Air purification disinfecting and killing system and disinfecting and killing method

Abstract

The invention belongs to the technical field of air purification, and discloses an air purification disinfection system which comprises an air quality data acquisition system, a big data storage and processing system, a federal learning system and a robot management system. The air quality acquisition system sends acquired air quality data to the big data storage and processing system, the big data storage and processing system sends the air quality data to the federated learning system, and the federated learning system analyzes and processes the data and synchronizes an analysis result to the robot management system; and the robot management system controls the disinfection robot to perform disinfection operation. The disinfection instruction is issued through the cloud brain, the whole process of receiving the instruction and completing the disinfection task by the disinfection robot does not need human intervention, the infection risk of disinfection personnel is reduced to the maximum extent, and the safety of the disinfection personnel is ensured; the air quality in the current or future area is predicted through federal learning, one or more robots are coordinated to cooperatively work for disinfection according to the pollution degree, and the purification efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of air purification, and in particular to an air purification and sterilization system and a sterilization and sterilization method. Background Art

[0002] Large indoor spaces, such as hospitals, schools, building lobbies, cinemas, and cultural and sports venues, have large personnel flows and are relatively closed, making them high-risk areas for virus transmission. In such scenarios, keeping the air relatively clean and sterile is the most direct and effective way to reduce the incidence of virus infection.

[0003] In the transmission of airborne viruses, the three main types of pollutants are microbial pollutants (bacteria, fungi and viruses), particulate pollutants (PM2.5), and volatile organic compounds (TVOC). At present, the means of disinfection for the above pollutants are mainly divided into manual disinfection and mechanical disinfection. Manual disinfection requires disinfection personnel to wear protective clothing and carry disinfection equipment to spray disinfectants in designated areas. It is time-consuming, labor-intensive, and costly. In addition, there is a risk of infection for disinfection personnel, and the safety of disinfection personnel cannot be fully guaranteed. Mechanical disinfection requires the disinfection device to be replenished with disinfectant, which is difficult to replenish each time. Disinfection devices are generally large in size and inconvenient to move. Disinfection operations cannot be carried out flexibly and timely in high-risk areas, and are inconvenient to use.

[0004] To solve the above problems, the publication number is CN 115671357 A, and the name of the invention is an invention of an automatic path planning system for ultraviolet disinfection combined with air purification. It discloses an automatic path planning system for ultraviolet disinfection combined with air purification, including an environmental disinfection control system, an air disinfection and purification system, UVC-LED, a human-computer interaction system, a task control system, an air quality detection system, a mobile control system, a human protection warning system, and a robot anti-fall system; including sensors such as laser radar, camera, ultrasonic wave, inertial measurement unit, etc., with an industrial control computer as a processing unit, an industrial tablet as a human-computer interaction interface, and UVC-LED, photocatalyst, potassium permanganate particles, and nanofiber HEPA filter to disinfect and purify the environment. The ultrasonic sensor is located at the bottom of the robot, which can enable the robot to sense whether there are steps or pits on the road surface, so as to construct an anti-fall system for the disinfection robot. The air quality detection system will also generate an air quality map as one of the important parameters for robot path planning. Although the invention discloses an automatic path planning system for ultraviolet disinfection combined with air purification, it does not specifically disclose the specific implementation method of path planning. It can be seen that how the disinfection device can realize the planning of the disinfection path is one of the important issues that need to be solved in the current field of air disinfection. Summary of the invention

[0005] In view of the above-mentioned problems existing in the prior art, the purpose of the present invention is to provide an air purification and disinfection system, which, through a federated learning mechanism, predicts the pollution degree and future pollution degree according to the acquired air quality indicators; and coordinates the disinfection robot to carry out disinfection at a designated location based on the prediction.

[0006] Another object of the present invention is to provide an air purification and disinfection method.

[0007] In order to solve the above problems, the present invention adopts the following technical solutions: an air purification and disinfection system, including an air quality data acquisition system, a big data storage and processing system, a federated learning system, and a robot management system;

[0008] The air quality collection system sends the acquired air quality data to the big data storage and processing system, and the big data storage and processing system sends the air quality data to the federated learning system. The federated learning system analyzes and processes the data and synchronizes the analysis results to the robot management system.

[0009] The robot management system controls the disinfection robot to perform disinfection operations.

