LED outdoor landscape lamp control system based on solar power supply

Abstract

The invention relates to the technical field of solar lighting, and discloses an LED outdoor landscape lamp control system based on solar power supply. The system comprises a solar power supply management module, an energy storage scheduling and state management module, an intelligent sensing and environment identification module, a light control decision module and a self-adaptive feedback and system optimization module. The method comprises the following steps: acquiring voltage, current and irradiance data of a solar module, and combining a maximum power point tracking algorithm to generate power supply data; outputting scheduling parameters based on the dynamic scheduling model, acquiring environment data by using a multi-modal sensor and an image recognition technology, and calculating an illumination control factor; fusing the scheduling parameter, the illumination factor and the time factor, and calculating a target control matrix; and monitoring an execution feedback signal, and iteratively optimizing the control model based on historical control effect data to realize closed-loop optimization control. According to the invention, multi-source data fusion and adaptive control are realized, and the energy efficiency and intelligent level of the LED landscape lamp in a complex environment are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of solar lighting, in particular to a LED outdoor landscape lamp control system based on solar power supply, which is suitable for urban landscape lighting, garden lighting, public space intelligent lighting and the like. BACKGROUND

[0002] The existing landscape lamp control system relies on mains power supply, has large energy consumption and complex wiring: the traditional LED outdoor landscape lamp adopts mains power supply mode, which needs a large number of cable laying and power distribution equipment support in large-scale deployment scenarios, which not only has high construction cost and complex maintenance, but also is difficult to implement in remote or infrastructure-free areas, which is not conducive to the promotion of green energy-saving concept.

[0003] Most systems lack multi-modal environmental perception ability and have single control strategy: the current landscape lighting system mainly relies on single sensor (such as photoresistor or timer) for control, lacks multi-modal perception means such as temperature, humidity, personnel behavior and image recognition, and cannot dynamically adjust the light strategy according to the real environment, resulting in energy waste and unreasonable lighting.

[0004] The control logic is fixed and cannot be self-optimized according to historical operation effect: the control logic in the existing technology mainly adopts preset rules or simple threshold control, lacks the ability to learn and optimize historical operation data, cannot automatically adjust the control model according to different scenes or seasonal differences, has low intelligent degree and poor adaptability.

[0005] There is a lack of efficient collaborative management mechanism of solar power supply and energy storage system: although part of the system integrates solar components and batteries, it mainly adopts simple power detection or fixed scheduling mode, which cannot fully utilize solar energy resources; in rainy weather or power peak, the power allocation efficiency is low, and the system reliability and continuous operation ability are insufficient. SUMMARY

[0006] In order to achieve the above application purpose, the present application provides the following technical scheme: a LED outdoor landscape lamp control system based on solar power supply, comprising the following modules:

[0007] A solar power supply management module is used for collecting voltage data V pan , current data I pan , irradiance data G ins of the solar component, and generating power supply data D pow for controlling the energy storage state;

[0008] An energy storage scheduling and state management module is used for receiving the power supply data D pow , and outputting dynamic scheduling parameters P dyn ;

[0009] Intelligent perception and environment recognition module, for collecting environment data including light L env , temperature T env , humidity H env , and extracting scene feature matrix F img by image recognition, generating light control factor C ill ;

[0010] Light control decision module, for fusing the dynamic scheduling parameter P dyn and light control factor C ill , and combining time factor T sch , calculating target control matrix M ctl ;

[0011] Adaptive feedback and system optimization module, for real-time monitoring of execution feedback signal F exe , and based on historical control effect data E his , iteratively optimizing the control model, outputting adjustment weight vector W adj feedback to the light control decision module, realizing closed-loop optimization control.

