacent points on the route. Figure 6 After deleting point 3, it forms a drone delivery route with points 4 and 7, becoming... Figure 7 .

[0038] Based on the above approach, this invention requires the following assumptions and definitions to construct a path planning model for an accompanying unmanned aerial vehicle (UAV), as detailed below:

[0039] The accompanying drone path planning model assumes the following: ① A drone can only complete one delivery per flight, while a transport vehicle can complete multiple deliveries; ② During delivery, the drone is assumed to fly at a constant speed, and must arrive at the rendezvous point before the transport vehicle, hovering in place; ③ The drone can take off from the rendezvous point but cannot return to the takeoff point; ④ The drone's return point must be the transport vehicle's delivery point; the drone cannot intersect with the transport vehicle before the rendezvous point, and the transport vehicle cannot repeatedly reach the delivery point to retrieve the drone; ⑤ Neither the drone nor the transport vehicle can stay at a non-delivery point, nor can they repeatedly arrive at the delivery point; ⑥ If the drone returns to the logistics center, the accompanying flight process ends. ⑥ When the weight of the delivered goods exceeds the drone's effective carrying capacity, the delivery requires a signature, or the delivery address cannot guarantee a smooth landing for the drone, the delivery point cannot be delivered by drone.

[0040] Symbol definition: G = {1, 2, ... g} is defined as all delivery points. Let G be a subset of the set of delivery points that meet the requirements for drone services, where the starting and ending points in logistics delivery are the same physical location, but are identified as 0 and g+1 respectively during the calculation process. Representing all network nodes, now defined Represents the set of starting points. This represents the set of points to reach. Once all nodes are determined, the travel time for the transport vehicle and the drone between these nodes can be calculated; therefore, a matrix is ​​defined. and The delivery time matrices are for transport vehicles and drones, respectively. The matrix defines... The time taken for the transport vehicle to travel from the i-th delivery point to the j-th delivery point. Let $T$ be the time it takes for the drone to travel from delivery point $i$ to delivery point $j$, where $T$ is the time taken to travel from delivery point $i$ to delivery point $j$. , , This means that the starting point and the ending point are the same; the time for the transport vehicle to arrive at the j-th delivery point is defined as... The time when the drone arrives at delivery point j is The transport vehicle needs a certain amount of time to prepare for launching and recovering the drone, with parameters defined as D. L D R The drone's endurance is defined by parameter e. Furthermore, a triplet is defined.<i,j,k> This is a representation of the drone delivery route, where i is the launch point (return point is prohibited due to battery life limitations); j is the delivery point. Due to logical consistency, selecting the launch point is prohibited; k is the recovery point, k k can be selected as the delivery point or the logistics destination, but ensure that the process from i->j->k does not exceed the drone's flight time.

[0041] Define transport vehicle route design variables 1 represents the transport vehicle traveling from the i-th delivery point to the j-th delivery point, where ,and Define the drone route design variables. 1 represents a drone carrying a package launching from delivery point i, arriving at delivery point j for delivery, and then flying to delivery point k to meet up with the transport vehicle. .

[0042] For the route planning problem of accompanying unmanned aerial vehicles (UAVs), the characterization equation is as follows:

[0043] (1)

[0044] (2)

[0045] (3)

[0046] (4)

[0047] (5)

[0048] (6)

[0049] (7)

[0050] (8)

[0051] (9)

[0052] (10)

[0053] (11)

[0054] (12)

[0055] (13)

[0056] (14)

[0057] (15)

[0058] (16)

[0059] (17)

[0060] (18)

[0061] (19)

[0062] (20)

[0063] (twenty one)

[0064] (twenty two)

[0065] Equation (1) is the objective function. The purpose of solving this problem is to minimize the arrival time at the destination. Min represents minimizing... Let g be the time when the transport vehicle arrives at g+1, which is the time it returns to the warehouse. Equation (2) is a constraint that ensures that any delivery point is delivered exactly once; Equation (3) means that there is one and only one delivery point starting from the starting point (0, i.e., the warehouse); Equation (4) means that there is one and only one delivery point to reach the destination (g+1, i.e., the warehouse). Equation (5) guarantees the continuity of the transport vehicle, that is, for any delivery point j, the transport vehicle arrives at delivery point j and also starts from delivery point j. Equation (6) means that the UAV can be launched from any delivery point i and at most once; Equation (7) means that the UAV can return to the transport vehicle from any delivery point k and at most once. In Equation (8), if the UAV is launched from delivery point i and retrieved at delivery point k for any ijk, the transport vehicle must pass through delivery point i and delivery point k; moreover, Equation (9) ensures that if the UAV is launched from the starting point (warehouse) and retrieved at delivery point k, the transport vehicle must pass through delivery point k.

[0066] Equations (10) and (11) ensure that when the drone is launched from delivery point i, the transport vehicle and the drone arrive at delivery point i almost simultaneously. Equations (12) and (13) ensure that when the drone returns from delivery point k, the transport vehicle and the drone arrive at delivery point k almost simultaneously. Equations (10) to (13) ensure that if the drone launches from delivery point i... Once launched, it cannot be retrieved at delivery point i. W is a sufficiently large value, not less than the maximum value required for the drone and transport vehicle to return to the destination.

[0067] Equation (14) assumes that the transport vehicle arrives at the delivery point k from the delivery point h, and the time it takes for the transport vehicle to arrive at the delivery point k is... The time it takes for the transport vehicle to arrive at the delivery point h must be considered. The time taken to travel from delivery point h to delivery point k In addition, the launch and reception of the drone must be considered. If the drone is launched from delivery point k, the preparation time required for launch is... Alternatively, if the drone is retrieved at delivery point k, the preparation time required for retrieval is... .

