Tobacco harvesting and transferring coordinated scheduling method and scheduling system

Abstract

The invention discloses a scheduling method and scheduling system for tobacco harvesting and transferring collaboration.The scheduling method comprises the steps that modeling is conducted on discrete operation states and state conversion in the harvesting process of harvesters and transferring robots on the basis of a finite-state machine and a difference equation, and transferring task information of the harvesters is predicted; and scheduling the transfer robot to complete the transfer task, constructing a predictive tobacco transfer scheduling model taking minimization of the non-productive operation time of the harvesters and the movement distance of the transfer robot as objective functions, setting constraint conditions, and performing accurate optimization solution by using a multi-objective optimization algorithm. According to the scheduling method, the tobacco harvesting and transferring processes can be effectively and cooperatively scheduled, so that the tobacco harvesting efficiency is improved, and the labor intensity of workers is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent tobacco transportation, and in particular to a scheduling method and a scheduling system for coordinated tobacco harvesting and transportation. Background Art

[0002] Tobacco is an important economic crop and occupies a vital position in my country's national economy. At the same time, my country is also the world's largest tobacco producer. However, the level of intelligence in the tobacco leaf harvesting process is not high.

[0003] At present, tobacco leaf harvesting in my country mainly adopts the method of manual picking and self-transportation. The harvesters need to go from the collection station where the tobacco leaves are stored to the planting rows to pick tobacco leaves. When the tobacco leaves they carry reach the upper limit of the weight, they return to the collection station to place the picked tobacco leaves, and then return to the original position to continue picking until the tobacco field is harvested. During the entire harvesting process, the time the harvesters spend transporting the tobacco leaves by themselves is non-productive working time, which accounts for a very large proportion of the total time, resulting in low tobacco leaf harvesting efficiency. Therefore, in order to complete the tobacco leaf harvesting activities within a limited time, the harvesters can only continuously increase the intensity of labor, which in turn affects the physical health of the harvesters. Summary of the Invention

[0004] In order to solve the technical defects existing in the prior art, the present invention provides a scheduling method and scheduling system for coordinated tobacco leaf harvesting and transportation. By effectively coordinating the processes of tobacco leaf harvesting and transportation, the problems of low efficiency and high labor intensity of existing tobacco leaf harvesting are solved.

[0005] In a first aspect, the present invention provides a method for coordinating tobacco leaf harvesting and transportation, which adopts the following technical solutions: A method for coordinating tobacco leaf harvesting and transportation, comprising the following steps: S1: Obtaining motion parameters of the harvester and the transfer robot and characteristic parameters of the tobacco field, and modeling the discrete operation states and state transitions of the harvester and the transfer robot during the harvesting process based on a finite state machine and a differential equation; S2: predicting the transport task information of the harvester, the transport task information including the time interval when the harvester reaches the upper limit of the load and the tobacco leaf transport location, and adding the information to the transport task set; S3: Allocating the transfer task according to the time when the transfer robot arrives at the tobacco transfer location, and generating scheduling time window information based on the allocation result; S4: setting the objective function and constraints of the predictive tobacco transportation scheduling model according to the scheduling time window information; S5: using a multi-objective optimization algorithm to accurately optimize and solve the predicted tobacco transportation scheduling model, and generating an optimal scheduling scheme for tobacco harvesting and transportation coordination.

[0006] Further, the finite state machine and the difference equation are used to model the discrete operation states and state transitions of the picker and the transportation robot during the harvesting process, including: The finite state machine is used to divide the operation state of the picker into nine types and the operation state of the transportation robot into eight types, and a prerequisite is required for the transition between states. At any discrete time, the activities of the picker in a certain operation state can be represented by a state vector, which is updated by the difference equation.

[0007] Further, the transportation task information of the picker is predicted, including: The weight of the tobacco carried by the picker and the tobacco picking speed are obtained in real time, and the time interval for the picker to reach the weight limit is calculated according to the upper limit of the tobacco weight carried by the picker. According to the current position of the picker, the tobacco picking speed and the time interval for reaching the weight limit, the tobacco transportation position of the picker is calculated, and if the tobacco transportation position is not in the current planting row, the position of the end of the row is set as the tobacco transportation position.

