Operating information for fire-fighting equipment

Abstract

Devices, systems, and methods for operating information from fire fighting devices are described herein. In some examples, one or more embodiments include a memory and a processor to execute instructions stored in the memory to: receive an input query for operational information of a fire fighting apparatus; providing the input query to a small language model, the small language model associated with the fire fighting equipment and trained to provide information about the fire fighting equipment; and generating a response to the input query including the operational information using the small language model.

Description

Technical Field

[0001] This disclosure relates to equipment, systems, and methods for providing operational information for fire protection equipment. Background Technology

[0002] Facilities such as commercial facilities, office buildings, hospitals, and campuses (e.g., including buildings and outdoor spaces) may have event detection systems that can be triggered during events such as emergencies, e.g., fires, to warn occupants to evacuate. Such event detection systems may include alarm systems with control panels and multiple event devices (e.g., sensors, sounders, pull-out fire alarm boxes, etc.) located throughout the facility (e.g., on different floors and / or in different rooms), which can act when an event (e.g., a hazardous event, a malfunction event, etc.) occurs in the facility. In the example of the event, the multiple event devices may notify occupants of the facility of the event via alarms and / or other mechanisms. Attached Figure Description

[0003] Figure 1 This is an example of a system for receiving operational information from fire-fighting equipment, according to one or more embodiments of this disclosure.

[0004] Figure 2 Examples of training small and large language models according to one or more embodiments of this disclosure are illustrated.

[0005] Figure 3 Examples of providing input queries to small and / or large language models according to one or more embodiments of this disclosure are illustrated.

[0006] Figure 4 This is an example of a fire protection device for receiving operational information from fire protection equipment, according to one or more embodiments of this disclosure. Detailed Implementation

[0007] This document describes devices, systems, and methods for processing operational information from fire protection equipment. In some examples, one or more embodiments include a memory and a processor for executing instructions stored in the memory to: receive an input query for operational information about the fire protection equipment; provide the input query to a small language model associated with and trained to provide information about the fire protection equipment; and use the small language model to generate a response to the input query that includes the operational information.

[0008] Facilities may utilize event detection systems to alert occupants to emergencies, such as fires. An event detection system can be a system of devices that operate to collect information about the facility and provide the collected information for analysis. Such event detection systems may also take action based on the collected information, such as providing auditory and / or visual warnings in an emergency. For example, an event detection system may utilize event devices to alert occupants to emergencies occurring in a space, such as fires. As used herein, the term "event device" refers to a device capable of receiving event-related inputs and / or generating event-related outputs. Such event devices may be part of an event detection system within the facility / the entire space within the facility and may include devices such as fire protection equipment including fire sensors, smoke detectors, heat detectors, carbon monoxide (CO) detectors, or combinations thereof; fire control panels; air quality sensors; interfaces; manual alarm points (MCPs); pull-out fire alarm boxes; input / output modules; aspirating devices; fire doors; and / or auditory / visual devices (e.g., speakers, emitters, flashers, buzzers, microphones, cameras, video displays, video screens, etc.), relay output modules, and other types of event devices. Such event devices may also include self-test capabilities.

[0009] In the event of an inquiry regarding fire-fighting equipment and / or incident equipment associated with that equipment, technical support is typically involved to determine a response to the inquiry. For example, if fire-fighting equipment malfunctions, the inquiry may include a request to understand why the malfunction occurred. Technical support can be contacted to troubleshoot the malfunction, determine its cause, and find solutions to fix it.

[0010] Determining the response to a query may involve a technical review and / or analysis of fire equipment operation data. For example, technical support personnel may have to manually review event logs, diagnostic data, configuration file information, source code information, etc., to determine the response to a specific query about fire equipment.

[0011] However, technical support personnel may not be located in the same geographic area as the facility and / or incident system. Due to these geographic limitations, meetings between technical support personnel and field staff at the facility with the incident system may be difficult due to time zone differences.