[0068] Equation (15) represents the time it takes for a drone to arrive at delivery point j if it is launched from delivery point i. The takeoff time of the drone needs to be considered. and flight time Similarly, equation (16) states that if the time required for the drone to be retrieved at delivery point k is... The delivery time of the drone needs to be considered. and flight time ,parameter The transport vehicle needs to arrive at the rendezvous point in advance and be ready to retrieve the drone.

[0069] The drone's endurance is reflected in equation (17). The second item represents the time it takes for the drone to arrive at delivery point k. e represents the time when the drone departs from delivery point i, and e represents the drone's flight time.

[0070] Assuming the drone launches from delivery point i, returns to delivery point k, and then launches from delivery point l, equation (18) prevents launch time at delivery point l. Exceeding the recycling time at delivery point k Equations (19)-(22) specify the range of values ​​for the variables.

[0071] Furthermore, such as Figure 5 As shown, the path planning method for the accompanying UAV is as follows:

[0072] Step 1: Parameter initialization. Input the geographical location information of the delivery points. Using only delivery vehicles, perform preliminary route planning. Based on the preliminary route planning scheme, initialize the arrival time array for each delivery point and the node connection relationship array A{. } and B{ The initial assignment is performed as follows:

[0073] Step 1-1: Enter the geographical location information of the delivery point;

[0074] Steps 1-2 include calculating the time consumption matrices T and T̀ between the transport vehicle and the drone at each delivery point, and setting the delivery point set G, etc.

[0075] Steps 1-3 involve constructing a delivery problem based on the information G of each delivery point and solving it using the simulated annealing algorithm to obtain the initial delivery route for the delivery vehicle (as shown in row 3 of Table 1), as detailed below:

[0076] Step 1-3-1: Initialize the route planning parameters from the following three aspects: a) Initial solution: Randomly generate a feasible solution, usually the initial delivery route; b) Initial temperature: Set a relatively high initial temperature to control the acceptance probability during the search process; c) Cooling rate: Set a cooling rate to determine the speed at which the temperature decreases.

[0077] Step 1-3-2, Iterative process, generating neighbor solutions: Randomly generate neighbor solutions based on the current solution. This can be achieved by swapping the locations of two delivery points or changing the order of visits;

[0078] Step 1-3-3, Evaluate the quality of the solution: Calculate the total delivery time of the current solution and its neighboring solutions;

[0079] Steps 1-3-4, Acceptance Strategy: If the neighbor's solution is better (i.e., shorter delivery time), then accept the neighbor's solution. If the neighbor's solution is not better, then accept the neighbor's solution with a certain probability P:

[0080]

[0081] It is the difference between the neighboring solutions and the current solution, where T is the current temperature;

[0082] Steps 1-3-5, Cooling: Reduce the temperature according to the cooling rate:

[0083]

[0084] Where, 0 < <1 is the cooling coefficient. This is the temperature after cooling down;

[0085] Steps 1-3-6: Termination conditions: The maximum number of iterations is reached, the temperature drops to a certain threshold, or the quality of the solution no longer improves. Then, the output result is the optimal delivery route and the corresponding delivery time.

[0086] Steps 1-4 involve creating arrays of arrival times for each delivery point and arrays of node connections A{ } and B{ Perform the initial assignment.

[0087] Step 2: Based on the results of the initial or iterative planning, determine the current path planning scheme, i.e.:

[0088] If it is the first iteration, the path planning scheme in step 1 is selected as the current path planning scheme; otherwise, the iterative planning result is used as the current path planning scheme.

[0089] Step 3: Based on the current route planning scheme, remove delivery points for transport vehicles and calculate the revenue generated by removing each point. In the current route planning scheme, select delivery points sequentially according to their order, perform the removal operation, and calculate the revenue generated by removing each point, as detailed below:

[0090] Step 3-1, start the iteration (each variable in the iteration process must meet the constraints in the path planning model, i.e., equation (2)-equation (22)), and select nodes in sequence from all the delivery points of the transport vehicles (row 7 of Table 1) to prepare for removal;

[0091] Step 3-2: Retrieve the locations before and after the delivery point, and calculate the efficiency difference (row 8 of Table 1) savings before and after removal, as detailed below:

[0092] Step 3-2-1, retrieve the delivery point Information on the preceding and following delivery points i and k along the transport vehicle's route (Table 2, row 1);

[0093] Step 3-2-2: Calculate the savings from removing the node based on the information of the preceding and following delivery points, i.e., the change in delivery time of the transport vehicle (Table 2, row 2), as shown in the following formula; if the delivery sub-route of the transport vehicle containing the delivery point is accompanied by a drone (Table 2, row 3), then execute step 3-3 to calculate the savings; otherwise, proceed to step 4.

[0094] savings=

[0095] Let i be the time it takes for the transport vehicle to travel from i to j. The time it takes for the transport vehicle to travel from j to k. Let be the time it takes for the transport vehicle to travel from i to k. The settlement result is the time gain of the transport vehicle due to the deletion of j.

[0096] Step 3-3 requires a re-estimation of the time it takes for the transport vehicle to arrive at the drone rendezvous point for this sub-route. And consider the time saved. Variation in delivery time with drones The numerical relationship between the two values ​​is used to determine the savings (row 8 of Table 2).

