System and method for determining travel intent of elevator passengers and providing responsive alerts

Abstract

An elevator system having an elevator car; at least one sensor, where the system is configured to: receive sensor data from the at least one sensor indicative of a first passenger calling the elevator car or being within the elevator car; determining a travel intention of the first passenger from the at least one sensor data; and taking an action when the travel intent triggers an alarm condition.

Description

Technical Field

[0001] Embodiments relate to elevator systems, and more particularly, to systems and methods for determining an elevator passenger's travel intent and providing a responsive alert. Background Art

[0002] It is possible that people with suspicious motives use elevators more frequently than other users. For example, a thief may use an elevator to steal from a passenger traveling alone. Over a period of time, although a thief may frequently ride an elevator, they may appear to have no purpose for riding the elevator. It is desirable to control elevators to minimize the associated risk to passengers. Summary of the invention

[0003] Disclosed is an elevator system comprising an elevator car; at least one sensor, wherein the system is configured to: receive sensor data from the at least one sensor indicating that a first passenger has called the elevator car or is within the elevator car; determine a travel intention of the first passenger based on the at least one sensor data; and take action when the travel intention triggers an alarm condition.

[0004] In addition to or as an alternative to one or more aspects of the system, in order to take the action, the system is configured to issue an alert or assign the first passenger to ride alone or with multiple people.

[0005] In addition to or as an alternative to one or more aspects of the system, to determine the travel intention of the first passenger, the system is configured to: process a time series model of the history of the first passenger's elevator calls and travel intentions as training data through one or more artificial intelligence (AI) models to thereafter identify the first passenger and the travel intention of the first passenger.

[0006] In addition to or as an alternative to one or more aspects of the system, the one or more AI models include: a user recognition model; and an intent recognition model.

[0007] In addition to or as an alternative to one or more aspects of the system, when the alarm condition is triggered, the system is configured to perform supervised learning by: soliciting feedback to confirm the accuracy of classifying the travel intention as abnormal; and applying the feedback to the one or more artificial intelligence (AI) models.

[0008] Additionally or alternatively to one or more aspects of the system, the time series model identifies, for a plurality of passengers including the first passenger, unique passenger identifiers and a history of elevator calls and travel intentions for each of the plurality of passengers.

[0009] In addition to one or more aspects of the system, or as an alternative, the system is configured to determine the travel intention of the first passenger for a call by feeding call data to the one or more artificial intelligence (AI) models, wherein, for the first passenger, the call data includes one or more of: call time; call floor; requested destination floor; observed destination floor; clothing worn; equipment carried; and the number of people traveling with the first passenger.

[0010] In addition to one or more aspects of the system, or as an alternative, when determining whether the travel intention of the first passenger triggers the alarm condition, the system is configured to: compare the travel intention of the first passenger with the time series model to determine whether to classify the travel intention as an abnormality of the first passenger; and determine whether the abnormality of the first passenger triggers the alarm condition.

[0011] Additionally or alternatively to one or more aspects of the system, the system is configured to trigger the alarm condition upon determining that the total number of anomalies for the first passenger exceeds a threshold.

[0012] In addition to one or more aspects of the system, or as an alternative, when the alarm condition is triggered, the system is configured to issue one or more of the following: a user isolation flag, so that the system will thereafter only serve the first passenger when no other passengers are served; a multi-person flag, so that the system will thereafter only serve the first passenger when at least two other passengers are already on the elevator car; or a denial of service flag, so that the system will thereafter not serve the first passenger.

[0013] Disclosed is a method of operating an elevator system, comprising: receiving sensor data from at least one sensor, wherein the sensor data indicates that a first passenger is calling the elevator car or is within the elevator car; determining a travel intention of the first passenger based on the sensor data; and taking action when the travel intention triggers an alarm condition.

[0014] In addition to or as an alternative to one or more aspects of the method, when the action is taken, the method includes issuing an alert or assigning the first passenger to ride alone or with multiple people.

[0015] In addition to or as an alternative to one or more aspects of the method, when determining the travel intention of the first passenger, the method includes: processing, by the system through one or more artificial intelligence (AI) models, a time series model of the history of the first passenger's elevator calls and travel intentions as training data to thereafter identify the first passenger and the travel intention of the first passenger.

[0016] In addition to or as an alternative to one or more aspects of the method, the one or more AI models include: a user recognition model; and an intent recognition model.