[0012] According to the present disclosure, the operation information of a fire protection device can enable a user to provide an input query regarding the operation information of the fire protection device. The fire protection device can include a trained small language model that can generate a response to the input query including the operation information. Compared with previous methods, using the small language model can enable the user to easily determine the operation information regarding the fire protection device, thereby allowing the user to more easily troubleshoot technical problems of the fire protection device and / or determine whether to take steps to remedy predicted technical problems without having to manually review the technical data of the fire protection device. Additionally, the small language model can be continuously trained using data recorded by the fire protection device, enabling the small language model to adapt as operating conditions change throughout the life cycle of the fire protection device. Thus, compared with previous methods, technical support time and effort can be reduced.

[0013] In the following detailed description, reference is made to the accompanying drawings that form a part of the detailed description. The drawings illustrate, by way of example, one or more embodiments in which the present disclosure may be practiced.

[0014] These embodiments are described in sufficient detail to enable one of ordinary skill in the art to practice one or more embodiments of the present disclosure. It should be understood that other embodiments may be utilized and process, electrical, and / or structural changes may be made without departing from the scope of the present disclosure.

[0015] It should be appreciated that elements shown in various embodiments herein may be added, exchanged, combined, and / or eliminated to provide several additional embodiments of the present disclosure. The proportions and relative dimensions of the elements provided in the drawings are intended to illustrate embodiments of the present disclosure and should not be limiting.

[0016] The accompanying drawings herein follow the following numbering convention: one or more first digits correspond to the drawing number, and the remaining digits identify an element or component in the drawing. Similar elements or components between different drawings may be identified by using similar numbers. For example, 102 may refer to element “02” in Figure 1 and a similar element may be referred to as 202 in Figure 2

[0017] As used herein, “a,” “an,” or “several” things may refer to one or more such things, and “multiple” things may refer to more than one such thing. For example, “several components” may refer to one or more components, and “multiple components” may refer to more than one component.

[0018] Figure 1This is an example of a system 100 for receiving operational information from fire-fighting equipment, according to one or more embodiments of this disclosure. System 100 may include fire-fighting equipment 104, remote fire-fighting equipment 105, event devices 106-1, 106-2, 106-N (collectively referred to herein as event device 106), remote event device 107, and remote server 110.

[0019] As described above, system 100 may be included in a facility, the space of the facility, etc. System 100 may include devices / families of devices for detecting events and / or processing and / or analyzing the detected events to determine whether to generate an alarm for the occupants of the facility.

[0020] For example, system 100 may include event device 106. Event device 106 may be a device that detects events and transmits the detected events for processing and / or analysis. As described above, event device 106 may include, for example, cameras, motion sensors, fire protection equipment including fire sensors, smoke detectors, heat detectors, carbon monoxide (CO) detectors, or combinations thereof; fire control panels; air quality sensors; interfaces; manual alarm points (MCPs); pull-out fire alarm boxes; input / output modules; aspirating devices; fire doors; and / or auditory / visual devices (e.g., speakers, emitters, flashers, buzzers, microphones, cameras, video displays, video screens, etc.), relay output modules, and other types of event devices. Additionally, event device 106 may also include self-test capabilities.

[0021] System 100 may further include fire protection equipment 104. In some examples, fire protection equipment 104 may be a fire control panel. Fire protection equipment 104 may be used to control various event devices 106 included in system 100.

[0022] Firefighting equipment 104 can be connected to event equipment 106, transmit multiple commands to event equipment 106, and / or provide power to event equipment 106. Firefighting equipment 104 can apply voltage to the event equipment loop to power event equipment 106 in the event equipment loop. This power supply can allow event equipment 106 to perform actions such as communication between event equipment 106 and firefighting equipment 104, self-test procedures and / or providing audible and / or visual warnings during an event, as well as other actions.

[0023] Firefighting equipment 104 can be further connected to remote server 110 via gateway 108. Gateway 108 can be a device (e.g., a building system gateway) that provides a communication link between firefighting equipment 104 and other devices, such as remote server 110. For example, gateway 108 can be able to transmit data (e.g., system device data, activation signals, etc.) from firefighting equipment 104 to remote server 110 and vice versa. Communication between firefighting equipment 104 and remote server 110 is further described herein.

[0024] As described above, a user may wish to determine information regarding the operation of fire protection equipment 104. As further described herein, a user may input an input query to fire protection equipment 104 to determine its operational information. As used herein, a query refers to a question or command that can be received by a computing device, and the computing device can execute instructions to perform computational tasks related to the question or command. Users may be, for example, customers, engineers, technical support personnel, building operators, and / or any other type of user.

