Intelligent control modular gardening combined lighting system

Abstract

The invention discloses an intelligent control modular gardening combined lighting system and belongs to the technical field of gardening lighting and intelligent control, and the system comprises a micro-area lighting execution module which applies differentiated lighting to a plurality of micro-areas of a plant planting area according to an instruction; the plant and environment data acquisition module is used for acquiring a plant population state and microcell environment data; the group benefit evaluation module is used for evaluating the group benefit generated by the current illumination strategy on the plant group based on the acquired data; the illumination parameter acquisition and instruction generation module is integrated with an artificial intelligence engine, the engine receives the plant state, the environment data and the group benefit evaluation result, and generates or iteratively updates an instruction for the illumination parameters of each micro-area based on an optimization target of maximizing the group benefit; through AI-driven micro-area heterogeneous illumination regulation and control, active interaction between plant populations is actively induced, overall gain exceeding individual simple optimization is achieved, and an efficient and intelligent illumination solution is provided for facility horticulture.

Description

Technical Field

[0001] The present invention relates to the field of horticultural lighting and intelligent control, and in particular to a modular horticultural combination lighting system with intelligent management and control. Background Art

[0002] Traditional horticultural artificial lighting systems, particularly those based on advanced light sources such as LEDs, have made significant progress in providing the light necessary for plant growth. However, existing lighting strategies often focus on providing a uniform and consistent light environment across plant populations, primarily optimizing for the physiological needs of individual plants. While this strategy can meet basic plant growth requirements to a certain extent, it often overlooks the complex interactions within plant populations and the heterogeneous nature of their responses to light environments.

[0003] In natural ecosystems, plant communities exhibit complex group behaviors and resource utilization strategies through competition and collaboration among individuals, as well as adaptive responses to microenvironmental changes. For example, interplant competition for shade and light, chemical communication through volatiles or root secretions, and perception and morphological adjustments in response to light cues from neighboring plants (such as changes in the far-red / red light ratio) are all important factors influencing community productivity and adaptability.

[0004] Existing technologies fail to fully tap into and exploit the potential for interactions within these plant communities. Simply pursuing uniform illumination may not optimally allocate light resources and may even inhibit the occurrence of certain beneficial community effects. Furthermore, there is a lack of systematic technical solutions and intelligent control methods for actively manipulating the heterogeneity of the light environment to positively guide plant community behavior and achieve a "the whole is greater than the sum of its parts" benefit. Summary of the Invention

[0005] The present invention provides an intelligently controlled modular horticultural combination lighting system to solve the technical problem that the lighting strategy in existing horticultural lighting systems is generally uniform lighting, which fails to fully utilize the interaction potential within the plant community to improve overall production efficiency and environmental adaptability.

[0006] In view of the above problems, the present invention provides a modular gardening combination lighting system with intelligent management and control, comprising:

[0007] The micro-area lighting execution module is configured with multiple lighting units capable of independently adjusting lighting parameters. The micro-area lighting execution module is used to apply differentiated lighting environments to multiple preset micro-areas within the plant planting area according to the first lighting parameter instruction received from the lighting parameter acquisition and instruction generation module;

[0008] The plant and environment data collection module collects at least one of the plant population status information and micro-area environment perception data in the plant planting area;

[0009] a group benefit evaluation module, which evaluates the group benefit of the current lighting strategy on the plant group based on at least one of the plant group status information and the micro-area environmental perception data received from the plant and environmental data acquisition module, and transmits the group benefit evaluation result to the lighting parameter acquisition and instruction generation module;

[0010] The lighting parameter acquisition and instruction generation module is integrated with an artificial intelligence engine, which is used to receive the plant population status information, the micro-area environment perception data and the population benefit evaluation results, and generate or iteratively update the first lighting parameter instruction based on this.

[0011] The technical solution provided by this application has at least the following technical effects or advantages:

[0012] By optimizing heterogeneous lighting in micro-areas through an artificial intelligence engine, we stimulate positive interactions among plants, enabling the total biomass, economic yield, or accumulation of specific quality components across the entire plant population to surpass the sum of traditional uniform lighting or simple individual optimization. By applying inductive lighting to specific micro-areas that induce benefits within a specific population, some plants are stimulated to produce stress resistance signals, which are then transmitted within the population. This improves the entire plant population's systemic resistance to adverse environmental factors such as pests and diseases, temperature fluctuations, and its recovery speed after stress, thereby reducing production risks.

