A dynamic directional irradiation control system and method for far-ultraviolet lamps based on source location and exposure constraints.
The dynamic directional irradiation control system for far-ultraviolet lamps, which integrates source location identification, environmental perception, and safety monitoring, solves the problem of insufficient light field distribution in far-ultraviolet lamp disinfection systems, improves disinfection efficiency and energy utilization, and ensures personnel safety.
Patent Information
- Application Number
- CN202611332512.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-31
- Publication Date
- 2026-11-10
AI Technical Summary
Existing far-ultraviolet lamp disinfection systems cannot dynamically adjust the light field distribution according to the location of the pollution source, resulting in low disinfection efficiency, difficulty in controlling personnel exposure safety risks, and low energy utilization.
A dynamic directional illumination control system for far-ultraviolet lamps based on source location and exposure constraints is adopted. It includes source location identification, environmental perception, safety monitoring, far-ultraviolet lamp array control, and adjustable optical component subsystems. The control is optimized through machine learning models to achieve dynamic adjustment of the power, illumination direction, and area of the far-ultraviolet lamps.
It significantly increased the inactivation dose in key areas, reduced ineffective irradiation in personnel areas, optimized energy consumption, ensured that the disinfection process was carried out within a safe threshold range, and adapted to the dynamic changes in the location of pollution sources.
Smart Images

Figure CN123067147A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of intelligent transportation environment technology, and in particular to a dynamic directional illumination control system and method for far-ultraviolet lamps based on source location and exposure constraints. Background Technology
[0002] With the increasing global awareness of public health and safety, the application of far-ultraviolet (FUL) disinfection technology in enclosed spaces is receiving growing attention. FULV light waves can effectively destroy the DNA or RNA structure of microorganisms such as bacteria and viruses, achieving rapid inactivation. Furthermore, its penetration depth into human skin and eyes is limited, allowing for safe use in occupied environments under strict dosage control. Based on this characteristic, FULV lamps have broad application prospects in enclosed cabin environments such as aircraft cabins, high-speed train carriages, and subway trains. They can be used for active disinfection of public transportation, reducing the risk of cross-infection and improving passenger safety.
[0003] However, existing far-ultraviolet (FUL) lamp control schemes generally employ fixed-power operation or only time-cycle start-stop modes, failing to dynamically adjust disinfection strategies based on actual pollution conditions. In practical applications, the distribution of bioaerosols exhibits significant spatial heterogeneity, with pollution source locations constantly shifting due to factors such as personnel movement and ventilation system operation. Fixed-power FUL arrays often struggle to accurately cover high-risk areas, resulting in insufficient irradiation doses in some areas and energy waste in others, hindering optimal disinfection effectiveness. Furthermore, the lack of dynamic control over light direction allows FUL energy to easily project onto areas where personnel are present, posing safety hazards and reducing the utilization rate of effective inactivation doses. Moreover, existing control methods do not adequately consider complex factors such as passenger density, seating arrangements, and the location of obstructions within the cabin, failing to optimize disinfection efficiency while ensuring safe exposure thresholds for personnel, leading to excessive energy consumption and unreliable disinfection results.
[0004] Therefore, how to achieve adaptive dynamic control of the power and direction of light of far-ultraviolet lamps, and improve the efficiency of disinfection while ensuring personnel safety, has become a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention
[0005] The purpose of this invention is to propose a dynamic directional irradiation control system and method for far-ultraviolet lamps based on source location and exposure constraints, in order to solve the problems of existing far-ultraviolet lamp disinfection systems, such as the inability to dynamically adjust the light field distribution according to the location of the pollution source, low disinfection efficiency, difficulty in effectively controlling personnel exposure safety risks, and low energy utilization.
[0006] To achieve the above objectives, this invention proposes a dynamic directional irradiation control system for far-ultraviolet lamps based on source location and exposure constraints. The system includes a source location identification subsystem, an environmental sensing subsystem, a safety monitoring subsystem, a far-ultraviolet lamp array control subsystem, an adjustable optical component subsystem, and an optimization control center. The source location identification subsystem acquires the location information of bioaerosol sources within the cabin. The environmental sensing subsystem acquires real-time environmental information from the cabin. The safety monitoring subsystem monitors the far-ultraviolet light dose and performs risk assessment. The far-ultraviolet lamp array control subsystem controls the output power, duty cycle, and activation sequence of the far-ultraviolet lamp array. The adjustable optical component subsystem adjusts the irradiation direction and area of the far-ultraviolet light. The optimization control center is connected to the source location identification subsystem, environmental sensing subsystem, safety monitoring subsystem, far-ultraviolet lamp array control subsystem, and adjustable optical component subsystem, and outputs optimal control parameters based on the source location information and safety constraints.
