Intelligent ship cabin system based on artificial intelligence

Abstract

The invention discloses a ship intelligent cabin system based on artificial intelligence, and the system comprises an intelligent environment adjustment module which collects cabin environment parameters in real time, predicts the temperature and humidity change trend, automatically controls an air conditioner and purification equipment, and adjusts the cabin environment; the equipment maintenance module is used for collecting operation parameters of ship equipment in real time, positioning fault causes and fault positions through a fault diagnosis algorithm and generating a maintenance scheme; the emergency intelligent response module is used for monitoring whether a fire or water leakage condition exists in the cabin in real time, automatically starting fire extinguishing or drainage equipment and planning an escape route; and the cabin space utilization module is used for analyzing personnel activity behaviors by using a camera in combination with an image recognition technology and an AI algorithm, and intelligently switching functions of the multifunctional cabin according to cabin use requirements and personnel activity conditions. Through the environment adjusting module, the equipment maintenance module, the emergency intelligent response module and the cabin space utilization module, the overall performance and the management level of the ship cabin are improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent control, and in particular to an artificial intelligence-based ship intelligent cabin system. Background Art

[0002] With the development of the shipping industry, the demand for intelligent, safe, comfortable, and efficient ship cabin management is growing. Traditional ship cabins have many problems. For example, cabin environment regulation is difficult to accurately adjust according to the needs of different personnel and real-time environmental changes. The management and maintenance of cabin equipment relies on manual inspections, which is inefficient and prone to missing faults. In the face of emergencies, there is a lack of fast and effective intelligent response mechanisms.

[0003] The Chinese patent publication number CN115242852A discloses "an intelligent cabin management system and an intelligent cabin management method". The passenger terminal is communicatively connected to the service processing platform and the intelligent cabin platform; the service processing platform receives the information request and identity information of the passenger terminal, and forwards the request to the intelligent cabin platform after completing the identity verification; the intelligent cabin platform collects the corresponding cabin information according to the instruction and feeds it back to the service processing platform, which then transfers it to the passenger terminal for display; the passenger generates an operation instruction based on the displayed information, which is forwarded to the intelligent cabin platform via the service processing platform to execute cabin environment adjustment or fee settlement, so as to accurately adjust the cabin environment where the passenger is located and improve the accuracy of multi-cabin management.

[0004] The above solution uses a service processing platform and passenger terminals to achieve manual adjustment of the cabin environment and fee settlement, but it still has the following limitations: 1) Environmental adjustment relies on user instructions and cannot actively adapt to real-time changes; 2) There is a lack of cabin equipment management; 3) In emergencies, only early warnings are provided, and there is no automated response mechanism. Therefore, it is urgent to design an artificial intelligence-based ship intelligent cabin system to solve the problems existing in the above-mentioned existing technologies. Summary of the Invention

[0005] In response to the above-mentioned defects or improvement needs of the prior art, the present invention provides a ship intelligent cabin system based on artificial intelligence, the purpose of which is to realize intelligent adjustment of the ship cabin environment, intelligent management and maintenance of equipment, intelligent response to emergency situations and intelligent utilization of cabin space, so as to effectively improve the overall performance and management level of the ship cabin.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions: A first aspect of the present invention provides an artificial intelligence-based ship intelligent cabin system, comprising: The intelligent environment adjustment module is used to collect cabin temperature, humidity, and air quality data in real time, predict temperature and humidity change trends through time series analysis algorithms or machine learning regression models, and automatically control the air conditioner, humidifier, dehumidifier, ventilation system, and air purification equipment to adjust the cabin environment; Equipment maintenance module: used to collect ship equipment operating parameters in real time, monitor and evaluate equipment operating status using data analysis algorithms, locate fault causes and locations through fault diagnosis algorithms, and generate corresponding maintenance plans; Emergency Intelligent Response Module: This module monitors cabin fires and water leaks in real time, analyzes fire conditions using a fire location algorithm, activates fire-fighting equipment, plans optimal escape routes, closes relevant pipeline valves and activates drainage pumps through the system, and employs a multi-objective optimization algorithm to comprehensively evaluate the best maintenance personnel. Cabin space utilization module: used to use the camera in the cabin combined with image recognition technology and AI algorithm to analyze personnel activities and intelligently switch multi-functional cabin functions according to cabin usage needs and personnel activities.

[0007] As an embodiment of the present application, the intelligent environment adjustment module includes a temperature and humidity control unit and an air quality adjustment unit; The temperature and humidity control unit collects cabin temperature and humidity data in real time through temperature and humidity sensors, compares it with a preset comfortable temperature and humidity range, and predicts the cabin temperature and humidity change trend through a time series analysis algorithm or a machine learning regression model based on the number of cabin occupants, activity type, and external ambient temperature. Based on the prediction results, control instructions are sent to the air conditioner, humidifier, and dehumidifier equipment to adjust the temperature and humidity. The air quality adjustment unit collects cabin air quality data in real time through harmful gas sensors and particulate matter sensors, compares and analyzes it with the air quality indicators of national standards, adjusts the working mode of the ventilation system and air purification equipment according to the pollutant type scenario, and adjusts the air quality.

[0008] As an embodiment of the present application, the working modes include a harmful gas purification mode, a particulate matter purification mode, and an energy-saving mode; The harmful gas purification mode is suitable for scenarios where formaldehyde, benzene, and TVOC harmful gases exceed the standard. The air purification equipment uses photocatalytic oxidation mode, activated carbon adsorption mode, or plasma purification mode to decompose harmful gases, and the ventilation system starts external circulation of fresh air to dilute the pollutant concentration. The particulate matter purification mode is suitable for scenarios where PM2.5, dust, smoke and other solid pollutants exceed the standard. The air purification equipment uses a high-efficiency filtration mode or a mixed purification mode to filter particulate matter, and the fresh air system turns off the external circulation and starts the internal circulation purification. The energy-saving mode is applicable to scenarios where the pollutant concentration drops below 50% of the national air quality index, and the power of the air purification equipment and the wind speed of the ventilation system are reduced accordingly.