[0013] Traditional uniform lighting may lead to insufficient or oversaturated lighting in some areas. The present invention uses AI to dynamically adjust the lighting parameters of each micro-area to more accurately match the light requirements of different areas and growth stages of the plant population, avoid unnecessary waste of light energy, and is expected to reduce overall energy consumption while ensuring or even increasing yields. At the same time, the present invention is not only a production tool, but also a research platform. By recording and analyzing the responses and benefit changes of plant populations under different heterogeneous lighting strategies, it is helpful to deeply understand the complex interaction mechanisms between plants and provide data and theoretical support for further optimization of population regulation strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 This is an architecture diagram of the modular horticultural combination lighting system with intelligent management and control according to the present invention. DETAILED DESCRIPTION

[0015] The present invention relates to an intelligently controlled modular horticultural combination lighting system to address the technical problem that existing horticultural lighting systems typically use uniform illumination strategies, failing to fully utilize the interaction potential within plant communities to improve overall production efficiency and environmental adaptability. This application, guided by artificial intelligence, actively and intelligently creates and dynamically regulates the heterogeneity of micro-region lighting environments within plant communities, aiming to stimulate positive interactions between plant communities. This effectively improves the overall productivity and environmental adaptability of the entire plant community without significantly increasing or even reducing total light input, achieving the goal of increasing efficiency by making the whole greater than the sum of its parts.

[0016] The positive interaction effects between plant individuals, that is, the group benefits referred to in this application, may be based on a variety of biological and ecological mechanisms. For example, by regulating the heterogeneity of micro-area illumination, the light competition and light signal perception between plants can be affected. Reasonable heterogeneous illumination may guide plants to form a more optimized canopy structure, reduce excessive shading between individuals, and improve the interception and utilization efficiency of light energy by the entire group. At the same time, specific spectral components (such as changes in the ratio of far-red light to red light) are important signals for plants to perceive the presence of neighbors. By regulating these light signals, the morphological construction (such as branching, internode elongation) and resource allocation strategies of plants can be affected, thereby achieving a better growth state at the group level.

[0017] Furthermore, complex chemical signaling communication occurs between plants. In response to specific stresses or environmental signals (including specific light qualities), some plants may release volatile organic compounds (VOCs) or transmit signaling molecules through root secretions. These signals can be sensed by other plants in the colony, triggering early warnings and defense responses or promoting growth. This system, by applying inductive light in specific micro-regions, exploits this interplant signaling mechanism to achieve group-wide stress resistance or growth promotion.

[0018] Furthermore, differential regulation of micro-area light intensity can also affect microenvironmental heterogeneity within a colony (e.g., temperature and humidity distribution within the canopy). This moderate heterogeneity can sometimes provide a more suitable growth environment for individuals with different physiological states or genotypes, thereby improving the stability and productivity of the entire colony. This application aims to learn and utilize these potential colony interaction mechanisms through an artificial intelligence engine to achieve colony benefits beyond simple individual optimization.

[0019] The above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods of the specification to better understand the above technical solution. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments of the present invention. It should be understood that the present invention is not limited to the example embodiments used only to explain the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. In addition, it should be noted that, for the convenience of description, only the parts related to the present invention, rather than all, are shown in the drawings.

[0020] like Figure 1 As shown in the architecture diagram of the intelligently controlled modular horticultural lighting system, the system includes the following modules:

[0021] Micro-area lighting execution module: It is equipped with multiple lighting units that can independently adjust lighting parameters. These lighting units can be arrays of intelligent LED lamps, and each unit or its sub-unit can be independently addressed and controlled to cover the entire plant planting area. The micro-area lighting execution module applies differentiated lighting environments to multiple preset micro-areas within the plant planting area based on the first lighting parameter instruction received from the lighting parameter acquisition and instruction generation module. This means that each micro-area (which can be as small as a few plants or even different canopy parts of a single plant) can receive unique lighting treatment.

[0022] Plant and Environmental Data Acquisition Module: This module is used to collect at least one of the following: plant population status information and micro-area environmental perception data within the plant planting area, and transmit this data to the lighting parameter acquisition and instruction generation module and the population benefit assessment module. Plant population status information may include canopy height, density, leaf area index (LAI) estimation, canopy color, and growth image information acquired through a multi-view camera array, or three-dimensional population structure data acquired through a laser radar (LiDAR) or depth camera. It may also optionally include leaf reflectance / transmission spectral information (indirectly reflecting chlorophyll, nitrogen, water, stress, etc.) perceived by a micro-spectrometer or hyperspectral / multispectral imaging unit, as well as physiological parameters such as chlorophyll fluorescence and leaf temperature. Micro-area environmental perception data may include temperature, humidity, CO2 concentration, and other parameters for each micro-area. These sensors may be integrated into the lighting unit or deployed independently.