[0012] Preferably, the solar power supply management module comprises:

[0013] Photovoltaic data acquisition unit, for collecting voltage data V pan , current data I pan , and irradiance data G ins of the current solar panel, and performing preliminary calibration processing;

[0014] Maximum power tracking unit, based on perturbation and observation algorithm and variable step adjustment strategy, calculating maximum power point P mppt , formula as follows:

[0015]

[0016] Wherein, P is the real-time power of the current solar panel, P=V pan ·I pan , ΔG ins is the irradiance change rate;

[0017] Power distribution control unit, combining predicted irradiance G pre and current state of charge SOC cur of the battery, dynamically outputting power supply data D pow , formula as follows:

[0018]

[0019] Wherein:

[0020] Gpre represents the expected irradiance, obtained from weather forecast;

[0021] SOC cur represents the current state of charge of the battery, obtained from solar cell readings;

[0022] SOC max represents the maximum chargeable capacity of the battery, which is a parameter inherent to the solar cell;

[0023] θ1, θ2, θ3 are power supply scheduling fusion weight coefficients, satisfying θ1+θ2+θ3=1.

[0024] Preferably, the energy storage scheduling and state management module comprises:

[0025] a dynamic scheduling algorithm unit, based on an improved hierarchical decision model, outputting dynamic scheduling parameters P dyn , with the following calculation formula:

[0026]

[0027] wherein P(t) is the output power of the solar module at time t, SOC cur represents the current state of charge of the battery, obtained from solar cell readings, SOC max represents the maximum chargeable capacity of the battery, which is a parameter inherent to the solar cell;

[0028] a backup mode switching unit, which automatically enters energy-saving mode when power supply is insufficient, and adjusts the dynamic scheduling parameters P dyn to ensure the minimum lighting demand.

[0029] Preferably, the intelligent perception and environment recognition module comprises:

[0030] a multi-modal perception unit for collecting outdoor environment data, including illumination L env , temperature T env , and humidity H env ;

[0031] an image recognition unit for scene recognition of monitoring images through a convolutional neural network, extracting scene features F img , such as crowd density, vegetation area, and terrain elevation difference;

[0032] an illumination factor calculation unit for calculating an illumination control factor C ill based on the above data, using the following formula:

[0033]

[0034] wherein L max is the maximum outdoor illumination collected, and Fimg is a scene feature matrix obtained by the image recognition unit, H opt is an optimal humidity constant, and is 0.25.

[0035] Preferably, the light control decision module comprises:

[0036] a time layering processing unit, configured to determine a LED light control time factor T sch according to a system preset, and to perform time layering on the light scene according to the LED light control time factor T

[0037] a control matrix calculation unit, configured to fuse a dynamic scheduling parameter P dyn , a light control factor C ill , and a time factor T sch to construct a multi-factor control target function, and to output a target control matrix M ctl for driving the operation state of the LED landscape light in different regions. ctl Each element m ij in the target control matrix M ij satisfies m cur ∈ [0, 1], and represents a duty cycle of the i-th light group in the j-th time period, and a calculation model thereof is:

[0038]

[0039] wherein, is a target control matrix of the i-th light group in the j-th time period, is a dynamic scheduling parameter of the i-th light group in the j-th time period, is a light control factor required by the i-th light group in the j-th time period, is a preset LED light control time of the i-th light group in the j-th time period, T cur is a current system time, T tot is a time period, and α1, α2, and α3 are normalization weight vectors, and satisfy α1+α2+α3=1.

[0040] an instruction issuing unit, configured to issue the target control matrix M ctl to each LED node to control the brightness, color, and switching state thereof.

[0041] Preferably, the adaptive feedback and system optimization module comprises:

[0042] an execution feedback collection unit, configured to collect an execution feedback signal of each control region in real time The execution feedback signal includes: an actual brightness value a response delay a current , and a user interaction behavior And compare with the previously issued instructions, construct multi-dimensional error vector:

[0043]

[0044] Wherein, The error vector of the ith lamp group in the j period, The brightness value of the ith lamp group in the j period, The response delay of the ith lamp group in the j period, The response delay of the ith lamp group in the j-1 period, pan The current data of the solar module, The current of the ith lamp group in the j period, The deviation between the user's expected behavior and the actual behavior of the ith lamp group in the j period.

