Escalator passenger fall detection device and method based on a shot sensor

Abstract

The application discloses an escalator passenger falling detection device based on a pair of sensors, which comprises a pair of sensors and a falling discrimination controller; at least two pairs of sensors are installed on the inner side plates on both sides of the escalator; wherein the installation height of the upper pair of sensors (H1) is 800mm-1000mm away from the step tread, and the installation height of the lower pair of sensors (H3) is 100mm-200mm away from the step tread; the falling discrimination controller monitors the time sequence mode that the shielding state of the upper pair of sensors (H1) changes from shielding to non-shielding within a preset time window (T_th) and the lower pair of sensors (H3) remains in the shielding state, and determines that a passenger falling event occurs when the time sequence mode appears. The application does not collect passenger images, has no privacy controversy, and has low hardware cost; the sensors are installed on fixed components, and power supply and maintenance are simple; the falling is accurately identified by using the time sequence shielding mode of multiple sensors, the environmental adaptability is strong, and the escalator can be controlled in linkage to ensure passenger safety.

Description

Technical Field

[0001] This invention relates to the field of escalator safety technology, specifically to an escalator passenger fall detection device and method based on a through-beam sensor. Background Technology

[0002] Escalators are widely used in public places such as shopping malls, train stations, and airports. Incidents of passengers falling while riding escalators due to instability, sudden illness, or loss of balance while carrying belongings are not common. If the escalator is not detected and stopped in time, fallen passengers may be crushed by subsequent steps or downstream passengers, causing secondary injuries.

[0003] Currently, existing technologies for escalator passenger fall detection mainly include the following categories: 1. Fall detection scheme based on computer vision / deep learning For example, patent document CN109726750B discloses a passenger fall detection device and method, which is based on image acquisition and processing, and determines whether a fall has occurred by analyzing changes in passenger posture through visual images. Another example is patent document CN108805093B, which discloses a deep learning-based method for detecting passenger falls on escalators, using a deep learning model to detect passenger fall behavior by acquiring video images of the escalator. Yet another example is patent document CN110472473A, which discloses a method for detecting falls on escalators based on posture estimation, using a human posture estimation algorithm to analyze the positions of key points on the human body and the trajectory of movement to detect falls.

[0004] The aforementioned vision-based solutions have the following drawbacks: they require the installation of high-definition cameras and the use of edge computing units to run image analysis algorithms, resulting in high hardware and deployment costs; they are greatly affected by environmental factors such as lighting conditions, occlusion, and angle, and lack robustness in outdoor or complex lighting escalator scenarios; the collection of passenger images raises privacy concerns; and they do not involve triggering the escalator control system to slow down or stop the escalator after detection.

[0005] 2. Scheme for installing vibration / pressure sensors in the ladder steps For example, patent document CN113916358B discloses a method for escalator hazard monitoring and response based on vibration signal analysis. This method involves setting up a data sampling device array on the escalator step treads to collect vertical vibration signals, using time-domain characteristic values ​​to determine if an object has fallen, and using frequency-domain characteristic values ​​to determine if a person has fallen. Another example is patent document CN107915111B, which discloses a passenger behavior monitoring method and device. This method monitors passenger behavior using pressure sensors installed on the inner wall of the escalator step surface and infrared sensors on both sides of the step surface.

[0006] The sensors in the above-mentioned scheme are installed on the step treads or kick plates, which are moving parts in a cyclic motion. They require independent power supply and wireless communication for each step, resulting in high system complexity and difficult maintenance. The vibration signals are easily interfered with by the mechanical vibration of the escalator itself, leading to a high false alarm rate. Furthermore, it only focuses on behavior monitoring and visual warnings, without addressing the control of escalator deceleration / stop after a fall.

[0007] 3. Passenger detection scheme based on infrared physical sensors For example, patent document CN104058325A discloses an escalator safety warning system that uses through-beam infrared sensors installed on both sides of the escalator. When a passenger enters, the infrared beam is blocked, and the system detects the passenger's presence and issues a warning. Another example is Korean patent document KR102308850B1, which discloses a safety operation control device for escalators and moving walkways. Infrared sensors are installed facing each other on the lower part of the handrails on both sides of the escalator exit side. Combined with a load sensor, the escalator's operating speed is controlled based on the detection results.