[0007] Furthermore, the source location identification subsystem includes a sound acquisition unit, a camera acquisition unit, an acoustic-visual multimodal recognition model, and a pollution source location identification module; the sound acquisition unit is used to acquire sound signals inside the cabin; the camera acquisition unit is used to acquire video images inside the cabin; the acoustic-visual multimodal recognition model is used to fuse sound features and visual features to identify behaviors that generate bioaerosols, such as coughing and sneezing, and determine the location of the pollution source; the pollution source location identification module is used to output the pollution source location identification result; the source location identification subsystem outputs source location information to the optimization control center through a data bus.
[0008] Furthermore, the acoustic-visual multimodal recognition model contains two parallel branches: an acoustic feature extraction network and a visual feature extraction network. The acoustic feature extraction network uses a convolutional neural network to perform spectral analysis on the acquired acoustic signals, while the visual feature extraction network uses a spatiotemporal convolutional network to process the video image sequence. The feature data extracted by the two branches are weighted and fused in the feature fusion layer and then input to the pollution source location identification module for location determination.
[0009] Furthermore, the environmental perception subsystem includes a personnel distribution perception module, a fixed object layout acquisition module, and a material reflectivity database. The personnel distribution perception module is used to acquire real-time distribution information of passengers in the cabin. The fixed object layout acquisition module is used to acquire layout information of fixed objects such as seats and partitions in the cabin. The material reflectivity database is used to store far-ultraviolet light reflectivity parameters of various materials in the cabin. The environmental perception subsystem outputs environmental parameters to the optimization control center via a data bus.
[0010] Furthermore, the material reflection characteristic database pre-stores reflectivity and scattering characteristic parameters of different material surfaces such as metal, plastic, fabric, and leather, providing a data basis for shading and reflection risk constraints.
[0011] Furthermore, the safety monitoring subsystem includes a light dose monitoring module, a risk assessment module, and an emergency protection module; the light dose monitoring module is used to monitor the cumulative far-ultraviolet radiation dose in various areas of the cockpit in real time; the risk assessment module is used to perform risk assessment based on real-time dose information and safety constraints; the emergency protection module is used to trigger emergency protection actions when an excessive risk is detected; the safety monitoring subsystem outputs real-time dose signals and early warning signals to the optimization control center via a data bus.
[0012] Furthermore, the far-ultraviolet lamp array control subsystem includes far-ultraviolet lamp units, a power drive circuit, a pulse modulation circuit, and a communication interface circuit; the far-ultraviolet lamp unit includes multiple far-ultraviolet lamp tubes or far-ultraviolet LED arrays; the power drive circuit is used to control the output power of the far-ultraviolet lamps to achieve continuous power adjustment; the pulse modulation circuit is used to achieve duty cycle control and timing control, enabling each lamp unit to work collaboratively according to a predetermined timing sequence; the communication interface circuit is used to communicate bidirectionally with the optimization control center, receive power setting instructions, duty cycle setting instructions, and timing control instructions, and provide feedback on the working status of each lamp unit.
[0013] Furthermore, the adjustable optical component subsystem includes a rotatable reflector, a Fresnel lens assembly, and a servo drive module; the rotatable reflector is used to change the reflection direction of the far-ultraviolet light, enabling the beam to point towards the target area; the Fresnel lens assembly is used to adjust the diffusion angle of the far-ultraviolet light, thereby adjusting the size of the irradiated area; the servo drive module is used to drive the rotatable reflector and the Fresnel lens assembly to adjust their angles; the adjustable optical component subsystem receives irradiation direction control commands and diffusion angle control commands from the optimization control center, and applies the control commands to the far-ultraviolet lamp array execution end.
[0014] Furthermore, the optimization control center includes a data input layer, an optimization decision layer, and a control output layer. The data input layer is used to collect source location information, environmental parameters, and real-time dose warning signals. The optimization decision layer has a built-in machine learning agent model or strategy library, which is used to perform optimization calculations based on the input information to construct an optimization control model with the goal of maximizing the inactivation dose in key areas and with constraints such as the permissible irradiation dose to the skin and eyes of passengers, restrictions on the occupied area of the cabin, and the risk of obstruction and reflection. The control output layer is used to issue control commands to the far-ultraviolet lamp array control subsystem and the adjustable optical component subsystem, including power setpoints, duty cycle setpoints, activation timing arrangements, and target angles of the adjustable optical components.
