Intelligent garden maintenance management system and method based on Internet of Things

Abstract

The invention relates to the technical field of garden management, and discloses a smart garden maintenance management system based on Internet of Things, which comprises a multi-mode sensing layer, an edge intelligent layer, a cloud management layer and a collaborative execution layer, and is characterized in that the multi-mode sensing layer is in communication connection with the edge intelligent layer, and the edge intelligent layer is in communication connection with the cloud management layer. According to the intelligent garden maintenance management system and method based on the Internet of Things, comprehensive, real-time and accurate monitoring of a garden environment and a plant growth state is realized by constructing a complete architecture integrating a multi-mode sensing layer, an edge intelligent layer, a cloud management layer and a collaborative execution layer, and the intelligent garden maintenance management system and method based on the Internet of Things realize comprehensive, real-time and accurate monitoring of the garden environment and the plant growth state through an innovative edge-cloud collaborative decision-making mechanism. The instant response capability of the edge side in an extreme scene (such as sudden drought) is ensured, big data analysis and long-term strategy optimization are carried out by using the powerful computing capability of the cloud, and efficient balance between response speed and decision intelligence is realized.

Description

Technical Field

[0001] This invention relates to the field of garden management technology, specifically to an intelligent garden maintenance management system and method based on the Internet of Things. Background Technology

[0002] Garden maintenance and management refers to the comprehensive, meticulous maintenance and overall management of vegetation, supporting facilities, and ecological environment in various garden spaces such as urban parks, community green spaces, and roadside landscape belts. Specifically, it encompasses daily care of plants, including irrigation, fertilization, pruning, pest and disease control according to the growth habits of different species, as well as seasonal replanting of seedlings and removal of dead branches, to maintain the healthy growth of vegetation and the layering of the landscape; it also involves the operation and maintenance of facilities and equipment, such as regular inspection, cleaning, and renewal of garden walkways, benches, lighting, and sprinkler systems to ensure their safety and usability; and it also includes the dynamic maintenance of the ecological environment, such as regulating soil moisture in green spaces, improving the microclimate, regulating visitor activity areas, and maintaining order in the garden. The core goal of this series of tasks is to continuously optimize the functions of ecological regulation and leisure services while maintaining the aesthetics and cleanliness of the garden landscape, ultimately achieving the long-term livability and sustainable operation of the garden space.

[0003] However, existing garden maintenance and management models generally suffer from the following core pain points: First, monitoring dimensions are singular and lagging. Most systems only focus on basic environmental parameters such as soil temperature and humidity, lacking multi-dimensional collaborative perception of plant physiological status and pest and disease characteristics, resulting in a lack of comprehensive data support for maintenance decisions. Second, data processing and decision response are disconnected. Traditional systems either rely on centralized cloud computing, leading to response delays in extreme scenarios, or only have simple local threshold judgment capabilities, failing to achieve a balance between long-term optimization and immediate response. Third, equipment coordination is poor. Irrigation, fertilization, pruning, and spraying equipment mostly operate independently, lacking a unified scheduling mechanism, which easily leads to overlapping or omissions in operations, resulting in resource waste. Fourth, management processes are not closed-loop, maintenance records are fragmented, making it impossible to achieve full life-cycle traceability of plant growth and dynamic iteration of maintenance plans, making it difficult to adapt to the needs of large-scale and diversified garden management. Summary of the Invention

[0004] (a) Technical problems to be solved To address the shortcomings of existing technologies, this invention provides an IoT-based smart garden maintenance management system and method, which has the advantages of intelligence, precision, and closed-loop operation, and solves the problems of incomplete monitoring dimensions, unbalanced decision-making and response, insufficient equipment coordination, and lack of closed-loop management processes in existing technologies.

[0005] (II) Technical Solution To achieve the aforementioned goals of intelligence, precision, and closed-loop operation, this invention provides the following technical solution: an IoT-based smart garden maintenance and management system, comprising a multimodal perception layer, an edge intelligence layer, a cloud management layer, and a collaborative execution layer. The multimodal perception layer is communicatively connected to the edge intelligence layer, the edge intelligence layer is communicatively connected to the cloud management layer, and the edge intelligence layer and the cloud management layer are communicatively connected to the collaborative execution layer. Furthermore, the edge intelligence layer and the cloud management layer work collaboratively to form an edge-cloud collaborative decision-making mechanism. The multimodal perception layer includes a core perception unit and an auxiliary perception unit; The edge intelligence layer includes a regional edge gateway, a data preprocessing module, a local decision engine, and a breakpoint resume unit. The cloud management layer includes an intelligent fusion analysis module, a global scheduling module, and a full lifecycle traceability module; The collaborative execution layer includes intelligent irrigation equipment, integrated water and fertilizer equipment, automatic pruning equipment, mobile spraying equipment, and mobile maintenance terminals.