[0008] The above solution uses through-beam infrared sensors to detect the presence of passengers (binary detection), which cannot identify changes in posture such as passengers falling; the infrared sensors are only installed at the exit to detect the passengers' descent, not for full-time monitoring; and they do not have fall recognition functionality. Summary of the Invention

[0009] The technical problem to be solved by the present invention is to provide a passenger fall detection device and method for escalators based on a through-beam sensor, which, without collecting passenger images and without privacy disputes, can accurately distinguish between normal standing and falling behavior of passengers and link the escalator control system to achieve deceleration and stopping; at the same time, the sensor is installed on a fixed component, which is convenient for power supply and maintenance and has strong environmental adaptability.

[0010] To solve the above-mentioned technical problems, the present invention provides an escalator passenger fall detection device based on a through-beam sensor, comprising: a through-beam sensor and a fall detection controller; At least two layers of through-beam sensors are installed on the inner side panels of both sides of the escalator; each through-beam sensor includes a transmitter and a receiver respectively installed on the inner side panels of both sides, and a detection beam spanning the width of the step is formed between the transmitter and the receiver; wherein the upper through-beam sensor (H1) is installed at a height of 800mm to 1000mm from the step tread, and the lower through-beam sensor (H3) is installed at a height of 100mm to 200mm from the step tread; The fall detection controller is electrically connected to the through-beam sensors on each layer and is configured to collect the occlusion status signals of the through-beam sensors on each layer in real time, and monitor the timing pattern in which the occlusion status of the upper through-beam sensor (H1) changes from occluded to unoccluded within a preset time window (T_th), while the lower through-beam sensor (H3) remains occluded. When the timing pattern occurs, it is determined that a passenger fall event has occurred.

[0011] Preferably, it also includes a middle layer through-beam sensor (H2), which is 400mm to 500mm away from the step surface, and the fall detection controller is configured to identify human posture changes based on the temporal changes in the occlusion state of the three layers of through-beam sensors.

[0012] Preferably, the at least two layers of through-beam sensors are arranged in multiple sets at intervals along the running direction of the escalator, with the spacing between adjacent sets being 1 to 3 steps.

[0013] Preferably, the fall detection controller is configured to further confirm a fall event when adjacent sensor groups along the running direction successively exhibit fall signal patterns and the signal propagation speed is greater than the speed of the ladder running.

[0014] Preferably, the preset time window (T_th) is 0.5s to 2s.

[0015] Preferably, it also includes an audible and visual alarm device, which is connected to the fall detection controller and is triggered by the fall detection controller when a passenger fall event is detected.

[0016] Preferably, the fall detection controller is connected to the escalator control system and is configured to issue a deceleration and stop command to the escalator control system after a fall event is detected, so that the escalator control system controls the main drive motor to decelerate and brake according to a preset deceleration curve through the frequency converter.

[0017] The present invention also provides a method for detecting passenger falls on escalators based on a through-beam sensor, comprising the following steps: Step S1: Install at least two layers of through-beam sensors vertically on the inner side panels on both sides of the escalator, such that the installation height of the upper through-beam sensor (H1) is 800mm to 1000mm from the tread surface, and the installation height of the lower through-beam sensor (H3) is 100mm to 200mm from the tread surface. Step S2: Real-time acquisition of occlusion status signals of the through-beam sensors on each layer; Step S3: Monitor whether the occlusion status of the upper-layer through-beam sensor (H1) changes from occluded to unoccluded within a preset time window (T_th); Step S4: When the upper-level through-beam sensor (H1) changes from being blocked to being unblocked within the preset time window, while the lower-level through-beam sensor (H3) remains blocked, a passenger fall event is determined to have occurred. Step S5: After determining that a fall has occurred, a deceleration and stop command is sent to the escalator control system to control the escalator to decelerate and stop.

[0018] Preferably, in step S1, a middle layer through-beam sensor (H2) is installed 400mm to 500mm away from the step surface; in steps S3 to S4, the change in posture of the human body from standing to falling is identified based on the temporal changes in the occlusion state of the three layers of through-beam sensors.

[0019] Preferably, multiple sets of through-beam sensors are spaced apart along the escalator's running direction. The method further includes: when adjacent sensor sets along the running direction successively produce a fall signal pattern, calculating the propagation speed of the fall signal along the running direction; when the propagation speed is greater than the speed of the steps, further confirming the fall event.

[0020] Preferably, the preset time window (T_th) is 0.5s to 2s.

[0021] Preferably, in step S5, the escalator control system controls the main drive motor to decelerate according to an S-shaped deceleration curve via a frequency converter, with the deceleration rate not exceeding 1.0 m / s². After decelerating to a low speed, the working brake is activated to bring the escalator to a complete stop.