[0015] This invention further proposes a method for dynamic directional irradiation control of far-ultraviolet lamps based on source location and exposure constraints, comprising the following steps: The system acquires bioaerosol source location information and real-time cabin environmental parameters. The source location information is obtained by fusing and analyzing the collected sound signals and video images through an acoustic and visual multimodal recognition model. The environmental parameters include personnel distribution information, fixed object layout information, and material reflection characteristics parameters. Establish a set of controllable variables for the far-ultraviolet lamp array, including the output power, duty cycle, turn-on sequence of each lamp unit, and the illumination direction and irradiation area corresponding to the adjustable optical components; An optimization model for light field exposure at the source location was constructed, with the objective function being to maximize the inactivation dose in the key area, and the constraints being the permissible irradiation dose to the passenger's skin and eyes, the restrictions on the occupied area of the cabin, and the risks of occupancy and reflection. Call the machine learning agent model or preset strategy library to obtain the optimal combination of control parameters, including the optimal power setting value, the optimal duty cycle setting value, the optimal turn-on timing arrangement, and the optimal angle of the illumination direction; Output the far-ultraviolet lamp power control command, duty cycle control command, turn-on timing control command and directional irradiation angle control command to the corresponding subsystem; The far-ultraviolet lamp array is controlled to perform directional irradiation according to the optimal parameter combination, so that far-ultraviolet energy preferentially covers high-risk pollution areas and reduces ineffective irradiation to personnel areas.
[0016] Compared with existing technologies, the advantages of this invention are as follows: By using dynamic directional irradiation control based on the location of the pollution source, far-ultraviolet energy can preferentially cover high-risk pollution areas, significantly increasing the inactivation dose in key areas; by comprehensively considering multi-dimensional safety constraints such as permissible irradiation dose to passengers' skin and eyes, restrictions on occupied areas in the cabin, and risks of obstruction and reflection, and by using personnel safety as a hard constraint in the optimization model, the far-ultraviolet disinfection process is ensured to always be carried out within the safe threshold range; through dynamic power regulation and directional irradiation, excessive irradiation of unoccupied or low-risk areas is avoided, significantly reducing system energy consumption; by introducing machine learning agent models or strategy libraries, intelligent closed-loop control of the entire process from pollution source identification to optimization decision-making to execution feedback is realized, which can adapt to scenarios with dynamic changes in the location of pollution sources and automatically revert to a safe standby mode after the event ends. Attached Figure Description
[0017] Figure 1 This is a schematic diagram of the structure of the far-ultraviolet lamp dynamic directional irradiation control system based on source location and exposure constraints according to an embodiment of the present invention; Figure 2This is a flowchart of a method for dynamic directional irradiation control of far-ultraviolet lamps based on source location and exposure constraints in an embodiment of the present invention. Detailed Implementation
[0018] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be further described below.
[0019] like Figure 1 As shown in the figure, an embodiment of the present invention provides a dynamic directional illumination control system for far-ultraviolet lamps based on source location and exposure constraints. This system includes a source location identification subsystem, an environmental sensing subsystem, a safety monitoring subsystem, a far-ultraviolet lamp array control subsystem, an adjustable optical component subsystem, an optimization control center, and a far-ultraviolet lamp array actuator. The subsystems are interconnected via an industrial control bus. The optimization control center, as the core control unit of the entire system, is connected to the source location identification subsystem, environmental sensing subsystem, safety monitoring subsystem, far-ultraviolet lamp array control subsystem, and adjustable optical component subsystem, and is used to output optimal control parameters based on source location information and safety constraints.
[0020] The source location identification subsystem is strategically positioned within the cabin to acquire information on the location of bioaerosol sources within the cabin. This subsystem includes a sound acquisition unit, a camera acquisition unit, an acoustic-visual multimodal recognition model, and a pollution source location identification module. The sound acquisition unit utilizes a high-sensitivity microphone array positioned at the four corners of the cabin ceiling to collect sound signals within the cabin. The placement and angle of the microphone array are precisely calculated to ensure coverage of all passenger seating areas within the cabin. The analog sound signals acquired by the sound acquisition unit are converted from analog to digital and then transmitted to the acoustic-visual multimodal recognition model for processing. The camera acquisition unit employs a combination of multiple infrared and visible light cameras. Infrared cameras are used for personnel detection and behavior recognition under low-light conditions, while visible light cameras capture clear video images. The camera acquisition unit's shooting range covers the entire space within the cabin, and the video images acquired by the camera acquisition unit are also transmitted to the acoustic-visual multimodal recognition model.