[0009] As an embodiment of the present application, the equipment maintenance module includes an equipment status monitoring unit, an intelligent maintenance decision unit, and an equipment life prediction unit; The device status monitoring unit is used to collect the operating parameters of the device in real time through sensors, evaluate the operating parameters of the device using data analysis algorithms, and determine whether the device is operating normally; The intelligent maintenance decision unit uses a fault diagnosis algorithm to quickly locate the cause and location of the fault based on the evaluation results of the equipment status monitoring unit and generates a corresponding maintenance plan; The equipment life prediction unit is used to predict the remaining service life of the equipment using a life prediction model based on the equipment's usage time, operating conditions and historical maintenance records.

[0010] As an embodiment of the present application, the life prediction model combines the physical model and the data-driven model for hybrid prediction, and generates the final remaining life prediction value by fusing the output results of the physical model and the data-driven model through the leaf-bayesian network, wherein, The physical model includes: Wear accumulation sub-model: Calculates equivalent operating time based on the wear curves of key components in the equipment design manual and real-time load data; Chemical reaction sub-model: Use the Arrhenius equation to calculate the effect of temperature on the life of battery and capacitor components; The data-driven model includes: Long Short-Term Memory Network: Input operating time, temperature, vibration, and maintenance history parameters, and output the remaining service life probability distribution; Weibull regression model: Use equipment failure time as the dependent variable, operating conditions and environmental parameters as independent variables, and calculate the failure probability density function.

[0011] As an embodiment of the present application, the emergency intelligent response module includes a fire processing unit and a water leakage processing unit; The fire handling unit monitors the presence of a fire in real time by installing smoke sensors, temperature sensors, and flame detectors. When any fire sensor detects a fire-related signal, it immediately triggers a fire alarm signal to the system. At the same time, the fire location algorithm analyzes the alarm information and distribution of the fire sensors to determine the location and size of the fire, automatically activates the fire extinguishing equipment in the corresponding area, and plans the best escape route through the escape route planning algorithm. The water leakage processing unit monitors whether there is water leakage in the cabin in real time through the water leakage sensor. When the water leakage sensor detects a leak, the system closes the relevant pipeline valves and starts the drainage pump to drain the water. A multi-objective optimization algorithm is used to comprehensively evaluate the best maintenance personnel.

[0012] As an embodiment of the present application, the escape route planning algorithm constructs a cabin topology map based on the ship CAD drawing, marks the emergency exit, stair location, passage width and obstacle distribution, and uses the Dijkstra shortest path algorithm or The algorithm performs multi-objective path optimization based on minimizing path distance, avoiding fire and smoke, and minimizing personnel density to generate the best escape route.

[0013] As an embodiment of the present application, the multi-objective optimization algorithm quantifies the current location, maintenance skills, and maintenance time factors of the maintenance personnel and uses a weighted summation formula to calculate the maintenance personnel matching score. The calculation formula is as follows:

[0014] in, Score the maintenance personnel matching degree, 、 、 In order to dynamically adjust the weight coefficient, is the maximum straight-line distance between all maintenance personnel and the leak point, is the straight-line distance between the maintenance personnel’s current location and the leak point, is the skill score, is the base time, Estimated completion time.

[0015] As an embodiment of the present application, the cabin space utilization module includes a personnel behavior analysis unit and a cabin function switching unit; The personnel behavior analysis unit uses the privacy protection camera in the cabin combined with image recognition technology and behavior analysis algorithm to identify and analyze personnel activities and behaviors, and generates a cabin space utilization report based on the personnel behavior analysis results; The cabin function switching unit intelligently switches the multifunctional cabin function according to a predetermined plan and the analysis result of the personnel behavior analysis unit.

[0016] As an embodiment of the present application, the intelligent switching multi-functional cabin function includes: Time trigger mechanism: obtain cabin reservation plan from the system and enter pre-switching state 30 minutes in advance; Behavior trigger mechanism: When a person is detected performing a specific action sequence, it is determined as a pre-signal for function switching and the switching process is started in advance; Conflict handling mechanism: If a conflict between real-time behavior and the scheduled plan is detected, the system will send a reminder to the administrator's terminal, who will manually confirm whether to delay the switch or force the start.

