A distributed scheduling method and system for intelligent irrigation

By using a collaborative request and response mechanism between control terminals and leveraging time difference tables and multi-hop extended search, the problem of insufficient water pump output capacity in smart agricultural irrigation systems in hilly areas was solved, improving the response speed and scheduling flexibility of the irrigation system and enhancing its robustness.

CN123117718APending Publication Date: 2026-11-17HANGZHOU DIANZI UNIV +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611585093.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-10-14
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

In smart agricultural irrigation systems in hilly areas, the water output capacity of a single control pump cannot meet the simultaneous water intake or high-flow irrigation needs of multiple outlets. The lack of a coordination mechanism between control terminals for pumping capacity leads to insufficient flow and low irrigation efficiency.

Method used

Through a collaborative request and response mechanism between control terminals, the pumping capacity of water pumps can be expanded on demand. By utilizing local time difference tables and multi-hop expansion search mechanisms, the optimal assisting party can be quickly selected for joint water supply, and the scheduling strategy can be dynamically adjusted to meet the demand.

Benefits of technology

It improves the response speed and scheduling flexibility of irrigation systems in hilly areas, enhances the system's robustness in dynamic environments, and ensures that irrigation tasks are not interrupted.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123117718A_ABST
    Figure CN123117718A_ABST
Patent Text Reader

Abstract

This invention belongs to the field of smart agricultural irrigation technology and discloses a distributed scheduling method and system for intelligent irrigation. It is applied to a system containing multiple irrigation service nodes, each node including a card-swiping terminal, a water outlet, a control terminal, and a water pump. The method includes: when the local water supply capacity is insufficient, the control terminal determines the flow gap and triggers collaborative scheduling; it broadcasts assistance requests to other control terminals, and determines the primary assisting party based on available water supply capacity and response latency; the primary assisting party starts the water pump to supplement water supply, and the requesting terminal measures the water flow time difference and updates its local time difference table; subsequently, it prioritizes quickly selecting an assisting party based on the local time difference table, and performs a multi-hop extended search through the neighbor table when an assisting party is unavailable. This invention realizes on-demand collaborative expansion of the pumping capacity among control terminals, and has advantages such as fast response speed, wide coverage, localized learning optimization, and strong dynamic robustness, effectively solving the problem of insufficient water output capacity of a single control terminal pump in complex hilly pipe network environments.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of smart agricultural irrigation technology, and particularly relates to a distributed scheduling method and system for smart irrigation. Background Technology

[0002] In smart agriculture irrigation scenarios in hilly areas, card-swiping terminals, water outlets, control terminals, and water pumps are typically deployed in a distributed manner. When farmers initiate a water request at the card-swiping terminal by swiping a card or scanning a QR code to pay, the control terminal connected to the terminal starts the local water pump to draw water, which is then transported to the corresponding water outlet through the pipeline network.

[0003] However, hilly areas have complex terrain and limited water supply network coverage, making it difficult to deploy a sufficient number of control terminals and pumps as densely as in plains areas. When multiple outlets under the same control terminal draw water simultaneously, or when large-volume irrigation demands exceed the rated pumping head and flow rate of local pumps, the output capacity of a single control terminal pump cannot meet the irrigation needs. Existing systems lack a coordination mechanism for pumping capacity between control terminals, making it impossible to quickly and dynamically schedule nearby idle pumps for joint water supply. This results in insufficient outlet flow, low irrigation efficiency, and severely impacts the automation response capability and service quality of smart irrigation systems in hilly areas. Summary of the Invention

[0004] The purpose of this invention is to provide a distributed scheduling method and system for intelligent irrigation, which realizes the on-demand expansion of water pumping capacity through a collaborative request and response mechanism between control terminals, and solves the problem of insufficient water output capacity of a single control terminal pump in a complex pipe network environment in hilly areas.

[0005] To address the aforementioned technical problems, the specific technical solution of the intelligent irrigation distributed scheduling method and system of the present invention is as follows:

[0006] To achieve the above objectives, this invention provides a distributed scheduling method for intelligent irrigation, applied to an irrigation system comprising multiple irrigation service nodes. Each node includes a card-swiping terminal, a water outlet, a control terminal, and a water pump. Each control terminal is connected to the same broadcast domain and maintains a local "target control terminal - water flow time difference" table. The method includes the following steps: S1. After receiving a water discharge request, the control terminal starts the local water pump to supply water, and when the local water supply capacity does not meet the water discharge demand, it determines the flow gap and triggers coordinated scheduling. S2. The requesting end control end sends an assistance request to other control ends and determines the main assisting party based on the available water supply capacity and response delay fed back by other control ends. S3. The main assisting party starts the local water pump to supplement the water supply to the requesting party. The requesting party updates the local time difference table according to the arrival of the supplemented water flow and synchronizes the time difference table entries to the main assisting party. The time difference table is used to characterize the water supply response time from the requesting party to each control terminal. S4. When the requesting end needs assistance later, it should prioritize selecting the assisting party based on the local time difference table, and if the local time difference table is unavailable, it should perform a multi-hop extended search through the neighbor table.