[0021] The acoustic-visual multimodal recognition model is the core processing unit of the source location identification subsystem. This model is built on deep learning algorithms and employs a multimodal fusion network structure, simultaneously receiving feature data from both the sound acquisition unit and the camera acquisition unit. Internally, the acoustic-visual multimodal recognition model contains two parallel branches: a sound feature extraction network and a visual feature extraction network. The sound feature extraction network uses a convolutional neural network to perform spectral analysis on the acquired sound signals, extracting sound feature parameters related to behaviors such as coughing and sneezing, including peak sound energy, frequency distribution characteristics, and duration characteristics. The visual feature extraction network uses a spatiotemporal convolutional network to process video image sequences, identifying behaviors that generate bioaerosols, such as coughing and sneezing, by analyzing changes in human posture and facial expression features. The feature data extracted by the two branches are weighted and fused in the feature fusion layer. The resulting comprehensive feature vector is input to the pollution source location identification module for location determination. Based on the analysis results of the fused feature vector, combined with the latency estimation of the microphone array and the image positioning information acquired by the camera, the pollution source location identification module calculates the three-dimensional spatial coordinates of the pollution source and transmits this location information to the optimization control center via a data bus.
[0022] The environmental perception subsystem is deployed at multiple locations within the cabin to acquire real-time environmental information, providing environmental parameter support for optimization decisions. This subsystem includes a passenger distribution perception module, a fixed object layout acquisition module, and a material reflectivity database. The passenger distribution perception module employs a multi-sensor fusion scheme based on millimeter-wave radar and infrared thermal imaging technology. Millimeter-wave radar accurately detects the number and location distribution of passengers within the cabin, while infrared thermal imaging assists in determining passenger facial orientation and body movements. The passenger distribution perception module outputs the detected passenger distribution information in a gridded format. Each grid cell contains information on passenger density, occupancy status, and activity status for that area. This information is periodically updated and transmitted to the optimization control center. The fixed object layout acquisition module performs a comprehensive scan during system initialization to acquire precise 3D layout information of fixed objects such as seats, partitions, armrests, and luggage racks within the cabin. This layout information is stored as 3D point cloud data and used as static environmental parameters during subsequent operation. The material reflection characteristics database pre-stores the far-ultraviolet light reflection characteristics parameters of various common materials in the cockpit, including the reflectivity and scattering characteristics of different material surfaces such as metal, plastic, fabric, and leather. These parameters are obtained through experimental measurements and provide a data basis for the occlusion and reflection risk constraints in subsequent optimization calculations.
[0023] The safety monitoring subsystem is located in multiple key positions within the cabin to monitor far-ultraviolet radiation dose and conduct risk assessments, ensuring personnel safety during the disinfection process. The safety monitoring subsystem includes a radiation dose monitoring module, a risk assessment module, and an emergency protection module. The radiation dose monitoring module employs multiple distributed ultraviolet sensors deployed throughout the cabin, including above seats, in aisle areas, and restrooms, at each key location. The sensors collect real-time far-ultraviolet radiation intensity at their respective locations and calculate the cumulative radiation dose based on the collected radiation intensity and exposure time. The radiation dose monitoring module transmits the real-time dose information from each area to the risk assessment module and the optimization control center via a data bus. The risk assessment module performs a comprehensive risk assessment based on the received real-time dose information and preset safety constraints. These constraints include permissible skin radiation dose thresholds, permissible eye radiation dose thresholds, and restrictions on occupied areas within the cabin. The module outputs the assessment results as warning signals. When the assessment results indicate that the cumulative dose in a certain area is approaching the warning threshold, a warning signal is generated and transmitted to the optimization control center to request adjustments to control parameters. When the assessment results indicate that the cumulative dose in a certain area exceeds the permissible exposure threshold, an emergency stop signal is generated and transmitted to the emergency protection module and the optimization control center. Upon receiving the over-limit risk signal, the emergency protection module immediately triggers emergency protection actions, issuing an emergency shutdown command to the far-ultraviolet lamp array control subsystem via the industrial control bus to cut off the power supply to the far-ultraviolet lamp array, and simultaneously sending a status change notification to the optimization control center.
[0024] The far-ultraviolet (FUL) lamp array control subsystem is located in the equipment compartment on the top of the cockpit. It controls the output power, duty cycle, and activation sequence of the FUL lamp array, enabling precise adjustment of the FUL light intensity. The FUL lamp array control subsystem includes FUL lamp units, a power drive circuit, a pulse modulation circuit, and a communication interface circuit. Each FUL lamp unit consists of multiple FUL lamp tubes or FUL LED arrays, installed on the top of the cockpit in a preset array arrangement, covering the main spatial areas within the cockpit. The wavelength range of the FUL lamp units is 222 nanometers. Far-ultraviolet light at this wavelength can effectively destroy the DNA or RNA structure of microorganisms, providing excellent disinfection and sterilization effects. The power drive circuit adopts a digitally controlled adjustable voltage regulator scheme, which can continuously adjust the output voltage according to control commands, thereby achieving continuous adjustment of the FUL lamp output power. The power adjustment range is continuously adjustable from zero-power standby to full-power operation. The pulse modulation circuit is used to implement duty cycle control and timing control functions. It generates high-frequency PWM pulse signals and adjusts the duty cycle of these signals to achieve precise control of the average output power. Simultaneously, the pulse modulation circuit controls the on-time sequence of each lamp unit, enabling multiple lamp units to turn on or off sequentially according to a predetermined time order, thus achieving multi-lamp collaborative operation timing control. The communication interface circuit uses an industrial-grade CAN bus or Ethernet interface to achieve bidirectional communication with the optimization control center. It receives power setting commands, duty cycle setting commands, and timing control commands, and simultaneously uploads feedback data such as the operating status, actual output power, and temperature information of each lamp unit to the optimization control center.
