An intelligent light period regulation intelligent light control system based on machine learning

Abstract

The application relates to the technical field of intelligent lighting control, and discloses an intelligent photoperiod regulation intelligent lamp control system based on machine learning. The system comprises an intelligent lamp device, an ambient light sensor, a first-aid call module and a machine learning processing module. The intelligent lamp device has dual functions, can be used as an independent night lamp and functions as a mobile phone charger when there is no photoperiod regulation requirement, and can be driven by the machine learning processing module to realize intelligent regulation through five processes when there is a regulation requirement. The photoperiod analysis process generates a requirement result based on ambient light sensor data, the scheduling process generates a task sequence based on the requirement result, the distribution process formulates a resource scheme, the execution process is converted into an execution instruction, and the monitoring process generates regulation feedback data. When a user has a first-aid requirement, the first-aid call module can quickly respond and process. The system combines machine learning and multi-module cooperation, realizes accurate photoperiod regulation and multi-element integration of functions, and improves use convenience and scene adaptability.

Description

Technical Field

[0001] This invention relates to the field of intelligent lighting control technology, specifically to an intelligent light cycle adjustment intelligent lamp control system based on machine learning. Background Technology

[0002] With the popularization of intelligent technologies, lighting equipment has evolved from traditional single-function lighting tools to multifunctional and intelligent systems. Photoperiod regulation, as a crucial factor influencing biological rhythms and quality of life, has gradually become a research focus in the field of intelligent lighting. In modern life, the pace of life is constantly accelerating. The indoor lighting environment not only needs to meet basic lighting requirements but also needs to be dynamically adjusted according to different scenarios, user needs, and changes in ambient light to adapt to human physiological rhythms and improve comfort and work efficiency. Whether it's sleep assistance in the home, lighting for children's growth, adapting to work conditions in the office, or the lighting needs of special groups, all place higher demands on the level of intelligence in photoperiod regulation.

[0003] Most smart lighting products currently on the market only offer simple brightness adjustment, color switching, or timer functions. Their adjustment logic is largely based on preset programs or manual user settings, lacking the ability to dynamically perceive and intelligently analyze changes in ambient light and the user's actual needs. For example, while some products are equipped with light sensors, they can only control the light on and off based on ambient light intensity, failing to generate personalized light cycle adjustment schemes that incorporate factors such as the user's lifestyle and activity level. This passive adjustment method struggles to accurately match the human body's dynamic needs for the light environment, often resulting in a mismatch between the light environment and the user's circadian rhythm. Long-term use may negatively impact the user's sleep quality and overall well-being.

[0004] Existing smart lighting devices are relatively limited in function, mostly focusing solely on illumination and failing to effectively integrate with other life services, resulting in a waste of equipment resources. In homes and dormitories, users often need to equip themselves with multiple devices such as nightlights and chargers, which not only takes up space but also increases usage costs and operational complexity. Furthermore, in sudden emergencies, traditional lighting devices cannot provide effective emergency assistance, requiring users to find additional communication devices to initiate emergency requests. This delay may cause them to miss the optimal rescue opportunity, posing a significant safety hazard, especially for vulnerable groups such as the elderly and children.

[0005] At the technical level, existing light cycle regulation technologies largely rely on fixed algorithms and lack deep application of artificial intelligence technologies such as machine learning. This prevents them from continuously learning user behavior and environmental changes to optimize regulation strategies. Consequently, lighting equipment exhibits poor adaptability and struggles to cope with complex and ever-changing usage scenarios. For example, it cannot autonomously optimize key parameters such as the start and end times of the light cycle and the light intensity variation curve based on seasonal changes, weather conditions, and user schedules. Furthermore, some systems lack robust feedback mechanisms, failing to monitor and dynamically correct the light regulation effect in real time, resulting in insufficient regulation precision and further impacting the user experience. Summary of the Invention

[0006] The purpose of this invention is to provide a machine learning-based intelligent light cycle adjustment intelligent lamp control system to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides a machine learning-based intelligent light cycle adjustment intelligent lamp control system, the system comprising:

[0008] The system includes an intelligent lighting device, an ambient light sensor, an emergency call module, and a machine learning processing module. When there is no need for light cycle adjustment, the intelligent lighting device is used as an independent night light and a mobile phone charger. When there is a need for light cycle adjustment, the machine learning processing module executes a light cycle analysis process, a light task scheduling process, a light resource allocation process, a light cycle execution process, and a real-time monitoring process to achieve intelligent light cycle adjustment.

[0009] When an emergency call is needed, the emergency call module processes the user's emergency request. The optical cycle analysis process generates optical demand analysis results based on ambient light sensor data, the optical task scheduling process generates optical task sequences based on the optical demand analysis results, the optical resource allocation process generates optical resource allocation schemes based on the optical task sequences, the optical cycle execution process generates optical cycle execution instructions based on the optical resource allocation schemes, and the real-time monitoring process generates optical adjustment feedback data based on the optical cycle execution instructions.

[0010] Preferably, the optical demand analysis results include optical intensity demand values, a list of available equipment statuses, and the matching degree between optical demand and equipment; the optical task sequence includes task start time points, task priority order, and a task resource demand table; the optical resource allocation scheme includes equipment resource allocation ratios, time conflict resolution results, and resource utilization parameters; the optical cycle execution instructions include an optical cycle timetable, a set of equipment control commands, and execution status identifiers; and the optical adjustment feedback data includes ambient light change records, equipment operating status data, and optical cycle deviation values.

[0011] Preferably, the steps for obtaining the light demand analysis results are as follows: collecting ambient light intensity data through an ambient light sensor, combining it with the status data of the smart lighting device, calculating the difference between the current light demand and the device capacity, and generating a light demand difference report; analyzing the light demand difference report, using machine learning algorithms to predict the trend of light demand changes, matching device resources with light demand, and generating a device-light demand matching table; adjusting the light demand allocation weights according to the device-light demand matching table, and generating the light demand analysis results.