[0023] The Colony Benefit Evaluation Module evaluates the colony benefit of the current lighting strategy on the plant population based on at least one of the plant population status information and micro-area environmental perception data received from the Plant and Environmental Data Acquisition Module. The colony benefit evaluation results are transmitted to the Lighting Parameter Acquisition and Instruction Generation Module. The colony benefit evaluation metric corresponds to the system objective and can be a single metric or a weighted combination of multiple objectives. In reinforcement learning applications, the output of this module directly or indirectly constitutes the reward signal.

[0024] Specifically, the assessment process may include the following steps: First, the collected raw plant status data (such as images, spectra, and point clouds) is processed and analyzed to extract key phenotypic parameters or physiological indicators. For example, canopy volume and leaf area index are calculated through image segmentation and three-dimensional reconstruction techniques, vegetation health is estimated through spectral indices (such as NDVI), or parameters related to photosynthetic efficiency are extracted from chlorophyll fluorescence data using specific algorithms. Second, the micro-area environmental data and optional preset plant growth models (such as mechanistic models based on photosynthesis, respiration, and material distribution, or empirical growth models trained based on historical data) are combined to predict plant growth trends and potential yields for the current and future periods. Then, one or more specific indicators are calculated to obtain a quantitative population benefit assessment value. For example, for yield indicators, it may be necessary to summarize the predicted yields of each micro-area and compare them with the total yield expected based on individual optimization to determine whether there is a super-additive effect. For stress resistance indicators, this may involve monitoring recovery time and survival rate after applying simulated or real stress. Evaluation indicators for community benefits correspond to system objectives and can be single indicators or weighted combinations of multiple objectives. These include total community biomass or economic yield (which can be estimated through model prediction or image analysis), total community light use efficiency (PUE), community resistance or resilience to specific stresses, and community structure optimization indicators (such as the vertical distribution of leaf area index).

[0025] Lighting parameter acquisition and instruction generation module: It integrates an artificial intelligence engine, is used to receive plant population status information, micro-area environmental perception data and population benefit evaluation results, and based on this, generates or iteratively updates the first lighting parameter instruction. The hardware can be an edge computing device, a local server or a cloud platform.

[0026] Furthermore, the artificial intelligence engine in the lighting parameter acquisition and instruction generation module generates or iteratively updates the first lighting parameter instruction based on a preset optimization goal of maximizing group benefit, where the group benefit is an overall gain effect that exceeds the simple sum of individual capabilities; the first lighting parameter instruction includes at least one parameter of light intensity, spectral composition, lighting duration or pulse mode customized for each or each group of micro-areas, which is used to actively create and dynamically regulate the heterogeneity of the lighting environment between micro-areas, inducing and stimulating the plant population to exhibit group benefit.

[0027] Specifically, the artificial intelligence engine in the lighting parameter acquisition and instruction generation module generates or iteratively updates the first lighting parameter instruction based on a preset optimization goal of maximizing group benefit, where group benefit is defined as an overall gain effect that exceeds the simple sum of individual capabilities. The first lighting parameter instruction includes at least one parameter customized for each micro-area or group of micro-areas: light intensity (measured by photosynthetic photon flux density (PPFD), spectral composition (including the intensity and relative proportions of different bands such as red light, blue light, far-red light, green light, and UV light), light duration (such as daily cumulative light duration), or pulse mode (characterized by frequency and duty cycle). These instructions are designed to induce and stimulate plant populations to exhibit group benefits by actively creating and dynamically controlling the heterogeneity of the lighting environment between micro-areas. Through specific heterogeneous lighting combinations, plant populations demonstrate collaboration and complementarity in resource utilization (such as light energy and space), information exchange, or stress resistance, thereby achieving an overall effect that cannot be achieved by the sum of simple individual optimizations.