[0017] In addition to or as an alternative to one or more aspects of the method, when the alarm condition is triggered, the method includes performing supervised learning by the system by: soliciting feedback to confirm the accuracy of classifying the travel intention as abnormal; and applying the feedback to the one or more artificial intelligence (AI) models.

[0018] Additionally or alternatively to one or more aspects of the method, the time series model identifies, for a plurality of passengers including the first passenger, unique passenger identifiers and a history of elevator calls and travel intentions for each of the plurality of passengers.

[0019] In addition to one or more aspects of the method, or as an alternative, the system determines the travel intention of the first passenger for a call by feeding call data to the one or more artificial intelligence (AI) models, wherein, for the first passenger, the call data includes one or more of the following: call time; call floor; requested destination floor; observed destination floor; clothing worn; equipment carried; and the number of people traveling with the first passenger.

[0020] In addition to one or more aspects of the method, or as an alternative, when determining whether the travel intention of the first passenger triggers the alarm condition, the method includes: the system comparing the travel intention of the first passenger with the time series model to determine whether to classify the travel intention as an abnormality of the first passenger; and determining whether the abnormality of the first passenger triggers the alarm condition.

[0021] Additionally or alternatively to one or more aspects of the method, the alarm condition is triggered by the system when it is determined that a total number of anomalies for the first passenger exceeds a threshold.

[0022] In addition to one or more aspects of the method, or as an alternative, when the alarm condition is triggered, the method includes the system issuing one or more of the following: a user isolation flag, so that the system will thereafter only serve the first passenger when no other passengers are served; a multi-person flag, so that the system will thereafter only serve the first passenger when at least two other passengers are already on the elevator car; or a refusal of service flag, so that the system will thereafter not serve the first passenger. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The present disclosure is illustrated by way of examples and not limitation in the accompanying figures, in which like references indicate similar elements.

[0024] Figure 1 is a schematic illustration of an elevator system in which various embodiments of the present disclosure may be employed;

[0025] Figure 2 Additional details of a system for detecting passenger intent and providing an alert are shown;

[0026] Figure 3 is a flow chart illustrating a method of detecting a passenger's travel intention;

[0027] Figure 4 is a flow chart showing additional details of determining a passenger's travel intention; and

[0028] Figure 5 is a flow chart showing additional details of sounding an alarm when a passenger's travel intention triggers an alarm condition. DETAILED DESCRIPTION

[0029] A detailed description of one or more embodiments of the disclosed apparatus and methods is presented herein by way of example and not limitation, with reference to the accompanying drawings.

[0030] Figure 1 1 is a perspective view of an elevator system 101 that includes an elevator car 103, a counterweight 105, a tension member 107, a guide rail (or track system) 109, a machine (or machine system) 111, an orientation reference system 113, and an electronic elevator controller (controller) 115. The elevator car 103 and the counterweight 105 are connected to each other by the tension member 107. The tension member 107 may include or be configured as, for example, a rope, a steel cable, and / or a coated steel belt. The counterweight 105 is configured to balance the load of the elevator car 103 and to facilitate movement of the elevator car 103 relative to the counterweight 105 in an elevator hoistway (or hoistway) 117 and along the guide rail 109 simultaneously and in opposite directions.

[0031] The tension member 107 engages a machine 111, which is part of the overhead structure of the elevator system 101. The machine 111 is configured to control movement between the elevator car 103 and the counterweight 105. The position reference system 113 can be mounted on a fixed portion of the top of the elevator shaft 117, such as on a support rail or a guide rail, and can be configured to provide a position signal related to the position of the elevator car 103 in the elevator shaft 117. In other embodiments, the position reference system 113 can be directly mounted to the moving components of the machine 111, or can be positioned in other positions and / or configurations as known in the art. As known in the art, the position reference system 113 can be any device or mechanism for monitoring the position of the elevator car and / or counterweight. For example and not limitation, as will be appreciated by those skilled in the art, the position reference system 113 can be an encoder, a sensor or other system, and can include speed sensing, absolute position sensing, etc.

[0032] The controller 115 is positioned in a controller room 121 of the elevator shaft 117 as shown, and is configured to control the operation of the elevator system 101 and in particular the operation of the elevator car 103. For example, the controller 115 can provide a drive signal to the machine 111 to control the acceleration, deceleration, leveling, stopping, etc. of the elevator car 103. The controller 115 can also be configured to receive a position signal from the position reference system 113 or any other desired position reference device. When moving up or down in the elevator shaft 117 along the guide rail 109, the elevator car 103 can stop at one or more landings 125 controlled by the controller 115. Although shown in the controller room 121, those skilled in the art will appreciate that the controller 115 can be positioned and / or configured in other positions or positions within the elevator system 101. In one embodiment, the controller can be positioned remotely or in the cloud.