[0012] Preferably, the steps for obtaining the optical task sequence are as follows: extracting the optical demand intensity and time distribution from the optical demand analysis results, calculating the start time and duration of each optical task using a task scheduling algorithm, and generating a preliminary task schedule; based on the preliminary task schedule, evaluating resource conflicts and priorities between tasks, optimizing the task order using a sorting model, and generating a task priority list; combining the task priority list and device resource constraints, calculating the task execution interval and resource occupancy rate, and generating the optical task sequence.

[0013] Preferably, the steps for obtaining the optical resource allocation scheme are as follows: deriving task resource requirements and time nodes from the optical task sequence, analyzing the available resources and constraints of the smart lighting device, and generating a resource requirement and constraint comparison table; using a resource allocation algorithm to process the resource requirement and constraint comparison table, analyzing conflicts and deficiencies in device resource allocation, and generating conflict analysis results; based on the conflict analysis results, reallocating device resources and time slots, optimizing resource utilization efficiency, and generating an optical resource allocation scheme.

[0014] Preferably, the step of obtaining the optical cycle execution command specifically includes: formulating a detailed optical cycle timetable based on the optical resource allocation scheme, including optical intensity adjustment points and equipment operation commands, and generating a draft optical cycle timetable; verifying the compatibility of the draft optical cycle timetable with the equipment status, adjusting time conflicts and resource allocation, and generating a verified optical cycle timetable; converting the verified optical cycle timetable into a set of control commands that the equipment can execute, and generating optical cycle execution commands.

[0015] Preferably, the step of acquiring the light adjustment feedback data specifically includes: during the light cycle execution process, monitoring ambient light changes in real time through an ambient light sensor, recording light intensity data and timestamps, and generating an ambient light change record; simultaneously monitoring the operating status of the smart light device, collecting device parameters and operation logs, and generating device operating status data; comparing the ambient light change record with the light cycle execution command, calculating the light cycle execution deviation, and generating a light cycle deviation value; and combining the ambient light change record, device operating status data, and light cycle deviation value to generate light adjustment feedback data.

[0016] Preferably, the system further includes an emergency call processing module, which interrupts the current optical cycle adjustment process based on the emergency request triggered by the user, prioritizes the emergency call, generates an emergency response command, and records emergency event data.

[0017] Preferably, the machine learning processing module uses historical photoperiod data and emergency event data to train an adaptive model and dynamically optimize the parameter settings in the photoperiod analysis process and the phototask scheduling process.

[0018] Preferably, the system integrates a user behavior analysis unit, which collects user activity patterns and data, and combines them with light regulation feedback data and emergency response instructions to adjust the light cycle regulation strategy and emergency call response priority.

[0019] Compared with the prior art, the beneficial effects of the present invention are:

[0020] The dual-function design of the smart lighting device enables efficient use of device resources, breaking the limitations of traditional lighting equipment with its single function. When there is no need for light cycle adjustment, the device can be used as an independent night light, providing users with soft and safe basic lighting for nighttime activities, avoiding strong light stimulation that could affect the body's physiological rhythms. Simultaneously, it also functions as a mobile phone charger, eliminating the need for additional charging equipment, reducing the number of devices in the room, saving space, and lowering user costs and operational complexity. This integrated design allows the device to adapt to various scenarios such as homes, dormitories, and offices, meeting diverse user needs at different times and enhancing its practicality and user acceptance.

[0021] The introduction of a machine learning processing module provides intelligent core support for photoperiod regulation, changing the passive regulation mode of traditional systems that rely on preset programs. Through a photoperiod analysis process, real-time data collected by ambient light sensors is deeply mined to accurately capture key information such as ambient light intensity and trends, generating realistic analysis results based on the user's potential light needs. The light task scheduling process constructs an ordered sequence of light tasks based on these results, ensuring that light regulation behavior matches the user's activity rhythm and physiological rhythm; the light resource allocation process rationally plans light resources according to the task sequence, avoiding waste and ensuring the stability and efficiency of the light regulation process. This dynamic regulation mechanism based on machine learning can continuously learn user habits and scene change patterns, constantly optimizing the photoperiod regulation scheme and gradually forming personalized regulation strategies. For example, it can automatically adjust the attenuation curve of pre-sleep light intensity based on the user's sleep habits and adapt the light color temperature to different times of day based on the user's work status.

[0022] The real-time monitoring process establishes a complete closed-loop adjustment mechanism, making the photoperiod adjustment process traceable and correctable. This process tracks the photoperiod adjustment effect in real time based on photoperiod execution commands. The generated feedback data promptly reflects the deviation between the adjustment scheme and actual needs, providing an update basis for the machine learning processing module and driving the continuous improvement of the adjustment strategy. This dynamic feedback mechanism effectively improves the accuracy of photoperiod adjustment, avoiding inaccuracies caused by sudden environmental changes or altered user requirements, ensuring that the lighting environment always meets user needs, and creating a more comfortable lighting experience for users.

[0023] The integrated emergency call module adds a crucial safety barrier for users, especially the elderly, children, and those living alone. In emergencies such as sudden illness or accidental injury, users can quickly initiate an emergency request through the system without needing to find additional communication devices, shortening emergency response time and providing strong support for protecting user safety. This design, which combines lighting and safety functions, expands the application value of smart lighting devices, making them not only lighting tools but also safety aids in users' lives, thus enhancing the overall competitiveness of the system. Attached Figure Description

[0024] Figure 1 This is a schematic diagram illustrating the working principle of the intelligent light cycle adjustment system based on machine learning as described in this invention.

[0025] Figure 2 A flowchart for obtaining the results of optical demand analysis;

[0026] Figure 3 This is a flowchart for obtaining the light mission sequence. Detailed Implementation

[0027] 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.