[0022] Preferably, the escalator enters a locked state after stopping and can only resume operation after being manually reset.

[0023] Preferably, in step S5, the audible and visual alarm and voice alarm are activated simultaneously with the deceleration and elevator stop command.

[0024] Preferably, the spacing between two adjacent sets of the through-beam sensors along the escalator's running direction is 1 to 3 steps.

[0025] Compared with the prior art, the present invention has the following beneficial effects: 1. Low cost and no privacy concerns: Using through-beam sensors instead of cameras, it does not collect passenger images, thus avoiding privacy concerns. The hardware cost is significantly lower than visual AI solutions, and there is no need to equip it with edge computing devices.

[0026] 2. Easy installation and maintenance of fixed components: The sensors are installed on the inner side plates (fixed components) on both sides of the escalator. They can be directly powered by fixed lines and have wired communication. There is no need to solve the problem of independent power supply and wireless communication for the moving steps. The system has low complexity and is easy to maintain.

[0027] 3. Accurate fall recognition based on temporal occlusion pattern: By utilizing the temporal occlusion pattern changes of multi-layer sensors (the upper layer of occlusion disappears quickly while the lower layer remains occluded), the system can recognize the change in posture of a person from standing to falling. This breaks through the limitation of existing through-beam infrared sensors, which can only detect the binary detection of "whether there is a person" and can accurately distinguish between normal standing and falling behavior.

[0028] 4. Strong environmental adaptability: Based on physical characteristics (changes in occlusion caused by rapid descent of human body height), it is not affected by lighting conditions and can work stably outdoors or in complex lighting scenarios.

[0029] 5. Multi-sensor collaborative verification reduces false alarms: Multiple sensors are deployed along the running direction. By analyzing the propagation speed of the fall signal along the running direction (greater than the speed of the ladder), the fall event is further confirmed, effectively reducing the false alarm rate.

[0030] 6. Interlocked escalator control ensures safety: Upon detecting a fall, the escalator control system is immediately activated to decelerate and stop the escalator in an S-shaped deceleration curve. At the same time, audible and visual alarms and voice warnings are activated, and the escalator enters a locked state that requires manual reset to avoid secondary injury. Detailed Implementation

[0031] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can fully understand other advantages and technical effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through different specific embodiments, and the details in this specification can also be applied based on different viewpoints, with various modifications or changes made without departing from the overall design concept of the invention. It should be noted that, unless otherwise specified, the following embodiments and features can be combined with each other. The following exemplary embodiments of the present invention can be implemented in many different forms and should not be construed as being limited to the specific embodiments set forth herein. It should be understood that these embodiments are provided to make the disclosure of the present invention thorough and complete, and to fully convey the technical solutions of these exemplary embodiments to those skilled in the art. Example

[0032] This embodiment provides an escalator passenger fall detection device based on a through-beam sensor, including a through-beam sensor and a fall detection controller.

[0033] (1) Basic configuration of through-beam sensor At least two layers of through-beam sensors are installed on the inner side panels of both sides of the escalator. The through-beam sensors are through-beam infrared sensors, including a transmitter and a receiver respectively installed on the inner side panels of both sides, forming a detection beam spanning the width of the escalator steps between the transmitter and the receiver. The through-beam sensor layout in this embodiment is as follows: • Upper-level through-beam sensor H1, installed at a height of 900mm from the step surface, corresponding to the passenger's waist position (generally within the range of 800mm to 1000mm). • The lower-level through-beam sensor H3 is installed at a height of 150mm from the tread surface of the step, corresponding to the passenger's ankle position (generally within the range of 100mm to 200mm).

[0034] When a passenger is standing on the step, the infrared beam is blocked by the body, and the receiver outputs a blocked status signal (low level); when there are no passengers or when there is no obstruction at that height after a passenger has fallen down, the receiver outputs an unblocked status signal (high level). The through-beam sensor is powered by a fixed line and communicates with the fall detection controller via a wired connection, eliminating the need to solve the problem of independent power supply and wireless communication for the moving steps.

[0035] Furthermore, in this embodiment, a middle layer of through-beam sensors H2 is added to the two layers of through-beam sensors (H1, H3), positioned 400mm–500mm from the step surface, forming a three-layer layout. The fall detection controller is configured to identify human posture changes based on the temporal changes in the occlusion states of the three layers of through-beam sensors: during a fall, the occlusion disappears in the order H1→H2→H3, and the time interval between the disappearance of each layer of occlusion exhibits a fall-specific pattern, thereby improving the accuracy and discriminative power of fall detection.