[0006] Preferably, the system supports dual control mode switching: under normal operating conditions, it adopts a mode in which centralized scheduling is performed by the global scheduling module; when the network is interrupted or the edge intelligence layer determines that it is an emergency scenario, it automatically switches to a local control mode triggered by the local decision engine.

[0007] Preferably, the core sensing unit integrates a soil temperature and humidity sensor, a soil fertility sensor, a plant stem flow sensor, an air temperature and humidity sensor, a light sensor, and a pest and disease monitoring module; the pest and disease monitoring module integrates a biosensor, an odor sensor, and a high-definition night vision camera.

[0008] Preferably, the data preprocessing module is used to perform noise reduction, normalization and outlier removal on the sensed data; the local decision engine is preset with an emergency threshold for environmental parameters, which is used to trigger immediate maintenance commands in extreme scenarios; the breakpoint resume unit is used to store data locally when the network is interrupted and to complete data retransmission after the network is restored.

[0009] Preferably, the intelligent fusion analysis module has a built-in library of multiple plant growth models and performs fusion analysis on multimodal data based on machine learning algorithms to dynamically generate personalized maintenance plans; the full life cycle traceability module is used to record monitoring data, maintenance operations, and changes in growth status of plants from planting to maturity, forming a traceable data chain; the global scheduling module adopts A... Algorithm optimizes job paths and prioritizes equipment collaboration.

[0010] Preferably, the devices in the collaborative execution layer are equipped with a positioning module, obstacle avoidance sensor and status feedback unit, and support short-range collaborative communication between devices; the intelligent irrigation equipment adopts a dual-mode design of drip irrigation and sprinkler irrigation, which can automatically switch according to plant type and soil moisture distribution.

[0011] Preferably, the sensing device of the multimodal sensing layer supports switching between at least two communication modes, namely LoRa, WiFi or 5G, and is powered by solar energy with a protection level of not less than IP67.

[0012] Preferably, the cloud management layer further includes a visual management platform and a fault early warning module; the visual management platform supports multi-terminal login and real-time mapping of 3D modeling of garden areas and equipment status; the fault early warning module is used to monitor the operating parameters of execution layer equipment in real time and issue an alarm when abnormalities occur.

[0013] A method for smart garden maintenance and management based on the Internet of Things (IoT) includes the following steps: S1: System Deployment and Initialization: Divide the garden into grids according to the functional zones, deploy the perception layer, edge layer and execution layer devices, and input plant growth parameters and maintenance rules into the cloud platform to build an initial maintenance rule library and plant growth model; S2: Multimodal data acquisition and preprocessing: Multi-dimensional environmental and plant data are acquired through the perception layer, and after preprocessing by the edge gateway, the effective data is synchronized to local storage and cloud database; S3: Edge-Cloud Collaborative Decision Making: The cloud-based intelligent fusion analysis module generates personalized maintenance plans based on real-time and historical data and by calling plant growth models. If the edge gateway detects that the data has reached the emergency threshold, the local decision engine directly triggers an emergency command and reports it to the cloud. The preset priority order is: pest and disease outbreak > extreme temperature > soil drought. The local decision engine triggers commands according to priority. If there is a conflict, the low-priority operation is suspended and the high-priority task is executed first. The conflict log is synchronously reported to the cloud. S4: Collaborative Operation Execution and Status Feedback: The cloud-based global scheduling module converts the maintenance plan into operation instructions and optimizes the path before sending them to the execution layer devices. The devices then work collaboratively and provide real-time feedback on their status information. S5: Full life cycle closed-loop optimization: The full life cycle traceability module archives data from the entire maintenance process, regularly evaluates the maintenance effect, and dynamically optimizes plant growth model parameters and maintenance plans to achieve closed-loop management of "monitoring-decision-execution-evaluation-optimization".

[0014] (III) Beneficial Effects Compared with existing technologies, the present invention provides an Internet of Things-based smart garden maintenance management system and method, which has the following beneficial effects: 1. This IoT-based smart garden maintenance management system and method, through the construction of a complete architecture integrating a multimodal perception layer, an edge intelligence layer, a cloud management layer, and a collaborative execution layer, achieves comprehensive, real-time, and accurate monitoring of the garden environment and plant growth status.