[0028] Furthermore, the artificial intelligence engine in the lighting parameter acquisition and instruction generation module adopts at least one of the following algorithms to achieve the generation or iterative update of the first lighting parameter instruction for the preset optimization goal of maximizing group benefits:

[0029] a) Evolutionary algorithm or genetic algorithm encodes the complete lighting parameter combination covering all micro-areas as individuals to be optimized, uses the group benefit evaluation result as the fitness function, and searches for the lighting parameter instruction set that maximizes the group benefit through selection, crossover and mutation operations;

[0030] b) A reinforcement learning algorithm defines the plant population and its dynamic response to light as the environment. The artificial intelligence engine acts as the agent, and the adjustment of the first light parameter instruction is the action. The group benefit evaluation result constitutes the reward signal. The agent learns through interaction with the environment and forms a control strategy that can output the first light parameter instruction that maximizes the long-term cumulative group benefit based on the current plant population state and environmental data.

[0031] c) An optimization algorithm based on a proxy model, in which an artificial intelligence engine uses historical data or simulated data to train a proxy model that predicts the impact of different first lighting parameter instructions on group benefits, and performs an optimization search on this proxy model to identify the first lighting parameter instructions that can improve group benefits.

[0032] Specifically, the artificial intelligence engine in the lighting parameter acquisition and instruction generation module uses at least one of the following algorithms to achieve the generation or iterative update of the first lighting parameter instruction based on the preset optimization goal of maximizing group benefits:

[0033] a) Evolutionary or genetic algorithms: In these algorithms, a complete set of lighting parameters covering all micro-zones (including parameters such as light intensity and wavelength ratios for each micro-zone) is encoded as individuals or genotypes in the algorithm. Through operations such as simulating natural selection (selection based on the results of group benefit evaluation as a fitness function), crossover (combining features of excellent lighting patterns), and mutation (introducing new lighting parameter changes), the algorithm iteratively searches for the first set of lighting parameter instructions that maximizes the global or near-global optimal group benefit.

[0034] The genotype of an individual can be directly composed of a numerical sequence of a series of lighting parameters set for each micro-area (such as the intensity value of each adjustable spectral channel in each micro-area, the duration of lighting, etc.). The design of the fitness function directly converts the results of the group benefit evaluation (such as the weighted comprehensive score of yield, quality, and stress resistance indicators) into a scalar value, which is used to evaluate the advantages and disadvantages of each lighting parameter combination. The selection operation can adopt methods such as roulette selection and tournament selection; the crossover operation can perform single-point crossover, multi-point crossover, or uniform crossover on the lighting parameter sequence to combine lighting strategy fragments of different outstanding individuals; the mutation operation performs small-scale random perturbations on certain lighting parameter values ​​of selected individuals or adjusts them according to preset rules to maintain population diversity and explore new solution space. The termination conditions of the algorithm can be set to reach a preset maximum number of iterations, the fitness function value has not significantly improved for multiple consecutive generations, or a solution that meets the preset benefit threshold is found.

[0035] b) Reinforcement learning algorithm: In this type of algorithm, the plant population and its dynamic response to light, such as growth changes and stress reactions, are defined as the environment. The action of the artificial intelligence engine is to adjust the lighting parameters of one or more micro-areas (i.e., the adjustment of the first lighting parameter instruction). The group benefit evaluation result provided by the group benefit evaluation module constitutes a reward signal, which can be set as a positive reward when the group benefit increases, and vice versa. It can be set as a negative reward or no reward. Through continuous trial and error and learning with the environment, the intelligent agent can adopt methods such as Q-learning, DeepQ-Network, PolicyGradient, etc. to gradually form a control strategy that can output the first lighting parameter instruction that can maximize the long-term cumulative group benefit based on the current plant population status and environmental data.

[0036] In a reinforcement learning framework, the state space can be composed of vectorized representations of plant population state information (such as canopy structural parameters and physiological indicators) and micro-region environmental sensory data (such as temperature, humidity, and CO2). These data require appropriate normalization and feature engineering. The action space corresponds to the adjustment range of the first light parameter instruction output by the AI ​​engine. For example, it can be a discrete set of actions (such as increasing / decreasing / maintaining a light parameter in a specific micro-region by a preset amount) or, in more advanced applications, a continuous parameter adjustment value. The reward function needs to map short-term or cumulative population benefit evaluation results (such as yield change rate, improved light energy utilization efficiency, and improved stress tolerance) into a scalar reward value to guide the learning direction of the agent. Learning algorithms such as Q-learning require maintaining a Q-value table or using function approximation (such as Deep Q-Network (DQN)) to estimate the value of state-action pairs; Policy Gradient methods directly learn a policy function that maps states to actions (or a probability distribution over actions). To address the slow response characteristics of plant growth and data sparsity, it may be necessary to combine offline reinforcement learning (RL) technology to use historically accumulated data for policy learning, or adopt model-based reinforcement learning (RL) to first learn a model of plant growth and response and then optimize the policy based on this model.