[0097] The estimation method is ,

[0098] The time value is the time when the transport vehicle arrives at the drone launch point a (data comes from the time array in steps 1-4). For drones launched from drone launch point A to delivery points visited by drones along the delivery sub-route of the transport vehicle. Time, Delivery points visited by drones along the delivery route from the transport vehicle The time to drone recovery point b For the time required to recover the drone, Let be the time it takes for the transport vehicle to arrive at the drone recovery point b. The difference between the two is the reduction in waiting time for the drone due to the deletion of j.

[0099] Step 4: Calculate the change in delivery efficiency cost based on the delivery points removed in Step 3. Insert the removed delivery points into the transport vehicle route or construct a drone delivery route using the removed delivery points and the transport vehicle route, and calculate the cost incurred by this insertion, as detailed below:

[0100] Step 4-1: Traverse all transport vehicle subroutines (the initial transport vehicle subroutines are the transport vehicle routes; new subroutines are formed based on the launch and recovery points of the drones, i.e., all transport vehicle routes between the launch and recovery points of the drones are defined as subroutines; all continuous transport vehicle routes that do not contain the launch and recovery points of the drones after the division of all subroutines of this type are also defined as subroutines) as shown in row 9 of Table 1.

[0101] Step 4-2: Determine if there is a drone accompanying the current transport vehicle route (i.e., including the drone launch point and recovery point, see row 10 of Table 1). If there is a drone accompanying the transport vehicle, adjust the transport vehicle route and proceed to step 4-3, see pseudocode table 3 (row 11 of Table 1). If not, plan the drone accompanying route and proceed to step 4-4, see pseudocode table 4 (row 12 of Table 1).

[0102] Step 4-3: Adjust the transport vehicle route within the current sub-route. After completion, proceed to step 4-5, which includes:

[0103] Step 4-3-1: Iterate through adjacent delivery points along the transport vehicle's route (rows 1 and 2 of Table 3) and calculate which delivery point removed in step 3 should be inserted between the two points. The insertion cost (row 4 of Table 3) is calculated. If the cost is less than the savings calculated in step 3, proceed to step 4-3-2; otherwise, proceed to step 4-3-4. The method for calculating the insertion cost is as follows:

[0104]

[0105] Among them, delivery points The insertion point is the delivery point i, and the delivery points i and k are two adjacent delivery points on the transport vehicle's route.

[0106] Step 4-3-2: Determine whether the drone's endurance in the current sub-route supports the insertion of this node (Table 3, row 6). If it does, insert the node and perform an efficiency improvement judgment, then proceed to step 4-3-3; otherwise, proceed to step 4-3-5.

[0107] Step 4-3-3: Calculate the difference, maxSavings, between the savings from removing the delivery point in Step 3 and the insertion cost, cost, from inserting the delivery point in Step 4-3-2. Compare this value with the maximum value of maxSavings in the historical iteration data (0 in the first iteration). If it is greater than the maximum value in the historical iteration data (row 7 of Table 3), proceed to Step 4-3-4; otherwise, cancel the insertion of the node and proceed to Step 4-3-4. The calculation method for the difference, maxSavings, is as follows:

[0108] maxSavings = savings - cost

[0109] Step 4-3-4: Temporarily save the current path planning scheme, update the maximum value of historical iteration data maxSavings (rows 8-10 of Table 3), and execute step 4-3-5.

[0110] Step 4-3-5: Check if all adjacent delivery points have been traversed. If so, complete step 4-5; otherwise, select the next pair of adjacent delivery points and execute step 4-3-1.

[0111] Step 4-4: Plan the drone escort route within the current sub-route. After completion, proceed to step 4-5, which includes:

[0112] Step 4-4-1: Traverse the delivery point pairs in the current transport vehicle route (Table 4, row 1) and combine them with the delivery points removed in step 3 to form the drone delivery route;

[0113] Step 4-4-2: Calculate whether the constructed drone delivery route is within the drone's endurance (Table 4, row 2). If the endurance supports delivery, proceed to step 4-4-3; otherwise, proceed to step 4-4-6.

[0114] Step 4-4-3: Calculate the time it takes for the transport vehicle to arrive at the drone recovery and delivery point on the drone delivery route, in order to calculate the cost of constructing the drone delivery route (Table 4, row 4).

[0115] cost = max{0, max{ }-( )}

[0116] in and These represent the arrival times of the transport vehicles at k and i, respectively. and These represent the time consumed by the drone to travel from i to j and the time consumed by the drone to travel from j to k, respectively. and These represent the preparation time for launching and recovering the drone, respectively, with max representing the maximum value. The original intention of the formula is that if the transport vehicle arrives at k first, the cost is the time difference between the drone and the transport vehicle, plus the launch and recovery preparation time; if the drone arrives at k first, the cost is the launch and recovery preparation time.

[0117] Step 4-4-4: Calculate the difference between the revenue (savings) from removing delivery points in Step 3 and the cost (cost) in Step 4-4-3, and compare it with the maximum value of maxSavings in the historical iteration data (0 in the first iteration). If it is greater than the maximum value in the historical iteration data (row 5 of Table 4), proceed to Step 4-4-5; otherwise, cancel the drone delivery route constructed in Step 4-4-1 and execute Step 4-4-6. The calculation method for the difference maxSavings is as follows:

[0118] maxSavings = savings - cost

[0119] Step 4-4-5: Temporarily save the current path planning scheme, update the maximum value of the historical iteration data maxSavings (rows 6-8 of Table 4), and execute step 4-4-6.

[0120] Step 4-4-6: Check if all delivery point pairs have been traversed. If so, proceed to step 4-5; otherwise, select the next delivery point pair and proceed to step 4-4-1.