[0028] Please see Figure 1This invention provides a machine learning-based intelligent light cycle adjustment control system for smart lights. The system includes a multi-functional smart light device. When there is no need for light cycle adjustment, the smart light device provides basic lighting as an independent night light and also integrates a mobile phone charging interface to support mobile device charging needs. When an ambient light cycle adjustment need arises, a machine learning processing module activates and coordinates the light cycle analysis process, the light task scheduling process, the light resource allocation process, the light cycle execution process, and the real-time monitoring process. The light cycle analysis process generates light demand analysis results based on real-time light intensity data collected by an ambient light sensor. The light task scheduling process receives the light demand analysis results and plans a sequence of light tasks. The light resource allocation process processes the light task sequence to allocate resources to the smart light device, forming a light resource allocation scheme. The light cycle execution process converts the light resource allocation scheme into operable light cycle execution instructions. The real-time monitoring process monitors environmental changes and device status during instruction execution, generating light adjustment feedback data. An emergency call module operates independently of the light cycle adjustment process. When a user triggers an emergency request, the current light cycle adjustment activity is interrupted, the emergency event is prioritized, and an emergency response instruction is generated. The machine learning processing module uses historical data and real-time feedback to continuously optimize the photoperiod adjustment strategy, achieving adaptive intelligent control.

[0029] Example 1: See Figure 2The light demand analysis results include light intensity demand values, a list of available device statuses, and the matching degree between light demand and device status. The light intensity demand value is a quantitative indicator representing the target illumination level to be achieved within a specific time and space range. The list of available device statuses is a dynamically updated data structure that enumerates the current operating status, resource idleness, and functional completeness of all smart lighting devices in the system. The matching degree between light demand and device status is a calculated parameter reflecting the degree of fit between the capabilities of one or more smart lighting devices and specific light demands. The acquisition of light demand analysis results is achieved by collecting ambient light intensity data through an ambient light sensor. The ambient light sensor continuously monitors the illumination intensity of the surrounding environment at a preset frequency, generating a series of timestamped light intensity readings. The collected ambient light intensity data is transmitted to the machine learning processing module, which simultaneously receives status data from each smart lighting device. This status data includes the device's on / off status, current brightness level, power consumption, network connection status, and fault codes. The differential calculation unit built into the machine learning processing module compares the current ambient light intensity data with a preset target light intensity range, which may be based on user settings, a schedule, or predictions from a machine learning model. The difference calculation unit outputs a light demand difference report, which details the gap between the current light level and the target level, including the magnitude of the difference, its duration, and spatial distribution. The analysis of the light demand difference report employs machine learning algorithms to predict light demand trends. These algorithms can be time series prediction models, such as Long Short-Term Memory networks or Autoregressive Integral Moving Average models. These models are trained using historical photocycle data, which includes records of ambient light changes over a past period, photocycle execution commands, and external factors such as weather and time. The model predicts future light demand intensity fluctuations by analyzing historical patterns. The prediction results are combined with real-time acquired ambient light intensity data to generate a comprehensive light demand prediction view. The matching of device resources with light demand is performed by the resource matching unit. This unit accesses the device availability list and evaluates the available light output capability, color adjustment range, beam angle, and physical location of each smart light device. The resource matching unit compares the device capabilities with the predicted light demand, calculates a suitability score for each device to meet specific needs, and generates a device-light demand matching table. The device-to-light-demand matching table is a matrix or list structure that lists the matching score between each smart light device and different light demands.

[0030] The optical demand allocation weights are adjusted based on the device-to-optical-demand matching table. These weights are adjustable parameters used to determine how optical demand tasks are allocated when multiple devices are available. Weight adjustments may be based on device efficiency, energy-priority strategies, or user preferences. The adjustment process may involve optimization algorithms to maximize overall matching or minimize total energy consumption. The final optical demand analysis result is a structured data object containing calculated optical intensity demand values, an updated list of available device states, and quantified optical demand-to-device matching. The optical task sequence includes task start times, task priority order, and a task resource requirement table. The task start time defines the precise moment each optical task begins execution; the task priority order is a sorted list that determines the execution order of multiple tasks during resource contention; and the task resource requirement table details the type, quantity, and configuration parameters of smart light device resources required to execute each task. The optical task sequence acquisition steps extract optical demand intensity and temporal distribution from the optical demand analysis results. Optical demand intensity data comes from optical intensity demand values, and temporal distribution information is derived from optical demand forecast trends. Task scheduling algorithms are applied to this data. These algorithms can be First-Come, First-Served, Shortest Job First, or priority-based scheduling algorithms to calculate the start time and duration of each light task. The calculation process considers the expected execution time of the task, equipment preparation time, and possible time constraints to generate a preliminary task schedule. The preliminary task schedule is a task execution plan arranged in chronological order. Based on the preliminary task schedule, resource conflicts and priorities between tasks are assessed. The resource conflict detection module scans the preliminary task schedule to identify tasks that request the same smart light device resources within the same time interval. The priority assessment module evaluates the importance of each task according to preset rules, which may consider the task's light intensity requirement, impact range, and user-defined urgency level. A sorting model is used to optimize the task order. The sorting model can be a weighted priority sorting or a machine learning-driven dynamic sorting algorithm. The optimization objective is to minimize the total waiting time, avoid resource conflicts, or balance the load, generating a task priority list.

[0031] The task execution interval and resource utilization rate are calculated by combining the task priority list and device resource constraints. Device resource constraints include the total number of smart light devices, the maximum light output capacity of each device, the minimum dimming level, physical location constraints, and energy budget. The task execution interval is calculated to ensure sufficient buffer time between consecutive tasks to prevent device overload. The resource utilization rate is calculated to assess the utilization level of smart light devices at different time periods. The generated light task sequence is a finalized and optimized task execution plan, including the start time of all tasks, a clear priority order, and a detailed task resource requirement table. The light resource allocation scheme includes device resource allocation ratios, time conflict resolution results, and resource utilization parameters. The device resource allocation ratio defines how the light output resources of a single smart light device are allocated among multiple competing tasks, such as allocating a percentage of the device's brightness to a specific task. The time conflict resolution results identify time overlap issues that still exist or have been resolved after conflict detection and resolution in the light task sequence. The resource utilization parameter is a metric reflecting the expected utilization efficiency of the smart light devices within the planned period. The process of obtaining a light resource allocation scheme involves deriving task resource requirements and time nodes from the light task sequence. Task resource requirements are directly obtained from the task resource requirement table, while time nodes are derived from the task start time. The available resources and constraints of the smart lighting equipment are analyzed. Available resources include the equipment's idle time slots, available light intensity adjustment range, color channels, and special function modes. Constraints include the minimum shutdown time, maximum continuous operating time, thermal limitations, and grid load limitations that the equipment must adhere to. The analysis process generates a resource requirement-constraint comparison table, which compares the requirements of each task with the actual limitations of the equipment, highlighting potential mismatches or overloads. A resource allocation algorithm is used to process the resource requirement-constraint comparison table. This algorithm can be linear programming, a greedy algorithm, or a constraint satisfaction algorithm, aiming to find a feasible resource allocation scheme. The algorithm resolves conflicts and deficiencies in equipment resource allocation, such as when two tasks simultaneously request the same equipment, or when task demands exceed the equipment's maximum capacity, generating conflict resolution results. The conflict resolution results detail the type of conflict, the tasks and equipment involved, and possible resolution directions.