[0007] Furthermore, in step S1, the determination that the local water supply capacity does not meet the water output demand includes any of the following situations: Scenario 1: The actual flow rate of any open outlet is lower than the set threshold for the requested flow rate of that outlet; Scenario 2: The total requested flow rate of all open outlets exceeds the total water output limit of this control terminal; when Scenario 1 and Scenario 2 are triggered simultaneously, the flow gap value under the two rules is calculated separately, and the larger of the two values ​​is taken as the final flow gap value; if the final flow gap value is less than or equal to zero, the coordinated scheduling is not triggered.

[0008] Furthermore, in step S1, the calculation of the flow gap includes: if scenario one is triggered, the difference between the requested flow and the actual flow of each outlet with insufficient actual flow is accumulated; if scenario two is triggered, the total upper limit of the total water output of this control terminal is subtracted from the sum of the requested flow of all open outlets; if scenario one and scenario two are triggered simultaneously, the larger value of the results obtained by the above two calculation methods is taken.

[0009] Furthermore, in step S2, the requesting control terminal broadcasts an assistance request message carrying the flow gap value and request timestamp to the entire network and starts a response waiting timer; after receiving the broadcast, the other control terminals compare their remaining available water volume with the flow gap value. If the remaining available water volume is sufficient to meet the gap value, they reply with a non-overload response; otherwise, they reply with an overload response. The requesting terminal records the arrival time of each response and calculates the response delay. After receiving a set number of non-overload responses or the response waiting timer expires, it proceeds to the main assisting party determination step.

[0010] Furthermore, in step S2, determining the primary assisting party includes: the requesting end sorts the non-overloaded response records in ascending order of response delay, and sends confirmation assistance messages to the sorted control ends in sequence; after receiving the confirmation assistance message, the selected control end re-checks whether its remaining available water volume is still sufficient to meet the flow gap value. If so, it replies that it agrees to assist and deducts the corresponding available remaining flow; otherwise, it replies that it refuses to assist; when the requesting end receives a refusal to assist, it selects the next non-overloaded control end for confirmation and saves the other non-overloaded response records as alternative assisting parties.

[0011] Furthermore, in step S3, the requesting end sends a start-up water discharge command carrying the target water discharge flow rate and start-up timestamp to the main assisting party, and records the local start-up time; the main assisting party starts the local water pump and delivers the supplementary water to the requesting end's pipeline network; the requesting end detects the arrival time of the supplementary water flow, calculates the actual time difference of the water flow and updates the local time difference table, and at the same time synchronizes the time difference table entry corresponding to the main assisting party to the main assisting party.

[0012] Furthermore, in step S3, when the requesting end synchronizes the time difference table entries with the primary assisting party, it attaches a request neighbor table identifier; after receiving it, the primary assisting party replies to the requesting end with the entries in its local time difference table excluding the requesting end; the requesting end adds the time difference in the reply entries to the measured time difference from the requesting end to the primary assisting party to obtain the multi-hop time difference estimate and stores it in the local time difference table, sorted by time difference; if a direct measured entry for the same target control end already exists locally, the direct measured value is retained instead of being overwritten by the estimated value.

[0013] Furthermore, in step S4, the requesting end queries its local time difference table and excludes the control ends currently assisting it, then unicasts a quick assistance request to the target control end in ascending order of time difference; the target control end checks its remaining available water supply, and if it is sufficient to meet the flow gap value, it agrees to assist and starts the water pump, otherwise it refuses to assist; after receiving the refusal, the requesting end marks the target control end as temporarily unavailable and selects the target control end with the second smallest time difference, and the temporarily unavailable mark is restored after the timer expires; if all available targets in the local time difference table fail to be reached, the requesting end rolls back to perform broadcast discovery.

[0014] Further, in step S4, the multi-hop extended search includes: the requesting end sending a request neighbor table message to the target control end that has been tried but is unavailable; the target control end sending an overload status detection message to the neighbor control ends in its local time difference table; if the remaining available water volume meets the flow gap value of the requesting end, the neighbor control end replies with a non-overload confirmation and locks itself; otherwise, it replies with an overload confirmation; the target control end replies with the non-overload confirmation entries to the requesting end; the requesting end adds the time difference in the entries to the time difference from the requesting end to the target control end to obtain a multi-hop time difference estimate and adds it to the candidate set, sorts it by the accumulated time difference, and attempts to quickly assist requests in sequence; if the assistance is successful, the measured time difference is measured and updated; if it fails, it falls back to execute the broadcast discovery.

[0015] This invention also discloses a distributed scheduling system for intelligent irrigation, applied to an irrigation system containing multiple irrigation service nodes. Each node includes a card reader terminal, a water outlet, a control terminal, and a water pump. Each control terminal is connected to the same broadcast domain and maintains a local time difference table. The control terminal is configured to execute the distributed scheduling method for intelligent irrigation. Furthermore, the control terminal is also configured to prioritize local water supply when receiving a new local irrigation task while acting as an assistant, and to send an assistance capacity change message to the requesting terminal when the remaining capacity is insufficient. The requesting terminal adjusts the water supply or recalculates the flow gap and requests other control terminals based on the assistance capacity change message.

[0016] The distributed scheduling method and system for intelligent irrigation of the present invention have the following advantages:

[0017] 1. Rapid Cooperative Scheduling: Through broadcast discovery and response delay sorting mechanism, the requesting end can quickly select the optimal target from multiple candidate assisting parties, which greatly shortens the decision time of cooperative scheduling and improves the response speed of irrigation systems in hilly areas.