[0037] c) Proxy-model-based optimization algorithms: In this type of algorithm, when directly evaluating the impact of a specific lighting pattern on colony benefits (e.g., through actual observation of plant growth) is too costly or time-consuming, the AI ​​engine first uses limited experimental data or high-quality simulation data to train a proxy model, such as a Gaussian process regression model or a neural network model, that can quickly predict the impact of different first-light parameter instructions on colony benefits. The AI ​​then performs an efficient optimization search on this lightweight proxy model, using Bayesian optimization or evolutionary algorithms, to identify potential first-light parameter instructions that can significantly improve colony benefits. These identified instructions can then be verified and further fine-tuned through actual experiments.

[0038] When constructing a surrogate model, the input features are different combinations of first-order illumination parameters, and the output is the corresponding predicted population benefit assessment value. Training the surrogate model requires a set of high-quality training samples. These can be obtained through small-scale real-world planting experiments, high-fidelity simulation models of plant growth and light environment interactions (such as those based on the Functional-Structural Plant Model (FSPM)), or historical planting data. Surrogate models such as Gaussian Process Regression (GPR) can provide uncertainty estimates of the predicted values, which is very useful for subsequent exploratory optimization (e.g., balancing exploration and exploitation through acquisition functions such as Expected Improvement or Upper Confidence Bound within a Bayesian optimization framework). Neural network models (such as multi-layer perceptrons (MLPs)) have the ability to fit complex nonlinear relationships. When performing optimization searches on surrogate models, since the evaluation cost of surrogate models is much lower than that of real-world experiments or complex simulations, computationally intensive global optimization algorithms (such as evolutionary algorithms and particle swarm optimization) or efficient sequential optimization strategies (such as Bayesian optimization) can be employed. The number and selection criteria (such as expected benefit improvement and prediction confidence) of the identified potential excellent lighting parameter instructions will be set according to the available experimental resources and optimization goals.

[0039] Furthermore, the first illumination parameter instruction generated by the illumination parameter acquisition and instruction generation module includes an instruction set for instructing the micro-area illumination execution module to divide and regulate functional micro-areas in the plant population, and the functional micro-areas include:

[0040] i) a growth-optimized micro-area, wherein the light parameters are configured to promote the biomass accumulation or formation of a specific economic yield of the plants in the micro-area;

[0041] ii) Community benefit induction micro-areas, whose lighting parameters are configured to induce plants in the micro-areas to produce specific physiological and biochemical responses, which enhance the comprehensive performance or stress resistance of the entire plant population.

[0042] Specifically, the first illumination parameter instruction generated by the illumination parameter acquisition and instruction generation module includes an instruction set for instructing the micro-region illumination execution module to divide and regulate functional micro-regions within the plant population. This functional differentiation is an important way to achieve population benefits. The functional micro-regions include:

[0043] i) Growth-optimized micro-area: Its lighting parameters (such as high intensity, optimized red-to-blue ratio light) are configured to mainly promote the biomass accumulation or the formation of specific economic yields of plants in the area, serving as the main force of group production.

[0044] ii) Community benefit inducing micro-areas: their light parameters, especially spectral composition (such as blue light, green light, far-red light or UV of specific wavelengths) or light intensity in specific bands, are specially configured to induce specific physiological and biochemical responses in plants in the area. This response is intended to enhance the comprehensive performance of the entire plant population, such as stress resistance, resource utilization efficiency or structural optimization, non-locally (i.e., the impact exceeds the scope of the micro-area) and systemically (i.e., affecting the entire plant population or most individuals).