[0036] (2) Arrangement of through-beam sensors along the running direction The at least two layers of through-beam sensors are arranged in multiple sets at intervals along the escalator's running direction, with the spacing between adjacent sets being 1 to 3 steps (approximately 400mm to 1200mm). In this embodiment, a set of three layers of through-beam sensors is arranged every 400mm (approximately 1 step) along the running direction, striking a balance between full coverage and control costs, and ensuring no blind spots in fall detection.

[0037] (3) Fall detection controller - temporal occlusion mode detection The fall detection controller is electrically connected to the through-beam sensors on each layer and is configured as follows: • Real-time acquisition of occlusion status signals from the through-beam sensors on each layer; • The system monitors a timing pattern where the upper-level through-beam sensor H1 changes from being blocked to being unblocked within a preset time window T_th, while the lower-level through-beam sensor H3 remains blocked. When this timing pattern occurs, a passenger fall event is determined to have occurred. In this embodiment, the preset time window T_th = 1.5s (range 0.5s to 2s).

[0038] Furthermore, the fall detection controller is configured to further confirm a fall event when adjacent sensor groups along the operating direction successively detect a fall signal pattern, and the signal propagation speed is greater than the elevator running speed. That is, if the downstream sensor group also detects a fall pattern shortly after the upstream sensor group determines a fall, and the signal propagation speed is greater than the elevator running speed (the fall signal propagation speed is abnormally fast compared to the normal elevator speed), the fall event is further confirmed, effectively reducing the false alarm rate.

[0039] (4) Audible and visual alarm device The device also includes an audible and visual alarm device connected to the fall detection controller, which is triggered by the fall detection controller when a passenger fall event is detected. In this embodiment, the audible and visual alarm device includes a warning light and a speaker, which can issue audible and visual alarms and voice warnings (such as "Emergency, the escalator is about to slow down").

[0040] (5) Linkage with escalator control system The fall detection controller is connected to the escalator control system and is configured to issue a deceleration and stop command to the escalator control system after detecting a fall event. This causes the escalator control system to control the main drive motor to decelerate and brake according to a preset deceleration curve via a frequency converter. In this embodiment, the escalator control system controls the main drive motor to decelerate according to an S-shaped deceleration curve via a frequency converter, with a deceleration rate not exceeding 1.0 m / s². After the frequency converter decelerates to a low speed, the working brake activates, completely stopping the escalator. After stopping, the escalator enters a locked state, requiring maintenance personnel to confirm safety on-site and manually reset it before operation can resume.

[0041] Working principle of the device When a passenger is standing normally on the escalator, their body simultaneously blocks the upper H1, middle H2, and lower H3 light beams, keeping all three sensors in an obstructed state. When a passenger falls, their height decreases rapidly. The upper H1 light beam is first no longer blocked (the obstruction state changes from blocked to unblocked), while the lower body continues to block the lower H3 light beam (remaining blocked). This temporal pattern of "upper layer obstruction disappearing rapidly, lower layer remaining blocked" is a typical physical characteristic of falls, significantly different from normal standing, bending, and squatting, and can serve as a reliable basis for fall detection. The addition of the middle H2 layer further refines the temporal characteristics of changes in body posture, improving the accuracy and discriminative power of fall detection. Example

[0042] This embodiment provides a method for detecting passenger falls on escalators based on through-beam sensors, including the following steps: Step S1: Install at least two layers of through-beam sensors vertically on the inner side panels 1 on both sides of the escalator, such that the installation height of the upper through-beam sensor H1 is 800mm to 1000mm from the tread surface and the installation height of the lower through-beam sensor H3 is 100mm to 200mm from the tread surface.

[0043] Furthermore, in step S1, a middle-layer through-beam sensor H2 is installed 400mm to 500mm from the tread surface of the step, forming a three-layer through-beam sensor layout consisting of an upper layer, a middle layer, and a lower layer.

[0044] Furthermore, the spacing between two adjacent sets of the through-beam sensors along the escalator's running direction is 1 to 3 steps.

[0045] In this embodiment, H1 is 900mm from the tread surface, H2 is 450mm from the tread surface, and H3 is 150mm from the tread surface. A set is set every 400mm (about 1 step spacing) along the running direction.