[0025] The adjustable optical component subsystem is paired with the far-ultraviolet lamp unit to adjust the direction and area of far-ultraviolet light illumination, achieving precise beam orientation. The adjustable optical component subsystem includes a rotatable reflector, a Fresnel lens assembly, and a servo drive module. The rotatable reflector, installed at the front of each far-ultraviolet lamp unit, is made of a high-reflectivity ultraviolet light reflective material. Under the control of the servo drive module, the rotatable reflector can rotate around the horizontal and vertical axes, thereby changing the reflection direction of the far-ultraviolet light and ensuring the beam is precisely aimed at the target area. The Fresnel lens assembly, installed in front of the rotatable reflector, is used to adjust the diffusion angle of the far-ultraviolet light, allowing for adjustment of the illumination area size. The Fresnel lens assembly is made of quartz glass with high ultraviolet light transmittance, maintaining high luminous flux while changing the beam diffusion angle. The servo drive module uses a high-precision digital servo, capable of precisely driving the angle adjustment of the rotatable reflector and Fresnel lens assembly according to control commands. The angle adjustment resolution reaches 0.1 degrees, and the response time is less than 50 milliseconds, meeting the requirements of real-time control. The adjustable optical component subsystem receives irradiation direction control commands and diffusion angle control commands from the optimization control center, and converts the control commands into corresponding mechanical actions that act on the far-ultraviolet lamp unit.
[0026] The Optimization Control Center, the core decision-making unit of the entire system, is integrated into the main control chassis on the top of the cockpit and uses a high-performance embedded processor to run optimization decision-making algorithms. The Optimization Control Center comprises a data input layer, an optimization decision-making layer, and a control output layer. The data input layer has multiple data receiving interfaces to collect source location information output by the source location identification subsystem, environmental parameters output by the environmental perception subsystem, and real-time dose and warning signals output by the safety monitoring subsystem. The data input layer standardizes the received data before transmitting it to the optimization decision-making layer. The optimization decision-making layer incorporates a machine learning proxy model or strategy library to perform optimization calculations based on the input information. It constructs an optimization control model with the objective of maximizing the inactivation dose in key areas, constrained by the permissible radiation dose to passenger skin and eyes, cockpit passenger area limitations, and occlusion and reflection risks. It quickly solves for the optimal combination of control parameters by calling pre-trained machine learning proxy models or preset strategy libraries. The machine learning agent model is trained using a deep reinforcement learning algorithm. During training, a digital model of the cockpit is constructed in a simulation environment, setting different personnel distribution scenarios, pollution source location scenarios, and combinations of environmental parameters. The reward function is a weighted value of the cumulative inactivation dose and the cumulative safety over-limit penalty. The current state is used as the input state, and the control parameters are used as the output action. The PPO algorithm or SAC algorithm is used for training until the model converges. The trained agent model can quickly output control actions based on the current state during runtime without complex online optimization calculations, meeting the timeliness requirements of real-time control. The control output layer has multiple control command output interfaces, used to send power setpoints, duty cycle setpoints, and activation timing instructions to the far-ultraviolet lamp array control subsystem, and simultaneously send target angle instructions for the adjustable optical components to the adjustable optical component subsystem.
[0027] The far-ultraviolet (UUV) lamp array actuator is positioned within the target disinfection area of the cockpit. Multiple UUV lamp units are arranged in a preset array on the top or side walls of the cockpit, covering the main spatial areas within the cockpit. The UUV lamp array actuator receives power and drive signals from the UUV lamp array control subsystem and optical adjustment actions from the adjustable optical component subsystem to complete the directional irradiation of far-ultraviolet light. The UUV lamp array actuator 7 is connected to the industrial control bus, enabling data communication with the optimization control center and uploading operational status feedback information to the optimization control center.