[0032] Based on conflict resolution results, device resources and time slots are reallocated. This reallocation process may involve adjusting task start times, assigning tasks to alternative devices, or reducing the light intensity requirements of tasks. Optimizing resource utilization efficiency is the core objective of reallocation, achieved through load balancing, reducing idle time, and avoiding bottlenecks. The generated light resource allocation scheme is a comprehensive plan that clearly defines when and how each smart light device serves which light task, including device resource allocation ratios, time conflict resolution results, and calculated resource utilization parameters. The light cycle execution instruction includes a light cycle timetable, a set of device control commands, and execution status identifiers. The light cycle timetable is a detailed timeline that specifies what type of illumination adjustment is performed at what time. The set of device control commands is a specific set of instructions that can be understood by the smart light device hardware, such as setting brightness values, changing color temperature, and turning on or off. Execution status identifiers are used to track the current status of each instruction internally, such as pending execution, in progress, completed, or failed.

[0033] The steps for obtaining photoperiod execution instructions are as follows: A detailed photoperiod schedule is developed based on the light resource allocation scheme. This schedule transforms the abstract tasks in the scheme into specific time sequences and light intensity adjustment points. Light intensity adjustment points define the specific times and target values ​​at which the illumination level needs to be changed within the photoperiod schedule. Device operation commands are generated based on the light intensity adjustment points and the specified smart light device, such as sending dimming commands to a specific device via a wireless communication protocol. The generated draft photoperiod schedule is a preliminary execution plan. The compatibility of the draft schedule with the device status is verified. This verification process checks whether the draft schedule matches the latest status of the smart light device, ensuring, for example, that the device scheduled to be turned on is not currently faulty. Time conflicts and resource allocations are adjusted, resolving any issues discovered during verification, such as reassigning tasks due to sudden device offline, generating a verified photoperiod schedule. The verified photoperiod schedule is then converted into a set of control commands executable by the device. This conversion involves translating high-level schedule instructions into low-level protocol commands that the smart light device driver layer can recognize, generating the final photoperiod execution instructions. Light adjustment feedback data includes ambient light change records, device operating status data, and photoperiod deviation values. Ambient light change records are sequences of actual illumination data measured by the ambient light sensor during photocycle execution. Device operation status data records detailed operating parameters of the smart lighting device during command execution. Photocycle deviation is a measure of the difference calculated by comparing the actual ambient light changes with the expected results of the photocycle execution command.

[0034] The acquisition of light regulation feedback data involves real-time monitoring of ambient light changes by an ambient light sensor during photocycle execution. The sensor operates continuously at a high sampling rate, recording light intensity data and adding a precise timestamp to each data point, generating an ambient light change record. Simultaneously, the system monitors the operating status of the smart lighting devices, collecting operating parameters from each device, such as real-time power consumption, internal temperature, LED drive current, network latency, and any error codes. It also records operation logs of the devices executing control commands, generating device operating status data. The ambient light change record is compared with the photocycle execution command. The comparison module compares the actual light intensity value read by the sensor with the expected light intensity value in the photocycle execution command at the same time point, calculating the difference. This difference may be expressed as an absolute error or a relative percentage error, generating a photocycle deviation value. The photocycle deviation value quantifies the deviation between the actual effect of photocycle regulation and the planned target. Finally, the data fusion unit integrates the ambient light change record, device operating status data, and photocycle deviation value, packaging this information into a structured data object to generate complete light regulation feedback data.

[0035] Example 2: See Figure 3 The acquisition of the optical task sequence begins with a deep analysis of the optical demand analysis results. The optical demand intensity and temporal distribution information contained in the analysis results are extracted. Optical demand intensity reflects the quantified value of the target illumination level, while the temporal distribution outlines the dynamic changes in optical demand over time. A task scheduling algorithm then intervenes to process this data. This algorithm can be a priority-based round-robin algorithm or a preemptive scheduling algorithm that considers deadlines. The algorithm receives the optical demand intensity and temporal distribution as input, calculates the start time and duration of each optical task, where the start time is a precise timestamp and the duration defines the length of time the task needs to occupy. This process generates a preliminary task schedule, which forms the initial framework of the optical task sequence. The evaluation based on the preliminary task schedule focuses on resource conflicts and priority determination among tasks. The resource conflict detection module scans all planned tasks in the preliminary task schedule, identifying task pairs or groups that compete for the same smart lighting device resources in the same or overlapping time periods. The priority evaluation module scores the importance of each task according to a preset set of rules, which may include factors such as the light intensity requirement of the task, the importance of the physical area affected by the task, and user-defined urgency flags. A sorting model is introduced to optimize the task order. The sorting model can be a weighted cost function model whose objective function is to minimize the total task completion time or maximize resource utilization. The model reorders the task list and outputs an optimized task priority list, which clarifies the execution order of tasks when resources are insufficient.

[0036] Combining the task priority list with device resource constraints is a crucial step in generating a feasible optical task sequence. Device resource constraints are a set of hard constraints, including the total number of smart lamp devices in the system, the maximum and minimum light output capacity of each smart lamp device, the illumination coverage limitations caused by the physical installation location of the devices, the dimming accuracy of the devices, the color reproduction index range, and the total energy consumption budget of the system. Calculating the task execution interval is to insert necessary buffer time between consecutively executed tasks, preventing smart lamp devices from aging prematurely or malfunctioning due to frequent switching or rapid brightness adjustments. Calculating resource utilization is to estimate the load of each smart lamp device over a future period and assess whether its workload is within the safe range allowed by the design. The final generated optical task sequence is a rigorously structured data object that fully includes the start time of all optical tasks, the optimized sorted task priority order, and a detailed task resource requirement table. The task resource requirement table clearly lists the specific smart lamp device identifiers, required brightness levels, color parameters, and expected energy consumption values ​​required to execute each task.