[0046] Step S2: Real-time acquisition of occlusion status signals from the through-beam sensors on each floor. When a passenger is standing normally on the steps, H1, H2, and H3 are all in an occlusion state.

[0047] Step S3: Monitor whether the occlusion state of the upper-layer through-beam sensor H1 changes from occluded to unoccluded within a preset time window T_th.

[0048] Furthermore, in step S3, the posture change of the human body from standing to falling is identified based on the temporal changes in the occlusion states of the three-layer through-beam sensors. When the occlusion state of the upper layer H1 of a certain group of sensors changes from occluded to unoccluded, the fall detection window is activated and the timestamp t0 is recorded.

[0049] Furthermore, the preset time window T_th is 0.5s to 2s. In this embodiment, T_th = 1.5s.

[0050] Step S4: When the upper beam sensor H1 changes from being blocked to being unblocked within the preset time window T_th, and the lower beam sensor H3 remains blocked (the human body has fallen on the step, and the lower body still blocks the beam from the lower layer), it is determined that a passenger fall event has occurred.

[0051] Furthermore, in steps S3 to S4, the posture change of the human body from standing to falling is identified based on the temporal changes of the occlusion state of the three layers of through-beam sensors: the order of occlusion disappearance during the fall is H1→H2→H3, and the time interval between the disappearance of each layer of occlusion has a unique pattern of falling.

[0052] Furthermore, multiple sets of through-beam sensors are spaced apart along the escalator's running direction. The method also includes: when adjacent sensor sets along the running direction successively detect a fall signal pattern, calculating the propagation speed of the fall signal along the running direction; if the propagation speed is greater than the step speed, further confirming the fall event. That is, if the downstream sensor set also detects a fall pattern shortly after the upstream sensor set determines a fall, and the signal propagation speed is greater than the step speed (the fall signal propagation speed is abnormally faster than the normal escalator speed), the fall event is further confirmed, effectively reducing the false alarm rate.

[0053] Step S5: After determining that a fall has occurred, a deceleration and stop command is sent to the escalator control system to control the escalator to decelerate and stop.

[0054] Furthermore, in step S5, the escalator control system controls the main drive motor to decelerate according to an S-shaped deceleration curve via a frequency converter. The deceleration rate does not exceed 1.0 m / s². After decelerating to a low speed, the working brake is activated to bring the escalator to a complete stop.

[0055] Furthermore, the escalator enters a locked state after stopping and requires manual reset before operation can resume. That is, maintenance personnel must confirm safety on-site and manually reset the escalator before operation can be restored.

[0056] Furthermore, in step S5, an audible and visual alarm and a voice alarm are activated simultaneously with the deceleration and escalator stop command. In this embodiment, the voice alarm content is "Emergency situation, the escalator is about to decelerate."

[0057] In this embodiment, upon confirmation of a fall, the fall detection controller immediately sends a deceleration and stop command to the escalator control system, simultaneously activating audible and visual alarms and voice alerts. The escalator control system uses a frequency converter to control the main drive motor to decelerate according to an S-shaped deceleration curve, with a deceleration rate not exceeding 1.0 m / s². After the frequency converter decelerates to a low speed, the working brake engages, completely stopping the escalator. After stopping, the escalator enters a locked state, requiring maintenance personnel to confirm safety on-site and manually reset it before operation can resume.

[0058] The present invention has been described in detail above through specific embodiments and examples, but these are not intended to limit the invention. Many modifications and improvements can be made by those skilled in the art without departing from the principles of the invention, and these should also be considered within the scope of protection of the present invention.

Claims

1. A passenger fall detection device for escalators based on a through-beam sensor, characterized in that, include: Through-beam sensor and fall detection controller; At least two layers of through-beam sensors are installed on the inner side panels of both sides of the escalator; each through-beam sensor includes a transmitter and a receiver respectively installed on the inner side panels of both sides, and a detection beam spanning the width of the step is formed between the transmitter and the receiver; wherein the upper through-beam sensor (H1) is installed at a height of 800mm to 1000mm from the step tread, and the lower through-beam sensor (H3) is installed at a height of 100mm to 200mm from the step tread; The fall detection controller is electrically connected to the through-beam sensors on each layer and is configured to collect the occlusion status signals of the through-beam sensors on each layer in real time, and monitor the timing pattern in which the occlusion status of the upper through-beam sensor (H1) changes from occluded to unoccluded within a preset time window (T_th), while the lower through-beam sensor (H3) remains occluded. When the timing pattern occurs, it is determined that a passenger fall event has occurred.