[0008] Further, the transportation task is allocated according to the time when the transportation robot reaches the tobacco transportation position, including: The time when the transportation robot reaches the tobacco transportation position is affected by the operation state and the current load of the transportation robot. The transportation task is preferentially allocated to the transportation robot that first reaches the tobacco transportation position.

[0009] Further, the global path planning algorithm used by the transportation robot when going to the tobacco transportation position is the improved A* algorithm, which adds the evaluation of node path characteristics to the evaluation function of the A* algorithm and increases the value of the search inflection point.

[0010] Further, the scheduling time window information includes all time nodes of the picker and the transportation robot in each harvesting and transportation time window.

[0011] Further, the objective function of the predictive tobacco transportation scheduling model includes: (1); wherein, is the first objective function, which represents the shortest non-productive operation time of the picker; is the second objective function, which means minimizing the moving distance of the transfer robot; Q is the number of harvesters; Wp is the number of time windows of harvester p; M is the number of transfer robots; is the non-productive operation time of harvester p in the kth time window; X rpk is a variable with the value 0 or 1, indicating whether the transfer robot r participates in the transfer task within the kth time window of the harvester p; represents the time when the transfer robot goes to the tobacco leaf transfer location within the k-th time window of the harvester; is the time it takes for the transfer robot r to return to the collection station or move to the next tobacco leaf transfer location within the kth time window of the harvester p; Vr is the moving speed of the transfer robot r.

[0012] Furthermore, the constraints set in the predictive tobacco transportation scheduling model include: 1) Each harvester can only have one transfer robot serving in each time window:

[0013] Among them, P is the set of harvesters; 2) The transfer robot is allowed to be dispatched to the next transfer location only after completing the current transfer task: (3); in, is the time when the transfer robot r is dispatched to the tobacco leaf transfer location within the kth time window of the harvester p; Complete the previous harvester for the transport robot r No. The moment of the transfer task within a time window.

[0014] 3) The load of tobacco leaves after the transfer robot completes the transfer task is not greater than the maximum capacity: (4); Among them, C rpk is the load of the transfer robot r after completing the transfer task of harvester p within the kth time window; Cr is the maximum load of the transfer robot r.

[0015] Furthermore, the multi-objective optimization algorithm adopts the NSGA-II algorithm, the core of which lies in fast non-dominated sorting, congestion distance calculation and elite selection strategy. In a second aspect, the present invention provides a scheduling system for coordinated tobacco harvesting and transportation, which adopts the following technical solutions: The information acquisition module is used to obtain data such as position and speed in real time through various sensors equipped by the harvester and transfer robot and transmit the data to other modules; The robot control module is used to receive the transfer task information assigned by the scheduling server and control the transfer robot to go to the tobacco transfer location to complete the transfer task; A transport task prediction module, configured to predict transport task information of the harvester based on the harvester data; The scheduling server module is used to allocate the transfer tasks in the transfer task set according to the time when the transfer robot arrives at the tobacco transfer location; The visualization module is used to display the status of the harvesters and transfer robots, the weight of the tobacco leaves picked by the harvesters, and the real-time scheduling diagram.

[0016] Compared with the prior art, the present invention can achieve at least one of the following beneficial effects: 1. The present invention proposes a scheduling method and system for coordinated tobacco leaf harvesting and transportation, which can effectively coordinate the tobacco leaf harvesting and transportation processes, thereby improving tobacco leaf harvesting efficiency and reducing manual labor intensity; 2. The present invention improves the A* algorithm by adding an evaluation of node path characteristics to the evaluation function and increasing the cost of inflection points during the search, which can effectively reduce the size of the search space and improve the path planning speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a flow chart of a method for coordinating tobacco leaf harvesting and transportation according to the present invention; Figure 2 Schematic diagram of the division of the finite state machine in the present invention; Figure 3 Schematic diagram of the scheduling time window in the present invention; Figure 4 This is a structural block diagram of a tobacco leaf harvesting and transportation coordinated scheduling system in the present invention; Figure 5 Schematic diagram of the improved A* algorithm in the present invention. DETAILED DESCRIPTION

[0018] The following is combined with Figure 1-4 The present invention is described in further detail.

[0019] The embodiment of the present invention discloses a scheduling method for coordinated tobacco leaf harvesting and transportation.