[0015] 2. This IoT-based smart garden maintenance management system and method, through an innovative edge-cloud collaborative decision-making mechanism, not only ensures the immediate response capability on the edge side in extreme scenarios (such as sudden drought), but also utilizes the powerful computing capabilities of the cloud for big data analysis and long-term strategy optimization, achieving an efficient balance between response speed and intelligent decision-making. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the overall system framework of the present invention; Figure 2 This is a schematic diagram of the method steps of the present invention; Figure 3 This is a schematic block diagram illustrating the system operation flow of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Please see Figure 1-3 A smart garden maintenance and management system based on the Internet of Things (IoT) includes a multimodal perception layer, an edge intelligence layer, a cloud management layer, and a collaborative execution layer. The multimodal perception layer communicates with the edge intelligence layer, the edge intelligence layer communicates with the cloud management layer, and the edge intelligence layer and the cloud management layer communicate with the collaborative execution layer. The edge intelligence layer and the cloud management layer work together to form an edge-cloud collaborative decision-making mechanism. The data synchronization frequency between the edge intelligence layer and the cloud management layer is once every 5 minutes. After an emergency scenario is triggered, the cloud completes data verification and solution iteration within 10 minutes after the network is restored. The multimodal sensing layer includes a core sensing unit and an auxiliary sensing unit. The core sensing unit collects soil / air data once per hour, while the pest and disease monitoring module collects data in real time. The auxiliary sensing unit serves as a redundant backup, collecting data once every two hours and cross-validating it with the core unit data to comprehensively collect data on the garden environment and the plants themselves. The edge intelligence layer includes a regional edge gateway, a data preprocessing module, a local decision engine, and a breakpoint resume unit, which are used to realize data preprocessing and local real-time intelligent decision-making. The cloud-based management layer includes an intelligent fusion analysis module, a global scheduling module, and a full lifecycle traceability module, which are used for global intelligent analysis, scheduling, and management. The collaborative execution layer includes intelligent irrigation equipment, integrated water and fertilizer equipment, automatic pruning equipment, mobile spraying equipment, and mobile maintenance terminals, which are used to accurately and collaboratively execute maintenance instructions and provide status feedback.

[0019] In the implementation of the case, the system supports switching between two control modes: under normal working conditions, it adopts a mode in which the global scheduling module performs centralized scheduling; when the network is interrupted or the edge intelligence layer determines that it is an emergency scenario, it automatically switches to a local control mode triggered by the local decision engine. Among them, the centralized scheduling mode takes the global perspective of the cloud management layer as the core, relies on the multimodal real-time data and historical data integrated by the intelligent fusion analysis module, and combines the built-in plant growth model library to generate personalized maintenance strategies covering the entire garden. Then, the global scheduling module optimizes the equipment operation path and allocates task priorities to achieve efficient collaborative operation of multiple devices. The local control mode focuses on the instant response to extreme scenarios. The local decision engine pre-stores the emergency thresholds of key environmental parameters such as soil drought and pest outbreaks, and can quickly trigger maintenance actions without waiting for cloud instructions. Through the breakpoint resume unit of the edge intelligence layer and the data synchronization mechanism of the cloud management layer, the data continuity of dual-mode switching is ensured—when the network is interrupted, the collected data and work records are stored locally, and automatically re-uploaded to the cloud after the network is restored, ensuring the integrity of the data chain of the full life cycle traceability module; at the same time, through the device's built-in status feedback unit, the work effect and device operating status are fed back to the corresponding control module in real time. In the centralized scheduling mode, it supports the dynamic optimization scheme of the cloud, and in the local control mode, it provides a basis for emergency command adjustment, realizing seamless connection and stable operation of the two modes.

[0020] In the implementation of the case, the core sensing unit integrates soil temperature and humidity sensors, soil fertility sensors, plant stem flow sensors, air temperature and humidity sensors, light sensors, and a pest and disease monitoring module; the pest and disease monitoring module integrates biosensors, odor sensors, and high-definition night vision cameras, among which; Among them, each sensor and monitoring module has a clear division of labor and complementary data: the soil temperature, humidity and fertility sensor focuses on the core indicators of the soil environment, accurately capturing the soil nutrients and water basis for plant growth; the plant stem flow sensor directly measures the water transmission status of plants, reflecting their real-time physiological needs; the air temperature, humidity and light sensor records external climate conditions, providing a basis for assessing the growth environment; the pest and disease monitoring module uses biosensors to identify specific biological signals of pests and diseases, odor sensors to capture metabolic gases of pests and diseases, and high-definition night vision cameras to achieve all-weather visual monitoring, ensuring early detection and identification of pests and diseases from three dimensions; Through multimodal data synchronous acquisition and collaborative sensing mechanisms, the core sensing unit can simultaneously acquire four major categories of key data: soil, plants, air, and pests and diseases, avoiding the limitations of single-dimensional monitoring. Combined with the sensing device's ability to switch between multiple communication modes such as LoRa, WiFi, or 5G, as well as its IP67-level protection and solar power supply design, it ensures the stability, comprehensiveness, and real-time performance of data acquisition in complex garden environments, providing high-quality, multi-dimensional raw data support for subsequent edge-side preprocessing and cloud-based fusion analysis.

