WO2025195778A1 - Passenger transport system with object detection device - Google Patents

Abstract

The invention relates to a passenger transport system (1) in the form of an escalator (3) or a moving walkway and to a method for operating same. The passenger transport system has a transition region (5) in which a movable transport belt (7) adjoins a stationary floor plate (11) provided with a comb plate (9). The passenger transport system (1) additionally has an object detection device (13) for detecting objects (15) in a monitoring volume (17). The monitoring volume (17) is located in the transition region (5) at a defined vertical distance (DH) from a tread surface plane (19) and above same, and the object detection device (13) is configured and arranged on the passenger transport system (1) so as to monitor the monitoring volume (17) contactlessly and three-dimensionally with respect to a position and an extent of at least one object (15) located therein and to exclude surrounding adjacent volumes (27) outside the monitoring volume (17) from the monitoring. In this way, monitoring of the transition region of the passenger transport system can be simplified.

Description

Passenger transport system with object detection device

The present invention relates to a passenger transport system in the form of an escalator or a moving walkway. Furthermore, the invention relates to a method for operating such a passenger transport system.

Escalators and moving walkways are used to transport people within structures or buildings. Escalators, which are sometimes also called moving walkways, are used to transport people along inclined travel paths between different height levels, in particular between different floors. For this purpose, the escalator comprises a conveyor belt in the form of a step belt, in which several tread units in the form of steps are coupled together one behind the other in the direction of travel and can be moved continuously along a steeply inclined travel path. A moving walkway, in contrast, is used to transport people along horizontal travel paths or, at most, slightly inclined travel paths. For this purpose, the moving walkway has a conveyor belt in the form of a pallet belt, in which several tread units in the form of pallets are coupled together and can be moved continuously along a non- or only slightly inclined travel path. The travel path extends between two access areas of the passenger transport system at opposite ends of the travel path. People can access the conveyor belt via the access areas and then climb onto one of the steps or pallets.

In the access areas, a stationary base plate, along with a stationary comb plate attached to it, borders the moving conveyor belt. The line along which the conveyor belt borders the comb plate and where the moving tread units enter or exit relative to the stationary comb plate is called the comb intersection line.

Due to this device-specific design, objects such as a piece of clothing, a shoe, or similar, or even waste such as a discarded bottle, can, etc., can accumulate in a transition area adjacent to the comb cutting line. It has been observed that potentially dangerous situations can arise in this case. In particular, the object can become trapped between the step or pallet and the comb plate, or remain in front of the comb plate and, for example, act as a tripping hazard. This could potentially cause damage to the object or, in the worst case, even injury to passengers. Furthermore, it could potentially cause significant damage to the escalator or moving walkway.

It was therefore considered to monitor the transition area adjacent to the comb intersection line on escalators or moving walkways for the presence of unwanted objects. For example, EP 3 299 330 B1 describes detecting an engagement state between a step and a comb plate of a passenger transport system. However, it was recognized that conventional approaches for monitoring this transition area can be complex in terms of the hardware and/or software required for this purpose. WO2015/090764 A1 describes various arrangements of monitoring sensors in the conveying area of an escalator in order to monitor people from the most ideal position possible.

Therefore, there may be a need, among other things, for a passenger transport system in the form of an escalator or moving walkway in which the transition area can be monitored for the presence of unwanted objects in a simplified manner, particularly with low requirements for the hardware and/or software to be provided. Furthermore, there may be a need for a method for operating such a passenger transport system, by means of which the transition area can be advantageously monitored for the presence of unwanted objects, particularly with low effort and/or high reliability.

Such a need can be met by the subject matter according to the independent patent claims. Advantageous embodiments are explained both in the dependent claims and in the following description and illustrated in the figures.