[0010] In some embodiments, the air quality data acquisition system is used to collect TVOC, CO2, PM25, and PM10 data, and transmit the data to a big data storage and processing system via the MQTT communication protocol.

[0011] In some embodiments, the big data storage and processing system is used to store the air quality data collected by the air quality collection system and synchronize it to the federated learning system.

[0012] In some embodiments, the federated learning system dynamically updates model parameters by learning air quality data; analyzes the acquired real-time data, and synchronizes the analysis results to the robot management system in real time.

[0013] In some embodiments, the robot management system includes a map management module, an area management module, a task management module, a report management module, an air detection module, a pollution source positioning module, a robot setting module, and a message center module; wherein,

[0014] The map management module is used to manage the map created for any disinfection robot, set a unique identifier for each disinfection robot, record the disinfection robot map information, record the disinfection robot position coordinates, and record the robot charging station position;

[0015] The area management module is used to manage any disinfection robot patrol, air pollutant detection, disinfection and purification area;

[0016] The task management module is used to manage the purification and disinfection tasks of any disinfection robot;

[0017] The report management module is used to manage the historical disinfection tasks, current disinfection tasks, and planned disinfection tasks of any disinfection robot;

[0018] The air detection module is used to manage the air pollutant indicators detected by any robot in the patrol area;

[0019] Pollution source positioning is used to locate the designated location of the corresponding disinfection robot patrol area based on the acquired air pollutant index data and type;

[0020] The robot setting module is used to set the patrol area, patrol route, and charging station location of the disinfection robot;

[0021] The message center module is used to view the online, offline, charging, disinfection and purification status of each disinfection robot in real time.

[0022] In some embodiments, in the federated learning system, the ID of a disinfection robot is in one-to-one correspondence with the stored air quality index data, and the system sets a learning model for the air quality index data collected by each ID.

[0023] Each set of air quality index data stored in the big data corresponds to a federated learning model. K sets of data have K federated learning models for model training and model parameter update respectively. Each federated learning model parameter is transmitted using homomorphic encryption technology.

[0024] In K (K≥2) federated learning models, each model parameter is input to the Arbiter in FATE based on homomorphic encryption technology. The Arbiter performs weighted average on the received encrypted model parameters, and then transmits the encrypted global model parameters back to each federated learning model to update the parameters of each model, thus realizing the training and modeling of N models.

[0025] After each participant trains the model locally, a set of neural network weights will be generated. These parameters are homomorphically encrypted and sent to the central server Arbiter for weighted averaging. The number of samples of a participant is ni, and the number of global samples is N. Then the weight of the participant is ni / N.

[0026]

[0027] Among them, Wglobal is the global model parameter, Wi is the model parameter of the i-th participant, and K is the number of participants.

[0028] Another object of the present application is to provide an air purification and disinfection method, comprising the steps of:

[0029] S1. Map creation and area selection: Based on the shape and area geographic information of the target purification and disinfection area, a free space path planning algorithm is used to establish an environmental model;

[0030] S2. Collection of air pollutant indicators: During patrol, the disinfection robot collects TVOC, CO2, PM25, PM10, temperature, and humidity indicators in real time through sensors, and uploads the indicator data to the big data storage and processing system through the MQTT protocol. The disinfection robot ID forms a one-to-one correspondence with the stored indicator data;

[0031] S3, the federated learning algorithm uses historical indicator data and multi-party secure computing-based federated learning to update model parameters and make predictions;

[0032] S4: Instructions are issued to assign specific disinfection tasks to each disinfection robot;

[0033] S5. Air purification and disinfection.

[0034] In some embodiments, the instruction is specifically issued as follows:

[0035] Including fixed-point purification and disinfection, based on real-time air quality index data, and based on multi-party secure computing federated learning to locate areas where air pollutant content is higher than the set value, the cloud brain platform sends instructions to coordinate at least one disinfection robot to the designated area to perform disinfection tasks;

[0036] It includes purification and disinfection. Based on real-time air quality index data and federated learning based on multi-party secure computing, it locates the time period when the air pollutant content is higher than the set value. The cloud brain platform creates disinfection and purification tasks and issues instructions to coordinate at least one disinfection robot to perform disinfection tasks in heavily polluted areas during the specified time period.