[0121] Step 4-5: Determine whether all transport vehicle routes have been traversed. If so, proceed to step 5; otherwise, select the next transport vehicle route and proceed to step 4-2.

[0122] Step 5: Determine whether the removal operation in Step 3 has been completed. If so, obtain the current path planning scheme, which includes the latest design variables, and proceed to Step 6. Otherwise, select the next delivery point and re-execute Step 3.

[0123] Step 6: Update the design variables using the current path planning scheme (see Table 5), compare costs and benefits. If the cost is lower than the benefit, then:

[0124] savings > cost

[0125] If the current path planning scheme is temporarily stored as the result of the iterative planning, step 2 is executed again; otherwise, the iteration is stopped and the overall planning is completed.

[0126] Example 1:

[0127] Based on the above technical solution, the specific procedures adopted are as follows: Figure 5The pseudocode for the main program is shown in Table 1. Parameter initialization includes delivery point information, suitable delivery point numbers for drone delivery, and determining the time taken by the drone and transport vehicle to travel between any pair of delivery points. First, it is assumed that `solveTSP` is a general solver for the Traveling Salesman Problem, capable of returning an array of node information representing the delivery order of the transport vehicles, `truckRoute`, and an array of time information `t` representing the arrival times of the transport vehicles at each delivery point. The second-order array `truckSubRoutes` is initially an array containing the order of the transport vehicle delivery points; as the program progresses, the transport vehicle path is divided into a certain number of sub-routes.

[0128] Next, we process delivery points suitable for drone services, as shown in row 8 of Table 1. We calculate the revenue incurred from removing these delivery points, as shown in the revenue calculation module `calcSavings` in Table 2. Row 9 of Table 1 iterates through all transport vehicle routes to calculate the cost of inserting delivery point j into different locations on the transport vehicle routes, or to calculate the cost of serving delivery point j via drone. The cost of inserting delivery point j is calculated using the `calcCostTruck` function in the cost calculation module (row 11 of Table 1), as shown in Table 3. The cost of serving delivery point j via drone is calculated using the `calcCostUAV` function in the cost calculation module (row 12 of Table 1), as shown in Table 4.

[0129] At the end of each iteration, it is determined whether to update the design variables. If the difference between the revenue and cost, maxSavings, is greater than 0, then the variables are updated; otherwise, the design variables remain unchanged. The design variable update module performUpdate contains functions for updating the design variables, as shown in Table 5.

[0130] Table 1. Main Program Pseudocode Table

[0131]

[0132] The pseudocode in Table 2 shows the revenue calculation process in line 2. However, if the sub-route involving the removal of delivery points involves drone delivery, the calculation becomes more complex. It requires considering the numerical relationship between the saved time and the drone delivery cost, as shown in line 8 of Table 2. A negative value indicates that removing delivery point j will cause the transport vehicle to wait for the drone at the rendezvous point; conversely, a positive value indicates that the drone is waiting for the transport vehicle, and removing delivery point j will shorten the waiting time. The planning method improves delivery efficiency by removing specific service points in the sub-path. The `calcSavings` pseudocode, shown in Table 2, is used to calculate the change in the target value caused by removing delivery point j.

[0133] Table 2 Pseudocode for calcSavings

[0134] Input: j (delivery point of the transport vehicle) and t (array of times when the transport vehicle arrives at each delivery point) 1： Retrieve nodes i and k (the nodes before and after delivery point j of the transport vehicle). 2： savings= 3： if (the sub-route containing j includes drones) 4： Retrieve the starting point of the delivery sub-route of the transport vehicle, i.e., the drone launch point. 5： Search for the endpoint of the delivery sub-route of the transport vehicle, i.e., the drone recovery point. 6： Search j_ / * Delivery points visited by drones along the delivery vehicle's sub-route * / 7： calculate 8： savings=min{savings, } 9： end if

[0135] Pseudocode Table 3 aims to insert delivery point j between adjacent i and k in the transport vehicle route. Even if inserting j is beneficial to improving the overall delivery efficiency (cost < savings), it is necessary to consider the impact of inserting j on the UAV's endurance due to the extension of the sub-route. Therefore, it is necessary to confirm that the UAV has sufficient energy to rendezvous with the transport vehicle at the end of the sub-route. If inserting j is the current optimal choice, save the adjacent nodes and mark delivery point j to prevent UAV delivery. The calcCostTruck pseudocode, as shown in Table 3, is used to calculate the cost of inserting delivery point j at different positions in the transport vehicle route.

[0136] Table 3 calcCostTruck Pseudocode Table

[0137]

[0138] The calcCostUAV function calculates the cost of UAV delivery. Traverse all pairs of nodes i and k in the sub-route that are not hinged to the UAV route. The nodes do not have to be adjacent, but i is delivered before k. Calculate the time delay from launching at i, delivering at j, and rendezvousing at k. First, the flight duration of the UAV cannot exceed its endurance. Since delivery point j is removed from the transport vehicle route, the time for the transport vehicle to reach each node needs to be recalculated. The cost caused by adding the UAV route is the larger value between the extra time for the transport vehicle to launch and recover the UAV at delivery point i and the time delay for the UAV to wait for recovery at delivery point k. The calcCostUAV pseudocode, as shown in Table 4, is used to calculate the cost of delivering j by UAV

[0139] Table 4 calcCostUAV Pseudocode Table

[0140]

[0141] The performUpdate function can update the design variables. As the program progresses, the truckRoute will split into multiple sub-routes, and the division of the sub-routes is determined by the launch and recovery of the UAV.