[0017] The beneficial effects of the present invention are: 1. The present invention provides an intelligent environment adjustment module including a temperature and humidity control unit and an air quality adjustment unit. The temperature and humidity control unit collects cabin temperature and humidity data in real time through temperature and humidity sensors, compares the data with a preset comfortable temperature and humidity range, and predicts the cabin temperature and humidity change trend through a time series analysis algorithm or a machine learning regression model, taking into account the number of cabin occupants, activity type, and external ambient temperature influencing factors. Based on the prediction results, control instructions are sent to the air conditioner, humidifier, and dehumidifier equipment to adjust the temperature and humidity. The air quality adjustment unit collects cabin air quality data in real time through harmful gas sensors and particulate matter sensors, compares and analyzes the data with the national standard air quality indicators, and adjusts the working mode of the ventilation system and air purification equipment according to the pollutant type scenario to adjust the air quality. The intelligent environment adjustment module can accurately adjust the cabin temperature, humidity, and air quality according to occupant needs and real-time environmental changes, providing a more comfortable cabin environment for crew members and passengers. 2. The present invention provides an intelligent equipment maintenance module comprising an equipment status monitoring unit, an intelligent maintenance decision unit, and an equipment life prediction unit. The equipment status monitoring unit collects equipment operating parameters in real time through sensors, evaluates the equipment operating parameters using a data analysis algorithm, and determines whether the equipment is operating normally. The intelligent maintenance decision unit uses a fault diagnosis algorithm to quickly locate the cause and location of a fault based on the evaluation results of the equipment status monitoring unit and generates a corresponding maintenance plan. The equipment life prediction unit uses a life prediction model to predict the remaining service life of the equipment based on the equipment's usage time, operating conditions, and historical maintenance records. The intelligent equipment maintenance module implements real-time monitoring of equipment status, early warning of faults, and intelligent maintenance decision-making, thereby greatly improving equipment management efficiency, reducing equipment failure rates, and reducing ship operating costs. 3. The present invention provides an intelligent emergency response module comprising a fire handling unit and a water leakage handling unit. The fire handling unit monitors the presence of a fire in real time by installing smoke sensors, temperature sensors, and flame detectors. When any fire sensor detects a fire-related signal, it immediately triggers a fire alarm signal to the system. Simultaneously, a fire location algorithm analyzes the alarm information and distribution of the fire sensors to determine the location and size of the fire, automatically activates fire-fighting equipment in the corresponding area, and plans the optimal escape route using an escape route planning algorithm. The water leakage handling unit monitors the presence of water leaks in the cabin in real time by using a water leakage sensor. When a water leakage sensor detects a leak, the system closes the relevant pipeline valves and activates a drainage pump for drainage. A multi-objective optimization algorithm is used to comprehensively evaluate the best maintenance personnel. The intelligent emergency response module can respond quickly to emergencies such as fires and water leakages, and take effective countermeasures to ensure the safety of personnel and property on board. 4. The present invention sets up a cabin space utilization module including a personnel behavior analysis unit and a cabin function switching unit, wherein the personnel behavior analysis unit uses the privacy protection camera in the cabin in combination with image recognition technology and behavior analysis algorithm to identify and analyze personnel activities, and generates a cabin space utilization report based on the personnel behavior analysis results. The cabin function switching unit intelligently switches the multi-functional cabin function according to the predetermined plan and the analysis results of the personnel behavior analysis unit. The cabin space utilization module realizes efficient utilization of cabin space and improves the use value of the ship cabin through personnel behavior analysis and intelligent switching of cabin functions. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 This is a general architecture diagram of an artificial intelligence-based ship intelligent cabin system provided in an embodiment of the present invention; Figure 2 This is a system block diagram of an intelligent environment adjustment module of an artificial intelligence-based ship intelligent cabin system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0020] Reference Figure 1-Figure 2 The first aspect of the present invention provides an artificial intelligence-based ship intelligent cabin system, comprising: The intelligent environment adjustment module is used to collect cabin temperature, humidity, and air quality data in real time, predict temperature and humidity change trends through time series analysis algorithms or machine learning regression models, and automatically control the air conditioner, humidifier, dehumidifier, ventilation system, and air purification equipment to adjust the cabin environment; Equipment maintenance module: used to collect ship equipment operating parameters in real time, monitor and evaluate equipment operating status using data analysis algorithms, locate fault causes and locations through fault diagnosis algorithms, and generate corresponding maintenance plans; Emergency Intelligent Response Module: This module monitors cabin fires and water leaks in real time, analyzes fire conditions using a fire location algorithm, activates fire-fighting equipment, plans optimal escape routes, closes relevant pipeline valves and activates drainage pumps through the system, and employs a multi-objective optimization algorithm to comprehensively evaluate the best maintenance personnel. Cabin space utilization module: used to use the camera in the cabin combined with image recognition technology and AI algorithm to analyze personnel activities and intelligently switch multi-functional cabin functions according to cabin usage needs and personnel activities.

[0021] As an embodiment of the present application, the intelligent environment adjustment module includes a temperature and humidity control unit and an air quality adjustment unit; The temperature and humidity control unit collects cabin temperature and humidity data in real time through a temperature and humidity sensor, and compares it with a preset comfortable temperature and humidity range, taking into account influencing factors such as the number of cabin occupants, activity type, and external ambient temperature. Among the influencing factors, the activity type includes static office work and dynamic exercise, which correspond to differences in heat production; the external ambient temperature is obtained through the ship's meteorological system; the cabin temperature and humidity change trend is predicted through a time series analysis algorithm or a machine learning regression model, and control instructions are sent to the air conditioner, humidifier, and dehumidifier equipment based on the prediction results to adjust the temperature and humidity; Among them, this application uses high-precision digital temperature and humidity sensors, such as the SHT30 series, which have a temperature accuracy of ±0.3℃ and a humidity accuracy of ±2%RH to meet the needs of ship cabin environment monitoring; the temperature and humidity sensors are evenly distributed on the cabin ceiling, the middle of the wall 1.5-2 meters from the ground, and areas with dense personnel activities, such as above the conference table, to ensure data coverage without dead angles; the temperature and humidity sensor collects temperature and humidity data every 1 minute, generates a data packet containing a timestamp, temperature value, and humidity value, and transmits it to the system control center, that is, the ship's central control room server, via the Modbus RTU bus or Wi-Fi wirelessly. The transmission delay is ≤5 seconds to ensure real-time data.

[0022] Preferably, in the present application, the time series analysis algorithm adopts the ARIMA model, and the machine learning regression model adopts the random forest. Based on the historical data training model, a mapping relationship between temperature and humidity changes and influencing factors such as the number of cabin personnel, activity type, and external ambient temperature is established; the real-time temperature and humidity data are compared with the preset comfort range, the temperature preset range is 22℃~26℃, and the humidity preset range is 40%~60%. When the real-time temperature and humidity data exceed the preset comfort range, the adjustment process is triggered. For example, if the collected temperature is 28℃ and the humidity is 35%, the system determines that cooling and humidification are required, and enters the next step of calculation. According to the current data and influencing factors, the future is predicted. The system calculates the temperature and humidity trends over the next 30 minutes to 2 hours. For example, when there are a lot of people and a lot of activity, the model predicts that the temperature will rise by 0.5°C and the humidity will drop by 2% every 10 minutes. The system combines historical data to calculate the equipment adjustment amount. For example, if the current temperature is 28°C and it is predicted to rise to 30°C in 1 hour, based on the prediction results, the system sends PID control instructions to the air conditioner to adjust the cooling capacity and wind speed, and sends start and stop instructions and humidification parameters to the humidifier until the temperature and humidity return to the comfortable range. Among them, when the system combines historical data to calculate the equipment adjustment amount, it mainly uses the proportional-integral-differential control algorithm to achieve quantitative adjustment of equipment such as air conditioners and humidifiers. The specific calculation formula is as follows:

[0023] in, The current equipment control quantity, such as air conditioning cooling capacity and humidifier humidification capacity; is the proportional coefficient, used to quickly respond to deviations; The deviation between the current temperature and humidity values ​​and the preset comfort range; is the integral coefficient, which is used to eliminate steady-state errors, such as long-term temperature fluctuations; Indicates at a point in time Error value at the moment; represents the differential of the time variable; is the differential coefficient, which is used to suppress overshoot, such as avoiding excessive temperature adjustment; Indicates the rate of change of error.