[0021] In the case implementation, the data preprocessing module is used to perform noise reduction, normalization and outlier removal on the sensed data; the local decision engine is preset with emergency thresholds for environmental parameters to trigger immediate maintenance instructions in extreme scenarios; the breakpoint resume unit is used to store data locally when the network is interrupted and to complete data retransmission after the network is restored. The specific emergency thresholds for environmental parameters are: soil moisture ≤15%, pest and disease biological signal intensity ≥80, and air temperature ≥38℃ or ≤-5℃. Furthermore, the three processing logics of the data preprocessing module each have their own focus: noise reduction can filter out noise data generated by sensor interference from the environment; normalization can map sensing data of different dimensions (such as soil moisture percentage, light intensity in lux) to the same numerical range; and outlier removal can identify and remove erroneous data that exceeds the reasonable range (such as extreme values ​​caused by sensor failure) through algorithms. The three work together to ensure the validity of the data. The emergency threshold of the local decision engine is precisely set for different garden scenarios. For example, the emergency threshold for soil moisture in arid areas is lower than that in humid areas, and the threshold for biological signals during the peak period of pests and diseases is more sensitive. The breakpoint resume unit adopts local encrypted storage to ensure the security and integrity of data during network outages and avoid the loss of critical monitoring data. Through the module collaboration mechanism within the edge intelligence layer, the clean data processed by the data preprocessing module provides accurate judgment basis for the local decision engine, enabling it to trigger emergency commands that meet actual needs based on reliable data. On the other hand, it is synchronized to the breakpoint resume unit for backup. When the network is restored, the complete data retransmitted by the breakpoint resume unit can be seamlessly connected with the historical data in the cloud, providing comprehensive data support for the intelligent fusion analysis module to optimize the plant growth model and adjust the maintenance plan, realizing efficient linkage between edge data processing and cloud data application.

[0022] In the case implementation, the intelligent fusion analysis module has a built-in library of multiple plant growth models and uses machine learning algorithms to fuse and analyze multimodal data to dynamically generate personalized maintenance plans. The learning algorithms employed include random forests and long short-term memory networks (LSTM). The full lifecycle traceability module records monitoring data, maintenance operations, and changes in growth status of plants from planting to maturity, forming a traceable data chain. The global scheduling module uses AI... Algorithm optimization of job paths and equipment coordination priority; Furthermore, the Random Forest algorithm focuses on feature association mining of multi-dimensional data, such as integrating soil fertility, air temperature and humidity, and pest and disease signals to accurately identify plant growth stress factors (identification accuracy ≥92%), providing a basis for decisions on pest and disease control and fertilization supplementation; the LSTM algorithm, relying on the advantages of time-series data processing, analyzes the long-term variation patterns of plant stem flow rate, growth cycle, and environmental parameters to predict subsequent growth needs (such as irrigation adjustments and fertilization timing 1-2 weeks later), supporting the forward-looking optimization of maintenance programs; A The algorithm constructs a cost function by quantifying indicators such as equipment operating distance, energy consumption cost, and task urgency, and plans the optimal operating path, which can reduce the invalid movement distance of equipment by ≥30% and improve the efficiency of multi-device collaborative operation.

[0023] Through the collaborative mechanism of the three modules, the personalized maintenance plan (including parameters such as irrigation amount, fertilizer ratio, and pruning cycle) generated by the intelligent fusion analysis module will be synchronized to the global scheduling module and the full life cycle traceability module: the global scheduling module is based on A The optimized algorithm breaks down the solution into operation instructions adapted to different devices and issues them to the collaborative execution layer in order of priority to avoid job conflicts and resource waste. The full life cycle traceability module archives the solution content, equipment execution logs and plant growth status feedback data in real time, forming a complete traceability chain of "data collection-solution generation-job execution-effect feedback". This archived data will then feed back into the intelligent fusion analysis module, providing continuously updated training samples for random forest and LSTM algorithms, and continuously improving the model analysis accuracy and the personalized adaptability of maintenance solutions.