2. The escalator passenger fall detection device based on a through-beam sensor according to claim 1, characterized in that, It also includes a middle layer through-beam sensor (H2), which is 400mm to 500mm away from the step surface. The fall detection controller is configured to identify changes in human posture based on the temporal changes in the occlusion state of the three layers of through-beam sensors.

3. The escalator passenger fall detection device based on a through-beam sensor according to claim 1, characterized in that, The at least two layers of through-beam sensors are arranged in multiple sets at intervals along the running direction of the escalator, with the spacing between adjacent sets being 1 to 3 steps.

4. The escalator passenger fall detection device based on a through-beam sensor according to claim 3, characterized in that, The fall detection controller is configured to further confirm a fall event when adjacent sensor groups along the running direction successively produce fall signal patterns and the signal propagation speed is greater than the speed of the ladder running.

5. The escalator passenger fall detection device based on a through-beam sensor according to claim 1, characterized in that, The preset time window (T_th) is 0.5s to 2s.

6. The escalator passenger fall detection device based on a through-beam sensor according to claim 1, characterized in that, It also includes an audible and visual alarm device, which is connected to the fall detection controller and is triggered by the fall detection controller when a passenger fall event is detected.

7. The escalator passenger fall detection device based on a through-beam sensor according to claim 1, characterized in that, The fall detection controller is connected to the escalator control system and is configured to send a deceleration and stop command to the escalator control system after a fall event is detected, so that the escalator control system controls the main drive motor to decelerate and brake according to a preset deceleration curve through the frequency converter.

8. A method for detecting passenger falls on escalators based on through-beam sensors, characterized in that, Includes the following steps: Step S1: Install at least two layers of through-beam sensors vertically on the inner side panels on both sides of the escalator, such that the installation height of the upper through-beam sensor (H1) is 800mm to 1000mm from the tread surface, and the installation height of the lower through-beam sensor (H3) is 100mm to 200mm from the tread surface. Step S2: Real-time acquisition of occlusion status signals of the through-beam sensors on each layer; Step S3: Monitor whether the occlusion status of the upper-layer through-beam sensor (H1) changes from occluded to unoccluded within a preset time window (T_th); Step S4: When the upper-level through-beam sensor (H1) changes from being blocked to being unblocked within the preset time window, while the lower-level through-beam sensor (H3) remains blocked, a passenger fall event is determined to have occurred. Step S5: After determining that a fall has occurred, a deceleration and stop command is sent to the escalator control system to control the escalator to decelerate and stop.

9. The method for detecting passenger falls on escalators based on through-beam sensors according to claim 8, characterized in that, In step S1, a middle layer through-beam sensor (H2) is installed 400mm to 500mm from the step surface; in steps S3 to S4, the change in posture of the human body from standing to falling is identified based on the temporal changes in the occlusion state of the three layers of through-beam sensors.

10. The method for detecting passenger falls on escalators based on through-beam sensors according to claim 8, characterized in that, The method further includes setting multiple sets of through-beam sensors at intervals along the running direction of the escalator, and calculating the propagation speed of the fall signal along the running direction when adjacent sensor sets along the running direction successively produce a fall signal pattern; and further confirming the fall event when the propagation speed is greater than the running speed of the steps.

11. The method for detecting passenger falls on escalators based on through-beam sensors according to claim 8, characterized in that, The preset time window (T_th) is 0.5s to 2s.

12. The method for detecting passenger falls on escalators based on through-beam sensors according to claim 8, characterized in that, In step S5, the escalator control system controls the main drive motor to decelerate according to an S-shaped deceleration curve via a frequency converter. The deceleration rate does not exceed 1.0 m / s². After decelerating to a low speed, the working brake is activated to bring the escalator to a complete stop.

13. The method for detecting passenger falls on escalators based on through-beam sensors according to claim 12, characterized in that, After the escalator stops, it enters a locked state and can only be restarted after being manually reset.

14. The method for detecting passenger falls on escalators based on through-beam sensors according to claim 8, characterized in that, In step S5, the audible and visual alarm and voice alarm are activated simultaneously with the deceleration and elevator stop command.

15. The method for detecting passenger falls on escalators based on through-beam sensors according to claim 8, characterized in that, The spacing between adjacent groups of the through-beam sensors along the escalator's running direction is 1 to 3 steps.