[0033] The machine 111 may include a motor or similar drive mechanism. According to an embodiment of the present disclosure, the machine 111 is configured to include an electric drive motor. The power supply for the motor may be any power source, including an electrical grid, which is supplied to the motor in combination with other components. The machine 111 may include a traction sheave, which transmits force to the tension member 107 to move the elevator car 103 within the elevator shaft 117.

[0034] Although shown and described by a roping system including tension members 107, elevator systems that employ other methods and mechanisms for moving an elevator car within an elevator hoistway may employ embodiments of the present disclosure. For example, embodiments may be employed in ropeless elevator systems that use linear motors to transfer motion to an elevator car. Embodiments may also be employed in ropeless elevator systems that use hydraulic lifts to transfer motion to an elevator car. Embodiments may also be employed in ropeless elevator systems that use self-propelled elevator cars (e.g., elevator cars equipped with friction wheels, pinch wheels, or traction wheels). Figure 1 These are non-limiting examples presented for purposes of illustration and explanation only.

[0035] In other embodiments, the system includes a transport system that moves passengers between floors and / or along a single floor. Such a transport system may include escalators, moving walkways, etc. Therefore, the embodiments described herein are not limited to systems such as Figure 1 An elevator system such as the elevator system shown in .

[0036] Go to Figure 2 , discloses additional details of a system 101 for detecting passenger intent and providing an alert. The system 101 includes an elevator car 103 and a controller 115A, which can be used with Figure 1 115A and is operatively coupled to the elevator car 103. A sensor (generally referenced as 230) is operatively coupled to the controller 115A and is configured to transmit sensor data 240 to the controller 115A. The sensor 230 may include imaging technology, non-limiting examples of which include one or more cameras 230A, 230B located in one or more of the elevator lobby 125A and within the elevator car 103. Other imaging technologies may be used for the sensor 230. Alternatively, the sensor 230 may be configured to detect an electronic device (e.g., a mobile phone) of a person as a proxy for the person. The sensor 230 may communicate with the controller 115A via a wired protocol or a wireless protocol through a wireless network 290. The sensor 230 and the controller 115A may communicate via a cloud service 295. The processing described below may be performed via edge computing on either the sensor 230 or the controller 115A or on the cloud 295. Portions of the processing may be performed separately, and the resulting data may be stitched together, such as by the controller 115A, for use in presenting the determinations identified below. For example, the controller 115A may have a non-transitory memory 250 storing a program 200 and a processor 260 executing the program, the program 200 being configured to perform the processing described below.

[0037] As discussed in more detail below, in the disclosed embodiment, each passenger (generally referenced as 210), including a first passenger 210A in the elevator car 103 and a second passenger 210B in the lobby 125A, is tagged when they use the elevator car 103. Tagging can be performed using a combination of car and lobby cameras 230 or other sensor tools as described above. An artificial intelligence (AI) program 200 executing a deep learning algorithm analyzes travel intentions through information collected from video footage, for example, when traveling to a gym, office, etc., the passenger 210 may be using the elevator car 103. The identified travel intentions are also characterized or tagged for the passenger 210. If there appears to be no purpose to use the elevator car 103 or the purpose is not detected for the passenger 210, an exception occurs, as further explained below.

[0038] When a threshold number of anomalies occur for passenger 210, system 101 determines that an alarm condition has occurred. Future elevator trips for passenger 210 will be assigned so that the number of people in elevator car 103 will be single or multiple. For example, a person with a larger number of anomalies may be assigned to ride alone compared to a person with a smaller number of anomalies who may be assigned to ride with multiple people. Alternatively, the type of anomaly can control the determination of whether a person should be assigned to ride alone or ride with multiple people. That is, the perceived risk level associated with the person due to the frequency and type of anomaly observed can control the decision about the type of exposure that the person may have to other people in elevator car 103. It is to be appreciated that, depending on the detected anomaly, alone or in conjunction with other factors such as time of day, the system can determine not to serve passenger 210 at that time, for a period of time, or until otherwise instructed by elevator service center 280. If future rides for passenger 210 are assigned as single, only passenger 210 may ride with the passenger in elevator car 103. If the future ride of the passenger 210 is designated as multiple, then before the passenger 210 can ride in the elevator car 103, multiple people (eg, 3 people) should be in the car.