[0020] like Figure 1 As shown, a scheduling method for coordinated tobacco leaf harvesting and transportation includes the following steps: S1: Obtain the motion parameters of the harvester and transfer robot and the characteristic parameters of the tobacco field, and model the discrete operation states and state transitions of the harvester and transfer robot during the harvesting process based on finite state machines and differential equations.

[0021] In an embodiment of the present invention, the harvester carries a GPS sensor to obtain the harvester's real-time movement speed, position and tobacco leaf load information; the transfer robot is equipped with RTK-GNSS, lidar, inertial measurement unit, pressure and other sensors to obtain the transfer robot's movement speed, position and tobacco leaf load information; the tobacco field characteristic parameters include the location of the collection station, the number of planting rows, size, average yield and other information.

[0022] Furthermore, the discrete operation states and state transitions of the harvester and the transfer robot during the harvesting process are modeled based on finite state machines and difference equations, which specifically includes the following steps: S101: Figure 2 As shown in the figure, the finite state machine is used to divide the harvester's operating states into nine types: start, enter the row, harvest, wait for transportation, transfer tobacco leaves, transfer row, self-transport, unload tobacco leaves, and end. The operating states of the transfer robot are divided into eight types: start, wait for scheduling, track target point, wait for harvest, transfer tobacco leaves, return to the collection station, unload tobacco leaves, and end. The conversion between states requires a prerequisite.

[0023] S102: At any discrete moment, the activities of the harvester in a certain operating state can be represented by a state vector. The state vector is updated in the form of a difference equation. Taking the harvester as an example, the specific expression is as follows: 1) Harvester state vector:

[0024] Among them, X p，k is the state vector of harvester p at discrete time k, X p，k , Y p，k The position coordinates of the harvester p at discrete time k, W p，k is the weight of tobacco leaves carried by the picker p at discrete time k, T p，k is the duration that the harvester p is in the current operating state at discrete time k.

[0025] 2) Update of difference equation form:

[0026] Among them, x p，k+1 ,y p，k+1 is the state vector of harvester p at discrete time k+1, W p，k+1 is the weight of tobacco leaves carried by the harvester p at discrete time k+1, T p，k+1 Vs is the duration of the harvester p in the current operating state at discrete time k+1, p is the moving speed of the harvester p, is the heading angle of the harvester, which is related to the starting position and ending position of the harvester's current operating state. is the tobacco leaf picking rate of harvester p.

[0027] S2: Predict the harvester's transfer task information, which includes the time interval when the harvester reaches the upper limit of the load and the tobacco leaf transfer location, and add it to the transfer task set.

[0028] The specific steps include: S201: Obtain the weight of tobacco leaves carried by the harvester and the speed of tobacco leaf picking in real time, and calculate the time interval for the harvester to reach the upper limit of tobacco leaf weight based on the upper limit of the harvester's tobacco leaf weight.

[0029] Specifically, it is calculated as the difference between the maximum weight of tobacco leaves that a harvester can carry and the weight of tobacco leaves that the harvester is currently carrying, divided by the tobacco leaf picking speed.

[0030] S202: Calculate the tobacco transfer position of the harvester based on the current position of the harvester, the tobacco picking speed and the time interval for reaching the upper limit of the load. If the tobacco transfer position is not within the current planting row, the position of the end of the field is set as the tobacco transfer position. The position of the end of the field refers to the fixed position at the junction of the two ends of the planting row and the ridge.

[0031] Specifically, the tobacco leaf transfer position is the x- and y-axis coordinate values ​​of the harvester's current position plus the product of the harvester's tobacco leaf picking speed in the x- and y-directions and the time interval for the harvester to reach the upper limit of the load.

[0032] S3: Assign transfer tasks according to the time when the transfer robot arrives at the tobacco transfer location, and generate scheduling time window information based on the assignment results.

[0033] In an embodiment of the present invention, the time it takes for the transfer robot to reach the tobacco leaf transfer position varies depending on the operating state and current load of the transfer robot.