[0037] In some embodiments, the disinfection adopts bipolar ion purification and disinfection. The disinfection robot receives the command, runs to the designated area, turns on the bipolar ion generator and emits high concentrations of positive and negative ions into the surrounding air. Bacteria, viruses, PM2.5, PM10 and TVOC gases in the air are charged with opposite polarities after colliding with positive and negative ions. Ultrafine particles are adsorbed on large particles, condense into large particles and fall into the dust discharge trough.

[0038] In some embodiments, when two or more disinfection robots go to the same designated area to perform disinfection tasks, each disinfection robot uses laser radar, ultrasonic sensor, and depth camera to achieve mutual recognition between the disinfection robots to prevent collisions.

[0039] Compared with the prior art, the beneficial technical effects of the present invention are:

[0040] 1. The air purification and disinfection system described in the present invention, the disinfection robot is linked to the cloud brain in communication, and the cloud brain includes a big data storage and processing system, a federated learning system, and a robot management system. The big data storage and processing system receives and synchronizes the air quality data to the federated learning system. The federated learning system analyzes the acquired air quality data, makes disinfection judgments, and synchronizes the data analysis results to the robot management system. The robot management system issues disinfection instructions, and the disinfection robot receives the instructions and completes the disinfection task. The disinfection system described in the present application does not require human intervention throughout the process, which minimizes the risk of infection for disinfection personnel and ensures the safety of disinfection personnel.

[0041] 2. Based on the acquired air quality data, the present invention uses federated learning to analyze the current or predict the degree of air pollution in a certain area in the future, coordinates robots to work together according to the degree of pollution, and effectively improves the efficiency of disinfection;

[0042] 3. The present invention uses a free space path planning algorithm to enable the disinfection robot to complete tasks more accurately, efficiently and energy-savingly, making disinfection operations fast and efficient;

[0043] 4. The present invention adopts a bipolar ion air disinfection and purification method, which does not require the replenishment of disinfectant and greatly reduces the maintenance cost of the disinfection equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 A schematic diagram of the structure of the air disinfection and purification system provided by the present invention;

[0045] Figure 2 A map management logic diagram provided by the present invention;

[0046] Figure 3 A schematic diagram of the robot disinfection path management logic provided by the present invention;

[0047] Figure 4 A flow chart of the robot disinfection method provided by the present invention;

[0048] Figure 5 Schematic diagram of the robot federated learning training model provided by the present invention. DETAILED DESCRIPTION

[0049] The technical solution of the present invention will be further described in detail below in conjunction with the embodiments and drawings.

[0050] Example

[0051] See also Figures 1 to 3 As shown, an air purification and disinfection system provided by the present invention includes an air quality data acquisition system, a big data storage and processing system, a federated learning system, and a robot management system;

[0052] The air quality collection system sends the acquired air quality data to the big data storage and processing system, and the big data storage and processing system sends the air quality data to the federated learning system. The federated learning system analyzes and processes the data and synchronizes the analysis results to the robot management system.

[0053] The robot management system controls the disinfection robot to perform disinfection operations.

[0054] When using the above system,

[0055] First, an air quality data acquisition system is used to collect air quality data, such as TVOC, CO2, PM25, PM10, temperature, humidity and other data in the air, and the data is transmitted to a big data storage and processing system through the MQTT communication protocol.

[0056] The big data storage and processing system is used to store the air quality data collected by the air quality collection system, and synchronize the real-time air quality data and historical air quality data to the federated learning system.

[0057] The federated learning system dynamically updates model parameters by learning from historical air quality data, analyzes the acquired real-time data, and synchronizes the analysis results to the robot management system in real time. In some embodiments, federated learning also includes third-party intelligent learning, such as object recognition, speech recognition, etc.

[0058] In some embodiments, the robot management system includes a map management module, an area management module, a task management module, a report management module, an air detection module, a pollution source location module, a robot setting module, and a message center module;

[0059] Among them, the map management module is used to manage the map created for any disinfection robot, that is, to set a unique identification ID for each disinfection robot, record the map information of the disinfection robot, record the location coordinates of the disinfection robot, and record the location of the robot charging pile; for example, according to the shape, area and other geographical information of the purification and disinfection area, a free space path planning algorithm is used to establish an environmental model; with path coverage and robot running time as optimization goals, the disinfection robot can complete the task more accurately, efficiently and energy-savingly.