[0024] In the implementation of the case, all devices in the collaborative execution layer are equipped with built-in positioning modules, obstacle avoidance sensors, and status feedback units, and support short-range collaborative communication between devices. The positioning module accuracy is ≤5m, the obstacle avoidance sensor detection distance is ≤3m, and the short-range communication between devices adopts the Bluetooth Mesh protocol with a communication latency of ≤500ms. The intelligent irrigation equipment adopts a dual-mode design of drip irrigation and sprinkler irrigation, which can automatically switch according to plant type and soil moisture distribution. The flow rate of the intelligent irrigation equipment in drip irrigation mode is 5-10L / h, and the flow rate in sprinkler irrigation mode is 20-30L / h. The switching threshold is that drip irrigation is activated when the soil moisture is ≤20%, and sprinkler irrigation is activated when the soil moisture is ≥20% and the plant is shallow-rooted. Furthermore, the positioning module, based on GPS / BeiDou dual-mode positioning, can accurately obtain the real-time location of the equipment, providing data support for the cloud-based global scheduling module to optimize the operation path; the obstacle avoidance sensor, through a combination of infrared detection and ultrasonic sensing, can identify obstacles such as trees, rocks, and pedestrians in the garden, avoiding collisions during equipment operation; the status feedback unit collects equipment operating parameters (such as irrigation flow, spraying pressure, and battery power) and operation progress in real time, ensuring that the control end can monitor the equipment's operating status in real time; the devices achieve collaboration through the Bluetooth Mesh short-range communication protocol, for example, automatic pruning equipment and mobile spraying equipment can operate in the same range, avoiding repetitive operations or omissions; the drip irrigation mode of the intelligent irrigation equipment is suitable for deep-rooted plants such as trees and shrubs, thereby accurately dripping water to reduce water evaporation and save water resources; the sprinkler irrigation mode is suitable for shallow-rooted plants such as lawns and flowers, which can quickly cover large areas and meet the needs of uniform water replenishment; Through the coordinated operation of equipment hardware functions and cloud scheduling, the positioning module and the global scheduling module work together to achieve multi-device zoned operation and optimal path planning; obstacle avoidance sensors are combined with the equipment control system to ensure safe and stable operation; the status feedback unit transmits data back to the cloud and edge in real time, providing a basis for adjusting maintenance plans and early warning of equipment failures; short-range collaborative communication enables multiple types of equipment to work together to improve maintenance efficiency; the dual-mode switching mechanism of the intelligent irrigation equipment, combined with soil moisture data and plant type information collected by the multimodal sensing layer, achieves "on-demand water supply", which ensures maintenance effect and avoids resource waste.

[0025] In the implementation case, the sensing devices of the multimodal sensing layer support switching between at least two communication modes, namely LoRa, WiFi, or 5G, and are powered by solar energy. The protection level is no less than IP67, and the communication protocols are Bluetooth Mesh and LoRaWAN. The LoRa communication distance is ≥3km, the WiFi adaptation distance is ≤100m, and the 5G transmission latency is ≤20ms. The solar power system is equipped with a 100Ah energy storage battery, and the continuous battery life is ≥72 hours in continuous cloudy and rainy weather. Furthermore, the communication mode switching is based on intelligent adaptation to the garden environment and network status: LoRa mode is suitable for scenarios with large garden areas and dispersed equipment, ensuring stable data transmission with its low power consumption and long-distance transmission characteristics; WiFi mode is used for networking in the core area of ​​the garden or for short-range equipment, enabling high-bandwidth data exchange; 5G mode meets the transmission needs of high-definition images (such as pest monitoring videos) and real-time monitoring data with large capacity and low latency. The device can automatically identify the network signal strength and transmission needs to switch modes. The solar power system, combined with energy storage batteries, ensures that the equipment in garden areas without external power supply can continue to work, and can also continue to operate through energy storage on rainy days; the IP67 protection rating can resist rainstorms, sandstorms, and insect corrosion, adapting to complex outdoor garden environments. By employing redundant communication mode design and highly reliable power supply and protection schemes, the pain points of "unstable communication, difficult power supply, and easy damage" of outdoor garden equipment are solved: switching between multiple communication modes avoids data transmission interruptions caused by a single network outage, solar power reduces cable laying costs and energy consumption, and IP67 protection level extends the service life of the equipment; the three work together to ensure that the sensing equipment operates continuously and stably in different garden scenarios (such as urban parks and suburban green spaces) and under different climatic conditions, providing uninterrupted multimodal data support for subsequent data preprocessing and decision analysis.

[0026] In the case implementation, the cloud management layer also includes a visual management and control platform and a fault early warning module; the visual management and control platform supports multi-terminal login and 3D modeling of the garden area and real-time mapping of equipment status; the fault early warning module is used to monitor the operating parameters of the execution layer equipment in real time and to issue alarms when abnormalities occur; The visualization management platform features multi-terminal login coverage across computers, mobile apps, and tablets, allowing managers to view the garden's status anytime, anywhere. The 3D modeling, built based on the actual terrain, plant distribution, and equipment deployment locations, accurately recreates the park's scene. Real-time equipment status mapping intuitively displays the operating status of each device (e.g., whether irrigation equipment is working, remaining pesticide dosage on spraying equipment), sensor data values ​​(e.g., soil moisture, air temperature), and work progress, achieving "what you see is what you get" visualization management. The fault warning module presets normal operating parameter ranges for equipment (e.g., working pressure of irrigation equipment, motor speed threshold of pruning equipment). By collecting real-time data such as voltage, current, and operating efficiency of the execution layer equipment, it accurately identifies equipment faults (e.g., motor failure, pipe blockage) or abnormal operating conditions (e.g., insufficient power, sensor malfunction). Through the linkage mechanism between the two major modules and other cloud-based functions, the visualization management platform synchronously displays the alarm information from the fault early warning module at the corresponding equipment location in the 3D model, and marks the fault type and urgency level, making it convenient for managers to quickly locate the fault point. At the same time, the platform supports directly issuing maintenance dispatch instructions based on the early warning information, and linking with the global dispatch module to adjust the work plans of other equipment to avoid the fault affecting the overall maintenance progress. In addition, the visualization management platform also records the fault occurrence time, handling process and results, and synchronizes them to the full life cycle traceability module, providing data support for equipment maintenance plan optimization and fault rate statistical analysis, and improving the level of refined management of garden maintenance.