[0039] The alarm condition is also optionally reported to the building manager station 280, which can be stored on the server 270 and viewed in real time, or played back at a later time so that necessary actions can be taken. The program 200 can receive feedback from the building manager station 280 that identifies unusual false detections to improve itself, for example, applying AI supervised learning techniques.

[0040] Go to Figure 3, the flowchart shows details of the method for detecting the travel intention of the first passenger 210A. As shown in box 110, the system 101 is configured to receive sensor data 240 from the sensor 230, which indicates that the first passenger 210A is calling from the lobby 125A or in the elevator car 103. As shown in box 120, the system 101 is configured to determine the travel intention of the first passenger 210A based on the sensor data 240. As shown in box 130, the system 101 is configured to take action when the travel intention triggers an alarm condition. This can include, for example, issuing an alarm to the dispatcher (box 132), or assigning the first passenger to ride alone or with multiple people (box 134).

[0041] Go to Figure 4 , the flowchart shows additional details of determining the travel intention of the first passenger 210A (box 120). As shown in box 120A, the system 101 is configured to process the time series model for the first passenger 210A as training data by the program 200 to thereafter identify the first passenger 210A and the travel intention of the first passenger 210A. For simplicity, such a model is equivalent to the list of entries in Table 1, for example. Table 1 passenger Elevator usage time Travel Intention A 8 a.m. Office Entry B 7 a.m. jogging A 12 p.m. Lunch B 8 a.m. Return Home B 8:02 AM Leave*1 B 8:10am Re-enter*2 B 9:30am walk B 9:35am Return Home B 9:50am walk B 9:55am Return Home B 10:10 am stroll B 10:15am Return home*3 B 3 p.m. shop A 5 p.m. Office departure

[0042] The time series model identifies, for example, a unique passenger identifier assigned to the passenger 210 (eg, passenger A or passenger B), the past travel time for each past elevator call made by the passenger 210 , and the past travel intentions for each past elevator call made by the passenger 120 .

[0043] At each elevator use by passenger 210, an intent is marked. Intent may be inferred based on various factors, including the time of call, the floor of the call, the requested destination floor, the observed destination (departure) floor, the clothing worn, the equipment carried, and the number of people traveling with passenger 210. For example, if passenger 210 is wearing running clothing and is exiting elevator car 103 at main lobby 125A, it may be inferred that passenger 210 is going for a jog. It is to be appreciated that the intent field in the training data (Table 1) may be manually entered for each passenger. Thus, a history of passenger-intent-time may be established for each passenger 210.

[0044] For training purposes, anomalies in a time series model can be manually identified from information in the training data. For example, Table 1 lists the activities of two passengers 210, including a first passenger 210A (or A) and a second passenger 210B (or B). At line entry *1, the entry can initially be classified or marked as an anomaly because the second passenger 210B appears to have left the elevator car 103 within two minutes of the entry. At line entry *2, based on the rapid subsequent re-entry, it can be inferred that line entry *1 is not an anomaly. That is, the re-entry may be due to the second passenger 210B forgetting an item when jogging back. At line entry *3, the intention can be classified as an anomaly due to repeated short-term use or calls of the elevator car 103. For example, within a span of 45 minutes, that is, between 9:30 am and 10:15 am, the second passenger 210B registered six uses, including three round trips, each spanning five minutes and separated by fifteen minutes. Thus, the training data (Table 1) can be used to teach the system 101 to detect when a passenger 210 is behaving in a manner that would be classified as abnormal, which can result in an alarm condition. This is because the behavior of the second passenger 210B may indicate suspicious behavior, including, for example, following other passengers 210. The following is further discussion of how the system can be trained to identify the expected or predicted intent of a passenger and thereby identify anomalies (i.e., actions that differ from the expected or predicted intent).

[0045] The time series model can be processed by one or more artificial intelligence (AI) models. For example, time series data can be analyzed by a time series machine learning algorithm to detect anomalies. Such a model can be trained based on the expected pattern of intent with timestamps to identify any anomalies. In one embodiment, the model workflow can include preparing training data with historical data of the intent of passengers / users of the building. Each unique passenger marked will have a series of associated data. Each intent is converted from text to numbers using known machine learning (ML) techniques for transforming categorical data. The data can be manually annotated, or unsupervised learning can be used to detect data clusters to mark whether there are anomalies. The ML model can then be trained using the time series model (Table 1). The ML model can then be made operational by deploying it in a real-world scenario.