[0018] 2. Multi-hop coverage expansion: By synchronizing table entries between target control terminals, each control terminal can obtain the estimated time difference of multi-hop neighbors, expanding the coverage of candidate assistants and improving the scheduling flexibility of the system in complex pipeline network environments.

[0019] 3. Localized learning optimization: Through the progressive learning and multi-hop extended search mechanism of the "target control terminal - water flow time difference" table, the control terminal can gradually accumulate water supply response speed data of each target control terminal without relying on global topology information, realize localized optimal scheduling decision, reduce dependence on broadcast requests, and reduce network overhead.

[0020] 4. Dynamic Robust Response: Through the assistance capacity change handling mechanism, when the assisting party's water supply capacity changes due to new local tasks, the requesting party can promptly detect and readjust the scheduling strategy to ensure that the irrigation task is not interrupted, thus enhancing the system's robustness in dynamic operating environments. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the system deployment of the intelligent irrigation distributed scheduling method of the present invention. Detailed Implementation

[0022] To better understand the purpose, structure, and function of this invention, the following detailed description of a distributed scheduling method and system for intelligent irrigation, in conjunction with the accompanying drawings, is provided.

[0023] System deployment and prerequisites:

[0024] like Figure 1As shown, this embodiment deploys multiple irrigation service nodes within a hilly irrigation area for smart agriculture. Each node includes a card reader terminal 10, a water outlet 20, a control terminal 30, and a water pump 40. The control terminal 30 connects the local water pump 40 to the solenoid valve 50 of the water outlet 20 via a 4G network, and has functions for flow monitoring, broadcast communication, and water pump frequency conversion adjustment. Each control terminal 30 accesses the same broadcast domain via an irrigation IoT network 60 (such as Wi-Fi or wired industrial Ethernet), and the network supports broadcast message forwarding.

[0025] In this embodiment, each node is identified by a single letter (e.g., nodes A, B, C, D, E), the corresponding control terminal is labeled 30A, 30B, 30C, 30D, 30E, the water pump is labeled 40A, 40B, 40C, 40D, 40E, the outlet is labeled 20A, 20B, 20C, 20D, 20E, and the card reader is labeled 10A, 10B, 10C, 10D, 10E.

[0026] Each control terminal 30 is pre-set with the rated maximum flow rate Q_rate of its local water pump 40, and collects the requested and actual flow rates of each outlet 20 in real time. Each control terminal sets a maximum total water output limit, which by default is the same as the rated maximum flow rate Q_rate of the local water pump. This limit is the maximum sum of water that the control terminal 30 can provide to each outlet per unit time. Each control terminal 30 maintains its remaining available water volume Q_avail, defined as Q_rate minus the current actual output flow rate (i.e., the sum of the actual flow rates of all currently open outlets). Each control terminal 30 locally maintains a "Target Control Terminal - Water Flow Time Difference" table, which is initially empty and dynamically updated and sorted as the number of coordinated irrigation cycles increases.

[0027] The present invention provides a distributed scheduling method for intelligent irrigation, comprising the following steps:

[0028] Step 1: Determining Insufficient Water Supply Capacity at the Water Demand Initiation and Control End

[0029] When farmers pay by swiping their cards or scanning QR codes at card-swiping terminal 10A (or card-swiping terminal 10A' or card-swiping terminal 10A''), card-swiping terminal 10A sends a water discharge request message to the connected control terminal 30A. The message carries the terminal identifier (such as 10A or 10A') and the requested flow rate parameters (such as "rated flow rate 2 cubic meters / hour"). The water discharge rate of each outlet (such as 20A, 20A', and 20A'') is precisely controlled by the control terminal 30A through a solenoid valve.

[0030] After receiving a water discharge request, the control terminal 30A starts the local water pump 40A to pump water and supplies water to the farmland through the outlets (such as 20A, 20A', or 20A''). The control terminal 30A monitors in real time the requested flow rate value of each outlet, the flow meter reading (actual flow rate value) of the outlet, and the total requested flow rate of all open outlets.

[0031] If the control terminal 30A determines that the local water pump's pumping capacity is insufficient under any of the following circumstances, it will trigger the collaborative scheduling process:

[0032] Scenario 1: If the actual flow rate of any open outlet is lower than the set threshold n (configurable, default value is 80%) of the requested flow rate of that outlet, it indicates that the outlet is not supplying enough water.

[0033] Scenario 2: The total flow rate requested by all open outlets exceeds the 30A total water output limit Q_rate of this control terminal, indicating that the simultaneous opening of multiple outlets has caused the total demand to exceed the limit.

[0034] If both scenario one and scenario two are triggered simultaneously, the control terminal 30A calculates the flow gap value under the two rules respectively, and takes the larger value as the final flow gap value ΔQ.

[0035] Step 2: The control terminal initiates a broadcast request for water discharge assistance and collects responses.

[0036] Step 2.1: Based on the specific circumstances triggered in Step 1, the control terminal 30A calculates the current required flow gap value ΔQ (unit: cubic meters per hour). The calculation rules are as follows:

[0037] If scenario one is triggered (the actual flow rate at a certain outlet is lower than the requested flow rate threshold), the control terminal 30A iterates through all the open outlets it is connected to, calculates the shortfall value (requested flow rate - actual flow rate) for each outlet with insufficient actual flow, and adds up all the shortfalls, i.e.: ΔQ = Σ(requested flow rate of the outlet - actual flow rate of the outlet). If multiple outlets have insufficient flow at the same time, all shortfalls are added up.