[0034] Specifically, if the transfer robot is in the "waiting for scheduling" operating state, the time it takes to reach the tobacco leaf transfer position is the Manhattan distance between the current position of the transfer robot and the tobacco leaf transfer position divided by the moving speed of the transfer robot. If the transfer robot is in the "tracking target point", "waiting for harvesting" or "transferring tobacco leaves" operating state, it is necessary to determine whether the difference between the sum of the current load of the transfer robot and the weight of the tobacco leaves that need to be transferred for the transfer task being performed and the maximum load of the transfer robot is greater than the weight of the tobacco leaves that need to be transferred for the transfer task to be assigned. If so, the time it takes to reach the tobacco leaf transfer position is the Manhattan distance between the tobacco leaf transfer position that is performing the transfer task and the tobacco leaf transfer position to be assigned the transfer task divided by the moving speed of the transfer robot plus the transfer time. The time it takes for the robot to complete the transfer task it is performing. If not, the transfer robot will not be considered. If the transfer robot is in the "return to collection station" operating state, it is necessary to determine whether the difference between the current load of the transfer robot and the maximum load of the transfer robot is greater than the weight of the tobacco leaves that need to be transferred for the assigned transfer task. If so, the time it takes to reach the tobacco leaf transfer position is the Manhattan distance between the current position of the transfer robot and the tobacco leaf transfer position divided by the moving speed of the transfer robot. If not, the transfer robot will not be considered. If the transfer robot is in the "unloading tobacco leaves" time, the time it takes to reach the tobacco leaf transfer position is the Manhattan distance between the current position of the transfer robot and the tobacco leaf transfer position divided by the moving speed of the transfer robot plus the time it takes to complete unloading the tobacco leaves.

[0035] The transfer task will be assigned first to the transfer robot that arrives at the tobacco transfer location first.

[0036] Furthermore, the global path planning algorithm adopted by the transfer robot when heading to the tobacco transfer location is the improved A* algorithm. The improved A* algorithm adds an evaluation of the node path characteristics to the evaluation function on the basis of the A* algorithm and increases the cost value of the inflection point during the search, which can effectively reduce the size of the search space and improve the path planning speed.

[0037] The implementation process of the A* algorithm is as follows Figure 5As shown, it contains two list sets: open list and closed list. The open list is used to save all the nodes whose total cost has been calculated; the closed list is used to save the nodes that have been visited or obstacle nodes. When the algorithm starts, the starting point is first put into the open list, the starting point is selected as the current node, then the total cost of the four or eight adjacent nodes near the current node is calculated, and all the adjacent nodes are added to the open list. Next, the current node is deleted from the open list and added to the closed list, then the node with the minimum total cost in the open list is selected as the current node, and the above operation is repeated, and the cycle is continuously performed until the target node is selected as the current node in the open list. Finally, the path is obtained by moving from the target node to its parent node in reverse until the starting point, and the path is output as the global path.

[0038] The order of selecting the minimum node in the basic A* algorithm among multiple nodes with the same value is not guaranteed, and the number of tobacco planting rows in the map is large, which will greatly interfere with the search. In this case, the search space of the algorithm will be very large, especially when the starting point and the target point are far apart, the algorithm will search almost every ridge (the passable area between planting rows) between them, resulting in a significant reduction in path planning speed and a lack of real-time performance. Therefore, the A* algorithm is improved by adding an evaluation of the path characteristics of the node in the evaluation function and increasing the value of the inflection point during the search to reduce the size of the search space and improve the path planning speed without affecting the total length of the planned path.

[0039] Further, the scheduling time window information generated by the assignment result is as shown in Figure 3 The scheduling time window information includes all time nodes of the picker and the transfer robot in each picking and transfer time window.

[0040] S4: According to the scheduling time window information, set the objective function and constraint conditions of the predictive tobacco transfer scheduling model.

[0041] In the embodiment of the present application, according to the analysis of the predictive tobacco transfer scheduling model, the less the waiting time of the picker, the less the non-productive operation time, the greater the proportion of the productive operation time in the total time, and the higher the harvesting efficiency; in addition, the transfer robot has limited power, and the less the moving distance, the longer the time participating in the harvesting process, and the harvesting efficiency will also be improved. Therefore, the main goal of the model is to minimize the non-productive operation time of the picker and the moving distance of the transfer robot.