[0024] Application scenario example 1: When the number of participants in the cabin conference room exceeds 20, the temperature and humidity sensors detect a rapid temperature rise, such as from 24°C to 26.5°C within 10 minutes. The system starts the air conditioning in high air volume mode in advance and turns on the humidifier to prevent the humidity from dropping.

[0025] Application scenario example 2: If the ship enters a high-temperature sea area and the external temperature is greater than 35°C, the temperature and humidity sensors continuously collect high-temperature data, and the system automatically switches the air conditioner to "energy-saving cooling mode", reducing energy consumption while maintaining a comfortable temperature. For example, the cooling capacity is increased by 20%, but the wind speed is reduced by 10% to reduce noise.

[0026] The air quality adjustment unit collects cabin air quality data in real time through harmful gas sensors and particulate matter sensors. The harmful gas sensors and particulate matter sensors sample once every 3 minutes and compare and analyze the data with the national standard air quality indicators. The working mode of the ventilation system and air purification equipment is adjusted according to the type of pollutants and the scene to adjust the air quality.

[0027] As an embodiment of the present application, the working modes include a harmful gas purification mode, a particulate matter purification mode, and an energy-saving mode; The harmful gas purification mode is suitable for scenarios where formaldehyde, benzene, and TVOC harmful gases exceed the standard. The air purification equipment uses photocatalytic oxidation mode, activated carbon adsorption mode, or plasma purification mode to decompose harmful gases, and the ventilation system turns on external circulation of fresh air to dilute the pollutant concentration.

[0028] The principle of the photocatalytic oxidation mode is to use photocatalysts such as TiO2 to produce strong oxidizing free radicals under ultraviolet light, decomposing harmful gases such as formaldehyde into CO2 and H2O. It is suitable for formaldehyde release source areas such as newly renovated cabins and storage rooms for chemicals. The air purification equipment turns on the ultraviolet lamp, adjusts the power to maximum, and sets the fan speed to high to ensure that the airflow passes through the catalytic module sufficiently. The principle of the activated carbon adsorption mode is to absorb organic pollutants such as benzene and odor through the pore structure of activated carbon. It is suitable for volatile organic compound (VOC) pollution scenarios such as kitchen fume odor and printed material storage areas. The air purification equipment turns off the UV lamp to prevent the activated carbon from failing due to high temperature, and the fan speed is set to medium to extend the contact time between the gas and the activated carbon. The principle of plasma purification mode is to generate plasma through high-voltage discharge, destroying the structure of bacteria and viruses and decomposing some harmful gases. It is applicable to medical cabins and areas with dense crowds and poor ventilation, such as crew restaurants. The air purification equipment uses a plasma generator with the voltage adjusted to a set value of 5KV. The wind speed is linked to the pollutant concentration. The higher the concentration, the higher the wind speed.

[0029] The particulate matter purification mode is suitable for scenarios where PM2.5, dust, smoke and other solid pollutants exceed the standard. The air purification equipment uses a high-efficiency filtration mode or a mixed purification mode to filter particulate matter. The fresh air system turns off the external circulation and starts the internal circulation purification to prevent external polluted air from entering. Among them, the principle of high-efficiency filtration mode is to use HEPA filter with a filtration efficiency of ≥99.97%@0.3μm to physically intercept particulate matter. It is suitable for scenes such as ships passing through hazy seas and dust spreading from machinery compartments to living cabins; the air purification equipment starts full-power filtration; The principle of the hybrid purification mode is to use a HEPA filter + electrostatic dust collection combination. The particles are first charged by the electrostatic field and then captured by the filter, thereby improving the filtration efficiency of large particles. The applicable scenarios are cabin protection in sandstorm weather and air purification after welding operations. The electrostatic dust collection voltage of the air purification equipment is set to 3KV, the fan speed is set to high, and the electrostatic module is automatically cleaned every 2 hours to prevent dust accumulation from affecting efficiency.

[0030] The energy-saving mode is applicable to scenarios where the pollutant concentration drops below 50% of the national air quality index. The power of the air purification equipment and the wind speed of the ventilation system are reduced accordingly. For example, the power of the photocatalytic oxidation mode of the air purification equipment is reduced to 70%, the fan speed is adjusted to medium, and the filtration mode is turned on for intermittent operation, such as working for 30 minutes and pausing for 10 minutes.

[0031] As an embodiment of the present application, the equipment maintenance module includes an equipment status monitoring unit, an intelligent maintenance decision unit, and an equipment life prediction unit; The equipment status monitoring unit is used to collect equipment operating parameters in real time through sensors, such as current and voltage of lighting equipment, temperature and vibration of ventilation equipment motors, etc., and transmit the data to the system database through a wireless transmission module. The data analysis algorithm is used to evaluate the equipment operating parameters and determine whether the equipment is operating normally. The sensor installation locations and collected data are shown in Table 1:

[0032] Table 1 Sensor installation location and collected data The wireless transmission module uses MQTT (Message Queuing Telemetry Transport Protocol) or Modbus TCP, supporting low-power, highly reliable data transmission and adapting to the signal penetration requirements of complex metallic environments within ship cabins. An edge computing gateway (such as the UNO series) with an integrated Wi-Fi 6 or 4G LTE module is deployed to aggregate and forward sensor data. Sensors connect to the gateway via Zigbee or Bluetooth Low Energy (BLE) short-range communication protocols. The gateway then transmits the data to the system database via wired (Ethernet) or wireless communication. Channel hopping technology and data encryption (AES-128) are used to prevent electromagnetic interference and data leakage from ship electrical equipment. The database uses a time series database, such as InfluxDB, with an optimized storage structure for the time series characteristics of device operating parameters, supporting high-concurrency writes and fast queries. Data content includes real-time parameters such as current and voltage of lighting equipment, temperature, vibration frequency, and speed of motors, and refrigerant pressure of air conditioners; metadata such as device number, sensor type, installation location, calibration time, and communication status; and historical data from the past three years or more, stored by timestamp, for trend analysis and fault tracing.