[0060] The area management module is used to manage any disinfection robot patrol, air pollutant detection, and disinfection and purification area; for example, the disinfection robot uploads the created map to the big data storage and processing system of the cloud brain platform through the HTTP protocol, and sets the area and time period that the disinfection robot wants to patrol in the robot management system.

[0061] The task management module is used to manage the purification and disinfection tasks of any disinfection robot; in some embodiments, the disinfection task includes the disinfection time period, disinfection frequency, the number of robots used in the disinfection area, etc.

[0062] The report management module is used to manage the historical disinfection tasks, current disinfection tasks, and planned disinfection tasks of any disinfection robot;

[0063] The air detection module is used to manage the air pollutant indicators detected in the patrol area reported by any robot; in some embodiments, the disinfection robot collects air quality data through its own sensors. For example, during the patrol process, the disinfection robot collects TVOC, CO2, PM25, PM10, temperature, humidity and other indicators in real time through its own sensors, and uploads the air quality indicators to the big data storage and processing system of the cloud brain platform through the MQTT protocol. It should be noted that there is a one-to-one correspondence between the ID of the disinfection robot and the stored indicator data in order to determine the area.

[0064] Pollution source positioning is used to locate the designated location of the corresponding disinfection robot patrol area based on the acquired air pollutant index data and type;

[0065] The robot setting module is used to set the patrol area, patrol route, and charging station location of the disinfection robot;

[0066] The message center module is used to view the online, offline, charging, and disinfection and purification status of each disinfection robot in real time. For example, in some embodiments, the robot is offline due to hardware or software failures and cannot perform disinfection tasks. The message center module can obtain the current status of each robot in a timely manner through real-time monitoring. When any robot cannot perform disinfection tasks, it can be dispatched in time through the robot management system to improve disinfection efficiency.

[0067] In some embodiments, in the federated learning system, there is a one-to-one correspondence between the ID of a disinfection robot and the stored air quality index data. The system sets a learning model for the air quality index data collected by each ID, and the model can be automatically updated through historical air quality data.

[0068] Each set of air quality index data stored in the big data corresponds to a federated learning model. K sets of data have K federated learning models for model training and model parameter update respectively. Each federated learning model parameter is transmitted using homomorphic encryption technology.

[0069] In K (K≥2) federated learning models, each model parameter is input to the Arbiter in FATE based on homomorphic encryption technology. The Arbiter performs weighted average on the received encrypted model parameters, and then transmits the encrypted global model parameters back to each federated learning model to update the parameters of each model, thus realizing the training and modeling of N models.

[0070] After each participant trains the model locally, a set of neural network weights will be generated. These parameters are homomorphically encrypted and sent to the central server Arbiter for weighted averaging. The number of samples of a participant is ni, and the number of global samples is N. Then the weight of the participant is ni / N.

[0071]

[0072] Among them, Wglobal is the global model parameter, Wi is the model parameter of the i-th participant, and K is the number of participants.

[0073] Due to the particularity of air flow, the federated learning model needs to be updated according to the actual situation: with the help of historical air quality index data, the federated learning based on multi-party secure computing can update the model parameters;

[0074] In some embodiments, air quality data has certain regularity. For example, in school, students have fixed physical activity time. The federated learning model can use real-time air quality index data and federated learning based on multi-party secure computing to predict the air quality of a certain area at a certain point in the future. Disinfection robots can be arranged in advance to perform operations, which can improve disinfection efficiency while also preventing the spread of the virus.

[0075] Example 1

[0076] See also Figure 4 , Figure 5 As shown, an air purification and disinfection method comprises the steps of:

[0077] S1. Map creation and area selection: Based on the shape and area geographic information of the target purification and disinfection area, a free-space path planning algorithm is used to establish an environmental model and upload it to the big data storage and processing system of the cloud brain;

[0078] S2. Collection of air pollutant indicators: During patrol, the disinfection robot collects TVOC, CO2, PM25, PM10, temperature, and humidity indicators in real time through sensors, and uploads the indicator data to the big data storage and processing system through the MQTT protocol. The disinfection robot ID forms a one-to-one correspondence with the stored indicator data;

[0079] S3, the federated learning algorithm uses historical indicator data and multi-party secure computing-based federated learning to update model parameters;

[0080] S4. Instructions are issued to issue designated disinfection tasks to each disinfection robot, and the configuration information of the disinfection robot is sent to the disinfection robot through the MQTT protocol to perform air purification and disinfection at a designated time and in a designated area;

[0081] S5. Air purification and disinfection.