[0025] For example, a user can provide an input query to fire protection device 104 in response to a fault occurring in event device 106-1 associated with fire protection device 104. The input query can include natural language. As used herein, the term "natural language" refers to a structured communication system with grammar and vocabulary found in human communities. For example, a user could provide an input query stating "Diagnose a fault occurring with Event Device 106-1".

[0026] Although the above description of input queries as being provided in response to a failure of event device 106-1 is not limited to this, the implementation is not limited to this. For example, input queries may be provided to generate reports, predict future failures, provide steps for debugging the device, and / or input queries may be provided for devices other than event device 106-1 (e.g., event device 106-2, fire equipment 104, remote event device, etc.).

[0027] In some examples, this can be achieved via a user interface on the display (e.g., for clarity and to avoid confusing embodiments of this disclosure). Figure 1Input queries can be provided directly to fire equipment 104 via input devices associated with fire equipment 104 (not illustrated in the text). The user interface can provide (e.g., display and / or present) information to the user of fire equipment 104 and / or receive information from the user of fire equipment 104 (e.g., input by the user). In some embodiments, the user interface can be a graphical user interface (GUI) that can provide information to and / or receive information from the user of fire equipment 104. The display can be, for example, a touchscreen (e.g., the GUI may include touchscreen functionality). Alternatively, the display can include a television, computer monitor, mobile device screen, other types of display devices, or any combination thereof, connected to fire equipment 104 and configured to receive video signals output from fire equipment 104.

[0028] In some examples, it can be a mobile device connected to the fire equipment 104 via a wired and / or wireless network relationship (e.g., for clarity and to avoid confusing the embodiments of this disclosure). Figure 1 (Not illustrated) to provide input queries. As used herein, mobile devices can include devices that a user carries and / or wears (or can be carried and / or worn by a user). Mobile devices can be telephones (e.g., smartphones), tablets, personal digital assistants (PDAs), smart glasses and / or wrist-worn devices (e.g., smartwatches), and other types of mobile devices. Examples of such network relationships can include local area networks (LANs), wide area networks (WANs), personal area networks (PANs), distributed computing environments (e.g., cloud computing environments), storage area networks (SANs), metropolitan area networks (MANs), cellular communication networks, Long Term Evolution (LTE), Visible Light Communication (VLC), Bluetooth, Global Microwave Access Interoperability (WiMAX), Near Field Communication (NFC), Infrared (IR) communication, Public Switched Telephone Network (PSTN), radio waves and / or the Internet, and other types of network relationships.

[0029] Therefore, fire protection equipment 104 can receive input queries for operational information of fire protection equipment 104. As used herein, operational information of fire protection equipment refers to data describing actions performed by computing devices to complete a given task. As further described herein, operational information may include, for example, operational parameters of fire protection equipment 104 (e.g., memory consumption, temperature data of central processing unit (CPU), network packet transmission information (e.g., latency, bandwidth, etc.), operational parameters of connected event devices 106 (e.g., connection status, alarm status, device health status, etc.), whether there are any faults in fire protection equipment 104 and / or associated event devices 106, installation and / or commissioning procedures, and other types of operational information.

[0030] To process input queries, fire equipment 104 can provide the input query to a small language model 102, which is associated with and trained to provide information about the fire equipment 104. As used herein, the term small language model refers to a lightweight generative artificial intelligence (AI) model configured to process input to generate output. Small language model 102 can be a machine learning model capable of generating text, images, videos, or other data in response to cues such as input queries, utilizing generative models. Small language model 102 can be trained using local fire equipment operation data from fire equipment 104, such as combining... Figure 2 As further described.

[0031] The small language model 102 can be an artificial neural network (ANN). An artificial neural network (ANN) is a network that processes information by modeling a network of neurons. A network of neurons can be modeled in this way to process information. For example, an ANN can include a multi-neuron topology, which can be referred to as an artificial neuron or unit. An ANN operation refers to the operation of using units to process input to perform a given task.