According to a first aspect, a passenger transport system in the form of an escalator or a moving walkway is described, wherein the passenger transport system has a transition area at which a movable conveyor belt of the passenger transport system adjoins a stationary floor plate of the passenger transport system, provided with a comb plate. The passenger transport system also has an object detection device for detecting objects in a monitoring volume. The monitoring volume is defined by Vertical dimensions, longitudinal horizontal dimensions, and transverse horizontal dimensions are defined. The monitoring volume is spaced a defined vertical distance from a tread plane and located above this tread plane in the transition area. In the tread plane, an upward-facing surface of a tread extends from a tread unit of the conveyor belt adjacent to the comb plate. The object detection device is configured and arranged on the passenger transport system in such a way as to monitor the monitoring volume contactlessly and three-dimensionally with regard to the position and extent of at least one object located therein and to exclude surrounding neighboring volumes outside the monitoring volume from monitoring. The feature "neighboring volumes outside the monitoring volume" comprises all volumes surrounding the monitoring volume that do not belong to the monitoring volume. This includes, in particular, the adjacent volume, which is created by the defined vertical distance, below the monitoring volume and thus between the monitoring volume and the tread level or conveyor belt and/or comb plate. The characteristic "localized" means, on the one hand, that the extents of the monitoring volume are spatially limited or defined, and, on the other hand, that the monitoring volume occupies a defined spatial position above the tread level in the transition area, or is arranged with a defined spatial position in the transition area.

According to a second aspect, a method for operating a passenger transport system according to an embodiment of the first aspect of the invention is described. The method comprises contactless and three-dimensional monitoring of the monitoring volume while maintaining a defined vertical distance of the monitoring volume from a tread plane with respect to the position and extent of at least one object located in the monitoring volume, wherein neighboring volumes outside the monitoring volume are excluded from monitoring.

By way of introduction, a basic idea for embodiments of the invention described herein will be briefly explained, whereby this explanation is to be interpreted as merely a rough summary and not as limiting the invention: As already indicated above, a transition area adjacent to the comb cutting line between the stationary comb plate and the moving conveyor belt is to be monitored for the presence of unwanted objects. For this purpose, camera systems were usually used to visually record the transition area. For example, the aforementioned EP 3 299 330 B1 describes an approach in which the engagement between the teeth of the comb plate, on the one hand, and grooves in the tread of the conveyor belt, on the other, is to be monitored in order to detect, for example, broken teeth.

However, it was found that evaluating images captured with conventionally deployed camera systems can be very complex. In particular, it was recognized as difficult to distinguish between, on the one hand, the moving conveyor belt and the stationary comb plate and base plate, and, on the other hand, between objects moving relative to both the conveyor belt and the comb plate. This often requires very complex image processing and analysis. In other words, when monitoring the transition area, the simultaneous detection of moving components of the passenger transport system can impair the quality of the detection signals (image quality, frames), increase the data volume, and complicate their evaluation.

The approach presented here proposes the use of a special object detection device for monitoring the transition area. This object detection device is specifically configured to monitor only a limited monitoring volume for objects located therein and to specifically exclude adjacent neighboring volumes from monitoring. In particular, the object detection device is configured such that the monitoring volume is located at least slightly above a tread plane formed by the conveyor belt. In other words, the monitoring volume should not reach up to or encompass the moving components of the conveyor belt, but should be spaced from them by at least a slight, defined vertical distance. Furthermore, the object detection device is configured to monitor the position and extent of objects located therein in a contactless and three-dimensional manner. For example, a 3D camera, a 3D lidar system, or a 3D laser scanner can be used for this purpose. By locating the monitoring volume monitored by the object detection device at a defined vertical distance above the conveyor belt, it can be ensured that the detection data recorded by the object detection device is not disturbed by the influence of the moving conveyor belt or is difficult to evaluate. Particularly in the case of monitoring with an object detection device in the form of a 3D camera, a 3D lidar system, or a 3D laser scanner, recorded image data or general detection data can thus only record objects within the monitoring volume, without the conveyor belt located below or, if applicable, parts of the comb plate and the adjacent floor plate also being recorded, which would complicate data evaluation. Furthermore, low-profile objects such as a tissue or a piece of cardboard, which pose no danger to users or the passenger transport system, do not protrude into the monitoring volume. Thus, objects lying on the conveyor belt are not detected and do not trigger the output of a trigger signal to change the operating mode of the passenger transport system.

Overall, this significantly simplifies the detection of unwanted objects in the transition area near the comb plate. A defined vertical distance can be understood as a constant distance from the tread surface level over time. It is also possible to vary the defined vertical distance during the monitoring process within the limits specified below (scanning the monitoring volume vertically).

In the following, possible configurations and advantages of embodiments of the passenger transport system as well as the method that can be used to operate the passenger transport system are described in more detail.