[0027] A method for smart garden maintenance and management based on the Internet of Things (IoT) includes the following steps: S1: System Deployment and Initialization: Divide the garden into grids according to the functional zones, deploy the perception layer, edge layer and execution layer devices, and input plant growth parameters and maintenance rules into the cloud platform to build an initial maintenance rule library and plant growth model; Furthermore, the grid-based division is divided into management units according to garden functions (such as viewing areas, leisure areas, and green plant cultivation areas) and plant types (tree areas, shrub areas, and lawn areas). The grid division is based on a 10m×10m unit, and each grid is equipped with dedicated sensing devices (such as encrypted soil moisture sensors in lawn areas), edge gateways, and suitable execution devices (such as high-altitude pruning equipment in tree areas) to ensure accurate monitoring and operation coverage. The input plant growth parameters include core indicators such as suitable temperature and humidity, fertility requirements, and growth cycles for different varieties. The maintenance rules cover basic standards such as irrigation frequency, fertilizer ratio, and pruning timing. Both the initial maintenance rule library and the plant growth model have reserved iteration interfaces to support subsequent optimization based on actual maintenance results. Through a collaborative initialization mechanism of partitioned deployment, data entry, and model building, grid-based division clarifies equipment deployment boundaries and management responsibilities, avoiding resource waste; standardized parameter and rule entry provides a unified decision-making benchmark for the system, ensuring the rationality of the initial maintenance plan; the initial maintenance rule base and plant growth model build the core data framework for system operation, laying the foundation for subsequent multimodal data collection, collaborative decision-making, and closed-loop optimization, achieving a smooth transition of the system from "deployment completed" to "precise operation".

[0028] S2: Multimodal data acquisition and preprocessing: Multi-dimensional environmental and plant data are acquired through the perception layer, and after preprocessing by the edge gateway, the effective data is synchronized to local storage and cloud database; Furthermore, multi-dimensional data collection focuses on three core dimensions: environment, plants, and equipment. Environmental data includes soil temperature and humidity, fertility, air temperature and humidity, and light intensity. Plant data covers stem flow rate, growth status, and pest and disease signals. Simultaneously, the initial operating parameters of the execution layer equipment are collected. The edge gateway's preprocessing includes three core operations: data denoising (filtering environmental interference noise), normalization (unifying the range of data with different dimensions), and outlier removal (screening out extreme values ​​of sensor faults). The data preprocessing module uses median filtering for noise reduction, with a normalization range of [0,1]. Outlier removal uses the 3σ principle (data exceeding the mean ± 3 times the standard deviation is considered outlier) to ensure data cleanliness. Local storage prioritizes the retention of critical and urgent data, while the cloud database archives all data, forming a dual storage architecture of "local backup + cloud aggregation". Through a collaborative process of "layered acquisition + intelligent preprocessing + dual-end synchronization", the perception layer dynamically samples at a preset frequency (such as collecting soil data once per hour and monitoring pest and disease signals in real time) to ensure data timeliness; the preprocessing stage greatly reduces the interference of invalid data on subsequent decisions; dual-end data synchronization supports real-time decision-making on the edge side through local storage, and provides a complete data foundation for global analysis and model optimization with the help of cloud database, realizing efficient connection between data acquisition and application.

[0029] S3: Edge-Cloud Collaborative Decision Making: The cloud-based intelligent fusion analysis module generates personalized maintenance plans based on real-time and historical data and by calling plant growth models. If the edge gateway detects that the data has reached the emergency threshold, the local decision engine directly triggers an emergency command and reports it to the cloud. The preset priority order is: pest and disease outbreak > extreme temperature > soil drought. The local decision engine triggers commands according to priority. If there is a conflict, the low-priority operation is suspended and the high-priority task is executed first. The conflict log is synchronously reported to the cloud. When generating personalized maintenance plans in the cloud, it deeply integrates the correlation patterns of multimodal data. For example, by matching the correspondence between historical soil fertility data and plant growth rate, it matches the current air temperature, humidity, and light conditions to accurately determine parameters such as irrigation amount, fertilization cycle, and pruning intensity for plants in different grid areas. The emergency threshold determination of the edge gateway focuses on "extreme scenarios that endanger plant growth," such as soil moisture below the survival threshold or excessive pest and disease signal strength. Random forest is used for pest and disease type identification, LSTM is used for plant growth trend prediction, and the emergency instructions triggered by the local decision engine are preset standardized emergency operations (such as immediate drip irrigation and targeted spraying to repel pests) to ensure rapid response. Through a collaborative mechanism of "normal cloud-based coordination + extreme edge backup," the cloud, under normal circumstances, leverages its powerful computing capabilities to plan the optimal global maintenance scheme, taking into account the differentiated needs of plants in different regions. In extreme scenarios, the edge device can directly execute emergency operations without waiting for instructions from the cloud, avoiding delays in the best maintenance time. At the same time, after triggering an emergency instruction, the edge device will report information such as the event type, processing actions, and data changes to the cloud in real time. The cloud will then update the maintenance plan and archive the relevant data synchronously, ensuring both the intelligence and comprehensiveness of decision-making and the timeliness of response in extreme scenarios.