[0032] ANN operations can involve applying various machine learning algorithms to process the input. For example, an ANN can perform machine learning tasks by performing a weighted combination of inputs (from network inputs or previous layers) at each unit to generate an output. Probabilistic weight associations can be provided by the multiple units that make up the ANN. Units, along with weights, biases, embeddings, and / or activation functions, can be used to generate the ANN's output based on its inputs. ANN units can be grouped to form layers of the ANN. An ANN can implement or represent an algorithm consisting of a series of connected layers that process signals based on the outputs from other connected layers in the series.

[0033] Small language model 102 can be a small language model based on the size of the neural network of the model, the number of parameters used by the small language model to make decisions, and the amount of data on which the small language model 102 is trained. For example, compared to large language models that can include more than 15,000,000 parameters (e.g., up to hundreds of billions of parameters) and can run on multiple graphics processing units (GPUs), small language model 102 can include fewer than 15,000,000 parameters and can run on a CPU such as a 1000 MHz CPU. Examples of small language models can include DistilBERT, Orca 2, Phi 2, BERT Mini / Small / Medium / Tiny, GPT-Neo, GPT-J, and other examples of small language models. Small language model 102 can be stored locally in the memory of fire equipment 104 and can be a multimodal small language model.

[0034] Firefighting equipment 104 can provide input queries to a small language model 102. Firefighting equipment 104 can use the small language model 102 to generate a response to the input query that includes operational information. For example, the small language model 102 can process the input query by applying activation functions to the inputs of each layer within the small language model 102. The activation functions can transform the inputs to each layer into outputs that can be passed to successive layers until the small language model 102 can finally generate a response to the input query that includes operational information.

[0035] For example, the input query could be "Diagnose a fault occurring with Event Device 106-1", and the small language model 102 could process this input query to generate a response. The response to the input query could also include natural language. For example, the response to the input query could be "The fault in Event Device 106-1 is occurring because of dust buildup on the device". Therefore, using the trained small language model 102, the fire-fighting equipment 104 can receive input queries and provide responses to the user of the fire-fighting equipment 104.

[0036] Although the input queries described above are directed to event devices 106-1 connected to fire equipment 104, the implementation is not limited to this. Since event devices 106 are connected to fire equipment 104, they can be considered local to fire equipment 104. However, in some cases, fire equipment 104 may receive input queries related to fire equipment or event devices remote from the local group of fire equipment 104 and event devices 106. Additionally, in some examples, fire equipment 104 may utilize information associated with devices remote from the local group of fire equipment 104 and event devices 106 to generate a response to the query. Such examples are further described herein.

[0037] like Figure 1 As shown, fire protection device 104 can be connected to remote server 110 via gateway 108. Fire protection device 104 can communicate with remote server 110, as further described herein, where fire protection device 104 can utilize remote fire protection device operation data to generate responses to inquiries about fire protection device 104 / event device 106 or about remote fire protection device / remote event device.

[0038] In some cases, fire protection device 104 may receive input queries regarding operational information about devices different from fire protection device 104 and / or event device 106. For example, fire protection device 104 may receive an input query about remote event device 107. The input query might be "What is the alarm sensitivity setting for remote event device 107?". Fire protection device 104 may determine that the input query is not associated with fire protection device 104 or event device 106. Therefore, fire protection device 104 may, in response to the input query being about a device different from fire protection device 104, transmit the input query to a remote server 110 with a large language model 112.

[0039] Large language model 112 can also be an ANN that operates similarly to small language model 102. However, large language model 112 can be a deep learning machine learning algorithm trained on a larger dataset than small language model 102 and utilizing far more parameters than small language model 102. For example, large language model 112 can utilize billions of parameters to process input queries.

[0040] Remote server 110 can generate responses to input queries regarding remote event device 107. For example, remote server 110 can use a large language model 112 trained with remote fire equipment operation data to generate responses to input queries regarding operation information of devices different from fire equipment 104 and / or event device 106. For example, the response to the input query may include sensitivity settings for natural language for remote event device 107. The response can be transmitted from remote server 110 to fire equipment 104 via gateway 108.