[0028] Based on the above system structure, this embodiment of the invention also provides a method for dynamic directional irradiation control of far-ultraviolet lamps based on source location and exposure constraints. The complete process of this method is as follows: Figure 1The second half is shown below. In the specific implementation process, the method first enters the baseline working mode and executes the steps of acquiring bioaerosol source location information and real-time cabin environmental parameters. The source location information is obtained by fusing and analyzing the sound signals collected by the sound acquisition unit and the video images collected by the camera acquisition unit through an acoustic-visual multimodal recognition model. The environmental parameters include personnel distribution information acquired by the personnel distribution perception module, fixed object layout information acquired by the fixed object layout acquisition module, and material reflection characteristic parameters stored in the material reflection characteristic database. Next, the step of establishing a set of controllable variables for the far-ultraviolet lamp array is executed. This step determines the output power, duty cycle, activation sequence of each lamp unit, and the illumination direction and irradiation area corresponding to the adjustable optical components as optimization variables. Then, the step of constructing a source location light field exposure constraint optimization model is executed. This optimization model takes maximizing the inactivation dose in the key area as the objective function and takes the allowable irradiation dose to the passenger's skin and eyes, the limitation of the occupied area in the cabin, and the risk of occupancy and reflection as constraints. Next, the step of calling the machine learning agent model or the preset strategy library is executed to obtain the optimal combination of control parameters, including the optimal power setting value, the optimal duty cycle setting value, the optimal activation sequence arrangement, and the optimal angle of the irradiation direction. Then, the steps of outputting far-ultraviolet lamp power control commands, duty cycle control commands, activation timing control commands, and directional irradiation angle control commands to the corresponding subsystems are executed. Finally, the step of controlling the far-ultraviolet lamp array to perform directional irradiation according to the optimal parameter combination is executed, so that far-ultraviolet energy preferentially covers high-risk pollution areas and reduces ineffective irradiation of personnel areas is reduced.
[0029] During the directional irradiation control process, the system simultaneously executes safety monitoring and feedback control steps. This process forms a dynamic closed-loop control mechanism, enabling dynamic directional irradiation control and safety monitoring of the far-ultraviolet lamp. The safety monitoring and feedback control steps include the following decision nodes: The first decision node determines whether the real-time cumulative dose has reached the warning threshold. The light dose monitoring module continuously monitors the cumulative far-ultraviolet radiation dose in various areas of the cockpit and transmits the real-time dose information to the optimization control center. When the real-time cumulative dose is determined to have reached the warning threshold, the optimization control center performs operations to reduce power or correct the light direction, updates the optimization constraints, and then returns to call the machine learning proxy model to resolve the optimal control parameters. This feedback adjustment mechanism enables dynamic adjustment of the warning threshold.
[0030] The second judgment node is used to determine whether the real-time cumulative dose exceeds the allowable exposure threshold. When the real-time cumulative dose exceeds the allowable exposure threshold, the risk assessment module outputs an over-limit signal, the emergency protection module triggers a safety protection action, and sends an emergency shutdown command to the far-ultraviolet lamp array control subsystem 4 through the industrial control bus to cut off the power supply of the far-ultraviolet lamp array. The system switches to the safe standby mode, and the control process ends. This mechanism realizes emergency protection against over-limit risks.
[0031] The third decision node is used to determine whether the location of the pollution source has changed or whether the disinfection event has ended. The optimization control center continuously analyzes the latest source location information output by the source location identification subsystem. When it is determined that the location of the pollution source has changed, the system switches to the baseline working mode and returns to the acquisition step to reacquire the source location and environmental parameters. When it is determined that the disinfection event has ended, the system switches to the safe standby mode and ends the control process. This mechanism enables intelligent response to changes in status.
[0032] In a specific application scenario, suppose a passenger A coughs in the cabin. The sound acquisition unit of the source location identification subsystem 1 collects the cough sound signal, and the camera acquisition unit captures passenger A's facial expression and body posture changes. The acoustic-visual multimodal recognition model identifies the cough behavior and calculates the three-dimensional spatial coordinates of the pollution source through the fusion analysis of sound and visual features. The pollution source location identification module transmits the location information to the optimization control center.
[0033] The optimization control center simultaneously receives current passenger distribution information from the environmental perception subsystem. Assuming there are 10 passengers in the cabin, passenger A is in the 3rd row by the window, and the other passengers are distributed across rows 1 through 5. Based on the pollution source location information, the optimization control center 6 determines a high-risk disinfection area centered on passenger A with a 1-meter radius. Simultaneously, based on the passenger distribution information, it determines the locations of other passengers as their activity areas, requiring avoidance or reduction of exposure to these areas.
[0034] The optimization decision layer invokes a pre-trained machine learning proxy model for optimization calculations. The proxy model's input includes the coordinates of the pollution source, the relative positions of each lamp unit to the pollution source, the distribution of personnel in each area, and the current cumulative dose value for each area. Its output includes the power settings for each far-ultraviolet lamp unit, the target angle of the rotatable reflector, and the target diffusion angle of the Fresnel lens assembly. The optimal control parameters output by the proxy model prioritize pointing the far-ultraviolet beam towards the high-risk area where the pollution source is located. Simultaneously, the angle of the rotatable reflector is adjusted to avoid areas with personnel activity, and the diffusion angle of the Fresnel lens assembly is adjusted to concentrate the beam energy in the high-risk area.