[0046] In one embodiment, the AI ​​model that processes the time series model includes a user identification model and an intent identification model. The user identification model is an ML model that can identify users of the system 101 (e.g., passenger 210) based on their behavior or characteristics. For example, the user identification model can use data from sensor 230 to identify users based on their body movements and daily activities. The intent identification model uses ML to associate statements with given intents. This technique is also called intent classification. The process involves providing the ML model with examples of text with intents, such as Table 1. The ML model is then trained based on the data (Table 1) and can operate based on the data to identify the user's intent.

[0047] For example, a single passenger's transportation list can be assembled over time, with time as a parameter, using Bayesian inference for probabilistic predictions of the passenger's expected destination. Such a system can learn that passenger A always goes to floor N in the early morning, usually goes to floor C (cafeteria) at noon, and always goes to the parking lot in the evening.

[0048] Furthermore, the transportation lists for multiple passengers can be combined over time, with time as a parameter, again using Bayesian inference. Such a system facilitates the determination of the statistical distribution of elevator usage throughout a building during a typical day as well as weekends, holidays, etc. This information can be used to pre-assign car runs (even purposefully skipping floors) for efficient parking, scheduling cars, etc.

[0049] Given the information of the traffic list, elevator optimization is achieved through techniques for solving optimization problems in real time. The traffic list information can also be used for other elevator related applications (such as daily elevator load estimation), to provide an accurate energy report for future energy saving, elevator system diagnosis based on abnormal traffic list information, modernization value proposition, etc.

[0050] More specifically, using a camera-based system, passengers can be detected. The system can be trained to detect whether a passenger is traveling alone or in a group. Group types can be further classified as children, couples, friends, colleagues, etc. The system can be trained to detect whether a passenger (and group) has any luggage. With a certain degree of confidence, a general profile that conforms to categories such as professional purposes, leisure purposes, personnel purposes, health purposes, etc. can be established. The system can be trained to profile passengers, which helps to improve how to best identify the purpose of a visit. For example, a video image of a professional visit by a sports person will be different from a professional visit by a business person. The system can then map the learned features from the video (e.g., clothing, luggage, single or group travel, etc.) to a type of building facilities to improve the confidence of the purpose of the identification visit. The training process results in identifying the intention of the passenger riding the elevator car with a certain degree of confidence.

[0051] Return to Figure 4 , as shown in box 120B, the system is configured to determine the travel intention of the first passenger 210A for the call by feeding the call data to one or more AI models. The call data includes one or more of the following: the boarding time of the first passenger A, the boarding floor of the first passenger 210A, the destination floor requested by the first passenger 210A, the observed destination (elevator) floor of the first passenger 210A, the clothing worn by the first passenger 210A, the equipment carried by the first passenger 210A, and the number of people traveling with the first passenger. If the call data is outside the expected range of the first passenger 210A, an abnormality is identified. For example, if the first passenger 210A normally travels between the first floor and the tenth floor, and then randomly repeats traveling between the second floor and the third floor, the travel intention can be classified as abnormal. If they usually wear a suit and travel during working hours, and then randomly travel in a dark suit of non-professional clothing during the late night time, the travel intention can be classified as abnormal. If they normally travel alone during working hours, and randomly travel with another person during non-working hours, the travel intention may be classified as abnormal.

[0052] As shown in block 120C, when determining whether the travel intention of the first passenger 210A triggers an alarm condition, the system 101 is configured to compare the travel intention of the first passenger 210A with the time series model ( Figure 1 ), for example, to determine whether there is an anomaly. As shown in block 120D, the system 101 is configured to determine whether the travel intention is classified as an anomaly if the system 101 cannot identify the travel intention based on the training data. As shown in block 120F, the system 101 is configured to trigger an alarm condition when it is determined that the total number of anomalies of the first passenger 210A is greater than a threshold.

[0053] Go to Figure 5 , the flowchart shows additional details of issuing an alarm when the first passenger's travel intention triggers an alarm condition (box 120). As shown in box 130A, when the alarm condition is triggered, the system 101 is configured to issue a user isolation flag to the service center 280, so that the system 101 will only serve the first passenger 210A when not serving other passengers 210 thereafter. Alternatively, the system 101 is configured to issue a multi-person flag, so that the system 101 will only serve the first passenger 210A when at least two other passengers 210 are already on the elevator car 103. Alternatively, the system 101 is configured to issue a denial of service flag, so that the system 101 will not serve the first passenger 210A thereafter.