[0045] Historical control evaluation unit, according to historical control effect data Construct historical evaluation function The formula is as follows:

[0046]

[0047] Wherein, The target control matrix of the ith lamp group in the j period, The real-time monitoring execution feedback signal of the ith lamp group in the j period, The evaluation value, the lower the value, the more stable the control effect, and the system quantifies the adaptability of each region control strategy according to this, and marks out the abnormal region for optimization.

[0048] Control weight optimization unit, taking error feedback and evaluation score as input, updating the control weight vector α = [α1, α2, α3] in the light control decision module through adaptive weight adjustment strategy, the update strategy adopts the following formula:

[0049]

[0050] Wherein, The newly generated normalized weight vector, The normalized weight vector before updating, The historical evaluation function The first-order partial derivative of the normalized weight vector α i ;

[0051] The new weight vector W adj = α (new) Will be retransmitted to the light control decision module as optimization feedback, realizing parameter self-tuning and closed-loop operation.

[0052] Compared with the prior art, the present application has the beneficial effects of:

[0053] Improving power supply efficiency and system adaptability: by integrating the maximum power point tracking and power supply scheduling fusion algorithm in the solar power supply management module, the current maximum power supply power can be accurately obtained, and the dynamic optimization output of the power supply data can be realized in combination with the battery state of charge, the response of the photovoltaic energy efficiency utilization rate and the energy storage management is improved, and the system can be stably operated under different climate conditions.

[0054] Realize multi-dimensional perception driven fine light control: the system integrates environmental perception data, image recognition features and time factors to build a multi-factor control matrix, realizes intelligent control of the running state of LED landscape lights in different regions and different time periods through the light control decision module, realizes the transformation from timing switch to scene response, and significantly improves the energy saving and intelligent interaction of the landscape lights.

[0055] Construct a closed-loop feedback and self-optimization control mechanism: through the adaptive feedback and system optimization module, real-time collection of brightness, current, user behavior and other execution signals is realized, an error vector is constructed, and an evaluation function is calculated based on historical control effect, the control weight is updated by gradient optimization, dynamic learning and adaptive adjustment of light control parameters are realized, the system stability and user experience are effectively improved, and strong robustness and self-evolution ability are possessed. BRIEF DESCRIPTION OF DRAWINGS

[0056] Figure 1 The system module schematic diagram provided for the present application;

[0057] Figure 2 The solar power supply management module schematic diagram provided for the present application;

[0058] Figure 3 The energy storage scheduling and state management module schematic diagram provided for the present application;

[0059] Figure 4 The intelligent perception and environmental recognition module schematic diagram provided for the present application;

[0060] Figure 5 The light control decision module schematic diagram provided for the present application;

[0061] Figure 6 The adaptive feedback and system optimization module schematic diagram provided for the present application. DETAILED DESCRIPTION

[0062] In order to make the person skilled in the art better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should belong to the scope of protection of the present application.

[0063] Reference Figures 1-6 The embodiment of the present application provides a solar-powered LED outdoor landscape lamp control system, comprising the following modules:

[0064] The solar power supply management module is used for collecting voltage data V pan , current data I pan , irradiance data G ins , and generating power supply data D pow for controlling the energy storage state.

[0065] In the solar power supply management module, first, the working state of the current solar cell panel is acquired by the photovoltaic data acquisition unit, including voltage data V pan , current data I pan and environmental irradiance data G ins . In this process, the system performs preliminary calibration on the collected raw voltage, current and irradiance data through the built-in temperature compensation and standard deviation filtering mechanism, to ensure the stability and accuracy of data input, and to provide a reliable basis for subsequent calculation.