[0041] As an additional example, fire device 104 can utilize remote fire device operation data to generate a response to inquiries about fire device 104. For example, fire device 104 can receive updated source code that may include defects / vulnerabilities. For example, the release version of the source code including defects / vulnerabilities could be version 12, but other fire devices (e.g., remote fire device 105) could be on version 11. Even if the user may not know the version number of remote fire device 105, the user can still enter, and fire device 104 can also receive, an input query stating “What is the cause of this bug in the fire device 104?”

[0042] In this example, fire-fighting equipment 104 can provide an input query to a small language model 102 and transmit that input query to a remote server 110. While the small language model 102 can be able to process the input query and generate a response indicating the cause of the vulnerability (e.g., a defect in a source code update), the large language model 112 can also process the input query and generate a response indicating that the remote fire-fighting equipment 105 is on version 11 of the source code and has not experienced any defects.

[0043] As described above, the small language model 102 and the large language model 112 can be trained using local fire equipment operation data and remote fire equipment operation data, respectively. Combined Figure 2 The training data used for the small language model 102 and the large language model 112 are further described.

[0044] Figure 2 Examples of training a small language model 202 and a large language model 212 according to one or more embodiments of this disclosure are illustrated. The small language model 202 may be included in fire-fighting equipment, while the large language model 212 may be included in a remote server, such as in combination with... Figure 1 As previously described.

[0045] Such as combination Figure 1As previously described, the small language model 202 can be a lightweight generative AI model trained to provide information about fire equipment. As further described herein, the small language model 202 can be trained using local fire equipment operation data 214 to generate responses to input queries.

[0046] Local fire equipment operation data 214 may include data describing actions performed by fire equipment in the local group and / or event devices connected to that fire equipment. For example, local fire equipment operation data 214 may include configuration file information 216, event logs 218, diagnostic data 220 and / or source code information 222, and other types of local fire equipment operation data 214, as further described herein.

[0047] For example, configuration file information 216 may include settings that specify the operation of fire protection equipment and / or event equipment. For example, configuration file information 216 may include settings such as which output to activate based on the detector type (e.g., a flashlight, a sounder, etc. for smoke, fire, CO2, or other types of event equipment) or other cause and effect logic settings.

[0048] Event log 218 may include records of abnormal situations that occur. For example, event log 218 may include records of events such as device malfunction or actual alarm events.

[0049] Diagnostic data 220 may include data used to investigate and / or diagnose computational problems, such as performance, errors, and error outputs. For example, diagnostic data 220 may include memory consumption, CPU temperature data, and network traffic levels.

[0050] Source code information 222 may include information describing programming statements that instruct the device to perform a specific task. For example, source code information 222 may include version information, change logs, language type, and deviations from the expected instructions for the device's operation (e.g., an alarm activation output is assumed to occur after 30 seconds but actually occurs after 2 minutes).

[0051] like Figure 2 As shown, local fire equipment operation data 214 can be provided to the small language model 202. The small language model 202 can be trained using the local fire equipment operation data 214.

[0052] Additionally, as the fire-fighting equipment is operated, it can record supplementary local fire-fighting equipment operation data. This supplementary local fire-fighting equipment operation data may include supplementary configuration file information 216, event logs 218, diagnostic data 220, source code information 222, and user-generated feedback regarding the operation of the small language model 202. This supplementary local fire-fighting equipment operation data can be used to continuously retrain the small language model 202.

[0053] Similarly, a large language model 212 can be trained using remote fire equipment operation data 224 to generate a response to an input query, as further described herein.

[0054] Remote fire protection equipment operation data 224 may include data describing actions performed by fire protection equipment and / or event devices remotely located by event devices connected to fire protection equipment in the local group and / or the local group. For example, remote fire protection equipment operation data 224 may include remote site data 226, remote event logs 228, and / or remote site configuration data 230, as further described herein.

[0055] For example, remote site data 226 may include configuration file information, diagnostic data, and source code information of devices remotely originating from local fire protection / event equipment. Additionally, remote event logs 228 may include records of anomalies occurring in these remote devices. Furthermore, remote site configuration data 230 may include control panel-level data and site-level data associated with devices remotely originating from local fire protection / event equipment.

[0056] like Figure 2 As shown, remote fire-fighting equipment operation data 224 can be provided to the large language model 212. The large language model 212 can be trained using the remote fire-fighting equipment operation data 224.