[0037] The steps for obtaining the optical resource allocation scheme derive task resource requirements and time nodes from the optical task sequence. Task resource requirements are directly derived from the task resource requirement table in the optical task sequence, which details the hardware resource expectations of each task. Time nodes are derived from the task initiation time of the optical task sequence, defining the temporal boundaries of task execution. Analyzing the available resources and constraints of intelligent lighting devices is a systematic project. Available resources refer to the idle capacity of intelligent lighting devices in the current and foreseeable future period, such as the time window when the device is not occupied, the unused portion within its light output intensity range, available color light channels, dynamic effects libraries, etc. Constraints are the physical or logical limitations that intelligent lighting devices must comply with, including the minimum shutdown time, maximum continuous operating time, maximum allowable casing temperature limited by the heat dissipation system, peak power limit of the power module, and the constraint of network communication bandwidth on the control command transmission rate. This analysis process generates a resource requirement and constraint comparison table. This table, in tabular or matrix form, juxtaposes the resource requests proposed by each task with the actual capabilities and limitations of the corresponding intelligent lighting device, clearly revealing the degree of matching between demand and supply and potential gaps.

[0038] Using resource allocation algorithms to process resource demand and constraint lookup tables is a core step in resolving resource contention. These algorithms can be heuristic search algorithms, such as genetic algorithms or simulated annealing, used to find near-optimal solutions under complex constraints. The algorithm traverses the resource demand and constraint lookup table, analyzing conflicts and deficiencies in device resource allocation. Conflicts typically manifest as multiple tasks simultaneously requesting the same device's resources at the same time, while deficiencies occur when a single task's resource demand exceeds the aggregate capacity of one or more devices. After execution, the algorithm generates conflict resolution results, a diagnostic report that precisely identifies the location of the conflict, the conflicting tasks, the type of conflict, and its severity. The reallocation operation based on the conflict resolution results is a dynamic adjustment process. Reallocating device resources may involve migrating tasks from overloaded devices to less loaded backup devices, or splitting a large task into smaller tasks and assigning them to different devices for parallel execution. Reallocating time slots involves adjusting the task's start time or duration to avoid peak resource usage. Optimizing resource utilization efficiency is the guiding principle for reallocation. The goal is to achieve the most balanced overall working time, minimize idle time, and optimize total energy consumption of the smart lighting device group while meeting the basic requirements of all tasks. The final generated light resource allocation scheme is a detailed resource configuration blueprint that clearly defines the specific working parameters allocated to each smart lighting device at each point in time throughout the entire cycle of the light task sequence execution. The scheme includes the calculated device resource allocation ratio, the resolved or labeled time conflict resolution results, and the estimated resource utilization parameters. The device resource allocation ratio describes the division of shared device resources among different tasks as a percentage or weight. The time conflict resolution results record which strategies were used to resolve which conflicts, and the resource utilization parameters are a quantitative assessment of the resource utilization efficiency under this allocation scheme.

[0039] Example 3: The acquisition of optical cycle execution commands begins with a detailed interpretation and conversion of the light resource allocation scheme. This scheme includes key information such as device resource allocation ratios, time conflict resolution results, and resource utilization parameters. Developing a detailed optical cycle schedule based on the light resource allocation scheme is a systematic planning process. The optical cycle schedule needs to map the abstract resource configuration in the scheme to specific time points and device operations. The light intensity adjustment point is a core element of the optical cycle schedule, defining the target light intensity value to be achieved at a specific time. The device operation commands describe the specific sequence of actions that the smart lighting device needs to perform to achieve this light intensity, such as adjusting the PWM duty cycle, changing the RGB channel mixing ratio, or activating a preset scene mode. This planning process outputs a draft optical cycle schedule, which is the initial version of the optical cycle execution commands. Verifying the compatibility between the draft photocycle schedule and the device status is a necessary validation step. This validation is achieved by querying the real-time status registers of the smart light devices. These registers contain the device's online / offline status, current operating mode, fault flags, and temperature alarm information. The compatibility check ensures that the devices scheduled for use in the draft photocycle schedule are in a healthy, responsive state and that their current configuration does not fundamentally conflict with the planned operations. Adjusting time conflicts and resource allocation are corrective actions performed when problems are discovered during validation. Time conflicts may arise from overlap between the task schedule in the draft and the device's required maintenance window. Resource allocation adjustments may involve reassigning tasks to redundant devices with similar functions or modifying task execution parameters to accommodate currently available resources. After adjustments, a validated photocycle schedule is generated, which has higher feasibility for execution. The final stage of instruction generation is converting the verified photoperiod schedule into a set of control commands that the device can execute. This conversion is done by the device driver layer or a dedicated command translator. The command translator understands the private communication protocols and instruction sets of different brands and models of smart light devices. It translates high-level instructions such as "set brightness to 50%" into specific network messages or serial commands for specific devices. The generated photoperiod execution instruction is a complete package containing the photoperiod schedule, the device control command set, and the execution status identifier. The photoperiod schedule is the timeline for execution, the device control command set is the underlying code that drives the device, and the execution status identifier is used to track the current lifecycle status of each instruction within the system.