[0030] S4: Collaborative Operation Execution and Status Feedback: The cloud-based global scheduling module converts the maintenance plan into operation instructions and optimizes the path before sending them to the execution layer devices. The devices then work collaboratively and provide real-time feedback on their status information. Among them, the instruction conversion of the cloud global scheduling module will adapt to the operation logic of different devices to avoid devices waiting for each other or going back and forth repeatedly. At the same time, the collaborative operation of devices follows the "priority sorting" principle, with emergency maintenance instructions (such as emergency spraying of pests and diseases) taking precedence over regular instructions (such as daily pruning). The same type of equipment is divided into areas for synchronous operation, and different types of equipment (such as irrigation and fertilization equipment) are operated in sequence (irrigation first, then fertilization). Through a closed-loop execution mechanism of precise instruction conversion, path optimization, and real-time feedback, the global scheduling module ensures the compatibility of operation instructions with equipment, while path optimization improves work efficiency and reduces energy consumption. During operation, the equipment transmits work progress, operating parameters, and work effect data in real time through the status feedback unit. The feedback information is synchronized to the cloud and edge. The cloud can dynamically adjust subsequent instructions based on the feedback, while the edge verifies in real time whether the work meets expectations, ensuring that the maintenance plan is implemented effectively and in line with the plan.

[0031] S5: Full life cycle closed-loop optimization: The full life cycle traceability module archives the entire maintenance process data, regularly evaluates the maintenance effect and dynamically optimizes the plant growth model parameters and maintenance plan to achieve closed-loop management of "monitoring-decision-execution-evaluation-optimization"; The data archived by the full lifecycle traceability module covers "key information across the entire chain": including multimodal raw data collected by the perception layer, decision records from the edge to the cloud, operation logs of the execution layer devices, and data on changes in plant growth status, forming a complete time-series data archive; the maintenance effect evaluation adopts the method of "quantitative indicators + comparative analysis". The quantitative indicators include plant survival rate, growth rate, and incidence of pests and diseases, while the comparative analysis compares the actual growth status with the model prediction value and the preset standard value to accurately locate the optimization space of the maintenance plan; Through a closed-loop mechanism of data archiving, effect evaluation, model iteration, and scheme upgrade, the full data of the traceability module provides an objective basis for the evaluation of maintenance effects. The evaluation results guide the adjustment of plant growth model parameters and the optimization of maintenance schemes. The optimized model and scheme will serve as the execution benchmark for the next round of maintenance, continuously adapting to changes in the garden environment and the needs of plant growth, so that the entire maintenance management process gradually approaches the optimal state and achieves the goal of long-term, stable, and intelligent maintenance.

[0032] In summary, this IoT-based smart garden maintenance management system and method, by constructing a complete architecture integrating a multimodal perception layer, an edge intelligence layer, a cloud management layer, and a collaborative execution layer, achieves comprehensive, real-time, and accurate monitoring of the garden environment and plant growth status. It solves the problems of single and lagging monitoring dimensions and lack of multi-dimensional collaborative perception of plant physiology and pests in existing garden maintenance management.

[0033] Furthermore, through an innovative edge-cloud collaborative decision-making mechanism, it not only ensures the immediate response capability on the edge side in extreme scenarios (such as sudden drought), but also utilizes the powerful computing capabilities of the cloud for big data analysis and long-term strategy optimization, achieving an efficient balance between response speed and decision intelligence. This solves the problem of data processing and decision response being disconnected in traditional systems, and the inability to balance long-term optimization and immediate response.