[0057] Figure 3 An example of providing an input query 332 to a small language model 302 and / or a large language model 312 according to one or more embodiments of this disclosure is illustrated. The fire-fighting equipment 304 may receive the input query 332, and the small language model 302 and / or the large language model 312 at a remote server 310 may generate a response 334, as further described herein.

[0058] For example, input query 332 may include a request to generate a report detailing the operating parameters of the fire-fighting equipment. Fire-fighting equipment 304 can determine that input query 332 is about fire-fighting equipment 304. Therefore, fire-fighting equipment 304 can provide input query 332 to a small language model 302. Small language model 302 can analyze local fire-fighting equipment operating data and determine various operating parameters of fire-fighting equipment 304. For example, various operating parameters may include memory consumption, CPU temperature data, network packet transmission information, operating parameters of connected event devices, and the presence of any faults in the fire-fighting equipment and / or associated event devices. Response 334 may include a report detailing the values ​​of the operating parameters requested in input query 332.

[0059] As another example, input query 332 may include a request to diagnose a fault in an event device. Fire equipment 304 can determine whether the request is for an event device associated with the fire equipment or for an event device remote from the fire equipment (e.g., not in the local group). In response to a request for an event device associated with the fire equipment (e.g., in the local group), a small language model 302 can determine the cause of the fault in the local event device. In response to a request for a different event device (e.g., a remote event device), fire equipment 304 can transmit input query 332 to a remote server 310, and a large language model 312 can determine the cause of the fault in the remote event device. A response 334 can be transmitted from the remote server 310 to fire equipment 304, and fire equipment 304 can provide a response 334 including the cause of the fault.

[0060] As another example, input query 332 may include a request to predict future failures in an event device. Fire equipment 304 may determine whether the request is for an event device associated with the fire equipment or for an event device remote from the fire equipment (e.g., not in the local group). In response to a request for an event device associated with the fire equipment (e.g., in the local group), a small language model 302 may predict future failures in the local event device (e.g., including the type of failure, what could cause the future failure, a timeline of when the future failure might occur, how to prevent the future failure, and / or how to remedy the future failure if / when it occurs). In response to a request for a different event device (e.g., a remote event device), fire equipment 304 may transmit input query 332 to a remote server 310, and a large language model 312 may predict future failures in the remote event device. A response 334 may be transmitted from the remote server 310 to fire equipment 304, and fire equipment 304 may provide a response 334 including the predicted future failures.

[0061] As another example, input query 332 may include a request for steps to debug fire equipment 304. Small language model 302 can analyze local fire equipment operation data and generate a response 334 that includes steps for debugging fire equipment 304.

[0062] Therefore, the operational information of the fire-fighting equipment disclosed herein enables the provision of input queries to the fire-fighting equipment, and the fire-fighting equipment can utilize a trained small language model to generate responses to the input queries. Compared to large language models, small language models can more easily run locally on the fire-fighting equipment by utilizing the available computing resources at the fire-fighting equipment. Therefore, users can more easily troubleshoot technical problems with the fire-fighting equipment. Furthermore, the fire-fighting equipment can communicate with a remote server possessing a large language model, which can utilize the connected system to achieve a more robust response to input queries. Therefore, compared to previous methods, technical support time and effort can be reduced.

[0063] Figure 4 This is an example of fire protection equipment 404 for receiving operational information from fire protection equipment, according to one or more embodiments of this disclosure. Figure 4 As shown, the fire-fighting equipment 404 may include a memory 442 and a processor 440, the processor being used to process operational information from the fire-fighting equipment according to the present disclosure.

[0064] The memory 442 can be any type of storage medium that can be accessed by the processor 440 to execute various examples of the present disclosure. For example, the memory 442 can be a non-transitory computer-readable medium on which computer-readable instructions (e.g., executable instructions / computer program instructions) are stored, which can be executed by the processor 440 for use with operational information from fire-fighting equipment according to the present disclosure.