[0040] The acquisition of light regulation feedback data is carried out synchronously with the light cycle execution process. During the light cycle execution process, the ambient light sensor, as the "eye" of the system, is continuously activated. The ambient light sensor monitors the changes in ambient light in real time at a fixed sampling frequency. Its photoelectric conversion unit converts the received light signal into an electrical signal, which is then converted into digital light intensity data by an analog-to-digital converter. Each data point is stamped with a high-precision timestamp, which records the exact moment of the light measurement. This continuous data stream constitutes the ambient light change record, which is the first-hand data reflecting the actual lighting conditions. Meanwhile, monitoring the operating status of smart lighting devices is a way to obtain feedback from the device side. The monitoring system collects detailed parameters of the smart lighting devices during operation through the built-in sensors and status reporting mechanisms. This device operating status data includes the real-time input power of the smart lighting devices, the estimated junction temperature of the LED chips, the fluctuation of the drive current, the fan speed, and the signal strength and latency of the network connection. The operation log records every control command received by the smart lighting devices, the execution time of the commands, and the response code returned by the devices after the commands are executed. This information together constitutes the device operating status data, which describes the health status and behavioral trajectory of the devices when executing instructions.

[0041] Comparing the recorded changes in ambient light with the light cycle execution command is the core step in evaluating the adjustment effect. This comparison is performed by a dedicated data processing module, which processes the actual light intensity value recorded by the ambient light sensor (denoted as...). The target light intensity value expected at the same time in the optical cycle execution instruction (denoted as ) and the target light intensity value expected at the same time. Perform point-by-point comparisons. Optical period deviation value (denoted as...) The effectiveness of execution is quantified by calculating the difference between the two. A typical calculation method uses the absolute error form, and its formula is expressed as:

[0042]

[0043] in: Indicates at a specific moment The optical period deviation value has the same dimensions as the light intensity unit (e.g., lux). Indicates the ambient light sensor at time The actual light intensity value measured. Indicates the time in the optical cycle execution instruction. The set target light intensity value. The calculated value. The sequence clearly demonstrates the degree of deviation and temporal distribution of the actual lighting effect from the planned target.

[0044] The final light regulation feedback data is generated by integrating ambient light change records, equipment operating status data, and photoperiod deviation values. This integration process is not simply data packaging; it involves data time alignment, validity verification, and feature extraction. For example, correlating a high photoperiod deviation value at a specific moment with a high-temperature alarm in the equipment operating status data at the same time helps analyze the cause of the deviation. The generated light regulation feedback data is a multi-dimensional dataset that comprehensively records the environmental response, equipment operating status, and the achievement of control objectives during photoperiod execution. This data is fed back to the machine learning processing module in real time, providing a data foundation for machine learning algorithms to optimize future light demand prediction models and task scheduling strategies. This forms a closed-loop control system from execution to perception to learning, driving continuous adaptive improvement of system performance. The accuracy and completeness of the light regulation feedback data directly affect the effectiveness of model iterative optimization by the machine learning processing module.

[0045] Example 4: The integrated emergency call processing module signifies deep integration of this module with the intelligent lighting control system at both the hardware and software levels. This module possesses independent signal detection circuitry and logic processing units for monitoring dedicated user input channels. The emergency call processing module executes interrupt responses based on user-triggered emergency requests. These requests originate from physical emergency buttons, emergency swipes in mobile device applications, or specific emergency commands from voice assistants. These request signals are transmitted to the emergency call processing module via wired or wireless communication links. Interrupting the current photoperiod adjustment process demonstrates the emergency call processing module's attainment of the highest system control priority. When an interruption occurs, the machine learning processing module suspends the current operations of the photoperiod analysis process, photoperiod scheduling process, photoperiod allocation process, photoperiod execution process, and real-time monitoring process, prioritizing system resources for emergency event processing. Prioritizing emergency calls and generating emergency response instructions is the core function of the emergency call processing module. This priority mechanism is implemented through a hardware interrupt controller and the task scheduler of the real-time operating system, ensuring that the response delay for emergency requests is controlled at the millisecond level. Generating emergency response commands involves a series of predefined action sequences. These commands might include controlling smart lighting devices to enter a high-frequency flashing mode to attract attention, sending an alert SMS message containing the user's location information to a pre-set emergency contact's mobile phone number, activating a local audible and visual alarm, or even linking with a community security system. Recording emergency event data is crucial for post-event analysis and system optimization. This data is written to a dedicated log area in non-volatile memory. The format of the recorded data is shown in Table 1.

[0046] Table 1: Emergency Event Data Recording Format

[0047]

[0048] The machine learning processing module trains an adaptive model using historical photoperiod data and emergency response data. Historical photoperiod data is a time-series dataset accumulated from the photoperiod analysis, execution, and real-time monitoring processes, containing long-term ambient light variations, equipment operation records, and light adjustment feedback data. Emergency response data comes from structured logs recorded by the emergency call processing module, providing key samples of user emergency behaviors. The adaptive model training process employs supervised learning or reinforcement learning frameworks. The model structure can be a deep neural network or a gradient boosting decision tree. The training objective is to establish a mapping relationship from environmental conditions and user habits to optimal photoperiod parameters and emergency response strategies. Dynamically optimizing parameter settings in the photoperiod analysis and light task scheduling processes is a concrete manifestation of the model's application. Based on learned patterns, the adaptive model dynamically adjusts the smoothing coefficient and trend weight threshold of the light demand prediction algorithm in the photoperiod analysis process, as well as the calculation rules for task priorities and heuristic rules for resource conflict resolution in the light task scheduling process.

[0049] This dynamic optimization allows the system to transcend fixed preset logic. For example, if the system analyzes historical data and discovers that a user has never triggered an emergency call and whose light cycle adjustment is stable during a specific time period, it may appropriately reduce the priority weight of emergency response during that period, allocating more resources to optimizing the precision of light cycle adjustment. Conversely, for another user who is frequently active at night and has sporadic emergency call records, the system will increase the priority of emergency preparedness tasks in the corresponding light task scheduling process and ensure that the emergency call processing module's response resources remain highly ready. The training and optimization of the machine learning processing module is a continuous background task. It uses new light adjustment feedback data and emergency event data to continuously fine-tune the adaptive model, enabling the machine learning-based intelligent light cycle adjustment lighting control system to learn from operational experience, gradually improving the personalization level of light environment adjustment and the intelligence level of emergency event handling. The parameter setting optimization of the light cycle analysis process directly affects the accuracy and foresight of the light demand analysis results, while the parameter setting optimization of the light task scheduling process is related to the efficiency of system resource scheduling and its fit with the actual needs of users. The close collaboration between the emergency call processing module and the machine learning processing module forms an adaptive intelligent system that ensures both safety and emergency response while prioritizing daily comfort. The system's behavior patterns gradually evolve over time and with the accumulation of data, better serving the personalized needs of users.