[0034] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0035] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An Internet of Things-based smart garden maintenance management system, comprising a multi-modal perception layer, an edge intelligence layer, a cloud management layer and a collaborative execution layer, characterized in that: The multi-modal perception layer is in communication connection with an edge intelligence layer, the edge intelligence layer is in communication connection with a cloud management layer, the edge intelligence layer and the cloud management layer are in communication connection with a collaborative execution layer, and the edge intelligence layer and the cloud management layer work collaboratively to form an edge-cloud collaborative decision mechanism; The multi-modal perception layer comprises a core perception unit and an auxiliary perception unit; The edge intelligence layer comprises a regional edge gateway, a data preprocessing module, a local decision engine and a breakpoint continuation unit; The cloud management layer comprises an intelligent fusion analysis module, a global scheduling module and a full life cycle tracing module; The collaborative execution layer comprises intelligent irrigation equipment, water and fertilizer integrated equipment, automatic pruning equipment, mobile pesticide spraying equipment and mobile maintenance terminals.

2. The smart garden maintenance management system based on the Internet of Things according to claim 1, characterized in that: The system supports double control mode switching: in normal working conditions, a mode of centralized scheduling by the global scheduling module is adopted; When network interruption or an emergency scenario is determined by the edge intelligence layer, the system automatically switches to a local control mode triggered by the local decision engine.

3. The smart garden maintenance management system based on the Internet of Things according to claim 1, characterized in that: The core perception unit is integrated with soil temperature and humidity sensors, soil fertility sensors, plant sap flow sensors, air temperature and humidity sensors, light sensors and pest and disease monitoring modules; the pest and disease monitoring modules integrate biological sensors, odor sensors and high-definition night vision cameras.

4. The smart garden maintenance management system based on the Internet of Things according to claim 1, characterized in that: The data preprocessing module is used for noise reduction, normalization and outlier rejection processing of perception data; The local decision engine is preconfigured with an environmental parameter emergency threshold for triggering immediate maintenance instructions in extreme scenarios; the breakpoint continuation unit is used for locally storing data when the network is interrupted, and completing data retransmission after the network is restored.

5. The smart garden maintenance management system based on the Internet of Things according to claim 1, characterized in that: The intelligent fusion analysis module is built-in with a multi-type plant growth model library, and based on a machine learning algorithm, multi-modal data are fused and analyzed to dynamically generate a personalized maintenance scheme; the whole life cycle tracing module is used to record the whole process monitoring data, maintenance operation and growth state change of the plant from planting to maturity, forming a traceable data chain, and the global scheduling module adopts an A algorithm to optimize the operation path and the device coordination priority.

6. The smart garden maintenance management system based on the Internet of Things according to claim 1, characterized in that: The devices in the collaborative execution layer are all built-in with positioning modules, obstacle avoidance sensors and state feedback units, and support short-distance collaborative communication between devices; the intelligent irrigation equipment adopts a drip irrigation and sprinkler irrigation dual-mode design, which can automatically switch according to plant types and soil humidity distribution.

7. The smart garden maintenance management system based on the Internet of Things according to claim 1, characterized in that: The perception devices of the multi-modal perception layer support at least two communication mode switching among LoRa, WiFi or 5G, and are powered by solar energy, with a protection level not lower than IP67. 8.The smart garden maintenance management system based on the Internet of Things according to claim 1, characterized in that: The cloud management layer further comprises a visual management and control platform and a fault warning module; the visual management and control platform supports multi-terminal login and real-time mapping of garden area three-dimensional modeling and device state; the fault warning module is used for real-time monitoring of execution layer device operating parameters and alarming when abnormal.

9. A method for smart garden maintenance management based on an Internet of Things, comprising the smart garden maintenance management system based on the Internet of Things in claims 1-8, characterized in that: The operation steps are as follows: S1: system deployment and initialization: according to the functional zoning of the garden, the grid is divided, the perception layer, edge layer and execution layer devices are deployed, and the plant growth parameters and maintenance rules are input on the cloud platform to build the initial maintenance rule library and plant growth model; S2: multi-modal data acquisition and preprocessing: multi-dimensional environmental and plant data are collected through the perception layer, and effective data are synchronized to the local storage and cloud database after preprocessing by the edge gateway; S3: edge-cloud collaborative decision: the cloud intelligent fusion analysis module generates a personalized maintenance scheme based on real-time and historical data, calling the plant growth model; If the edge gateway detects that the data reaches the emergency threshold, the local decision engine directly triggers the emergency instruction and reports to the cloud. The default priority order is: pest outbreak > extreme temperature > soil drought. The local decision engine triggers the instruction according to the priority. If there is a conflict, the low-priority operation is suspended, and the high-priority task is executed first. The conflict log is synchronized to the cloud. S4: Cooperative operation execution and status feedback: The cloud global scheduling module converts the maintenance scheme into operation instructions and optimizes the path, and then sends it to the execution layer device. The device cooperates with the operation and feeds back the status information in real time. S5: Full life cycle closed-loop optimization: The full life cycle tracing module archives the maintenance full process data, regularly evaluates the maintenance effect, and dynamically optimizes the plant growth model parameters and the maintenance scheme, realizing the closed-loop management of "monitoring-decision-execution-evaluation-optimization".

Images