[0066] Next, the system inputs the calibrated V pan and I pan to the maximum power tracking unit, adopts the perturbation and observation algorithm and combines the variable step adjustment strategy to dynamically adjust the solar panel output. The system calculates the current power P = V pan ·I pan in real time, and after detecting the change rate ΔG ins of irradiance G ins , adjusts the perturbation direction and step length to approach the optimal working point, and finally determines the maximum power point P mppt . This strategy effectively avoids power loss under drastic fluctuations in environmental light intensity, and improves solar conversion efficiency.

[0067] After power tracking is completed, the power distribution control unit schedules the power supply strategy according to the predicted irradiance G pre provided by the current meteorological information platform and the current state of charge SOC cur of the battery. By comparing the maximum power point P mppt, the expected irradiance and the current state of charge are input into the control model, and combined with weight coefficients θ1, θ2, θ3 (which satisfy θ1+θ2+θ3=1), the final dynamic power supply data D is calculated pow , which is used to control the illumination intensity and working time of the LED landscape lamp. If the system monitors that the SOC cur is much lower than the SOC max , the power supply scheduling will automatically reduce the output power to prevent the battery from over-discharging and ensure that the device still has basic lighting function in continuous rainy and other extreme weather conditions.

[0068] The energy storage scheduling and state management module is used to receive the power supply data D pow and output the dynamic scheduling parameter P dyn .

[0069] In the energy storage scheduling and state management module, first, the dynamic scheduling algorithm unit continuously monitors the real-time output power P(t) of the solar component through the time period set by the system (every 5 minutes). P(t) is calculated by V pan and I pan obtained in the aforementioned solar power supply management module. Subsequently, the unit performs dynamic analysis based on the improved hierarchical decision model, which not only considers the power fluctuation characteristics, but also integrates factors such as the current state of charge SOC cur and the maximum capacity SOC max of the battery. By calculating the integral value of P(t) in the scheduling period at the current time t, i.e. accumulating and quantifying the total output of solar power in the period, and multiplying the integral value by the ratio of the state of charge, the dynamic scheduling parameter P dyn is finally output, and the calculation formula is The dynamic scheduling parameter P dyn not only represents the current power supply capability judgment result, but also directly affects the setting of the lighting brightness level in the subsequent system. In the case of good lighting conditions and SOC cur close to SOC max , the value of P dyn is relatively high, and the system allows the landscape lamp to run in the set high brightness mode. Conversely, when continuous rainy weather causes P(t) to continuously decrease, or the battery state of charge is low (i.e. SOC cur is small), P dyn automatically decreases, thereby guiding the system to reduce the lighting level and achieve supply-demand balance. In addition, the dynamic scheduling algorithm unit also has a historical data memory mechanism, which will give priority to the average lighting power change trend in the past 48 hours to deal with short-term instability problems such as sunset delay and strong light mutation, further improving the adaptability and accuracy of the scheduling response.

[0070] In the extreme condition of serious shortage of solar power supply or continuous bad weather, to ensure the minimum function of the landscape lighting system, the system automatically triggers the backup mode switching unit. The unit continuously monitors the calculation results of P dyn and the set critical threshold, and makes a decision in combination with the current lighting task intensity and the historical power consumption model. When the judgment result shows that the current P dyn is not enough to support the current lighting task, the system automatically switches to the energy-saving lighting mode, reduces the lighting brightness and shortens the lighting time period on the premise of not completely closing the lighting, and ensures the basic lighting demand. At the same time, the backup mode switching unit further adjusts the set parameters of P dyn , such as shortening the integral period of P(t) or temporarily reducing the upper limit of the lighting load, so as to maintain limited but continuous energy output in the low-power state and avoid the complete stop of the system. This strategy is particularly suitable for night walkways, parks, landscape bridges and other scenes with strong dependence on lighting for public safety, ensuring that the system has basic service capability in the most unfavorable energy environment, and embodying the practicality and robustness of the application in energy storage management and energy scheduling.