[0082] The above is the basic disinfection form of the air purification and disinfection system. The system creates a map and uploads it to the big data storage and processing system of the cloud brain platform. The federated learning system upgrades the model based on the data, configures the disinfection time and disinfection area, and sends the configuration information of the disinfection robot to the disinfection robot through the MQTT protocol for air purification and disinfection at the specified time and in the specified area.

[0083] Example 2

[0084] An air purification and disinfection method comprises the following steps:

[0085] S1. Map creation and area selection: Based on the shape and area geographic information of the target purification and disinfection area, a free-space path planning algorithm is used to establish an environmental model and upload it to the big data storage and processing system of the cloud brain;

[0086] S2. Collection of air pollutant indicators: During patrol, the disinfection robot collects TVOC, CO2, PM25, PM10, temperature, and humidity indicators in real time through sensors, and uploads the indicator data to the big data storage and processing system through the MQTT protocol. The disinfection robot ID forms a one-to-one correspondence with the stored indicator data;

[0087] S3, the federated learning algorithm uses historical indicator data and multi-party secure computing-based federated learning to update model parameters;

[0088] S4. Instructions are issued to assign specific disinfection tasks to each disinfection robot. For example, if a span of one month is configured on the cloud brain platform, the configuration information of the disinfection robot is sent to the disinfection robot through the MQTT protocol. The disinfection robot will perform air purification and disinfection in the designated area at the designated time every day within the set month.

[0089] S5. Air purification and disinfection: Here, a disinfection robot is required to perform air purification and disinfection in designated areas at designated times every day.

[0090] In some embodiments, based on real-time big data federated learning of air pollution indicators, the robot management system responds by predicting whether a certain area will have severe air pollution now or at a certain point in the future, and immediately coordinates the disinfection robot to go to the area to perform purification and disinfection, or creates a purification and disinfection task and immediately coordinates the disinfection robot to go to the area to perform purification and disinfection after the time point arrives.

[0091] In some embodiments, based on real-time big data federated learning of air pollution indicators, the robot management system responds by predicting whether a certain area will have severe air pollution now or at a certain point in the future, and immediately coordinates multiple disinfection robots to go to the area to perform purification and disinfection, or creates a purification and disinfection task and immediately coordinates multiple disinfection robots to go to the area to perform purification and disinfection after the task arrives at the time point.

[0092] Example 3

[0093] The disinfection described in this application is specifically disinfection using bipolar ion purification, including the steps of:

[0094] 100. The disinfection robot receives instructions from the robot management system and moves to the designated area;

[0095] 200. Turn on the bipolar ion generator and emit high concentrations of positive and negative ions into the surrounding air. Bacteria, viruses, PM2.5, PM10 and TVOC gases in the air will be charged with opposite polarities after colliding with positive and negative ions. Ultrafine particles will be adsorbed on large particles, condense into large particles and fall into the dust exhaust trough.

[0096] Finally, it is necessary to point out here that the above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto, and any changes or substitutions that can be easily thought of by any technician familiar with the present invention within the technical scope disclosed in the present invention should be covered within the protection scope of the present invention.

Claims

1. An air purification and disinfection system, characterized in that: Including air quality data collection system, big data storage and processing system, federated learning system, and robot management system; The air quality collection system sends the acquired air quality data to the big data storage and processing system, and the big data storage and processing system sends the air quality data to the federated learning system. The federated learning system analyzes and processes the data and synchronizes the analysis results to the robot management system. The robot management system controls the disinfection robot to perform disinfection operations.

2. An air purification and disinfection system according to claim 1, characterized in that: The air quality data acquisition system is used to collect TVOC, CO2, PM25, and PM10 data, and transmit the data to a big data storage and processing system through the MQTT communication protocol.

3. An air purification and disinfection system according to claim 1, characterized in that: The big data storage and processing system is used to store the air quality data collected by the air quality collection system and synchronize it to the federated learning system.

4. The air purification and disinfection system according to claim 1, characterized in that: The federated learning system dynamically updates model parameters by learning air quality data; analyzes the acquired real-time data, and synchronizes the analysis results to the robot management system in real time.