[0054] As shown in block 130B, when an alarm condition is triggered, the system 101 is configured to perform supervised learning techniques. That is, the system 101 is configured to solicit feedback from the dispatcher at the service center 280 to determine whether the abnormal classification is an error, and apply the feedback to one or more artificial intelligence (AI) models, i.e., to the program 20.

[0055] In the above embodiments, sensor data may be acquired and processed separately, or acquired and processed simultaneously and stitched together, or a combination thereof, and may be processed in raw or compiled form. Sensor data may be processed on the sensor (e.g., via edge computing), by a controller identified or implied herein, on a cloud service, or by a combination of one or more of these computing systems. The sensor may transmit data via a wired or wireless transmission line, applying one or more protocols as indicated below.

[0056] Wireless connection can apply the protocol including local area network (LAN, or WLAN for wireless LAN) protocol. LAN protocol includes WiFi technology based on the standard of Section 802.11 of the Institute of Electrical and Electronics Engineers (IEEE). Other applicable protocols include low power WAN (LPWAN), which is a wireless wide area network (WAN) designed to allow remote communication at low bit rates so that the final device can operate in an extended period (several years) using battery power. Long-range WAN (LoRaWAN) is a type of LPWAN maintained by the LoRa Alliance, and is a media access control (MAC) layer protocol for transmitting management and application messages between a network server and an application server, respectively. LAN and WAN protocols can generally be regarded as TCP / IP protocols (Transmission Control Protocol / Internet Protocol), which are used to manage the connection of computer systems to the Internet. Wireless connection can also apply the protocol including private area network (PAN) protocol. PAN protocol includes, for example, Bluetooth Low Energy (BTLE), which is a wireless technology standard designed and sold by the Bluetooth Special Interest Group (SIG) for exchanging data across short distances using short wavelength radio waves. PAN protocols also include Zigbee, a technology based on the Section 802.15.4 protocol from the IEEE, which represents a set of advanced communication protocols for creating personal area networks with small, low-power digital radios for low-power, low-bandwidth requirements. Such protocols also include Z-Wave, a wireless communication protocol supported by the Z-Wave Alliance that uses mesh networking to apply low-energy radio waves to communicate between devices such as appliances, allowing wireless control of the devices.

[0057] The wireless connection may also include radio frequency identification (RFID) technology, which is used to communicate with an integrated chip (IC) on, for example, an RFID smart card. In addition, sub-1Ghz RF devices operate in the ISM (industrial, scientific and medical) spectrum band below sub-1Ghz - typically in the frequency ranges of 769-935MHz, 315Mhz and 468Mhz. This spectrum band below 1Ghz is particularly useful for RF IOT (Internet of Things) applications. The Internet of Things (IoT) describes a network of physical objects - "things" - embedded with sensors, software and other technologies, for connecting and exchanging data with other devices and systems via the Internet. Other LPWAN-IOT technologies include narrowband Internet of Things (NB-IOT) and category M1 Internet of Things (Cat M1-IOT). The wireless communication of the disclosed system may include cellular, such as 2G / 3G / 4G (etc.). Other wireless platforms based on RFID technology include near field communication (NFC), which is a set of communication protocols for low-speed communication, for example, for exchanging data between electronic devices across short distances. NFC standards are defined by the ISO / IEC (defined below), the NFC Forum and the GSMA (Global System for Mobile Communications) groups.The above is not intended to limit the scope of applicable wireless technologies.

[0058] Wired connections may include connections (cables / interfaces) under RS ​​(Recommended Standard)-422, also known as TIA / EIA-422, which is a technical standard supported by the Telecommunications Industry Association (TIA) and initiated by the Electronic Industries Alliance (EIA) that specifies the electrical characteristics of digital signaling circuits. Wired connections may also include connections (cables / interfaces) under the RS-232 standard for serial communication transmission of data, which formally defines the signals connected between DTE (data terminal equipment) such as computer terminals and DCE (data circuit-terminating equipment or data communications equipment) such as modems. Wired connections may also include connections (cables / interfaces) under the Modbus serial communication protocol managed by the Modbus organization. Modbus is a server / client protocol designed for use with its programmable logic controller (PLC), and it is a commonly available means of connecting industrial electronic devices. Wireless connections may also include connectors (cables / interfaces) under the PROFibus (Process Field Bus) standard managed by PROFIBUS & PROFINET International (PI). PROFibus is a standard for fieldbus communication in automation technology and is published as part of IEC (International Electrotechnical Commission) 61158. Wired communication is also possible via the Controller Area Network (CAN) bus. CAN is a vehicle bus standard that allows microcontrollers and devices to communicate with each other in applications without a host computer. CAN is a message-based protocol published by the International Standards Organization (ISO). The above is not intended to limit the scope of applicable wired technologies.