[0142] Table 5 performUpdate Pseudocode Table

[0143]

[0144] Example 2:

[0145] This embodiment details a path planning system and method for efficient cargo delivery using an accompanying drone, based on a specific case. The system includes a parameter initialization module, a revenue calculation module, a cost calculation module, and a design variable update module. Based on the geographical location information of each delivery point, the system obtains an optimized path for the accompanying drone after multiple iterations of calculations by each module. The parameter initialization module solves the traveling salesman problem and supports the initialization of transport vehicle routes and design variables. The revenue calculation module calculates the efficiency gains from removing delivery points along the transport vehicle path. The cost calculation module calculates the cost reductions in delivery efficiency caused by adding transport vehicle delivery nodes at different locations along each sub-path. The design variable update module determines whether an optimized path has been obtained or whether to continue iteration.

[0146] The method includes the following steps:

[0147] Step 1: Parameter initialization. Input the geographical location information of the delivery points and assign initial values ​​to the arrival time array of each delivery point, the node connection relationship array α, and β.

[0148] Step 1-1: Enter the geographical location information of the delivery point;

[0149] Steps 1-2 calculate the time-consuming matrices T and T̀ of the transport vehicle and drone between each delivery point, and the set of delivery points G. There are currently 8 delivery points awaiting delivery. After data processing based on the delivery point information, the path planning for the accompanying drone is initialized. The calculated time for the transport vehicle between each delivery point is shown in Table 6, and the time for the drone between each delivery point is shown in Table 7. In the tables, delivery point 0 represents the logistics warehouse, and the 8 delivery points are numbered sequentially from 1 to 8. The value between any two delivery points represents the time consumed by the vehicle. For example... This means the time required from the logistics warehouse to delivery point 2 is 14.9 seconds. Similarly, the time required from delivery point 2 to the logistics warehouse is also 14.9 seconds. Therefore... ,and Meaningless.

[0150] Table 6. Transportation Time Information Between Delivery Points

[0151]

[0152] Table 7. Drone travel time information between delivery points

[0153]

[0154] Steps 1-3 involve constructing and solving a traveling salesman problem based on the information from each delivery point to obtain the initial delivery route for the transport vehicle, such as... Figure 6As shown, the dashed circles at 0 and 9 represent logistics warehouses, while the dashed circles at 8 and 6 represent delivery points unsuitable for drone services. Solid circles represent delivery points suitable for drone services. In this case, truckRoute = {0,8,4,3,7,2,6,5,1,9} and Gprime = {4,3,7,2,5,1}.

[0155] Steps 1-4 involve initial assignment of values ​​to the arrival time array for each delivery point, and the node connection relationship arrays α and β.

[0156] Step 2: In the current route planning scheme, select the delivery points of the transport vehicles in sequence according to the order of the delivery points, remove them, and calculate the revenue caused by removing the delivery point.

[0157] Step 2-1, start the iteration (each variable in the iteration process must satisfy equation (2)-equation (22)), traverse all delivery points (row 7 of Table 1), and select any node to prepare for removal;

[0158] Step 2-2: Retrieve the locations before and after the delivery point and calculate the efficiency difference before and after removal (row 8 of Table 1). Details are as follows:

[0159] Step 2-2-1: Retrieve the locations before and after the delivery point (Row 1 of Table 2);

[0160] Step 2-2-2: Calculate the savings from removing the node, i.e., the time change (Row 2 of Table 2); if the delivery point is associated with the drone sub-route (Row 3 of Table 2), then proceed to step 2-3; otherwise, go to step 3.

[0161] Steps 2-3 require a re-estimation of the time it takes for the drone to arrive at this sub-route. And consider the time saved. Variation in delivery time with drones The numerical relationship between the two values ​​is used to determine the savings (row 8 of Table 2).

[0162] Step 3: Calculate the change in delivery efficiency costs. Insert the removed delivery point into the transport vehicle route or construct a drone delivery route by removing the delivery point and the transport vehicle route. Calculate the cost caused by this insertion, as follows:

[0163] Step 3-1: Traverse all sub-routes (the sub-routes in the first iteration are the routes) as shown in row 9 of Table 1;

[0164] Step 3-2: Determine whether the sub-route is hinged to the drone delivery route (Table 1, line 10). If it is hinged to the drone delivery route, proceed to step 3-3 (see pseudocode Table 3, line 11 of Table 1). If not, proceed to step 3-5 (see pseudocode Table 4, line 12 of Table 1).

[0165] Step 3-3: Traverse the two adjacent delivery points in the sub-route (rows 1 and 2 of Table 3) and calculate the cost of inserting the delivery point removed in step 2-1 into the two points (row 4 of Table 3). If the cost is less than the savings calculated in step 2, proceed to step 3-4; otherwise, end step 3.

[0166] Step 3-4: Determine if the drone's endurance supports adding nodes (Table 3, row 6). If it does, proceed to steps 3-7 and 8; otherwise, proceed to step 3-3 and continue iterating. If the iteration is complete, end step 3.

[0167] Steps 3-5: Traverse any pair of delivery points in this sub-route (Row 1 of Table 4) and combine them with the delivery points removed in Step 2 to form a drone delivery route;

[0168] Steps 3-6: Calculate whether the drone delivery route is within the drone's battery life (Table 4, row 2). If the battery life supports the delivery, proceed to step 3-7; otherwise, end the process and proceed to step 3-5 to continue iterating. If the iteration is complete, end step 3.