[0071] The intelligent perception and environment recognition module is used to collect environment data including illumination L env , temperature T env , humidity H env , and generate illumination control factor C img by image recognition to extract scene feature matrix F ill .

[0072] In the intelligent perception and environment recognition module, first, the multi-modal perception unit uses the environment sensor array integrated on the lamp pole to collect raw data such as L env environmental light intensity), T env (environmental temperature) and H env (environmental humidity) in real time. The sensor sampling frequency is set to once every 30 seconds, and the system performs dynamic sliding average filtering on the data to remove abnormal values caused by sudden interference (such as car lights, short-time shadow, etc.), so as to ensure the stability and effectiveness of the input data. Among them, L env data is used to reflect the current lighting conditions, and T env and H env are used to assist in inferring the weather state and humidity fluctuation to provide basic parameters for subsequent algorithm judgment.

[0073] The image recognition unit collects real-time images through the high-resolution wide-angle camera arranged at the top of the lamp pole, and the image sampling interval is 1 frame every 10 seconds. The system uses a trained convolutional neural network model to quickly process the image and extract scene features F imgThe model supports classification and judgment of key elements such as crowd density, vegetation area proportion, and terrain elevation in the image. For example, when a significant increase in crowd density (such as gathering in a park square at night on weekends) or a low-lying area (light-shielded area) is identified, the convolutional neural network model will generate a specific scene feature matrix F img The matrix is then input as a weight factor to the light factor calculation unit for the final operation of the light adjustment factor. The entire image recognition and data generation process is completed locally by the edge AI chip, ensuring a response speed of less than 500 milliseconds and guaranteeing real-time perception ability in dynamic environments.

[0074] Finally, the light factor calculation unit fuses the aforementioned multi-modal environment data with the image feature matrix according to the formula The current light control factor C ill In this formula, L env is first normalized with the maximum light value L max historically collected to obtain the current relative light intensity, which is then multiplied by the scene feature F img matrix obtained through image recognition to quantify the potential demand for lighting intensity in the current area. Subsequently, it is multiplied by the humidity correction factor H When the humidity rises above H opt , the result value of C ill will be further reduced to avoid providing too strong lighting in high-humidity environments, reducing light pollution and energy consumption. In this embodiment, through the dynamically calculated C ill value, the system can automatically adjust the output brightness of the LED luminaire according to the actual scene state. This strategy has been used in multiple pilot areas, reducing night-time energy consumption by more than 25% while improving public satisfaction with lighting comfort to more than 92%, effectively achieving the dual goals of energy saving and intelligent control.

[0075] The light control decision module is used to fuse the dynamic scheduling parameter P dyn and the light control factor C ill , and combine the time factor T sch to calculate the target control matrix M ctl .

[0076] In the light control decision module, the time layering processing unit system intelligently divides the daily 24-hour operation period according to the pre-configured LED light control time factor T sch . Taking the deployment of this project as an example, the system divides the daily time into:

[0077] 1. Sunset transition period 17:30-22:30, with the most concentrated flow of people, the light needs to ensure high brightness and dynamic color change to attract attention;

[0078] 2. Energy-saving period at night 22:30-5:30 the next day, the flow is rare, the system will automatically reduce the brightness and turn off part of the area light group to save energy. This hierarchical structure not only optimizes the energy use efficiency, but also significantly improves the scene adaptation ability of the light application, meeting the dual needs of city beauty and energy consumption control.