[0050] Example 5: The system integration of a user behavior analysis unit signifies continuous observation and interpretation of user activities at both the hardware sensor network and software analysis levels. The user behavior analysis unit acquires raw information by collecting user activity patterns and data. This collection relies on various sensors deployed in the environment. For example, passive infrared sensors detect human movement and stillness, millimeter-wave radar sensors accurately track user movements while protecting privacy, pressure pad sensors embedded in floors or seats sense user sitting and standing behaviors, and ambient light sensors and the switching and dimming records of smart lighting devices indirectly reflect the user's presence and activity intensity. This sensor data undergoes timestamp alignment and fusion processing to abstract user activity patterns. These patterns include quantitative characteristics such as the user's habitual wake-up time, the distribution of time spent in a specific room, nighttime activity frequency, and reaction speed to changes in light intensity. Combining light regulation feedback data and emergency response commands is a key step in the user behavior analysis unit's in-depth analysis. The light regulation feedback data comes from the real-time monitoring process, recording the actual effects and environmental responses of historical light cycles. The emergency response commands come from the emergency call processing module, marking user-initiated emergency events. The user behavior analysis unit correlates and analyzes user activity patterns, light regulation feedback data, and emergency response commands. For example, the system observes that user A's living room light intensity usually stabilizes below 50 lux and activity decreases after 10 PM. However, on a certain day, light regulation feedback data shows multiple sharp fluctuations in living room light intensity at 11 PM, and the user's activity pattern shows abnormally frequent movement. Although no emergency response command is recorded, the user behavior analysis unit will still identify this pattern as a potential anxiety signal. Adjusting the light cycle regulation strategy and emergency call response priority is a direct application of the analysis results. Adjusting the light cycle regulation strategy may be reflected in modifying the target light intensity parameter in the light cycle analysis process. For example, for elderly users with regular sleep patterns, one hour before their usual bedtime, the user behavior analysis unit may suggest that the light cycle analysis process gradually reduce the color temperature to a warm tone and slowly decrease the light intensity demand value to the night light level, rather than following a general linear decrease model.

[0051] Adjusting the priority of emergency call responses is based on correlation analysis of user behavior patterns and historical emergency events. For example, if User B has a history of Parkinson's disease, their activity pattern data shows higher limb instability during the morning wake-up period, and their historical emergency response command records show several low-priority calls triggered during this time due to brief dizziness. The user behavior analysis unit will suggest that the emergency call processing module automatically raise the priority of User B's emergency call responses triggered in the morning by one level. The system will prioritize the use of brighter warning lights, louder alarm sounds, and simultaneously notify multiple emergency contacts. The application of the user behavior analysis unit transforms the system from passively responding to commands to proactively adapting to users' personalized needs and status changes.

[0052] A specific example occurred in the living room of a senior apartment building equipped with a machine learning-based intelligent light cycle adjustment system. The user, Mr. Zhang, is an elderly man living alone with mild cognitive impairment. The user behavior analysis unit continuously collected Mr. Zhang's activity patterns and data through millimeter-wave radar sensors and the status of the intelligent lighting devices in the living room. Over three weeks of observation, the unit learned Mr. Zhang's activity patterns: he typically left his bedroom and entered the living room between 7:00 and 7:30 AM; in the morning, he liked to read in an armchair by the window, habitually turning on a floor lamp and setting the light intensity to 300 lux; in the afternoon, he watched TV on the sofa, requiring only 100 lux of ambient light. One Thursday afternoon, the light adjustment feedback data showed that the living room light intensity was repeatedly and manually adjusted from 100 lux to maximum brightness and then quickly dimmed within a short period, a pattern deviating from the learned user activity pattern. Simultaneously, the user behavior analysis unit analyzed the real-time user activity patterns and data, discovering that the millimeter-wave radar sensor showed Mr. Zhang moving violently and irregularly in the sofa area. Based on the abnormal light regulation feedback data and unusual user activity patterns, the user behavior analysis unit determined that there was a potential risk, even though the emergency call processing module had not yet received a clear emergency request. The user behavior analysis unit immediately sent a signal to the machine learning processing module to adjust the light cycle adjustment strategy. Based on this signal, the machine learning processing module dynamically optimized the light cycle execution process, generating special light cycle execution instructions to control the living room main light to slowly switch to a calm blue dimming mode and automatically brighten the corridor lights, providing a clear path for possible movement. The user behavior analysis unit also suggested that the emergency call processing module adjust the emergency call response priority, marking the potential risk level of Mr. Zhang's current session as "increased." The emergency call processing module accordingly entered a high-alert state, preloaded emergency contact information, and shortened the waiting time threshold for automatically dialing emergency numbers. This time, Mr. Zhang did not trigger the emergency button; he gradually stabilized in the calm lighting environment, and the system recorded this complete event sequence. Afterwards, the user behavior analysis unit associated and stored this "atypical event of not triggering emergency assistance" with user activity patterns and light regulation feedback data for future training of a more accurate adaptive model. This example demonstrates how the user behavior analysis unit, through subtle behavioral deviations and system feedback, can proactively intervene before a user explicitly requests assistance. It adjusts the light cycle regulation strategy to create a supportive environment and prepares emergency resources, showcasing the system's deep-seated personalization and proactiveness. The user behavior analysis unit enables the machine learning-based intelligent light cycle regulation system to not only respond to explicit commands but also interpret implicit needs, seamlessly integrating environmental regulation with safety monitoring.