[0038] If scenario two is triggered (total requested flow exceeds the maximum total water output of this control terminal), control terminal 30A calculates: ΔQ = Σ(request flow of all open water outlets) - Q_rate.

[0039] If both scenario one and scenario two are triggered simultaneously, control terminal A will calculate the gap value according to the two rules mentioned above, and take the larger value as the final ΔQ to ensure that the gap is fully filled. If the actual calculated ΔQ ≤ 0, control terminal 30A will not trigger a coordination request (indicating that the current water supply capacity can actually meet the demand).

[0040] Calculation Example: Control terminal 30A requests 2.0 m³ / h from outlet 20A and 1.0 m³ / h from outlet 20A'. The local pump's Q_rate = 2.0 m³ / h. In actual operation, due to the limited capacity of pump 40A, control terminal 30A distributes the flow according to the requested ratio: outlet 20A's measured flow is 1.2 m³ / h, and outlet 20A''s measured flow is 0.8 m³ / h (total measured flow = 2.0 m³ / h, equal to Q_rate). Scenario 1: Outlet 20A shortfall = 2.0 - 1.2 = 0.8 m³ / h (below 80% of the threshold, i.e., below 1.6 m³ / h, triggering); Outlet 20A' shortfall = 1.0 - 0.8 = 0.2 m³ / h (exactly reaches 80% of the threshold, i.e., 0.8 m³ / h, not triggering); Total flow shortfall in Scenario 1 = 0.8 m³ / h. Scenario 2: Total demand = 2.0 + 1.0 = 3.0 m³ / h, Shortfall = 3.0 - 2.0 = 1.0 m³ / h. Scenario 1 and Scenario 2 are triggered simultaneously, so the larger of the total flow shortfall is taken as ΔQ, and ultimately ΔQ = 1.0 m³ / h.

[0041] Step 2.2: The control terminal 30A broadcasts a "Water Outflow Assistance Request" message (destination address is the broadcast address) to the entire network. The message carries the device ID of the control terminal 30A, the final calculated flow gap value ΔQ, the request sequence number, the timestamp T0, and the assistance request tag (set to 1). At the same time, the control terminal 30A starts a response wait timer T_wait (configurable, default 3 seconds) as a timeout protection.

[0042] Step 2.3: After receiving the broadcast "Water Supply Assistance Request" message from controller 30A, the other controllers in the network (such as 30B, 30C, 30D, and 30E) check whether their current remaining available water volume Q_avail is sufficient to meet ΔQ. If Q_avail ≥ ΔQ, the "Overload Status" tag is set to 0 (indicating no overload), and the controller includes its own device ID and the current Q_avail value in the response message, and immediately unicasts a reply to controller 30A; if Q_avail < ΔQ, it is set to 1 (indicating overload), and the controller also includes its own device ID in the reply to controller 30A.

[0043] Step 2.4: After sending the "Water Discharge Assistance Request" message, the control terminal 30A initiates the response collection process. The control terminal 30A receives unicast response messages from each target control terminal in real time, storing each received response record in a local temporary "Response Control Terminal Record Table." When a total of three non-overloaded (overload status tag set to 0) response records are received, the process immediately proceeds to Step 3 to begin assistor selection. Simultaneously, a response waiting timer T_wait (configurable, default 3 seconds) serves as timeout protection—if the timer expires first and the total number of non-overloaded responses is less than three, the selection process proceeds to Step 3 based on the actual collected response data. The control terminal 30A records the arrival time T1 of each response message and calculates the response delay Δt = T1 - T0 (this delay value reflects the network hop count and physical distance between the control terminals to some extent; the closer the distance, the smaller the response delay, and the closer the water pipe connection). All received response message information is temporarily stored in a local temporary cache "Response Control Terminal Record Table." Each record includes: the response control terminal identifier (e.g., 30B, 30C, 30D, 30E), the overload status tag (0 or 1), and the response delay Δt. Table 1 structure is as follows (example):

[0044] Table 1: "Response Control Terminal Record Table" of Control Terminal 30A

[0045]

[0046] Step 3: Select the assisting party, establish the local "Target Control Terminal - Water Flow Time Difference" table, and synchronize the table data with the assisting party.

[0047] Step 3.1: In step 2.4, controller 30A has already stored the received non-overload response records in the "Response Control Terminal Record Table". Now, the non-overload records in this table are sorted in ascending order of response delay Δt. Controller 30A selects the first record (the one with the smallest Δt) from the sorted records, such as controller 30B, and unicasts a "Confirm Assistance" message to controller 30B. The message carries the flow gap value ΔQ, the requesting pump ID, and the confirmation assistance tag (set to 1). If there are no non-overload records in the "Response Control Terminal Record Table" (i.e., no non-overload response is received before the timer expires), the assistance request fails, and the original retry logic is followed (starting a timer with a period of M; after M expires, step 2.2 is repeated, with a maximum of 3 retries).