[0042] Specifically, the objective function set by the predictive tobacco transportation scheduling model is expressed as follows:

[0043] in, is the first objective function, which means minimizing the non-productive operation time of the harvester; is the second objective function, which means minimizing the moving distance of the transfer robot; Q is the number of harvesters; Wp is the number of time windows of harvester p; M is the number of transfer robots; is the non-productive operation time of harvester p in the kth time window; is a variable with the value 0 or 1, indicating whether the transfer robot r participates in the transfer task within the kth time window of the harvester p; represents the time when the transfer robot goes to the tobacco leaf transfer location within the k-th time window of the harvester; is the time it takes for the transfer robot r to return to the collection station or move to the next tobacco leaf transfer location within the kth time window of the harvester p; Vr is the moving speed of the transfer robot r.

[0044] Specifically, the constraints set by the predictive tobacco transportation scheduling model are as follows: 1) Each harvester can only have one transfer robot serving in each time window:

[0045] Among them, P is the set of harvesters.

[0046] 2) The transfer robot is allowed to be dispatched to the next transfer location only after completing the current transfer task:

[0047] in, is the time when the transfer robot r is dispatched to the tobacco leaf transfer location within the kth time window of the harvester p; Complete the previous harvester for the transport robot r No. The moment of the transfer task within a time window.

[0048] 3) The load of tobacco leaves after the transfer robot completes the transfer task is not greater than the maximum capacity:

[0049] Among them, Crpk is the load of the transfer robot r after completing the transfer task of harvester p in the kth time window; Cr is the maximum load of the transfer robot r.

[0050] S5: Use a multi-objective optimization algorithm to accurately optimize and solve the predictive tobacco transportation scheduling model to generate the optimal scheduling plan for the coordination of tobacco harvesting and transportation.

[0051] Specifically, the multi-objective optimization algorithm adopts the NSGA-II algorithm. The core of the NSGA-II algorithm lies in fast non-dominated sorting, congestion distance calculation and elite selection strategy.

[0052] Based on the above method, an embodiment of the present invention further discloses a scheduling system.

[0053] Combine Figure 4 , the scheduling system includes the following modules: The information acquisition module is used to obtain data such as position and speed in real time through various sensors equipped by the harvester and transfer robot and transmit the data to other modules; The robot control module is used to receive the transfer task information assigned by the scheduling server and control the transfer robot to go to the tobacco transfer location to complete the transfer task; A transfer task prediction module is used to predict the transfer task information of the harvester based on the harvester data; The scheduling server module is used to allocate the transfer tasks in the transfer task set according to the time when the transfer robot arrives at the tobacco transfer location; The visualization module is used to display the status of the harvesters and transfer robots, the weight of the tobacco leaves picked by the harvesters, and the real-time scheduling diagram.

[0054] The embodiments of this specific implementation method are all preferred embodiments of the present invention and are not intended to limit the scope of protection of this application. Therefore, any equivalent changes made based on the structure, shape, and principle of this application should be included in the scope of protection of the present invention.

Claims

1. A method for coordinating tobacco leaf harvesting and transportation, characterized in that: The following steps are involved: S1: Obtaining motion parameters of the harvester and the transfer robot and characteristic parameters of the tobacco field, and modeling the discrete operation states and state transitions of the harvester and the transfer robot during the harvesting process based on a finite state machine and a differential equation; S2: predicting the transport task information of the harvester, the transport task information including the time interval when the harvester reaches the upper limit of the load and the tobacco leaf transport location, and adding the information to the transport task set; S3: Allocating the transfer task according to the time when the transfer robot arrives at the tobacco transfer location, and generating scheduling time window information based on the allocation result; S4: setting the objective function and constraints of the predictive tobacco transportation scheduling model according to the scheduling time window information; S5: Utilize a multi-objective optimization algorithm to accurately optimize and solve the predictive tobacco leaf transportation scheduling model to generate an optimal scheduling plan for coordinated tobacco leaf harvesting and transportation.

2. The method for coordinated tobacco leaf harvesting and transportation according to claim 1, characterized in that: The modeling of the discrete operation states and transitions between states during the harvesting process of the harvester and the transfer robot based on a finite state machine and a differential equation includes: The finite state machine is used to classify the harvester's operating states into nine types, and the transport robot's operating states into eight types, and each state transition requires a prerequisite; At any discrete moment, the activity of the harvester in a certain operating state can be represented by a state vector, and the state vector is updated in the form of the difference equation.