[0033] The data analysis algorithm evaluates the operating parameters of the equipment through threshold warning, trend analysis, and fault feature extraction. The threshold warning sets the normal parameter range of the equipment and compares the collected real-time operating parameters with the set normal parameter range of the equipment. If the parameters exceed the range, an early warning is triggered; the trend analysis identifies abnormal trends by comparing the current parameters of the equipment with historical data; the fault feature extraction module analyzes the vibration data spectrum using Fourier transform to identify mechanical fault characteristics such as bearing wear and gear meshing abnormalities; and transient anomalies in the current signal are detected through wavelet analysis.

[0034] Based on the evaluation results of the equipment status monitoring unit, the intelligent maintenance decision-making unit uses a fault diagnosis algorithm to quickly locate the cause and location of the fault and generate a corresponding maintenance plan. The fault diagnosis algorithm performs layered troubleshooting based on data fusion, which includes integrating real-time equipment parameters, historical maintenance records such as component replacement time, and design drawings such as circuit schematics. The layered troubleshooting includes first-level troubleshooting, second-level troubleshooting, and third-level troubleshooting. The first-level troubleshooting determines whether the fault is a sensor fault based on parameter out-of-bounds. For example, when a single sensor data is abnormal, the fault is verified by comparing it with data from adjacent devices. The second-level troubleshooting uses a fault tree model (FTA) to reversely deduce possible causes from abnormal parameters. For example, if the motor temperature is too high, it will cause poor heat dissipation (fan failure) or excessive load (bearing jam). The third-level troubleshooting uses case-based reasoning (CBR) to match historical failure cases. For example, if a certain model of motor has previously experienced temperature increases due to lack of bearing oil, this can quickly locate the fault point.

[0035] The equipment life prediction unit is used to predict the remaining service life of the equipment using a life prediction model based on the equipment's usage time, operating conditions, and historical maintenance records.

[0036] As an embodiment of the present application, the life prediction model combines the physical model and the data-driven model for hybrid prediction. The output results of the physical model and the data-driven model are integrated through the leaf-bayesian network to generate the final remaining life prediction value. The model parameters are regularly updated according to the new maintenance data to improve the prediction accuracy, wherein, The physical model includes: Wear accumulation sub-model: Based on the wear curves of key components in the equipment design manual, such as the relationship between bearing wear and operating time, combined with real-time load data, the equivalent operating time is calculated using the following formula:

[0037] in, is the equivalent running time, reflecting the actual amount of wear; The actual running time of the equipment; is the acceleration factor, which is an acceleration coefficient related to the load rate. The higher the load rate, the greater the acceleration factor. For example, the motor has a design life of 100,000 hours. When the load rate is 80%, the equivalent acceleration factor is 1.5. The actual operation time of 3 years (26,000 hours) is equivalent to 40,000 hours of wear.

[0038] Chemical reaction sub-model: Uses the Arrhenius equation to calculate the effect of temperature on the life of battery and capacitor components. For example, for every 10°C increase in temperature, the life of the capacitor is halved. The data-driven model includes: Long Short-Term Memory Network: Input operating time, temperature, vibration, and maintenance history parameters, and output the remaining service life probability distribution; Weibull regression model: Using equipment failure time as the dependent variable and operating conditions and environmental parameters as independent variables, the failure probability density function is calculated, expressed as follows:

[0039] in, The equipment running time; is a scale parameter, which is related to the average life of the equipment; is a shape parameter that reflects the failure mode, such as early failure, accidental failure or loss failure; 、 The operating conditions and environmental parameters are affected by the regression coefficient, thereby changing the failure probability.

[0040] As an embodiment of the present application, the emergency intelligent response module includes a fire processing unit and a water leakage processing unit; The fire handling unit monitors the presence of a fire in real time by installing smoke sensors, temperature sensors, and flame detectors. When any fire sensor detects a fire-related signal, it immediately triggers a fire alarm signal to the system. At the same time, the fire location algorithm analyzes the alarm information and distribution of the fire sensors to determine the location and size of the fire, automatically activates the fire extinguishing equipment in the corresponding area, and plans the best escape route through the escape route planning algorithm. The threshold standards for fire sensors to detect fires and issue alarms are as follows: Smoke sensor: uses the light scattering principle, and the threshold is set at a smoke concentration of ≥ 5% obs / m (optical density 5% per meter). That is, when the smoke causes the light attenuation to exceed 5%, the alarm is triggered; Temperature sensor: includes temperature rise rate alarm and absolute temperature alarm. The temperature rise rate alarm triggers the alarm mechanism when the temperature rise rate is ≥5℃ / min, excluding normal ambient temperature rise. The absolute temperature alarm is based on the ignition point of the ship's cabin materials and the tolerance limit of personnel, and the threshold is set to ≥70℃.

[0041] Flame detector: By detecting the intensity of infrared or ultraviolet radiation, the threshold is set to the flame thermal radiation power density ≥ 10kW / m², and combined with a flashing frequency of 4-8Hz to eliminate light interference.

[0042] The fire location algorithm in this application combines multi-source data fusion, spatial cluster analysis, and confidence assessment for precise positioning; Multi-source data fusion: The physical location coordinates of alarm sensors are collected and pre-annotated using the ship's BIM model, such as the coordinates of cabin A (X=10m, Y=5m, Z=2m). Alarm types and timestamps are extracted, such as smoke sensor S1 alarming at 14:05:10 and temperature sensor T2 alarming at 14:05:12.

[0043] Spatial cluster analysis: Using the DBSCAN density clustering algorithm, sensors within the same area (e.g., within a radius of ≤3 meters and with an alarm time difference of ≤5 seconds) are grouped into an "alarm cluster," with the center of the cluster determined to be the fire location. For example, sensors S1 (X=10, Y=5), S2 (X=10.5, Y=5.2), and T2 (X=9.8, Y=5.5) that continuously alarm within 3 seconds are clustered into the area with coordinates (10.1, Y=5.2), located at the northwest corner of cabin A.