[0048] Step 3.2: After receiving the "Confirm Assistance" message, Control Terminal 30B immediately checks whether it is still in an unoverloaded state (i.e., it still has enough remaining capacity to meet ΔQ, because after responding in Step 2 and before receiving the "Confirm Assistance" message, Control Terminal 30B may have been selected and had some capacity occupied by other requesters). If Control Terminal 30B is still in an unoverloaded state, it replies with a "Agree to Assistance" confirmation message to Control Terminal 30A and deducts ΔQ from its available remaining traffic (i.e., the remaining capacity is reduced by ΔQ). However, Control Terminal 30B can still continue to accept assistance requests from other control terminals (as long as the remaining capacity can still meet the other party's ΔQ). If Control Terminal 30B has become overloaded (the remaining capacity is insufficient to meet ΔQ), it replies with a "Reject Assistance" message, carrying the current status information in the message.

[0049] Step 3.3: When controller 30A receives a reply message from controller 30B, if it says "Agree to assist," then controller 30B is officially designated as the primary assisting party, and the process proceeds to step 5. If it says "Reject assist," then controller 30A deletes controller 30B's record from the local temporary cache "Response Controller Record Table," selects the controller with the second-lowest response latency that is not overloaded (such as 30C), and repeats the confirmation process of steps 3.2 to 3.3 above. If all candidate controllers are rejected, then this assistance request fails, and controller 30A performs retry logic processing according to step 3.1.

[0050] Step 3.4: After the primary assisting party is confirmed, control unit 30A performs recording and backup operations. Control unit 30A updates the local "Response Control Unit Record Table" (the "Whether it has been selected as the primary assisting party" field has a value of 0 indicating it is currently a candidate, 1 indicating it is currently selected, and 2 indicating it was previously selected but is not currently the assisting party), as shown in Table 2:

[0051] Table 2: "Response Control Terminal Record Table" of Control Terminal 30A

[0052]

[0053] From the sorted response records, control terminal 30A saves all non-overloaded response records (except for the first record selected as the primary assister) as alternative assisters (with the "Whether it has been selected as the primary assister" field set to 0) to the local "Response Control Terminal Record Table". These alternative records will be used in the round-robin scheduling in subsequent step 5. If all alternatives are exhausted, control terminal 30A re-executes the broadcast discovery in step 2.

[0054] Step 3.5: Start coordinated water output from control terminal 30A and measure the "water flow time difference":

[0055] Control terminal 30A sends a "start water discharge" command message to the main assisting control terminal 30B. The message includes the target water discharge flow rate (ΔQ) and the start timestamp T2. After receiving the command, control terminal 30B starts the local water pump 40B and delivers the supplementary water to the irrigation network under the jurisdiction of control terminal 30A through the irrigation network.

[0056] While sending the "Start Water Discharge" command, control terminal 30A records local time T2. When the flow meter at the inlet of the common water pipe connected to pump 40A detects the arrival of supplementary water flow from the assisting party (pump 40B) (the flow meter reading shows a step increase of approximately ΔQ—the error is configurable, such as within ±3%), time T3 is recorded. Control terminal 30A calculates the actual measured time difference of water flow for this collaboration, ΔT_AB = T3 - T2. This value reflects the sum of the physical pipeline transmission delay and the pump response delay from issuing the start command to actually receiving the supplementary water flow. Control terminal A saves the data to the local "Target Control Terminal - Water Flow Time Difference" table, as shown in Table 3:

[0057] Table 3: "Target Control Terminal - Water Flow Time Difference Table" for Control Terminal 30A

[0058]

[0059] The controller 30A sends the data record "Target controller 30B, water flow time difference = ΔT_AB" from the "Target controller - water flow time difference" table to the controller 30B via a unicast "time difference synchronization message". After receiving the message, the controller 30B saves it to its local "Target controller - water flow time difference" table, as shown in Table 4: (Note: If the controller 30B already has an entry for the controller 30A, it will update it with the new ΔT value; otherwise, it will add it).

[0060] Table 4: "Target Control Terminal - Water Flow Time Difference Table" for Control Terminal 30B

[0061]

[0062] Step 3.6: While sending a "time difference synchronization message" to control terminal 30B, control terminal 30A appends a "request neighbor table" tag (set to 1) to the message, requesting control terminal 30B to send all entries in its local "target control terminal - water flow time difference" table, excluding those of control terminal 30A, to control terminal 30A. Upon receiving this request, control terminal 30B includes the corresponding table entry data in its reply.

[0063] After receiving the table entry from control terminal 30B, control terminal 30A adds the time difference in each record to the measured ΔT_AB from control terminal 30A to control terminal 30B, obtaining the estimated multi-hop time difference (ΔT_AC = ΔT_AB + ΔT_BC) from control terminal 30A to its neighbor (e.g., control terminal 30C). This estimated value is stored in the local time difference table, sorted by ΔT from smallest to largest, and the default availability status is "available". If control terminal 30A already has an entry for the target control terminal locally (e.g., control terminal 30A has previously collaborated directly with control terminal 30C), control terminal 30A should retain its directly measured data without overwriting it, as shown in Table 5.

[0064] Table 5: "Target Control Terminal - Water Flow Time Difference Table" for Control Terminal 30A

[0065]

[0066] When an entry in the local time difference table of the control terminal 30B is added or the ΔT of the entry is updated, the control terminal 30B queries the local historical records and unicasts the updated entry data to all control terminals (such as 30A) that have requested the entry from it. The push scope is limited to the control terminals that have requested it and is not broadcast to the entire network.