[0169] Steps 3-7: Calculate the time it takes for the transport vehicle to arrive at the drone recovery and delivery point on the drone delivery route, in order to calculate the cost of constructing the drone delivery route (Table 4, row 4).

[0170] Steps 3-7: Calculate the difference between savings and cost, and compare it with the previous iteration data maxSavings. If it is better than the previous iteration data (Table 3, row 7; Table 4, row 5), then proceed to step 3-9; otherwise, end step 3.

[0171] Steps 3-9: Back up the design variable, update the revenue and cost difference (rows 8-10 of Table 3, rows 6-8 of Table 4), and end Step 3. In Gprime, iterate through the delivery points. For delivery point 4, removing it and setting it as a drone route resulted in reduced efficiency; therefore, delivery point 4 is not removed.

[0172] When delivery point 3 is removed, delivery points 4 and 7 are its preceding and following points, respectively. The difference in the optimization function is then calculated using Algorithm 2. =18. In Gprime, the drone delivery path is constructed by traversing node 2 and delivery point 3. If the two nodes are far apart, the drone's flight time will exceed its range. For example, if delivery points 8 and 2 are selected, starting from 8, the drone's time to reach 2 will be... The drone's flight time exceeds its endurance e=10; delivery points 4 and 7 are selected, and the drone's flight time... Within the drone's flight range, and with a calculated cost of 6.6, the efficiency improvement is 11.4. After one iteration, it was found that removing delivery point 3 and setting it as the drone's accompanying flight route could improve delivery efficiency. Therefore, delivery point 2 was removed from the route, and as the delivery point for drone service, it was used to construct an accompanying flight path with delivery points 4 and 7, as follows. Figure 7 As shown, the subroutines are {0,8,4}, {4,7} and {7,2,6,5,1,9}, and Gprime={2,5,1}.

[0173] Step 4: Update the design variables, compare costs and benefits. If the costs are lower than the benefits, update the design variables and return to Step 2.

[0174] Step 4-1: Determine whether the sub-route has been traversed. If not, go to step 3-1.

[0175] Step 4-2: Determine whether the delivery points have been traversed. If not, proceed to step 2-1.

[0176] Step 4-3: Determine if the revenue is greater than the cost (Table 1, row 16). If not, proceed to step 2-1. If the delivery point traversal is complete, end the iteration.

[0177] Step 4-4: Determine whether to construct a drone delivery route (Table 5, row 1). If not, proceed to step 4-7.

[0178] Steps 4-5: Add the drone route from step 3-3 (row 2 of Table 5);

[0179] Steps 4-6: Remove drone delivery points from each route of the transport vehicle (Table 5, rows 3 and 4).

[0180] Steps 4-7: Remove the delivery point from step 2-1 and insert it into the position of step 3 (Table 5, rows 8-10).

[0181] Steps 4-8: Update the arrival time at each delivery point (rows 6 and 10 of Table 5).

[0182] The final path planning for the accompanying drone is as follows: Figure 8 As shown in Table 8, the transportation efficiency of the three delivery methods is compared. The delivery method using drones with accompanying flight improves efficiency by 33.92% compared to the traditional delivery method.

[0183] Table 8. Comparison of Transportation Efficiency of Three Delivery Methods

[0184]

[0185] In its specific implementation, this application provides a computer storage medium and a corresponding data processing unit. The computer storage medium is capable of storing a computer program, which, when executed by the data processing unit, can run the invention's content regarding an efficient cargo delivery accompanying drone path planning system and method, as well as some or all of the steps in various embodiments. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0186] Those skilled in the art will clearly understand that the technical solutions in the embodiments of the present invention can be implemented using computer programs and their corresponding general-purpose hardware platforms. Based on this understanding, the technical solutions in the embodiments of the present invention, or the parts that contribute to the prior art, can be embodied in the form of computer programs, i.e., software products. These computer program software products can be stored in a storage medium and include several instructions to cause a device containing a data processing unit (which may be a personal computer, server, microcontroller, MCU, or network device, etc.) to execute the methods described in various embodiments or certain parts of the embodiments of the present invention.

[0187] This invention provides a method and approach for efficient cargo delivery accompanied by unmanned aerial vehicle (UAV) path planning. Many methods and approaches exist for implementing this technical solution; the above description is merely a preferred embodiment. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of this invention, and these improvements and modifications should also be considered within the scope of protection of this invention. All components not explicitly stated in this embodiment can be implemented using existing technologies.

Claims

1. A path planning system for an efficient cargo delivery drone, characterized in that, include: The parameter initialization module, revenue calculation module, cost calculation module, and design variable update module obtain the accompanying flight drone path through iterative calculation based on the geographical location information of each delivery point. The parameter initialization module is used to construct a cargo delivery route planning model under the condition of drone-accompanied transport vehicle, initialize the environmental parameters of the cargo delivery route planning model, and perform preliminary delivery route planning. The revenue calculation module is used to calculate the change in delivery efficiency revenue caused by removing each delivery point in the transport vehicle route; The cost calculation module is used to calculate the change in delivery efficiency cost caused by adding delivery nodes of transport vehicles at different locations in each sub-path; The design variable update module is used to determine whether an optimized path has been obtained or to continue iteration.