[0044] Confidence assessment: If a single sensor alarms, it is marked as "suspected fire" and triggers camera linkage confirmation, using image recognition to determine whether there is smoke or flames. If ≥2 types of sensors, such as a smoke sensor and a temperature sensor, alarm in the same area and the confidence level is ≥90%, it is directly determined to be "fire occurrence."

[0045] This application estimates the fire severity by using smoke concentration gradient, temperature field distribution, and flame area. The smoke concentration gradient is calculated based on the smoke concentration value of the alarm sensor and the inverse distance weighted (IDW) principle to estimate the concentration at the center of the fire source. The closer to the fire source, the higher the concentration. The expression formula is as follows:

[0046] in, For the The smoke concentration detected by the sensor, is the concentration attenuation coefficient, is the distance between the sensor and the center of the suspected fire source, calculated by the fire sensor coordinates and the cluster center.

[0047] The temperature field distribution is calculated by fitting the thermal imaging image with multi-node temperature data, calculating the area of ​​the high-temperature area, collecting the coordinates of each temperature sensor and the corresponding temperature value, constructing a two-dimensional temperature field, and calculating the area of ​​all continuous areas with a two-dimensional temperature field ≥80°C. When the area of ​​the continuous area exceeding 80°C is 5m²≤≤10m², it is judged as medium fire.

[0048] The flame area is calculated by using the YOLOv5 algorithm to identify the pixel ratio of the flame area and convert it into the actual area. The calculation formula is as follows:

[0049] in, Indicates the number of pixels in the flame area identified by the YOLOv5 algorithm, Indicates the actual area of ​​the cabin, Indicates the total number of pixels in the cabin image. A flame area greater than 2m² is considered a large fire.

[0050] In this application, the fire severity levels are divided into: Level I (small fire): Only a single sensor is alarming, and the smoke and temperature have not spread significantly. For example, only the smoke concentration is greater than 5% obs / m, but the high temperature area is less than 5m² and there is no flame. Level II (medium fire): Multiple sensors alarm, such as smoke sensor and temperature sensor alarm, and the high temperature area is 5m²≤≤10m², regardless of the size of the flame area; Level III (large fire): The flame area is greater than 2m² or the smoke spreads to adjacent compartments.

[0051] As an embodiment of the present application, the escape route planning algorithm constructs a cabin topology map based on the ship CAD drawing, marks the emergency exit, stair location, passage width and obstacle distribution, and uses the Dijkstra shortest path algorithm or The algorithm performs multi-objective path optimization based on minimizing path distance, avoiding fire and smoke, and minimizing personnel density to generate the best escape route.

[0052] The escape route planning algorithm also dynamically marks obstacles according to the fire location, marks the fire location and adjacent cabin passages as inaccessible nodes, and uses the Dijkstra shortest path algorithm or The algorithm searches for the optimal path to meet the rapidity required for path planning in the early stages of a fire. It dynamically adjusts the path based on the crowd density (people / m²) monitored in real time by cameras, addressing the inability of traditional algorithms to cope with dynamic obstacles such as the flow of people. It also introduces a multi-objective optimization mechanism to prioritize routes for areas with special needs, such as pregnant women and the elderly. The cabin reservation system identifies the locations of these special groups and prioritizes their escape routes.

[0053] The water leakage processing unit monitors whether there is water leakage in the cabin in real time through the water leakage sensor. When the water leakage sensor detects a water leakage, it immediately sends a water leakage signal to the system, along with the water leakage location information. After receiving the water leakage signal, the system first closes the pipe valve related to the water leakage area to prevent the water leakage from further expanding. At the same time, the system starts the drainage pump to drain the water, and uses a multi-objective optimization algorithm to comprehensively evaluate the best maintenance personnel to go to the water leakage point for repairs based on the water leakage location and possible causes.

[0054] Specifically, each water leakage sensor is bound to a unique identifier during installation, such as a UUID or Modbus address, corresponding to a specific location in the ship's piping system diagram, such as "Bathroom pipe connection - sensor No. 03"; the sensor location information is pre-entered into the system database, associated with the three-dimensional coordinates (X, Y, Z) and the piping system to which it belongs, such as "Water supply system - 2nd deck - Bathroom branch"; when the sensor detects a water leak, such as an electrode-type sensor triggering a signal due to water conductivity, a data packet containing "sensor ID + status + timestamp" is sent to the system via wired (RS485) or wireless (Zigbee) means. The system queries the database based on the sensor ID and automatically parses the physical coordinates of the leak location, the area to which it belongs, such as "Bathroom, Area A, 2nd deck", and the associated pipeline valves, such as "Bathroom water supply main valve V-02".

[0055] As an embodiment of the present application, the multi-objective optimization algorithm quantifies the current location, maintenance skills, and maintenance time factors of the maintenance personnel and uses a weighted summation formula to calculate the maintenance personnel matching score. The calculation formula is as follows:

[0056] in, Score the maintenance personnel matching degree, 、 、 In order to dynamically adjust the weight coefficient, is the maximum straight-line distance between all maintenance personnel and the leak point, is the straight-line distance between the maintenance personnel’s current location and the leak point, is the skill score, is the base time, Estimated completion time.

[0057] Among them, the current position of the maintenance personnel is quantified by using the ship's internal positioning system, such as UWB ultra-wideband positioning, to obtain coordinates in real time, and calculate the maximum straight-line distance between all maintenance personnel and the leak point and the straight-line distance between the maintenance personnel's current position and the leak point; the maintenance skills of the maintenance personnel are quantified by establishing a skill label library, such as "pipeline maintenance", "electrical maintenance", and "welding", with each maintenance personnel having a predefined skill matching score, such as 0-100 points, and pipeline maintenance skills ≥80 points are qualified; the maintenance time of the maintenance personnel is quantified by estimating the benchmark time to complete the task based on the leak type and historical maintenance records, and adjusting the time value in combination with the personnel's current task load; 、 、 In order to dynamically adjust the weight coefficient, if the water leak involves electrical equipment or may cause secondary disasters (such as in a nearby distribution room), the "skill matching" weight will be automatically increased to 0.6 to ensure that personnel with electrical safety qualifications are dispatched first. If all maintenance personnel are busy, cross-regional scheduling will be triggered, and the next best personnel will be selected from adjacent decks or spare teams. The navigation path, that is, the shortest safe channel, will be sent through the system.