[0035] The optimized control center sends power control commands to the power drive circuit of the far-ultraviolet lamp array control subsystem via the control output layer, and sends irradiation direction control commands to the servo drive module of the adjustable optical component subsystem. Upon receiving the control commands, the far-ultraviolet lamp array actuator 7 performs directional irradiation, precisely pointing the far-ultraviolet beam at the high-risk disinfection area surrounding passenger A for focused disinfection, while avoiding direct irradiation of other passenger areas.
[0036] During the disinfection process, the light dose monitoring module of the safety monitoring subsystem continuously monitors the cumulative dose in each area. When the cumulative dose in a certain area approaches the warning threshold, the risk assessment module generates a warning signal and transmits it to the optimization control center. The optimization control center automatically reduces the output power of the corresponding lamp unit in that area or adjusts the angle of the rotatable reflector to transfer some of the light beam to other high-risk areas. At the same time, it updates the optimization constraints and calls the surrogate model again to solve for new optimal control parameters. When the disinfection event ends, i.e., after passenger A stops coughing, and the source location identification subsystem does not detect any new coughing or sneezing behavior for a period of time, the optimization control center determines that the disinfection event has ended, shuts down the far-ultraviolet lamp array, and switches to a safe standby mode, waiting for the next disinfection task to be triggered.
[0037] The above are merely preferred embodiments of the present invention and do not constitute any limitation on the present invention. Any equivalent substitutions or modifications made by those skilled in the art to the technical solutions and content disclosed in the present invention without departing from the scope of the present invention shall be deemed to have remained within the protection scope of the present invention.
Claims
1. A dynamic directional irradiation control system for far-ultraviolet lamps based on source location and exposure constraints, characterized in that, The system includes a source location identification subsystem, an environmental perception subsystem, a safety monitoring subsystem, a far-ultraviolet lamp array control subsystem, an adjustable optical component subsystem, and an optimization control center. The source location identification subsystem acquires the location information of bioaerosol sources within the cabin. The environmental perception subsystem acquires real-time environmental information of the cabin. The safety monitoring subsystem monitors the far-ultraviolet light dose and performs risk assessment. The far-ultraviolet lamp array control subsystem controls the output power, duty cycle, and activation sequence of the far-ultraviolet lamp array. The adjustable optical component subsystem adjusts the irradiation direction and area of the far-ultraviolet light. The optimization control center is connected to the source location identification subsystem, environmental perception subsystem, safety monitoring subsystem, far-ultraviolet lamp array control subsystem, and adjustable optical component subsystem, and outputs optimal control parameters based on the source location information and safety constraints.
2. The far-ultraviolet lamp dynamic directional irradiation control system based on source location and exposure constraints according to claim 1, characterized in that, The source location identification subsystem includes a sound acquisition unit, a camera acquisition unit, an acoustic-visual multimodal recognition model, and a pollution source location identification module. The sound acquisition unit is used to acquire sound signals inside the cabin; the camera acquisition unit is used to acquire video images inside the cabin; the acoustic-visual multimodal recognition model is used to fuse sound features and visual features to identify behaviors that generate bioaerosols, such as coughing and sneezing, and determine the location of the pollution source; the pollution source location identification module is used to output the pollution source location identification results; the source location identification subsystem outputs source location information to the optimization control center via a data bus.
3. The far-ultraviolet lamp dynamic directional irradiation control system based on source location and exposure constraints according to claim 2, characterized in that, The acoustic-visual multimodal recognition model contains two parallel branches: an acoustic feature extraction network and a visual feature extraction network. The acoustic feature extraction network uses a convolutional neural network to perform spectral analysis on the acquired acoustic signals, while the visual feature extraction network uses a spatiotemporal convolutional network to process the video image sequence. The feature data extracted by the two branches are weighted and fused in the feature fusion layer and then input to the pollution source location identification module for location determination.
4. The far-ultraviolet lamp dynamic directional irradiation control system based on source location and exposure constraints according to claim 1, characterized in that, The environmental perception subsystem includes a personnel distribution perception module, a fixed object layout acquisition module, and a material reflectivity database. The personnel distribution perception module is used to acquire real-time distribution information of passengers in the cabin. The fixed object layout acquisition module is used to acquire layout information of fixed objects such as seats and partitions in the cabin. The material reflectivity database is used to store far-ultraviolet light reflectivity parameters of various materials in the cabin. The environmental perception subsystem outputs environmental parameters to the optimization control center through a data bus.