The object detection device of the passenger transport system proposed herein is specifically configured to detect objects within a spatially limited monitoring volume. The object detection device can detect objects without contact. Object detection occurs depending on whether an object is located at least partially within or completely outside the monitoring volume. For this purpose, the object detection device is capable, for example, of being directed at a larger overall volume encompassing the monitoring volume, but only recording data that indicates the presence of an object or parts of an object within the monitoring volume, whereas no data that indicates the presence of objects outside the monitoring volume is recorded, or such data is ignored. For this purpose, the object detection device is configured to record monitoring data that indicates the presence of objects in its surroundings three-dimensionally, i.e. spatially. In particular, the object detection device is configured to detect only objects that are located within a predetermined monitoring volume or that extend into this monitoring volume as potentially critical or undesirable objects. The object detection device can recognize both a position and a spatial extent of such an object based on the monitoring data it records.

The object detection device can be configured to record the monitoring data in various ways. For example, monitoring data can be recorded optically, acoustically, electromagnetically, or in another contactless manner. In particular, an object detection device designed for optical monitoring, for example, using visible light or infrared light, may be preferred.

For example, according to one embodiment, the object detection device may comprise a 3D camera, a 3D lidar system and/or a 3D laser scanner for monitoring the monitoring volume.

Such object detection devices can monitor the surveillance volume optically, i.e., visually, and in doing so, also determine relative positions, in particular distances, between an object located within the surveillance volume and the object detection device. Based on the information thus determined, it is possible to determine, among other things, whether the object is located inside or outside the surveillance volume and, if necessary, the current position within the surveillance volume of the object and/or its dimensions. Furthermore, the direction and/or speed at which the object is currently moving can be determined. A 3D camera can capture three-dimensional images of a monitored environment. Various techniques can be used for this. For example, cameras that can measure distances using a time-of-flight method (so-called TOF cameras) can be used. For example, the environment can be illuminated using light pulses, and the camera can then determine, for each captured image point or pixel, how long the emitted light pulse takes to reflect off an object and then return to a detector in the camera. This allows a distance to the captured object to be determined for each image point.

Similarly, a 3D lidar system can be used to determine distances by aiming a laser at an object or surface and measuring the time it takes for the reflected light to return to the receiver. Lidar (an acronym for "light detection and ranging") can operate in a fixed direction (e.g., vertically) or scan multiple directions; in this case, it is also referred to as lidar scanning or 3D laser scanning, a special combination of 3D scanning and laser scanning.

Due to its ability to determine the positions of objects within a monitored area in three dimensions, the object detection device can be specifically configured to monitor only the spatially limited monitoring volume for the presence of an object located therein, without also monitoring neighboring volumes located outside the monitoring volume. The monitoring volume is defined by vertical dimensions, longitudinal horizontal dimensions, and transverse horizontal dimensions, which are assumed to run in three mutually orthogonal directions. The vertical dimensions refer to a vertical direction, the longitudinal horizontal dimensions refer to a longitudinal direction parallel to a longitudinal extension direction and direction of movement of the conveyor belt within the transition area of the passenger transport system, and the transverse horizontal dimensions refer to a transverse direction orthogonal to both the longitudinal direction and the vertical direction.

The monitoring volume can be cuboid-shaped, ie it has a constant thickness as a vertical dimension, a constant length as a longitudinal horizontal dimension, and a constant width as a transverse horizontal dimension. Alternatively, the monitoring volume can also have other shapes or contours, for example, with rounded edges and/or corners, concave or convex surfaces, or similar.

According to one embodiment, the monitoring volume is located horizontally between and spaced at a transverse horizontal distance from side boundary planes in which opposite surfaces of side boundary elements of the passenger transport system extend on opposite sides of the transition area.

In other words, the monitoring volume should preferably not only be vertically spaced downwards from the conveyor belt, but also horizontally spaced from side boundary elements in a direction parallel to the transverse horizontal dimension of the monitoring volume. The corresponding distance is referred to herein as the transverse horizontal distance and, similar to the vertical distance defined above, should be selected to be sufficiently large that the object detection device is not disturbed by the side boundary elements when monitoring the monitoring volume, since these are located outside of and spaced from the monitoring volume.

Side boundary elements can be adjacent to the conveyor belt, the comb plate, and the floor plate, and can extend parallel to the conveyor belt, for example. In this context, side boundary elements can be considered, for example, balustrade bases of balustrades extending along both sides of the escalator or moving walkway.