[0079] The control matrix calculation unit accesses multiple real-time running parameters and constructs a control target function To generate the target control matrix M ctl The system first obtains dynamic scheduling parameters from the energy storage scheduling module This parameter reflects the available energy distribution ability of each light group at each time period in real time; then synchronously obtains the illumination control factor from the intelligent sensing module Reflects the actual demand degree of brightness by the current environment; finally, combines the preset time factor And the current system time T cur Calculate the time period correlation index and comprehensively construct the control function to accurately set the duty cycle m ij of different light groups at different time periods. For example, in the night peak period (such as light group i=5, j=2), if the area is densely populated, the ambient light is weak, and the system energy is sufficient, the input values in the formula will collectively drive the target m 52 Tend to 1, representing that the light group needs to work with a duty cycle close to 100%; on the contrary, in the late night period (such as light group i=3, j=3), if the current power is low, the environmental humidity is high, and the target brightness requirement is not high, then m 33 Will be automatically compressed to below 0.3, only maintaining the basic lighting function. The three normalized weights α1, α2, and α3 can be dynamically adjusted according to the actual use scene, such as appropriately increasing the weight of α1 in the energy-saving priority area to enhance the influence of P dyn ; while in the beautification priority area, the proportion of α2 or α3 can be increased to enhance the influence of environmental factors and time strategy on control output, realizing flexible adaptation of system strategy.

[0080] After the control matrix calculation is completed, the instruction issuing unit takes all m ctl values in the matrix M ij as the basis for regulation and control, and issues them to each LED terminal node through the central control node in a standardized communication protocol. In this embodiment, the terminal node is controlled by a microcontroller with PWM driving function, and each node parses M ctlcorresponding target duty cycle in the matrix, dynamically adjusts the PWM output period, and realizes the brightness precision control within ±5%. At the same time, the system also supports the synchronous issuance of color temperature and color adjustment instructions. During a specific holiday or event, the control matrix can introduce color light mode parameters to realize color gradient and synchronous flicker control in the whole area or local area. In addition, the asynchronous confirmation mechanism is adopted in the instruction issuance process. The control node must wait for the terminal feedback signal after issuance to ensure the successful execution of the scheduling instruction of each landscape lamp. When the node is offline or the data is abnormal, the control center will activate the redundant lamp group or adjust the m ij value to avoid local dark areas and ensure overall visual continuity and safety lighting requirements.

[0081] The adaptive feedback and system optimization module is used to monitor the execution feedback signal F exe based on historical control effect data E his , iteratively optimizes the control model, and outputs the adjustment weight vector W adj feedback to the light control decision module to realize closed-loop optimization control.

[0082] In the adaptive feedback and system optimization module, the execution feedback acquisition unit is deployed with multiple LED lamp groups. At 19:10 for j = 1 period, the 4th lamp group receives the target duty cycle m 41 = 0.85, that is, it is expected to maintain 85% brightness output.

[0083] After the end of this period, the execution feedback acquisition unit collects feedback signals through multiple source data channels, including: the actual brightness value read by the sensor (with slight deviation), the system calls the log record response delay (last period τ is 0.9s), the current sensor measures the actual consumption (compared with the photovoltaic panel power supply current I pan = 0.50A), at the same time, the users of the square report their opinions on the over-brightness of the area through the interactive interface, forming the user behavior deviation index Based on the above information, the system calculates the error vector of this lamp group:

[0084]

[0085] Through the integration of the error vector and historical data, basic data input is provided for subsequent control weight adjustment.

[0086] The system inputs the error vector into the historical control evaluation unit, combines the historical control effect to construct a comprehensive evaluation function The calculation formula is Still taking the 4th lamp group as an example, its error vector set in the first 5 periods is saved as Control target matrix history data With feedback data The complete adjustable. History evaluation function formula expansion, the error square and control instruction ratio of history stage are accumulated, and the following is obtained:

[0087]

[0088] Evaluation result The value of 0.022, the evaluation threshold set by the system is 0.015, which is significantly high. The system identifies that the light group has brightness deviation, response fluctuation and user satisfaction decrease in the past five periods, which belongs to the sub-stable state. Therefore, the area is labeled as a priority optimization label and marked as the object of this round of weight optimization.