[0059] When data is transmitted over a network between final processors as identified herein, the data may be transmitted in raw form, or may be processed in whole or in part in any of the final processors or intermediate processors, such as in a cloud service (e.g., where at least a portion of the transmission path is wireless) or other processors. The data may be parsed, partially or completely processed or compiled in any of the processors, and then may be spliced ​​together or maintained as separate information packets. Each processor or controller identified herein may be, but is not limited to, a single processor or multi-processor system of any architecture in a variety of possible architectures, including field programmable gate arrays (FPGAs), central processing units (CPUs), application specific integrated circuits (ASICs), digital signal processors (DSPs), or graphics processing units (GPUs) hardware arranged homogeneously or heterogeneously. The memory identified herein may be, but is not limited to, random access memory (RAM), read-only memory (ROM), or other electronic, optical, magnetic, or any other computer-readable medium.

[0060] In addition to the processor and non-volatile memory, the controller may further include one or more input and / or output (I / O) device interfaces, which are coupled in communication via an onboard (local) interface to communicate between other devices. The onboard interface may include, for example, but not limited to, an onboard system bus, including a control bus (for inter-device communication), an address bus (for physical addressing), and a data bus (for transmitting data). That is, the system bus can realize electronic communication between the processor, the memory, and the I / O connection. The I / O connection may also include a wired connection and / or a wireless connection identified herein. The onboard interface may have additional elements omitted for simplicity, such as a controller, a buffer (cache), a driver, a repeater, and a receiver to realize electronic communication. The memory may execute a program, access data or a lookup table, or a combination of each item to facilitate its processing, all of which may be pre-stored or received by other computing devices during the execution of its process, for example, via a cloud service or other network connections with other processors identified herein.

[0061] Embodiments may take the form of processor-implemented processes and means (such as a processor) for practicing those processes. Embodiments may also take the form of computer code-based modules, such as computer program code (e.g., computer program products) containing instructions embodied in tangible media (e.g., non-transitory computer-readable media), such as floppy disks, CD ROMs, hard drives, processor registers (such as firmware), or any other non-transitory computer-readable media, wherein when the computer program code is loaded into a computer and executed by the computer, the computer becomes a means for practicing the embodiments. Embodiments may also take the form of computer program code, such as whether stored in a storage medium, loaded into a computer and / or executed by a computer, or transmitted through some transmission medium (such as by wire or cable, by optical fiber, or via electromagnetic radiation), wherein when the computer program code is loaded into a computer and executed by a computer, the computer becomes a means for practicing the exemplary embodiments. When implemented on a general-purpose microprocessor, the computer program code segments configure the microprocessor to create specific logic circuits.

[0062] The terms used herein are only for the purpose of describing specific embodiments and are not intended to limit the present disclosure. As used herein, unless the context clearly indicates otherwise, the singular forms "a", "a", and "the" are intended to include the plural forms as well. It will be further understood that the terms "include" and / or "comprise" when used in this specification specify the presence of the features, integers, steps, operations, elements, and / or parts, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, parts, and / or groups thereof.

[0063] Those skilled in the art will appreciate that various example embodiments are shown and described herein, all with certain features in specific embodiments, but the present disclosure is not limited thereby. On the contrary, the present disclosure may be modified to incorporate any number of variations, changes, substitutions, combinations, sub-combinations, or equivalent arrangements not previously described but commensurate with the scope of the present disclosure. In addition, although various embodiments of the present disclosure have been described, it is to be understood that aspects of the present disclosure may include only some of the described embodiments. Therefore, the present disclosure is not to be considered limited by the foregoing description, but only by the scope of the appended claims.

Claims

1. An elevator system, comprising: Elevator car; at least one sensor, Wherein, the system is configured as follows: receiving sensor data from the at least one sensor indicating that a first passenger is calling the elevator car or is within the elevator car; determining the first passenger's travel intention based on the at least one sensor data; and When the travel intention triggers an alarm condition, an action is taken.