2. A method for efficient cargo delivery accompanying flight drone path planning, characterized in that, The method includes: Step 1: Construct a cargo delivery route planning model under the condition of drone-accompanied transport vehicle; Step 2: Initialize the environmental parameters of the goods delivery route planning model and perform preliminary delivery route planning to determine the current route planning scheme; Step 3: Based on the current route planning scheme, remove the delivery points of the transport vehicles in sequence and calculate the revenue caused by removing the delivery point; Step 4: Based on the delivery points removed in Step 3, perform the insertion operation and calculate the cost caused by the insertion to obtain the latest route planning scheme. Step 5: Determine whether all delivery points of the transport vehicles have been removed. If yes, proceed to step 6; otherwise, proceed to step 3. Step 6: Judge the latest path planning scheme. If the stopping condition is met, stop the iteration and complete the path planning. Otherwise, take the path planning scheme as the current path planning scheme and repeat step 3.

3. The method for efficient cargo delivery accompanying drone path planning according to claim 2, characterized in that, The cargo delivery route planning model under the condition of drone-accompanied transport vehicle described in step 1, that is, the route planning problem of drone-accompanied transport vehicle, is represented by the following mathematical model: Step 1-1, define environment parameters to represent initial conditions, as follows: For all delivery points, use and This indicates that the logistics center serves as both the starting and ending point of the route planning. This represents all network nodes in the path planning problem; defined accordingly. For the starting point set, To reach the set of points; Calculate the delivery time matrix for the transport vehicle and the drone based on their physical properties. and ,in, For transport vehicles from the distribution point Arrival at the delivery point The time consumed The time taken for the drone to travel from the i-th delivery point to the j-th delivery point; Step 1-2: Define design variables to represent path planning schemes, as follows: Define transport vehicle route design variables 1 represents the transport vehicle departing from the distribution point. Drive to the delivery point Define the variables for drone route design. 1 represents a drone carrying a package from a delivery point. Launch, arrival at the delivery point To make the delivery and fly to the delivery point. Meet with the transport vehicle; Steps 1-3: Define the objective function, as follows: ； in, To minimize, For the transport vehicle to arrive at the delivery node The time, i.e. the time it takes to return to the logistics center; Steps 1-4: Define the constraints, as follows: For any delivery point, a delivery is made exactly once, as shown below: ； For any delivery point The transport vehicle arrived at the delivery point. Also starting from delivery point j, it is represented as follows: ； Drones can be delivered from any point Launched, and at most once, the drone can launch from any delivery point. Returning to the transport vehicle, and at most once, is indicated as follows: ； ； Based on the arrival of the delivery vehicle at the delivery point time From the distribution point Heading to the delivery point Time spent Drones from delivery points Preparation time required for launch Or drones at delivery points Preparation time required for recycling Calculate the transport vehicle from the distribution point Arrival at the delivery point time , As a preset constant, it is represented as follows: ； ； Based on the takeoff time of the drone and flight time Calculate the drone from the delivery point Launch, arrival at the delivery point time , means as follows: ； Drone-based delivery time Flight time The transport vehicle should arrive at the meeting point in advance and allow sufficient time for the recovery of the drone. Calculate the time it takes for the drone to be retrieved at delivery point k. , means as follows: ； Limitations on drone battery life This represents the time it takes for the drone to arrive at delivery point k. Representing drones from delivery points Departure time The drone's flight time is expressed as follows: 。 4. The method for efficient cargo delivery accompanying drone path planning according to claim 3, characterized in that, Step 2, which involves preliminary delivery route planning, includes: Step 2-1: Set the geographical location information for all delivery points; Step 2-2: Calculate the time consumption matrix of the transport vehicle and the drone between each delivery point. and Set up a collection of delivery points ; Steps 2-3: Based on the collection of each delivery point Based on the cargo delivery route planning model constructed in step 1, the delivery problem of the transport vehicle is constructed and solved to obtain the initial delivery route of the transport vehicle (as shown in the third row of Table 1). Steps 2-4: Based on the initial delivery route of the transport vehicle, generate the arrival time array for each delivery point and the design variable array A{ } and B{ }, perform the initial assignment, and determine the delivery route as the current path planning scheme.

5. The method for efficient cargo delivery accompanying drone path planning according to claim 4, characterized in that, Solving the transportation vehicle delivery problem as described in steps 2-3 includes: Step 2-3-1, route planning parameter initialization, as follows: A feasible solution is randomly generated as the initial delivery route; an initial temperature is set to control the acceptance probability during the search process; a cooling rate is set to determine the rate at which the temperature decreases. Step 2-3-2: Randomly generate neighbor solutions based on the current solution by swapping the locations of two delivery points or changing the access order; Step 2-3-3: Calculate the total delivery time of the current solution and its neighboring solutions; Steps 2-3-4: If the neighbor's solution is better (i.e., the delivery time is shorter), then accept the neighbor's solution; otherwise, accept it based on probability. Accept neighbor solutions; where the probability is... The calculation method is as follows: ； in, It is the difference between the neighboring solutions and the current solution. This is the current temperature; Steps 2-3-5, reducing the temperature according to the cooling rate, are represented as follows: ； Where, 0 < <1 is the cooling coefficient. This refers to the time after cooling. Step 2-3-6: If the maximum number of iterations is reached, or the temperature drops to the threshold, or the current solution is optimal, then output the optimal delivery route and the corresponding delivery time; otherwise, repeat step 2-3-2.