[0089] After identifying that the No. 4 light group has control deviation, the control weight optimization unit starts the optimization algorithm to real-time fine-tune the control weight vector a = [a1, a2, a3] in the aforementioned light control decision module. The initial weight currently adopted by the system is a (old) =[0.4, 0.35, 0.25], which respectively corresponds to dynamic scheduling, light control factor and time factor. Combined with the partial derivative value of the history evaluation function (calculated from the error trend and feedback index), the system completes a round of weight correction:

[0090]

[0091] The new weight vector W adj = a (new) =[0.37, 0.38, 0.25] is obtained after updating. The adjustment result shows that the system appropriately reduces the dependence on the dynamic scheduling parameter and strengthens the feedback ability of the light control factor. The updated weight vector will be automatically transmitted back to the light control decision module, and will take effect in the next round of control matrix M ctl Calculation, realizing the self-tuning of the control logic.

[0092] It should be noted that the embodiments in the present application and the features and technical solutions in the embodiments can be combined with each other without conflict.

[0093] Obviously, the above-described embodiments are only some embodiments but not all the embodiments of the present application, the preferred embodiments of the present application are shown in the drawings, but do not limit the patent scope of the present application. The present application can be implemented in many different forms, and conversely, the purpose of providing these embodiments is to make the disclosure of the present application more thorough and comprehensive. Although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions recorded in the foregoing specific embodiments, or make equivalent replacements to some technical features therein. Any equivalent structure made by using the content of the present application specification and drawings, directly or indirectly applied to other related technical fields, is also within the patent protection scope of the present application.

Claims

1. A solar-powered LED outdoor landscape light control system, characterized in that: Includes the following modules: Solar power supply management module, used to collect voltage data V of solar panels pan , current data I pan , irradiance data G ins , and generate power supply data D for controlling energy storage status pow ; Energy storage scheduling and status management module, used to receive power supply data D pow , and output the dynamic scheduling parameter P dyn ; Intelligent perception and environment recognition module, used to collect information including light L env , temperature T env 、Humidity H env The environmental data including the scene feature matrix F is extracted through image recognition img , generate the light control factor C ill ; Lighting control decision module, used to integrate the dynamic scheduling parameters P dyn and light control factor C ill , and combined with the time factor T sch , calculate the target control matrix M ctl ; Adaptive feedback and system optimization module, used to monitor the execution feedback signal F in real time exe , and based on historical control effect data E his Iteratively optimize the control model and output the adjustment weight vector W adj Feedback is sent to the lighting control decision module to achieve closed-loop optimization control.

2. A solar-powered LED outdoor landscape light control system according to claim 1, characterized in that: The solar power supply management module includes: Photovoltaic data acquisition unit, used to collect the current voltage data V of the solar panel pan , current data I pan , irradiance data G ins , and perform preliminary calibration processing; The maximum power tracking unit calculates the maximum power point P based on the disturbance observation algorithm and variable step size adjustment strategy mppt , the formula is as follows: Where P is the current real-time power of the solar panel, P = V pan I pan , ΔG ins is the irradiance change rate; Power distribution control unit, combined with the expected irradiance G pre and the battery's current state of charge SOC cur , dynamically output power supply data D pow , the formula is as follows: in: G pre Indicates the expected irradiance, obtained from the weather forecast; SOC cur Indicates the current state of charge of the battery, obtained based on the solar cell reading; SOC max Indicates the maximum rechargeable capacity of the battery, which is an inherent parameter of the solar cell; θ1, θ2, and θ3 are power supply scheduling fusion weight coefficients, satisfying θ1+θ2+θ3=1.