2. The system according to claim 1, wherein: To take the action, the system is configured to issue an alert or assign the first passenger to ride alone or with multiple people.

3. The system according to claim 1, wherein: To determine the travel intention of the first passenger, the system is configured to: A time series model of the history of the first passenger's elevator calls and travel intentions is processed as training data by one or more artificial intelligence (AI) models to thereafter identify the first passenger and the travel intentions of the first passenger.

4. The system according to claim 3, wherein: The one or more AI models include: User identification model; and Intent recognition model.

5. The system of claim 4, wherein: When the alarm condition is triggered, the system is configured to perform supervised learning by: soliciting feedback to confirm the accuracy of classifying the travel intention as abnormal; and applying the feedback to the one or more artificial intelligence (AI) models.

6. The system of claim 3, wherein: The time series model identifies unique passenger identifiers for a plurality of passengers including the first passenger and a history of elevator calls and travel intentions for each of the plurality of passengers.

7. The system of claim 6, wherein: The system is configured to determine the travel intention of the first passenger for a call by feeding call data to the one or more artificial intelligence (AI) models, Wherein, for the first passenger, the call data includes one or more of the following: call time; call floor; requested destination floor; observed destination floor; clothing worn; equipment carried; and the number of people traveling with the first passenger.

8. The system of claim 7, wherein: When determining whether the travel intention of the first passenger triggers the alarm condition, the system is configured to: comparing the travel intention of the first passenger with the time series model to determine whether to classify the travel intention as abnormal for the first passenger; as well as A determination is made as to whether the anomaly of the first passenger triggers the alarm condition.

9. The system of claim 8, wherein: The system is configured to trigger the alarm condition upon determining that the total number of anomalies for the first passenger exceeds a threshold.

10. The system of claim 9, wherein: When the alarm condition is triggered, the system is configured to issue one or more of the following: A user isolation flag, so that the system will only serve the first passenger when not serving other passengers; a multi-person flag so that the system will thereafter only serve the first passenger when at least two other passengers are already on the elevator car; or A denial of service flag is set so that the system will not serve the first passenger thereafter.

11. A method of operating an elevator system, comprising: receiving sensor data from at least one sensor, wherein the sensor data indicates that a first passenger is calling the elevator car or is within the elevator car; determining the first passenger's travel intention based on the sensor data; and When the travel intention triggers an alarm condition, an action is taken.

12. The method according to claim 11, wherein: When the action is taken, the method includes issuing an alert or assigning the first passenger to ride alone or with multiple people.

13. The method according to claim 11, wherein: When determining the travel intention of the first passenger, the method includes: A time series model of the history of the first passenger's elevator calls and travel intentions is processed by the system through one or more artificial intelligence (AI) models as training data to thereafter identify the first passenger and the travel intentions of the first passenger.

14. The method according to claim 13, wherein: The one or more AI models include: User identification model; and Intent recognition model.

15. The method according to claim 14, wherein: When the alarm condition is triggered, the method includes performing supervised learning by the system by: soliciting feedback to confirm the accuracy of classifying the travel intention as abnormal; and applying the feedback to the one or more artificial intelligence (AI) models.

16. The method of claim 13, wherein: The time series model identifies unique passenger identifiers for a plurality of passengers including the first passenger and a history of elevator calls and travel intentions for each of the plurality of passengers.

17. The method according to claim 16, comprising: determining, by the system, the travel intention of the first passenger for a call by feeding call data to the one or more artificial intelligence (AI) models, Wherein, for the first passenger, the call data includes one or more of the following: call time; call floor; requested destination floor; observed destination floor; clothing worn; equipment carried; and the number of people traveling with the first passenger.

18. The method of claim 17, wherein: When determining whether the travel intention of the first passenger triggers the alarm condition, the method includes: comparing, by the system, the travel intention of the first passenger with the time series model to determine whether to classify the travel intention as abnormal for the first passenger; and A determination is made as to whether the anomaly of the first passenger triggers the alarm condition.

19. The method according to claim 18, comprising: The alarm condition is triggered by the system when it is determined that the total number of anomalies of the first passenger exceeds a threshold.

20. The method of claim 19, wherein: When the alarm condition is triggered, the method includes issuing, by the system, one or more of the following: A user isolation flag, so that the system will only serve the first passenger when not serving other passengers; a multi-person flag so that the system will thereafter only serve the first passenger when at least two other passengers are already on the elevator car; or A denial of service flag is set so that the system will not serve the first passenger thereafter.

Images