[0067] Step 4: Coordinated Irrigation Execution and Status Recovery

[0068] Step 4.1: After receiving the "Start Water Discharge" command, the main assisting control terminal 30B starts the local water pump 40B. It adjusts the inverter output frequency according to the flow gap value ΔQ in the request message. Based on the water pump 40A information carried by the control terminal 30A in the "Confirm Assistance" message, it delivers the supplementary water to the water pump 40A through the irrigation network. After receiving the supplementary water, the control terminal 30A, based on the actual demand of each local outlet (such as the difference between the requested flow rate and the measured flow rate at outlet 20A and / or outlet 20A'), autonomously controls the opening of the corresponding solenoid valve to distribute the supplementary water to each outlet as needed.

[0069] Step 4.2: Control terminal 30A monitors the total water supply flow in real time (local water pump 40A flow + auxiliary supplement flow). If, during irrigation, a certain water outlet closes in advance, resulting in a reduction in total demand, but other water outlets are still irrigating, control terminal 30A immediately sends a "regulate flow" message to control terminal 30B (carrying the requester control terminal ID, request sequence number, new target flow value and regulation flow tag set to 1), and control terminal 30B reduces the water supply flow accordingly.

[0070] Step 4.3: During the collaborative water supply process, if the assisting control terminal 30B receives a new local irrigation task (such as a local farmer swiping a card to collect water), control terminal 30B first assesses whether its current remaining capacity (excluding the portion currently assisting control terminal 30A) is sufficient to meet local demand. If the remaining capacity is sufficient, the local water outlet is opened normally, without affecting assistance to control terminal 30A. If the remaining capacity is insufficient to meet local demand, control terminal 30B prioritizes local water supply (because control terminal 30B primarily serves local farmers, this is a business priority), and immediately sends an "Assistance Capacity Change" message (carrying the new actual available remaining capacity) to control terminal 30A. Upon receiving this message, control terminal 30A adjusts the allocated water supply to control terminal 30B or recalculates the gap accordingly, and initiates supplementary requests to other control terminals as needed. After the local task is completed and capacity is restored, control terminal 30B sends a capacity change notification to control terminal 30A again, allowing control terminal 30A to readjust water allocation.

[0071] Step 4.4: After the irrigation task is completed (i.e., all outlets are closed), control terminal 30A sends a "Stop Assistance" message to control terminal 30B (carrying the requesting control terminal ID, request sequence number, and stop assistance tag set to 1). Upon receiving the message, control terminal 30B restores its local remaining capacity to the amount of ΔQ used in this operation (i.e., the remaining capacity increases by ΔQ; if the actual usage differs from ΔQ, the actual usage will prevail) to continue providing services to other control terminals. Control terminal 30A writes the collaboration record (assisting device ID, flow gap value, success / failure flag) to its local log for querying and analysis by the agricultural management platform.

[0072] Step 5: Rapid scheduling and multi-hop extended search based on the "target control terminal - water flow time difference table"

[0073] Step 5.1: Rapid Scheduling: As the number of coordinated irrigation sessions increases, the local "Target Control Terminal - Water Flow Time Difference" table on each control terminal is gradually improved. Taking control terminal 30B as an example, its local table may contain items as shown in Table 6 (sorted by water flow time difference from smallest to largest):

[0074] Table 6: "Target Control Terminal - Water Flow Time Difference Table" for Control Terminal 30B

[0075]

[0076] When control terminal 30B subsequently detects that the local water pump's pumping capacity is insufficient (triggering condition is the same as in step 1.3), control terminal 30B first queries the local "Target Control Terminal - Water Flow Time Difference" table, excluding the control terminal currently being assisted by control terminal 30B (i.e., control terminal 30B is currently the assisting party and the requesting party is control terminal 30A). If the table is not empty and contains available entries, control terminal 30B skips the broadcast request discovery process in step 2, and then selects the target control terminal (such as 30C) with the smallest time difference (i.e., the closest physical distance and the fastest water supply response) as the assisting party according to the ascending order of the water flow time difference in the table.

[0077] Control terminal 30B unicasts a "Quick Assistance Request" message to control terminal 30C. The message carries the device ID of control terminal 30B, the calculated flow gap value ΔQ, the request sequence number, and the quick assistance request tag (set to 1). Upon receiving the message, control terminal 30C checks its current remaining capacity. If its remaining capacity is sufficient to meet ΔQ, it directly replies "Agree to assist" and starts pump 40C to supply water to pump 40B, while deducting its own corresponding capacity. If its current remaining capacity is insufficient to meet ΔQ (e.g., it is assisting another control terminal or its own pumping capacity is insufficient), control terminal 30C replies with a "Refuse to assist" message, carrying the current overload tag (set to 1) in the message. After receiving a rejection, controller 30B automatically marks controller 30C as temporarily unavailable in the local "Target Controller - Water Flow Time Difference" table, and selects the next neighbor with the smallest water flow time difference in the table (such as controller 30D) as a backup, repeating the above fast request process. The "temporarily unavailable" mark is automatically restored to "available" after the timer T_unavailable (configurable, default 5 minutes) expires. If all neighbors in the table refuse to assist, controller 30B returns to step 2, re-executes the broadcast discovery mechanism to discover new available idle controllers, and appends the newly discovered controllers and their measured time differences to the local table to keep the table dynamically updated.