[0045] One implementation of the group benefit induction micro-area is to configure it as a stress-resistant pioneer area. When the system predicts or detects a stress risk (such as high temperature, disease), the artificial intelligence engine will instruct the application of specific pretreatment or induction light (such as an appropriate amount of UV-B, or blue light of a specific wavelength) in the stress-resistant pioneer area. The light parameters it receives are used to induce the plants in the area to produce and accumulate stress-resistant signal molecules (such as salicylic acid, jasmonate, etc.) or stimulate the plant's systemic acquired resistance (SAR) or induce systemic resistance (ISR). These stress-resistant signal molecules or systemic acquired resistance may be transmitted within the group through volatile organic compounds, root secretions or internal signal pathways, so that other plants that are not directly exposed to the induced light also show enhanced stress resistance. Therefore, the systemic resistance of the entire plant group to specific environmental stresses or the recovery speed after stress is improved, and its effect is designed to be better than the effect of uniformly applying the same stress pretreatment light to all micro-areas, which may cause unnecessary growth inhibition or energy waste due to universal treatment. Overall protection is achieved through local induction.

[0046] Furthermore, the group benefit evaluation module quantitatively evaluates the group benefit by using at least one of the following indicators, and uses the quantitative evaluation results as feedback for the artificial intelligence engine's optimization decision:

[0047] a) Under conditions where total light input energy remains constant or is lower than a baseline, the growth rate or improvement in total biomass, economic yield, or specific target quality components of a plant population is greater than that expected from the simple sum of individual contributions;

[0048] b) The observed or predicted light use efficiency or radiation use efficiency of the plant population as a whole is higher than the expected theoretical value based on the weighted average performance of the individual plants in each micro-plot under uniform optimized light;

[0049] c) the degree of improvement in the level of systemic acquired resistance induced by the population benefit-inducing micro-area, or the degree of acceleration in the overall recovery rate of the population after a specific stress, or the reduction in the population loss rate.

[0050] Specifically, the group benefit assessment module quantifies group benefits by monitoring, calculating, or predicting at least one of the following indicators, and uses this quantitative assessment result as feedback for the AI ​​engine's optimization decisions. These indicators are intended to objectively reflect "the overall benefit effect that exceeds the simple sum of individual capabilities":

[0051] a) Under the premise that the total light input energy (such as the daily total light integral (DLI) or the instantaneous total PPFD) remains unchanged or is reduced compared to the reference benchmark (such as a traditional uniform lighting scheme), the superadditive growth rate or nonlinear improvement of the total biomass (dry weight or fresh weight) of the plant population, the economic yield (such as the yield of harvestable parts such as fruits, flowers, and leaves), or specific target quality components (such as vitamins, anthocyanins, and specific pharmaceutically active substances) of the plant population; superadditive growth means that if the total expected yield of individual optimization is X, the yield achieved by this system is X + ΔX (ΔX>0), and this increase cannot be simply explained by individual effects.

[0052] b) The observed or predicted values ​​of the plant population's overall light use efficiency (PUE, defined as biomass produced per unit of photosynthetically active radiation) or radiation use efficiency (RUE) are higher than the theoretical value expected based on a simple weighted average of individual plant performance across microplots under uniformly optimized illumination. This suggests that heterogeneous illumination allows the population to capture and convert light energy more efficiently, reducing light energy waste (e.g., light loss due to overshading or light-saturated areas).

[0053] c) The degree of improvement in systemic acquired resistance throughout the plant population, induced by the community-benefit-inducing micro-region (evaluated by measuring defense-related gene expression, defense enzyme activity, or antioxidant content), or the acceleration of the overall population recovery rate after a specific stress (manifested by a shortened wilting recovery time and faster growth rate recovery), or a reduction in population loss rate (usually referring to morbidity or mortality). These are all direct manifestations of synergistic stress resistance at the population level.

[0054] Furthermore, the artificial intelligence engine in the lighting parameter acquisition and instruction generation module conducts continuous learning and model iteration based on historical plant population status information, micro-area environmental perception data, executed first lighting parameter instructions and corresponding population benefit evaluation results to optimize its strategy for generating the first lighting parameter instructions.

[0055] Specifically, the artificial intelligence engine in the lighting parameter acquisition and instruction generation module conducts continuous learning and model iteration based on historical plant population status information (such as canopy images and physiological parameters at different growth stages), micro-area environmental perception data (such as temperature, humidity, and CO2 concentration changes), the executed first lighting parameter instructions (i.e., what kind of light was applied to which micro-areas and when), and the corresponding group benefit evaluation results (such as yield, quality, and stress resistance). This learning can be offline (regularly retraining and updating the AI ​​model using accumulated data) or online (real-time fine-tuning of model parameters and control strategies during system operation). Through this mechanism, the artificial intelligence engine can continuously optimize its strategy for generating the first lighting parameter instructions, making it more adaptable to the growth characteristics of specific plant varieties, current environmental conditions, and even subtle differences in specific planting batches, thereby continuously improving the group benefit within the plant growth cycle or across multiple generations of planting cycles, giving the system the ability to adapt and self-evolve.