[0065] Memory 442 may be volatile or non-volatile memory. Memory 442 may also be removable (e.g., portable) memory or non-removable (e.g., internal) memory. For example, memory 442 may be random access memory (RAM) (e.g., dynamic random access memory (DRAM) and / or phase-change random access memory (PCRAM)), read-only memory (ROM) (e.g., electrically erasable programmable read-only memory (EEPROM) and / or optical disc read-only memory (CD-ROM)), flash memory, laser disc, digital versatile disc (DVD) or other optical storage devices and / or magnetic media (such as cassette tape, magnetic tape, or disk) and other types of memory.

[0066] Furthermore, although memory 442 is illustrated as being located within fire-fighting equipment 404, embodiments of this disclosure are not limited thereto. For example, memory 442 may also be located within another computing resource (e.g., enabling computer-readable instructions to be downloaded via the Internet or another wired or wireless connection).

[0067] Processor 440 may be a central processing unit (CPU), a semiconductor-based microprocessor, and / or other hardware device suitable for retrieving and executing machine-readable instructions stored in memory 442.

[0068] Although specific embodiments have been illustrated and described herein, those skilled in the art will understand that any arrangement calculated to achieve the same technology may replace the specific embodiments shown. This disclosure is intended to cover any and all modifications or variations of the various embodiments of this disclosure.

[0069] It should be understood that the above description is given by way of illustration and not limitation. Combinations of the above embodiments, as well as other embodiments not specifically described herein, will be apparent to those skilled in the art upon reading the above description.

[0070] The scope of the various embodiments of this disclosure includes any other application using the structures and methods described above. Therefore, the scope of the various embodiments of this disclosure should be determined with reference to the appended claims and the full scope of their equivalents.

[0071] In the foregoing detailed description, various features are combined in the exemplary embodiments illustrated in the accompanying drawings for the purpose of simplifying this disclosure. This approach should not be construed as reflecting an intention to require more features than expressly recited in each claim.

[0072] Conversely, as reflected in the following claims, the subject matter of the invention lies in fewer than all the features of a single disclosed embodiment. Therefore, the following claims are hereby incorporated into the detailed description, wherein each claim exists independently as a separate embodiment.

Claims

1. A fire-fighting device (104, 304, 404), said fire-fighting device comprising: Memory (442); and Processor (440), the processor being configured to execute executable instructions stored in the memory to: Receive input queries for operation information of the fire-fighting equipment (332); The input query (332) is provided to a small language model (102, 202, 302), which is associated with and trained to provide information about the fire-fighting equipment (104, 304, 404); as well as The small language models (102, 202, 302) are used to generate a response (334) to the input query (332) that includes the operation information.

2. The fire-fighting equipment (104, 304, 404) according to claim 1, wherein, The small language model (102, 202, 302) is trained using local fire equipment operation data to generate the response (334) to the input query.

3. The fire-fighting equipment (104, 304, 404) according to claim 2, wherein, The local fire-fighting equipment operation data includes at least one of the following: Configuration file information (216); The event log of the fire-fighting equipment (218); The diagnostic data (220) of the fire-fighting equipment; and The source code information (222) of the fire-fighting equipment.

4. The fire-fighting equipment (104, 304, 404) according to claim 1, wherein, The input query (332) includes natural language.

5. The fire-fighting equipment (104, 304, 404) according to claim 1, wherein, The response (334) to the input query (332) includes natural language.

6. The fire-fighting equipment (104, 304, 404) according to claim 1, wherein, The small language models (102, 202, 302) are multimodal small language models.

7. The fire-fighting equipment (104, 304, 404) according to claim 1, wherein, The small language models (102, 202, 302) are locally stored in the memory (442) of the fire-fighting equipment.

8. The fire-fighting equipment (104, 304, 404) according to claim 1, wherein, The fire-fighting equipment (104, 304, 404) is a fire control panel.

9. The fire-fighting equipment (104, 304, 404) according to claim 1, wherein: The input query (332) includes a request to generate a report detailing the operating parameters of the fire-fighting equipment; The small language models (102, 202, 302) are configured to analyze local fire equipment operation data; and The response to the input query includes a report detailing the operating parameters of the fire-fighting equipment.

10. The fire-fighting equipment (104, 304, 404) according to claim 1, wherein: The input query (332) includes a request for steps for commissioning the fire-fighting equipment; The small language model is configured to analyze local fire equipment operation data; and The response to the input query includes the steps for debugging the fire-fighting equipment.

Images