5. An air purification and disinfection system according to claim 4, characterized in that: The robot management system includes a map management module, an area management module, a task management module, a report management module, an air detection module, a pollution source positioning module, a robot setting module, and a message center module; wherein, The map management module is used to manage the map created for any disinfection robot, set a unique identifier for each disinfection robot, record the disinfection robot map information, record the disinfection robot position coordinates, and record the robot charging station position; The area management module is used to manage any disinfection robot patrol, air pollutant detection, disinfection and purification area; The task management module is used to manage the purification and disinfection tasks of any disinfection robot; The report management module is used to manage the historical disinfection tasks, current disinfection tasks, and planned disinfection tasks of any disinfection robot; The air detection module is used to manage the air pollutant indicators detected by any robot in the patrol area; Pollution source positioning is used to locate the designated location of the corresponding disinfection robot patrol area based on the acquired air pollutant index data and type; The robot setting module is used to set the patrol area, patrol route, and charging station location of the disinfection robot; The message center module is used to view the online, offline, charging, disinfection and purification status of each disinfection robot in real time.

6. An air purification and disinfection system according to claim 1, characterized in that: In the federated learning system, the ID of a disinfection robot is in one-to-one correspondence with the stored air quality index data. The system sets a learning model for the air quality index data collected by each ID. Each set of air quality index data stored in the big data corresponds to a federated learning model. K sets of data have K federated learning models for model training and model parameter update respectively. Each federated learning model parameter is transmitted using homomorphic encryption technology. In K (K≥2) federated learning models, each model parameter is input to the Arbiter in FATE based on homomorphic encryption technology. The Arbiter performs weighted average on the received encrypted model parameters, and then transmits the encrypted global model parameters back to each federated learning model to update the parameters of each model, thus realizing the training and modeling of N models. After each participant trains the model locally, a set of neural network weights will be generated. These parameters are homomorphically encrypted and sent to the central server Arbiter for weighted averaging. The number of samples of a participant is ni, and the number of global samples is N. Then the weight of the participant is ni / N. Among them, Wglobal is the global model parameter, Wi is the model parameter of the i-th participant, and K is the number of participants.

7. An air purification and disinfection method according to the air purification and disinfection system of claim 1, characterized in that: Included steps S1. Map creation and area selection: Based on the shape and area geographic information of the target purification and disinfection area, a free space path planning algorithm is used to establish an environmental model; S2. Collection of air pollutant indicators: During patrol, the disinfection robot collects TVOC, CO2, PM25, PM10, temperature, and humidity indicators in real time through sensors, and uploads the indicator data to the big data storage and processing system through the MQTT protocol. The disinfection robot ID forms a one-to-one correspondence with the stored indicator data; S3, the federated learning algorithm uses historical indicator data and multi-party secure computing-based federated learning to update model parameters and make predictions; S4: Instructions are issued to assign specific disinfection tasks to each disinfection robot; S5. Air purification and disinfection.

8. An air purification and disinfection method according to claim 7, characterized in that: The instruction is issued Including fixed-point purification and disinfection, based on real-time air quality index data, and based on multi-party secure computing federated learning to locate areas where air pollutant content is higher than the set value, the cloud brain platform sends instructions to coordinate at least one disinfection robot to the designated area to perform disinfection tasks; It includes purification and disinfection. Based on real-time air quality index data and federated learning based on multi-party secure computing, it locates the time period when the air pollutant content is higher than the set value. The cloud brain platform creates disinfection and purification tasks and issues instructions to coordinate at least one disinfection robot to perform disinfection tasks in heavily polluted areas during the specified time period.

9. The air purification and disinfection method according to claim 7, characterized in that: The disinfection adopts bipolar ion purification and disinfection. After receiving the command, the disinfection robot runs to the designated area, turns on the bipolar ion generator and emits high concentrations of positive and negative ions into the surrounding air. Bacteria, viruses, PM2.5, PM10 and TVOC gases in the air are charged with opposite polarities after colliding with the positive and negative ions. Ultrafine particles are adsorbed on large particles, condense into large particles and fall into the dust discharge trough.

10. An air purification and disinfection method according to claim 7, characterized in that: When two or more disinfection robots go to the same designated area to perform disinfection tasks, each disinfection robot uses lidar, ultrasonic sensors, and depth cameras to achieve mutual recognition between the disinfection robots to prevent collisions.

Images