3. The solar-powered LED outdoor landscape light control system according to claim 1, characterized in that: The energy storage scheduling and status management module includes: The dynamic scheduling algorithm unit, based on an improved hierarchical decision model, outputs the dynamic scheduling parameter P dyn , the calculation formula is as follows: Among them, P(t) is the output power of the solar panel at time t, SOC cur Indicates the current state of charge of the battery, obtained based on the solar cell reading, SOC max Indicates the maximum rechargeable capacity of the battery, which is an inherent parameter of the solar cell; The standby mode switching unit automatically enters the energy-saving mode when the power supply is insufficient, and adjusts the dynamic scheduling parameter P dyn To ensure minimum lighting requirements.

4. The solar-powered LED outdoor landscape light control system according to claim 1, characterized in that: The intelligent perception and environment recognition module includes: Multimodal sensing unit, used to collect outdoor environment data, the environment data includes light L env , temperature T env 、Humidity H env ; The image recognition unit uses a convolutional neural network to perform scene recognition on the surveillance image and extract scene features F img , identifying crowd density, vegetation areas, and terrain height differences; The illumination factor calculation unit calculates the illumination control factor C based on the above data ill , using the following formula: Among them, L max is the maximum value of outdoor light collected, F img H is the scene feature matrix obtained by the image recognition unit. opt is the optimal humidity constant, which is 0.

25.

5. The solar-powered LED outdoor landscape light control system according to claim 1, characterized in that: The lighting control decision module includes: Time layered processing unit, based on the system preset LED light control time factor T sch Perform hierarchical management of lighting scenes, including sunset, night peak, and late night; Control matrix calculation unit, used to integrate dynamic scheduling parameters P dyn , light control factor C ill and the time factor T sch , in order to construct a multi-factor control objective function and output the target control matrix M used to drive the operating state of LED landscape lights in different areas ctl , where the target control matrix M ctl Each element m in ij ∈[0, 1], the m ij represents the duty cycle of the i-th lamp group in the j-th time period, and its calculation model is: in, is the target control matrix of the i-th lamp group in the j-th period, is the dynamic scheduling parameter of the i-th lamp group in the j-th period, is the illumination control factor required by the i-th lamp group in the j-th period, T is the preset LED light control time of the i-th light group in the j-th period, cur is the current system time, T tot is the time period, α1, α2, and α3 are normalized weight vectors, and satisfy α1+α2+α3=1. The instruction issuing unit sends the target control matrix M ctl Send it to each LED node to control its brightness, color and switch status.

6. The solar-powered LED outdoor landscape light control system according to claim 1, characterized in that: The adaptive feedback and system optimization module includes: Execution feedback acquisition unit, used to collect execution feedback signals of each control area in real time The execution feedback signal Includes: actual brightness value Response delay Current and user interaction behaviors And compare it with the previously issued instructions to construct a multi-dimensional error vector: in, is the error vector of the i-th lamp group in the j-th period, is the brightness value of the i-th lamp group in the j-th period, is the response delay of the i-th lamp group in the j-th period, is the response delay of the i-th lamp group in the j-1th period, I pan is the current data of the solar panel, is the current of the i-th lamp group in the j-th period, is the deviation between the user's expected behavior and actual behavior of the i-th lamp group in the j-th period. Historical control evaluation unit, based on historical control effect data Constructing a history evaluation function The formula is as follows: in, is the target control matrix of the i-th lamp group in the j-th period, is the real-time monitoring execution feedback signal of the i-th lamp group in the j-th period, The lower the value, the more stable the control effect. The system quantitatively scores the adaptability of the control strategy for each area based on this value and marks abnormal areas for priority optimization. The control weight optimization unit takes the error feedback and evaluation score as input and updates the control weight vector α = [α1, α2, α3] in the lighting control decision module through an adaptive weight adjustment strategy. The update strategy adopts the following formula: in, is the newly generated normalized weight vector, is the normalized weight vector before updating, is the historical evaluation function Normalized weight vector α i The first partial derivative of ; The adjusted new weight vector W adj =α (new) It will be retransmitted to the lighting control decision module as optimization feedback to achieve parameter self-tuning and closed-loop operation.

Images