[0078] Step 5.2: Extended Discovery: If all available records in the local "Target Control End - Water Flow Time Difference" table have been tried and are all unavailable, Control End 30B unicasts a "Request Neighbor Table" message to each of these unavailable target control ends (such as Control End 30C). The message carries the source device ID (30B), target device ID (30C), request neighbor table tag (set to 1), and the flow gap value ΔQ_B of Control End 30B, requesting its local available table entries (excluding Control End 30B itself). After receiving the request, Target Control End 30C iterates through the target control ends corresponding to each record in its local table and sends a unicast "Overload Status Detection Message" to each target control end (taking Control End 30D as an example). The message carries: source device ID (30C), target device ID (30D), overload status detection tag (set to 1), recommender device ID (30B), and the gap value ΔQ_B of 30B, querying Control End 30D whether its current remaining capacity is sufficient to meet ΔQ_B.

[0079] After receiving the "overload detection message," the target control terminal 30D checks its current Q_avail. If Q_avail ≥ ΔQ_B, it replies to the control terminal 30C with a "no overload confirmation" message. This message carries its own ID device (30D), current Q_avail value, and a "no overload" tag (set to 1). It also marks itself as "locked by control terminal 30B" and starts a lock timeout timer T_lock (configurable, default 10 seconds). During the lock period, control terminal 30D does not accept assistance requests from other control terminals. If Q_avail < ΔQ_B, it replies to control terminal 30C with an "overload" message.

[0080] After collecting all responses, controller 30C only sends the controller entries (i.e., the controller ID, corresponding ΔT, and current Q_avail) that reply "No overload confirmation" to controller 30B. Upon receiving the entries from controller 30C, controller 30B adds the time difference in each entry to the existing ΔT_BC from controller 30B to controller 30C to obtain an estimated multi-hop time difference from controller 30B to the target controller (e.g., ΔT_BD = ΔT_BC + ΔT_CD), and stores it in a local time difference table (sorted by the accumulated ΔT from smallest to largest). The availability status is "available" by default. If controller 30B already has a directly measured entry from controller 30D (i.e., controller 30B has previously requested water supply from controller 30D as the requesting party), the directly measured value is used, and the estimated value is not overwritten.

[0081] For target control terminals that are available and not overloaded in the response table, control terminal 30B continues to repeat the above request operation, requesting their respective neighbor tables to continuously expand the candidate set. Control terminal 30B sorts the newly added candidate set by cumulative ΔT in ascending order and sends unicast "quick assistance requests" sequentially. If a candidate control terminal is available, it requests water supply to it. After the water flow arrives, it measures and records the measured ΔT, overwriting the original ΔT record of that target control terminal in the table with the new value (keeping the available status "available"), and sends a "release lock" notification to all temporarily locked control terminals.

[0082] If all newly added candidate controllers are unavailable, controller 30B will fall back to step 2, re-execute the broadcast discovery mechanism to discover new available idle controllers, and append the newly discovered controllers and their measured time differences to the local table to keep the table dynamically updated.

[0083] Step 5.3: Each control terminal periodically (e.g., every 24 hours) or triggered (e.g., after pipeline maintenance) clears entries that have not been updated for more than 30 days, and remeasures by actively sending probe broadcast messages to ensure the timeliness of the table data.

[0084] It is understood that the present invention has been described through some embodiments, and those skilled in the art will recognize that various changes or equivalent substitutions can be made to these features and embodiments without departing from the spirit and scope of the invention. Furthermore, under the teachings of the present invention, these features and embodiments can be modified to adapt to specific situations and materials without departing from the spirit and scope of the invention. Therefore, the present invention is not limited to the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are within the protection scope of the present invention.

Claims

1. A distributed scheduling method for intelligent irrigation, applied to an irrigation system comprising multiple irrigation service nodes, each node including a card reader terminal, a water outlet, a control terminal, and a water pump, characterized in that, Includes the following steps: S1. After receiving a water discharge request, the control terminal starts the local water pump to supply water, and when the local water supply capacity does not meet the water discharge demand, it determines the flow gap and triggers coordinated scheduling. S2. The requesting end control end sends an assistance request to other control ends and determines the main assisting party based on the available water supply capacity and response delay fed back by other control ends. S3. The main assisting party starts the local water pump to supplement the water supply to the requesting party. The requesting party updates the local time difference table according to the arrival of the supplemented water flow and synchronizes the time difference table entries to the main assisting party. The time difference table is used to characterize the water supply response time from the requesting party to each control terminal. S4. When the requesting end needs assistance later, it should prioritize selecting the assisting party based on the local time difference table, and if the local time difference table is unavailable, it should perform a multi-hop extended search through the neighbor table.

2. The distributed scheduling method for intelligent irrigation according to claim 1, characterized in that, In step S1, the determination that the local water supply capacity does not meet the water output demand includes any of the following situations: Scenario 1: The actual flow rate of any open outlet is lower than the set threshold for the requested flow rate of that outlet; Scenario 2: The total requested flow rate of all open outlets exceeds the total water output limit of this control terminal; when Scenario 1 and Scenario 2 are triggered simultaneously, the flow gap value under the two rules is calculated separately, and the larger of the two values ​​is taken as the final flow gap value; if the final flow gap value is less than or equal to zero, the coordinated scheduling is not triggered.