3. The method for coordinated tobacco leaf harvesting and transportation according to claim 1, characterized in that: The predicted transport task information of the harvester includes: The weight of tobacco leaves carried by the harvester and the speed of tobacco leaf picking are obtained in real time, and the time interval for the harvester to reach the upper limit of tobacco leaf weight is calculated based on the upper limit of tobacco leaf weight carried by the harvester; The harvester's tobacco transfer position is calculated based on the harvester's current position, tobacco picking speed and time interval for reaching the upper limit of the load. If the tobacco transfer position is not within the current planting row, the position at the end of the field is set as the tobacco transfer position.

4. The method for coordinated tobacco leaf harvesting and transportation according to claim 1, characterized in that: The allocating the transfer task according to the time when the transfer robot arrives at the tobacco transfer location includes: The time it takes for the transfer robot to arrive at the tobacco leaf transfer position varies depending on the operating state and current load of the transfer robot; The transfer task is preferentially assigned to the transfer robot that arrives at the tobacco transfer position first.

5. The method for coordinated tobacco leaf harvesting and transportation according to claim 4, characterized in that: The global path planning algorithm adopted by the transfer robot when heading to the tobacco transfer location is the improved A* algorithm. The improved A* algorithm adds an evaluation of node path characteristics to the evaluation function on the basis of the A* algorithm and increases the cost value of the inflection point during the search.

6. The method for coordinated tobacco leaf harvesting and transportation according to claim 1, characterized in that: The scheduling time window information includes all time nodes of the harvester and the transfer robot in each harvesting and transfer time window.

7. The method for coordinated tobacco leaf harvesting and transportation according to claim 1, characterized in that: The objective function set by the predictive tobacco transportation scheduling model includes: (1)； in, is the first objective function, which means minimizing the non-productive operation time of the harvester; is the second objective function, Q represents the minimum moving distance of the transfer robot; Q is the number of harvesters; W P is the number of time windows of harvester p; M is the number of transfer robots; is the non-productive operation time of harvester p in the kth time window; is a variable with the value 0 or 1, indicating whether the transfer robot r participates in the transfer task within the kth time window of the harvester p; represents the time when the transfer robot goes to the tobacco leaf transfer location within the k-th time window of the harvester; V is the time it takes for the transfer robot r to return to the collection station or go to the next tobacco leaf transfer location within the kth time window of the harvester p; r is the moving speed of the transfer robot r.

8. The method for coordinated tobacco leaf harvesting and transportation according to claim 1, characterized in that: The constraints set by the predictive tobacco transportation scheduling model include: 1) Each harvester can only have one transfer robot serving in each time window: Among them, P is the set of harvesters; 2) The transfer robot is allowed to be dispatched to the next transfer location only after completing the current transfer task: (3)； in, is the time when the transfer robot r is dispatched to the tobacco leaf transfer location within the kth time window of the harvester p; Complete the previous harvester for the transport robot r No. The moment of the transfer task within a time window; 3) The load of tobacco leaves after the transfer robot completes the transfer task is not greater than the maximum capacity: (4)； Among them, C rpk C is the load of the transfer robot r after completing the transfer task of the harvester p in the kth time window; r is the maximum load of the transfer robot r.

9. The method for coordinated tobacco leaf harvesting and transportation according to claim 1, characterized in that: The multi-objective optimization algorithm adopts the NSGA-II algorithm, the core of which lies in fast non-dominated sorting, congestion distance calculation and elite selection strategy.

10. A coordinated dispatching system for tobacco leaf harvesting and transportation, characterized in that: include: The information acquisition module is used to obtain data such as position and speed in real time through various sensors equipped by the harvester and transfer robot and transmit the data to other modules; The robot control module is used to receive the transfer task information assigned by the scheduling server and control the transfer robot to go to the tobacco transfer location to complete the transfer task; A transport task prediction module, configured to predict transport task information of the harvester based on the harvester data; The scheduling server module is used to allocate the transfer tasks in the transfer task set according to the time when the transfer robot arrives at the tobacco transfer location; The visualization module is used to display the status of the harvesters and transfer robots, the weight of the tobacco leaves picked by the harvesters, and the real-time scheduling diagram.

Images