[0053] 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. A machine learning-based intelligent light cycle adjustment intelligent lamp control system, characterized in that, The system includes an intelligent lighting device, an ambient light sensor, an emergency call module, and a machine learning processing module. When there is no need for light cycle adjustment, the intelligent lighting device is used as an independent night light and a mobile phone charger. When there is a need for light cycle adjustment, the machine learning processing module executes a light cycle analysis process, a light task scheduling process, a light resource allocation process, a light cycle execution process, and a real-time monitoring process to achieve intelligent light cycle adjustment. When an emergency call is needed, the emergency call module processes the user's emergency request. The optical cycle analysis process generates optical demand analysis results based on ambient light sensor data, the optical task scheduling process generates optical task sequence based on optical demand analysis results, the optical resource allocation process generates optical resource allocation scheme based on optical task sequence, the optical cycle execution process generates optical cycle execution instructions based on optical resource allocation scheme, and the real-time monitoring process generates optical adjustment feedback data based on optical cycle execution instructions. The optical demand analysis results include optical intensity demand values, a list of available equipment statuses, and the matching degree between optical demand and equipment. The specific steps for obtaining the light demand analysis results are as follows: Ambient light intensity data is collected using an ambient light sensor, combined with smart lamp device status data, to calculate the difference between the current light demand and device capabilities, generating a light demand difference report; the light demand difference report is analyzed, and machine learning algorithms are used to predict the trend of light demand changes, matching device resources with light demand to generate a device-light demand matching table; based on the device-light demand matching table, the light demand allocation weights are adjusted to generate the light demand analysis results; the device resources are the current working status, available light output capability, color adjustment range, beam angle, and physical location of each smart lamp device; the light demand is the target illumination level to be achieved within a specific time and space range; the device-light demand matching table is a matrix or list structure that lists the matching score between each smart lamp device and different light demands; the light demand allocation weight is an adjustable parameter used to determine how to allocate light demand tasks when multiple devices are available, and the weight adjustment is based on device efficiency, energy consumption priority strategy, or user preference; the light demand analysis results are a structured data object containing the calculated light intensity demand value, the updated list of available device statuses, and the quantified light demand-device matching degree. The optical resource allocation scheme includes equipment resource allocation ratio, time conflict resolution results, and resource utilization parameters. The specific steps for obtaining the optical resource allocation scheme are as follows: Extracting task resource requirements and time nodes from the optical task sequence; analyzing the available resources and constraints of the smart lighting equipment; generating a resource requirement and constraint comparison table; using a resource allocation algorithm to process the resource requirement and constraint comparison table, analyzing conflicts and deficiencies in equipment resource allocation, and generating conflict analysis results; based on the conflict analysis results, reallocating equipment resources and time slots, optimizing resource utilization efficiency, and generating an optical resource allocation scheme; the available resources include the equipment's idle time slots, available light intensity adjustment range, color channels, and special function modes; the constraints include the minimum shutdown time, maximum continuous operating time, thermal limits, and grid load limits that the equipment must comply with; the resource requirement and constraint comparison table is a table that compares the requirements of each task with the actual limitations of the equipment to highlight potential mismatches or overloads; the reallocation of equipment resources and time slots involves adjusting the task start time, allocating the task to alternative equipment, or reducing the light intensity requirement of the task. The built-in difference calculation unit of the machine learning processing module compares the current ambient light intensity data with the preset target light intensity range, which is determined based on user settings, schedules, or machine learning model predictions. The difference calculation unit outputs a light demand difference report, which details the gap between the current illuminance level and the target level, including the magnitude of the difference, duration, and spatial distribution information.

2. The intelligent light period adjustment system based on machine learning according to claim 1, characterized in that, The optical task sequence includes task start time, task priority order, and task resource requirement table; the optical cycle execution instruction includes optical cycle time table, device control command set, and execution status identifier; and the optical adjustment feedback data includes ambient light change records, device operating status data, and optical cycle deviation value.

3. The intelligent light cycle adjustment intelligent lamp control system based on machine learning according to claim 2, characterized in that, The specific steps for obtaining the optical task sequence are as follows: extract the optical demand intensity and time distribution from the optical demand analysis results, calculate the start time and duration of each optical task using a task scheduling algorithm, and generate a preliminary task schedule. Based on the preliminary task schedule, resource conflicts and priorities among tasks are assessed, a sorting model is used to optimize the task order, and a task priority list is generated. Combining the task priority list and equipment resource constraints, the task execution interval and resource utilization rate are calculated to generate an optical task sequence.

4. The intelligent light cycle adjustment system based on machine learning according to claim 3, characterized in that, The specific steps for obtaining the optical cycle execution command are as follows: Based on the optical resource allocation scheme, formulate a detailed optical cycle timetable, including optical intensity adjustment points and equipment operation commands, and generate a draft optical cycle timetable; verify the compatibility of the draft optical cycle timetable with the equipment status, adjust time conflicts and resource allocation, and generate a verified optical cycle timetable; convert the verified optical cycle timetable into a set of control commands that the equipment can execute, and generate the optical cycle execution command.

5. The intelligent light cycle adjustment system based on machine learning according to claim 4, characterized in that, The specific steps for acquiring the light adjustment feedback data are as follows: during the execution of the light cycle, the ambient light sensor monitors the changes in ambient light in real time, records the light intensity data and timestamp, and generates an ambient light change record; at the same time, the operating status of the smart light device is monitored, device parameters and operation logs are collected, and device operating status data is generated. By comparing the ambient light change records with the light cycle execution instructions, the light cycle execution deviation is calculated and the light cycle deviation value is generated. By combining ambient light change records, equipment operating status data, and photoperiod deviation values, light adjustment feedback data is generated.

6. The intelligent light cycle adjustment system based on machine learning according to claim 5, characterized in that, The system also includes an emergency call processing module, which interrupts the current optical cycle adjustment process based on the emergency request triggered by the user, prioritizes the emergency call, generates an emergency response command, and records emergency event data.

7. The intelligent light cycle adjustment system based on machine learning according to claim 6, characterized in that, The machine learning processing module uses historical photoperiod data and emergency event data to train an adaptive model and dynamically optimize parameter settings in the photoperiod analysis process and the phototask scheduling process.

8. The intelligent light cycle adjustment system based on machine learning according to claim 7, characterized in that, The system also includes a system-integrated user behavior analysis unit, which collects user activity patterns and data, and combines them with light regulation feedback data and emergency response instructions to adjust the light cycle regulation strategy and emergency call response priority.

Images