3. The distributed scheduling method for intelligent irrigation according to claim 1, characterized in that, In step S1, the calculation of the flow gap includes: if scenario one is triggered, the difference between the requested flow and the actual flow of each outlet with insufficient actual flow is accumulated; if scenario two is triggered, the total upper limit of the total water output of this control terminal is subtracted from the sum of the requested flow of all open outlets; if scenario one and scenario two are triggered simultaneously, the larger value of the results obtained by the above two calculation methods is taken.

4. The distributed scheduling method for intelligent irrigation according to claim 1, characterized in that, In step S2, the requesting control terminal broadcasts an assistance request message carrying the flow gap value and request timestamp to the entire network and starts a response waiting timer. After receiving the broadcast, the other control terminals compare their remaining available water volume with the flow gap value. If the remaining available water volume is sufficient to meet the gap value, they reply with a non-overload response; otherwise, they reply with an overload response. The requesting terminal records the arrival time of each response and calculates the response delay. After receiving a set number of non-overload responses or the response waiting timer expires, it proceeds to the main assisting party determination step.

5. The distributed scheduling method for intelligent irrigation according to claim 1, characterized in that, In step S2, determining the primary assisting party includes: the requesting end sorts the non-overloaded response records in ascending order of response delay, and sends confirmation assistance messages to the sorted control ends in sequence; after receiving the confirmation assistance message, the selected control end re-checks whether its remaining available water volume is still sufficient to meet the flow gap value. If so, it replies that it agrees to assist and deducts the corresponding available remaining flow; otherwise, it replies that it refuses to assist; when the requesting end receives a refusal to assist, it selects the next non-overloaded control end for confirmation and saves the other non-overloaded response records as alternative assisting parties.

6. The distributed scheduling method for intelligent irrigation according to claim 1, characterized in that, In step S3, the requesting end sends a start-up water discharge command carrying the target water discharge flow rate and start-up timestamp to the main assisting party, and records the local start-up time; the main assisting party starts the local water pump and delivers the supplementary water to the requesting end's pipeline network; the requesting end detects the arrival time of the supplementary water flow, calculates the actual time difference of the water flow and updates the local time difference table, and at the same time synchronizes the time difference table entry corresponding to the main assisting party to the main assisting party.

7. The distributed scheduling method for intelligent irrigation according to claim 1, characterized in that, In step S3, when the requesting end synchronizes the time difference table entries with the primary assisting end, it attaches a request neighbor table identifier; after receiving the request, the primary assisting end replies to the requesting end with the entries in its local time difference table except for those of the requesting end; the requesting end adds the time difference in the reply entries to the measured time difference from the requesting end to the primary assisting end to obtain the estimated multi-hop time difference value and stores it in the local time difference table, sorted by time difference; if a direct measured entry for the same target control end already exists locally, the direct measured value is retained instead of being overwritten by the estimated value.

8. The distributed scheduling method for intelligent irrigation according to claim 1, characterized in that, In step S4, the requesting end queries its local time difference table and excludes the control ends currently assisting it. It then unicasts a quick assistance request to the target control end in ascending order of time difference. The target control end checks its remaining available water supply. If it is sufficient to meet the flow gap, it agrees to assist and starts the water pump; otherwise, it refuses to assist. After receiving the refusal, the requesting end marks the target control end as temporarily unavailable and selects the target control end with the second smallest time difference. The temporarily unavailable mark is restored after the timer expires. If all available targets in the local time difference table fail to be reached, the requesting end rolls back to perform a broadcast discovery.

9. The distributed scheduling method for intelligent irrigation according to claim 1, characterized in that, In step S4, the multi-hop extended search includes: the requesting end sending a request neighbor table message to the target control end that has been tried but is unavailable; the target control end sending an overload status detection message to the neighbor control ends in its local time difference table; if the remaining available water volume meets the flow gap value of the requesting end, the neighbor control end replies with a non-overload confirmation and locks itself; otherwise, it replies with an overload confirmation; the target control end replies with the non-overload confirmation entries to the requesting end; the requesting end adds the time difference in the entries to the time difference from the requesting end to the target control end to obtain a multi-hop time difference estimate and adds it to the candidate set, sorts it by the accumulated time difference, and attempts to quickly assist requests in sequence; if the assistance is successful, the measured time difference is measured and updated; if it fails, it falls back to execute broadcast discovery.

10. A distributed scheduling system for intelligent irrigation, applied to an irrigation system comprising multiple irrigation service nodes, each node including a card reader terminal, a water outlet, a control terminal, and a water pump, wherein each control terminal is connected to the same broadcast domain and maintains a local time difference table, characterized in that, The control terminal is configured to execute the distributed scheduling method for intelligent irrigation as described in any one of claims 1 to 9; and the control terminal is further configured to prioritize local water supply when receiving a new local irrigation task while acting as an assistant, and to send an assistance capacity change message to the requesting terminal when the remaining capacity is insufficient, so that the requesting terminal can adjust the water supply or recalculate the flow gap and request other control terminals according to the assistance capacity change message.