[0056] Furthermore, the multiple lighting units in the micro-area lighting execution module that can independently adjust lighting parameters independently adjust at least two lighting parameters selected from the following groups for each micro-area according to the first lighting parameter instruction: photosynthetic photon flux density, the intensity and relative proportion of different bands in the spectral composition, the light cycle, the daily cumulative light duration, and the light pulse mode, wherein the different bands include at least one of red light, blue light, far-red light, green light, and ultraviolet light; so as to form a lighting parameter gradient, combination mode or dynamic change mode determined by the artificial intelligence engine based on the group benefit optimization goal between adjacent micro-areas or functional micro-areas.

[0057] Specifically, the multiple lighting units in the micro-area lighting execution module that can independently adjust lighting parameters, such as modular intelligent LED panels or light strips, can adjust the lighting parameters of each micro-area according to the first lighting parameter instruction generated by the artificial intelligence engine. These lighting units can independently adjust at least two lighting parameters selected from the following group for each micro-area: photosynthetic photon flux density PPFD (with an adjustment range of 0 to 2000 μmol·m -2 ·s -1range, the PPFD setting within this range can meet the photosynthetic response threshold of most chlorophyll, so that different crops can obtain effective photosynthetic drive under reasonable experimental conditions, with specific reference to the literature on agricultural photosynthetic response curves), the intensity and relative proportion of different bands in the spectral composition (wherein the different bands include at least one of red light (usually peaking at 630-660nm), blue light (usually peaking at 430-470nm), far-red light (usually peaking at 700-780nm), green light (usually peaking at 500-570nm), and ultraviolet light (UV-A, UV-B), and the dynamic construction and adjustment of the spectrum are achieved by independently controlling the output of LED chips in each band), light cycle (such as daily light duration, which can achieve a custom cycle other than 24 hours), daily cumulative light duration (DLI), or light pulse mode (by adjusting the pulse frequency and duty cycle to achieve specific physiological effects or energy saving). This multi-dimensional, high-precision control capability is the basis for achieving complex lighting heterogeneity, enabling the system to form lighting parameter gradients, complex combination patterns, or lighting patterns that change dynamically over time, determined by the artificial intelligence engine based on group benefit optimization goals, between adjacent micro-areas or functional micro-areas, such as simulating the movement of natural light spots or lighting changes caused by edge effects.

[0058] Furthermore, the artificial intelligence engine in the lighting parameter acquisition and instruction generation module, which is used to generate the internal operating logic and optimization target setting of the first lighting parameter instruction, is based on the effect of "actively inducing plant groups to exhibit group benefits that exceed the simple sum of individual capabilities through controllable, beneficial, and dynamically optimized lighting unevenness by artificial intelligence, thereby achieving overall productivity or environmental adaptability greater than the sum of the independent parts under traditional optimization methods."

[0059] In summary, the intelligently controlled modular horticultural combination lighting system provided in this application actively stimulates the benefits of plant groups through micro-area lighting differentiation driven by artificial intelligence, and is expected to bring significant technological progress in increasing yield, improving quality, enhancing stress resistance, and saving energy and reducing consumption, providing modern facility agriculture with a smarter, more efficient and more sustainable lighting solution.

[0060] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. Intelligently controlled modular gardening lighting system, characterized by: include: The micro-area lighting execution module is configured with multiple lighting units capable of independently adjusting lighting parameters. The micro-area lighting execution module is used to apply differentiated lighting environments to multiple preset micro-areas within the plant planting area according to the first lighting parameter instruction received from the lighting parameter acquisition and instruction generation module; The plant and environment data collection module collects at least one of the plant population status information and micro-area environment perception data in the plant planting area; a group benefit evaluation module, which evaluates the group benefit of the current lighting strategy on the plant group based on at least one of the plant group status information and the micro-area environmental perception data received from the plant and environmental data acquisition module, and transmits the group benefit evaluation result to the lighting parameter acquisition and instruction generation module; The lighting parameter acquisition and instruction generation module is integrated with an artificial intelligence engine, which is used to receive the plant population status information, the micro-area environment perception data and the population benefit evaluation results, and generate or iteratively update the first lighting parameter instruction based on this.