[0058] As an embodiment of the present application, the cabin space utilization module includes a personnel behavior analysis unit and a cabin function switching unit; The personnel behavior analysis unit uses the privacy protection camera in the cabin to capture an image of personnel activities in the cabin every 10 minutes, and combines image recognition technology and behavior analysis algorithms to identify and analyze the personnel activities in the image. For example, by identifying the posture, movement and position changes of the personnel, the activity trajectory of the personnel is analyzed; by counting the number of people in different areas and calculating the population density, a cabin space utilization report is generated based on the results of the personnel behavior analysis; for example, the report shows that the population density in a public area is relatively high between 2 and 4 pm every day, and the personnel activities are mainly communication and discussion. Based on this report, the management personnel can adjust the facility layout of the area, such as adding movable tables and chairs to facilitate personnel communication.

[0059] Among them, privacy protection cameras protect people's privacy through anonymization processing and area masking technology. Anonymization processing is to blur the face or desensitize the features of the images collected by the camera, retaining only non-privacy information such as human outline and posture to ensure that personal identity is not involved; area masking technology defines analysis areas, such as public activity areas, and excludes private spaces such as bedrooms and bathrooms from image collection, thereby protecting privacy from the physical deployment level.

[0060] The image recognition technology in this application includes personnel detection and tracking, density statistics and regional analysis. The personnel detection and tracking adopts the YOLO target detection algorithm to identify human targets in the image in real time, generate bounding box annotations, realize cross-frame tracking through the DeepSORT algorithm, and record the movement trajectory of a single target; the density statistics and regional analysis divides the cabin into grid areas, calculates the number of people in each grid through the feature pyramid network (FPN), and generates a real-time density heat map. The darker the color, the denser the people.

[0061] The behavioral analysis algorithm in this application includes activity type recognition, space utilization evaluation, and layout optimization suggestions. The activity type recognition is based on the temporal action detection model I3D network, which analyzes human posture sequences, such as standing, walking, sitting, and communication gestures, and identifies activity types, such as "meeting discussion" and "rest and relaxation"; the space utilization evaluation calculates the area occupancy rate by the ratio of the actual practical area to the total area, and calculates the time utilization rate by the ratio of the active time of the area to the total time, and generates an analysis report; the layout optimization suggestion uses the clustering algorithm DBSCAN to count the number of people in different areas, calculate the density value, analyze the personnel gathering pattern, and automatically generate facility adjustment plans. For example, if group discussions often occur in a certain area, it is recommended to add movable round tables and soundproof screens; if the passage is frequently congested, it is recommended to widen the path or adjust the direction of furniture placement.

[0062] The cabin function switching unit intelligently switches the multifunctional cabin function according to a predetermined plan and the analysis result of the personnel behavior analysis unit.

[0063] As an embodiment of the present application, the intelligent switching multi-functional cabin function includes: Time trigger mechanism: obtain cabin reservation plan from the system and enter pre-switching state 30 minutes in advance; Behavior trigger mechanism: When a person is detected performing a specific action sequence, it is determined as a pre-signal for function switching and the switching process is started in advance; Conflict handling mechanism: If a conflict between real-time behavior and the scheduled plan is detected, the system will send a reminder to the management terminal, who will manually confirm whether to delay the switch or force the start.

[0064] For example, if the booking plan shows that the cabin will be used as a rest cabin after 7 pm, and it is detected at 6:30 pm that the people in the cabin begin to prepare for rest, the AI ​​system will start the cabin function switching program. The system first adjusts the lighting to a soft warm tone and reduces the brightness of the light; then lowers the blackout curtains to create a quiet rest environment; then adjusts the air-conditioning temperature to 24°C, which is suitable for sleep. At the same time, the system controls the automated facilities in the cabin and converts the conference tables and chairs into beds and other rest facilities through mechanical transmission devices. During the conversion process, the system monitors the facility conversion status in real time through sensors to ensure that the conversion process is safe and smooth.

[0065] The above description is merely an illustration of some preferred embodiments of the present disclosure and the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present disclosure is not limited to the technical solutions formed by the specific combination of the above-mentioned technical features, but also includes other technical solutions formed by any combination of the above-mentioned technical features or their equivalents without departing from the above-mentioned inventive concept.

Claims

1. An artificial intelligence-based ship intelligent cabin system, characterized in that: include: The intelligent environment adjustment module is used to collect cabin temperature, humidity, and air quality data in real time, predict temperature and humidity change trends through time series analysis algorithms or machine learning regression models, and automatically control the air conditioner, humidifier, dehumidifier, ventilation system, and air purification equipment to adjust the cabin environment; Equipment maintenance module: used to collect ship equipment operating parameters in real time, monitor and evaluate equipment operating status using data analysis algorithms, locate fault causes and locations through fault diagnosis algorithms, and generate corresponding maintenance plans; Emergency Intelligent Response Module: This module monitors cabin fires and water leaks in real time, analyzes fire conditions using a fire location algorithm, activates fire-fighting equipment, plans optimal escape routes, closes relevant pipeline valves and activates drainage pumps through the system, and employs a multi-objective optimization algorithm to comprehensively evaluate the best maintenance personnel. Cabin space utilization module: used to use the privacy protection camera in the cabin combined with image recognition technology and AI algorithm to analyze personnel activities and intelligently switch multi-functional cabin functions according to cabin usage needs and personnel activities.