6. The method for efficient cargo delivery accompanying drone path planning according to claim 5, characterized in that, Step 3, which involves sequentially removing delivery points from transport vehicles and calculating the benefits arising from removing each delivery point, includes: Step 3-1: Based on the current route planning scheme, sequentially select the delivery points of the transport vehicles as the current delivery points and prepare to remove them; Step 3-2: Retrieve the locations before and after the current delivery point, and calculate the efficiency difference (i.e., revenue) before and after removal. The details are as follows: Step 3-2-1, retrieve the delivery point Information on delivery points before and after the transport vehicle's route and ; Step 3-2-2: Calculate the revenue from removing the node based on the information of the preceding and following delivery points. The changes in delivery time for transport vehicles are represented as follows: ； in, For transport vehicles from the distribution point Heading to the delivery point Time, For transport vehicles from the distribution point Heading to the delivery point Time, For transport vehicles from the distribution point Heading to the delivery point Time; Step 3-3: If there are drones accompanying the current delivery point and its surrounding locations, then the revenue will be updated as follows: ； in, For the transport vehicle to arrive at the drone launch site Time, For drones from drone launch point Launched to drone delivery point Time, For drones from drone delivery points To the drone recycling point Time, For the time required to recover the drone, For the transport vehicle to arrive at the drone recycling point Time; The recalculated time for the transport vehicle to arrive at the drone rendezvous point is calculated as follows:

7. The method for efficient cargo delivery accompanying drone path planning according to claim 6, characterized in that, The calculation of the insertion-induced cost described in step 4 includes: Step 4-1: Based on the delivery points removed in Step 3, form sub-routes for transport vehicles; define transport vehicle routes based on the launch and recovery points of drones, that is, define the transport vehicle route between the launch and recovery points of drones as a sub-route, and define the remaining continuous transport vehicle routes that do not include the launch and recovery points of drones as sub-routes; sequentially select sub-routes as the current transport vehicle route. Step 4-2: Determine if there are drones accompanying the current transport vehicle route, including drone launch and recovery points. If so, proceed to step 4-3 to adjust the transport vehicle route; otherwise, proceed to step 4-4 to plan the drone escort route. Step 4-3: Adjust the transport vehicle route in the current sub-route, insert the removed delivery points into the current transport vehicle sub-route, calculate the cost caused by the insertion, obtain the latest route planning scheme, and then execute step 4-5. Step 4-4: Plan the drone escort route in the current sub-route, remove the delivery point and transport vehicle sub-route to build the drone delivery route, calculate the resulting cost, obtain the latest path planning scheme, and then execute step 4-5. Step 4-5: Determine whether all transport vehicle routes have been traversed. If so, proceed to step 5; otherwise, select the next transport vehicle route and proceed to step 4-2.

8. The method for efficient cargo delivery accompanying drone path planning according to claim 7, characterized in that, Step 4-3 specifically includes: Step 4-3-1: Iterate through adjacent delivery points along the current transport vehicle route and calculate which delivery point removed in Step 3 should be inserted between the two points. Insertion cost If insertion cost Less than the profit calculated in step 3 If the condition is met, proceed to step 4-3-2; otherwise, proceed to step 4-3-4. Wherein, the insertion cost... The calculation method is as follows: ； Among them, delivery points It's an insertion point, a delivery point. and These are adjacent delivery points along the transport vehicle's route; Step 4-3-2: Determine whether the drone's endurance in the current sub-route supports inserting the delivery point. If it does, insert the delivery point and execute step 4-3-3 to determine efficiency improvement; otherwise, execute step 4-3-5. Step 4-3-3: Calculate the revenue from removing the delivery point in Step 3. The insertion cost of inserting the delivery point in step 4-3-2 The difference Difference from historical data If the maximum value is greater than the maximum value in the historical iteration data, proceed to step 4-3-4; otherwise, cancel the insertion of the delivery point and proceed to step 4-3-5. Step 4-3-4: Temporarily save the current path planning scheme and update the differences in historical data. Find the maximum value and proceed to step 4-3-5; Step 4-3-5: Check if all adjacent delivery points in the current transport vehicle route have been traversed. If so, proceed to step 4-5; otherwise, select the next pair of adjacent delivery points and proceed to step 4-3-1.

9. The method for efficient cargo delivery accompanying unmanned aerial vehicle (UAV) path planning according to claim 8, characterized in that, Step 4-4 specifically includes: Step 4-4-1: In the current transport vehicle route, sequentially traverse the delivery point pairs to form a drone delivery route with the delivery points removed in step 3; Step 4-4-2: Calculate whether the drone delivery route is within the drone's battery life. If yes, proceed to step 4-4-3; otherwise, proceed to step 4-4-6. Step 4-4-3: Calculate the time it takes for the transport vehicle to arrive at the drone recovery and delivery point along the drone delivery route, and calculate the insertion cost of constructing the drone delivery route. The details are as follows: ； in, and The delivery vehicle arrived at the distribution point. With delivery point Time, and The drones from the delivery point Arrival at the delivery point Time consumption and from delivery point Arrival at the delivery point Time consumption, and These are the preparation times for launching and recovering the drone, respectively. To obtain the maximum value; Step 4-4-4: Calculate the revenue from removing the delivery point in Step 3. Insert cost in step 4-4-3 The difference Difference from historical data If the maximum value is greater than the maximum value in the historical iteration data, proceed to step 4-4-5; otherwise, cancel the drone delivery route constructed in step 4-4-1 and proceed to step 4-4-6. Step 4-4-5: Temporarily save the current path planning scheme and update the differences in historical data. Find the maximum value and proceed to step 4-4-6; Step 4-4-6: Check if all delivery point pairs in the current transport vehicle route have been traversed. If so, proceed to step 4-5; otherwise, select the next pair of adjacent delivery points and proceed to step 4-4-1.

10. The method for efficient delivery of goods by accompanying unmanned aerial vehicle (UAV) path planning according to claim 9, characterized in that, The stopping conditions described in step 6 include: Calculate the current insertion cost and benefits If the following conditions are met: ； Then the stopping condition is met.

Images