5. The far-ultraviolet lamp dynamic directional irradiation control system based on source location and exposure constraints according to claim 4, characterized in that, The material reflection characteristics database pre-stores reflectivity and scattering characteristics parameters of different materials such as metals, plastics, fabrics, and leather, providing a data foundation for shading and reflection risk constraints.
6. The far-ultraviolet lamp dynamic directional irradiation control system based on source location and exposure constraints according to claim 1, characterized in that, The safety monitoring subsystem includes a light dose monitoring module, a risk assessment module, and an emergency protection module; the light dose monitoring module is used to monitor the cumulative far-ultraviolet radiation dose in each area of the cockpit in real time; the risk assessment module is used to perform risk assessment based on real-time dose information and safety constraints. The emergency protection module is used to trigger emergency protection actions when an excessive risk is detected; the safety monitoring subsystem outputs real-time dose signals and early warning signals to the optimization control center through the data bus.
7. The far-ultraviolet lamp dynamic directional irradiation control system based on source location and exposure constraints according to claim 1, characterized in that, The far-ultraviolet lamp array control subsystem includes a far-ultraviolet lamp unit, a power drive circuit, a pulse modulation circuit, and a communication interface circuit; the far-ultraviolet lamp unit includes multiple far-ultraviolet lamp tubes or far-ultraviolet LED arrays; the power drive circuit is used to control the output power of the far-ultraviolet lamps to achieve continuous power adjustment. The pulse modulation circuit is used to realize duty cycle control and timing control, so that each lamp unit can work together in a predetermined timing sequence. The communication interface circuit is used for bidirectional communication with the optimization control center, receiving power setting instructions, duty cycle setting instructions and timing control instructions, and providing feedback on the working status of each lamp unit.
8. The far-ultraviolet lamp dynamic directional irradiation control system based on source location and exposure constraints according to claim 1, characterized in that, The adjustable optical component subsystem includes a rotatable reflector, a Fresnel lens assembly, and a servo drive module. The rotatable reflector is used to change the reflection direction of the far-ultraviolet light, enabling the beam to point towards the target area. The Fresnel lens assembly is used to adjust the diffusion angle of the far-ultraviolet light, thereby adjusting the size of the irradiated area. The servo drive module is used to drive the rotatable reflector and the Fresnel lens assembly to adjust their angles. The adjustable optical component subsystem receives irradiation direction control commands and diffusion angle control commands from the optimization control center and applies these control commands to the far-ultraviolet lamp array execution end.
9. The far-ultraviolet lamp dynamic directional irradiation control system based on source location and exposure constraints according to claim 1, characterized in that, The optimization control center includes a data input layer, an optimization decision layer, and a control output layer. The data input layer is used to collect source location information, environmental parameters, and real-time dose warning signals. The optimization decision layer has a built-in machine learning agent model or strategy library, which is used to perform optimization calculations based on the input information to construct an optimization control model with the goal of maximizing the inactivation dose in key areas and with constraints such as the permissible radiation dose to the skin and eyes of passengers, the restrictions on the occupied area of the cabin, and the risks of obstruction and reflection. The control output layer is used to issue control commands to the far-ultraviolet lamp array control subsystem and the adjustable optical component subsystem, including power setpoints, duty cycle setpoints, activation timing arrangements, and target angles of the adjustable optical components.
10. A method for dynamic directional irradiation control of far-ultraviolet lamps based on source location and exposure constraints, characterized in that, Includes the following steps: The system acquires bioaerosol source location information and real-time cabin environmental parameters. The source location information is obtained by fusing and analyzing the collected sound signals and video images through an acoustic and visual multimodal recognition model. The environmental parameters include personnel distribution information, fixed object layout information, and material reflection characteristics parameters. Establish a set of controllable variables for the far-ultraviolet lamp array, including the output power, duty cycle, turn-on sequence of each lamp unit, and the illumination direction and irradiation area corresponding to the adjustable optical components; An optimization model for light field exposure at the source location was constructed, with the objective function being to maximize the inactivation dose in the key area, and the constraints being the permissible irradiation dose to the passenger's skin and eyes, the restrictions on the occupied area of the cabin, and the risks of occupancy and reflection. Call the machine learning agent model or preset strategy library to obtain the optimal combination of control parameters, including the optimal power setting value, the optimal duty cycle setting value, the optimal turn-on timing arrangement, and the optimal angle of the illumination direction; Output the far-ultraviolet lamp power control command, duty cycle control command, turn-on timing control command and directional irradiation angle control command to the corresponding subsystem; The far-ultraviolet lamp array is controlled to perform directional irradiation according to the optimal parameter combination, so that far-ultraviolet energy preferentially covers high-risk pollution areas and reduces ineffective irradiation to personnel areas.