2. The artificial intelligence-based ship intelligent cabin system according to claim 1, characterized in that: The intelligent environment adjustment module includes a temperature and humidity control unit and an air quality adjustment unit; The temperature and humidity control unit collects cabin temperature and humidity data in real time through temperature and humidity sensors, compares it with a preset comfortable temperature and humidity range, and predicts the cabin temperature and humidity change trend through a time series analysis algorithm or a machine learning regression model based on the number of cabin occupants, activity type, and external ambient temperature. Based on the prediction results, control instructions are sent to the air conditioner, humidifier, and dehumidifier equipment to adjust the temperature and humidity. The air quality adjustment unit collects cabin air quality data in real time through harmful gas sensors and particulate matter sensors, compares and analyzes it with the air quality indicators of national standards, adjusts the working mode of the ventilation system and air purification equipment according to the pollutant type scenario, and adjusts the air quality.

3. The artificial intelligence-based ship intelligent cabin system according to claim 2, characterized in that: The working modes include harmful gas purification mode, particulate matter purification mode and energy-saving mode; The harmful gas purification mode is suitable for scenarios where formaldehyde, benzene, and TVOC harmful gases exceed the standard. The air purification equipment uses photocatalytic oxidation mode, activated carbon adsorption mode, or plasma purification mode to decompose harmful gases, and the ventilation system starts external circulation of fresh air to dilute the pollutant concentration. The particulate matter purification mode is suitable for scenarios where PM2.5, dust, smoke and other solid pollutants exceed the standard. The air purification equipment uses a high-efficiency filtration mode or a mixed purification mode to filter particulate matter, and the fresh air system turns off the external circulation and starts the internal circulation purification. The energy-saving mode is applicable to scenarios where the pollutant concentration drops below 50% of the national air quality index, and the power of the air purification equipment and the wind speed of the ventilation system are reduced accordingly.

4. The artificial intelligence-based ship intelligent cabin system according to claim 1, characterized in that: The equipment maintenance module includes an equipment status monitoring unit, an intelligent maintenance decision unit and an equipment life prediction unit; The device status monitoring unit is used to collect the operating parameters of the device in real time through sensors, evaluate the operating parameters of the device using data analysis algorithms, and determine whether the device is operating normally; The intelligent maintenance decision unit uses a fault diagnosis algorithm to quickly locate the cause and location of the fault based on the evaluation results of the equipment status monitoring unit and generates a corresponding maintenance plan; The equipment life prediction unit is used to predict the remaining service life of the equipment using a life prediction model based on the equipment's usage time, operating conditions and historical maintenance records.

5. The artificial intelligence-based ship intelligent cabin system according to claim 4, characterized in that: The life prediction model combines the physical model and the data-driven model for hybrid prediction, and generates the final remaining life prediction value by fusing the output results of the physical model and the data-driven model through the leaf-bayesian network, where: The physical model includes: Wear accumulation sub-model: Calculates equivalent operating time based on the wear curves of key components in the equipment design manual and real-time load data; Chemical reaction sub-model: Use the Arrhenius equation to calculate the effect of temperature on the life of battery and capacitor components; The data-driven model includes: Long Short-Term Memory Network: Input operating time, temperature, vibration, and maintenance history parameters, and output the remaining service life probability distribution; Weibull regression model: Use equipment failure time as the dependent variable, operating conditions and environmental parameters as independent variables, and calculate the failure probability density function.

6. The artificial intelligence-based ship intelligent cabin system according to claim 1, characterized in that: The emergency intelligent response module includes a fire handling unit and a water leakage handling unit; The fire handling unit monitors the presence of a fire in real time by installing smoke sensors, temperature sensors, and flame detectors. When any fire sensor detects a fire-related signal, it immediately triggers a fire alarm signal to the system. At the same time, the fire location algorithm analyzes the alarm information and distribution of the fire sensors to determine the location and size of the fire, automatically activates the fire extinguishing equipment in the corresponding area, and plans the best escape route through the escape route planning algorithm. The water leakage processing unit monitors whether there is water leakage in the cabin in real time through the water leakage sensor. When the water leakage sensor detects a leak, the system closes the relevant pipeline valves and starts the drainage pump to drain the water. A multi-objective optimization algorithm is used to comprehensively evaluate the best maintenance personnel.

7. The artificial intelligence-based ship intelligent cabin system according to claim 6, characterized in that: The escape route planning algorithm constructs a cabin topology map based on the ship's CAD drawings, marking the emergency exits, stair locations, passage widths, and obstacle distribution, and uses the Dijkstra shortest path algorithm or The algorithm performs multi-objective path optimization based on minimizing path distance, avoiding fire and smoke, and minimizing personnel density to generate the best escape route.

8. The artificial intelligence-based ship intelligent cabin system according to claim 1, characterized in that: The multi-objective optimization algorithm quantifies the maintenance personnel's current location, maintenance skills, and maintenance time factors, and uses a weighted sum formula to calculate the maintenance personnel matching score. The calculation formula is as follows: in, Score the maintenance personnel matching degree, 、 、 In order to dynamically adjust the weight coefficient, is the maximum straight-line distance between all maintenance personnel and the leak point, is the straight-line distance between the maintenance personnel’s current location and the leak point, is the skill score, is the base time, Estimated completion time.

9. The artificial intelligence-based ship intelligent cabin system according to claim 1, characterized in that: The cabin space utilization module includes a personnel behavior analysis unit and a cabin function switching unit; The personnel behavior analysis unit uses the privacy protection camera in the cabin combined with image recognition technology and behavior analysis algorithm to identify and analyze personnel activities and behaviors, and generates a cabin space utilization report based on the personnel behavior analysis results; The cabin function switching unit intelligently switches the multifunctional cabin function according to a predetermined plan and the analysis result of the personnel behavior analysis unit.

10. The artificial intelligence-based ship intelligent cabin system according to claim 9, characterized in that: The intelligent switching multifunctional cabin function includes: Time trigger mechanism: obtain cabin reservation plan from the system and enter pre-switching state 30 minutes in advance; Behavior trigger mechanism: When a person is detected performing a specific action sequence, it is determined as a pre-signal for function switching and the switching process is started in advance; Conflict handling mechanism: If a conflict between real-time behavior and the scheduled plan is detected, the system will send a reminder to the administrator's terminal, who will manually confirm whether to delay the switch or force the start.

Images