2. The intelligently controlled modular gardening lighting system according to claim 1, wherein: The artificial intelligence engine in the lighting parameter acquisition and instruction generation module generates or iteratively updates the first lighting parameter instruction based on a preset optimization goal of maximizing group benefit, where the group benefit is an overall gain effect that exceeds the simple sum of individual capabilities. The first lighting parameter instruction includes at least one parameter of light intensity, spectral composition, lighting duration or pulse mode customized for each or each group of micro-areas, and is used to actively create and dynamically regulate the heterogeneity of the lighting environment between micro-areas, thereby inducing and stimulating the plant population to exhibit group benefit.

3. The intelligently controlled modular gardening lighting system according to claim 1, wherein: The artificial intelligence engine in the lighting parameter acquisition and instruction generation module uses at least one of the following algorithms to achieve the generation or iterative update of the first lighting parameter instruction for the preset optimization goal of maximizing group benefits: a) Evolutionary algorithm or genetic algorithm encodes the complete lighting parameter combination covering all micro-areas as individuals to be optimized, uses the group benefit evaluation result as the fitness function, and searches for the lighting parameter instruction set that maximizes the group benefit through selection, crossover and mutation operations; b) A reinforcement learning algorithm defines the plant population and its dynamic response to light as the environment. The artificial intelligence engine acts as the agent, and the adjustment of the first light parameter instruction is the action. The group benefit evaluation result constitutes the reward signal. The agent learns through interaction with the environment and forms a control strategy that can output the first light parameter instruction that maximizes the long-term cumulative group benefit based on the current plant population state and environmental data. c) An optimization algorithm based on a proxy model, in which an artificial intelligence engine uses historical data or simulated data to train a proxy model that predicts the impact of different first lighting parameter instructions on group benefits, and performs an optimization search on this proxy model to identify the first lighting parameter instructions that can improve group benefits.

4. The intelligently controlled modular gardening lighting system according to claim 1, wherein: The first illumination parameter instruction generated by the illumination parameter acquisition and instruction generation module includes an instruction set for instructing the micro-area illumination execution module to divide and regulate functional micro-areas in the plant population, and the functional micro-areas include: i) a growth-optimized micro-area, wherein the light parameters are configured to promote the biomass accumulation or formation of a specific economic yield of the plants in the micro-area; ii) Community benefit induction micro-areas, whose lighting parameters are configured to induce plants in the micro-areas to produce specific physiological and biochemical responses, which enhance the comprehensive performance or stress resistance of the entire plant population.

5. The intelligently controlled modular gardening lighting system according to claim 1, wherein: The group benefit evaluation module quantitatively evaluates group benefits by using at least one of the following indicators and uses the quantitative evaluation results as feedback for the AI ​​engine's optimization decision-making: a) Under conditions where total light input energy remains constant or is lower than a baseline, the growth rate or improvement in total biomass, economic yield, or specific target quality components of a plant population is greater than that expected from the simple sum of individual contributions; b) The observed or predicted light use efficiency or radiation use efficiency of the plant population as a whole is higher than the expected theoretical value based on the weighted average performance of the individual plants in each micro-plot under uniform optimized light; c) the degree of improvement in the level of systemic acquired resistance induced by the population benefit-inducing micro-area, or the degree of acceleration in the overall recovery rate of the population after a specific stress, or the reduction in the population loss rate.

6. The intelligently controlled modular gardening lighting system according to claim 1, wherein: The artificial intelligence engine in the lighting parameter acquisition and instruction generation module performs continuous learning and model iteration based on historical plant population status information, micro-area environmental perception data, executed first lighting parameter instructions, and corresponding population benefit evaluation results to optimize its strategy for generating the first lighting parameter instructions.

7. The intelligently controlled modular gardening lighting system according to claim 1, wherein: The multiple lighting units in the micro-area lighting execution module that can independently adjust lighting parameters independently adjust at least two lighting parameters selected from the following group for each micro-area according to the first lighting parameter instruction: photosynthetic photon flux density, the intensity and relative proportion of different bands in the spectral composition, the light cycle, the daily cumulative light duration, and the light pulse mode, wherein the different bands include at least one of red light, blue light, far-red light, green light, and ultraviolet light; so as to form a lighting parameter gradient, combination mode, or dynamic change mode determined by the artificial intelligence engine based on the group benefit optimization goal between adjacent micro-areas or functional micro-areas.

Images