According to one embodiment, an upwardly directed surface of a tread of a floor plate extends into a floor plate plane, bordering the comb plate. The defined vertical distance between the tread plane and the monitoring volume is greater than the distance between the tread plane and the floor plate plane.

In other words, the monitoring volume should preferably be designed and located in such a way that it is not only spaced apart from the upwardly directed surface of the tread of a tread unit of the moving conveyor belt, but also from the upwardly directed surface of the tread of the conveyor belt and the comb plate adjacent floor plate. In other words, the defined vertical distance between the monitoring volume and the surface of the tread on the conveyor belt should be selected such that the monitoring volume is also spaced from the surface of the tread on the floor plate, which typically protrudes vertically above the surface of the conveyor belt or extends further up than it.

Accordingly, the monitoring volume is selected so that neither parts of the moving conveyor belt nor parts of the stationary floor plate protrude into the monitoring volume. This ensures that neither reflections from the conveyor belt nor reflections from the floor plate interfere with monitoring of the transition area.

According to one embodiment, the defined vertical distance is between 1 mm and 100 mm, preferably between 3 mm and 50 mm. It was recognized, on the one hand, that such a defined vertical distance ensures that reflections on the surface of the conveyor belt do not interfere with monitoring or that negligibly small objects located on the conveyor belt are not detected as unwanted objects in the monitoring volume. On the other hand, it was recognized that such a defined vertical distance can ensure that objects located on the conveyor belt whose height exceeds a minimum protrude into the monitoring volume and are thus detected there by the object detection device.

Alternatively or additionally, according to one embodiment, the transverse horizontal distance can be between 1 mm and 100 mm, preferably between 3 mm and 50 mm. With such a transverse horizontal distance, it can be ensured that reflections from side boundary elements next to the conveyor belt do not interfere with the monitoring of the transition area.

According to one embodiment, the monitoring volume extends vertically above a comb cutting line at which the conveyor belt adjoins the comb plate, and/or in a region vertically above a partial region of the conveyor belt adjoining the comb cutting line in the longitudinal horizontal direction. In other words, the monitoring volume should preferably be located above the area known as the comb intersection line, where the conveyor belt enters the comb plate and where objects tend to accumulate due to the relative movement between these two components. This allows monitoring at least in a volume above the comb intersection line to determine whether one or more objects are accumulating there, for example, objects that were transported along with the conveyor belt and then become stuck on the stationary comb plate.

Alternatively or additionally, the monitoring volume should also monitor an area above the conveyor belt, extending longitudinally adjacent to the comb cutting line. This allows objects located slightly spaced longitudinally from the comb cutting line on the conveyor belt to be detected within the monitoring volume.

In particular, the height of the monitoring volume in the direction of the vertical dimensions can be less than or equal to 200 mm. In other words, the thickness of the monitoring volume can be limited to the specified dimensions. Due to such a limited height or thickness of the monitoring volume, on the one hand, the evaluation of data and signals provided by the object detection device when monitoring the monitoring volume can be limited, or the amount of data can be reduced. On the other hand, it can be ensured that objects located, for example, further above the conveyor belt do not interfere with monitoring of the transition area or are not detected as critical or undesirable objects.

Furthermore, the length of the monitoring volume in the direction of the longitudinal horizontal dimensions can be less than or equal to 400 mm. In other words, the length of the monitoring volume can be limited to the specified dimensions. Such a limitation of the length of the monitoring volume can also simplify the evaluation of data and signals or reduce the amount of data.

According to one embodiment, the object detection device can be arranged on and/or in one of the side boundary elements of the passenger transport system.

In other words, the object recognition device can, for example, be located in a the balustrade extending along the conveyor belt or in its balustrade base or be mounted thereon. Alternatively, the object detection device can be integrated into or mounted on a headpiece of a base brush or in a handrail guide profile, for example. The object detection device should be arranged longitudinally close to the comb cutting line, for example at a longitudinal distance of less than 50 cm, preferably less than 30 cm. In this case, the object detection device is located in the immediate vicinity of the transition area to be monitored. For example, a minimum distance between the object detection device and the monitored volume can be less than 100 cm, preferably less than 50 cm, or less than 20 cm, or even less than 10 cm.

This ensures, among other things, that when monitoring the monitoring volume, no objects located, for example, between the monitoring volume and the object detection device can cause shadowing and thus inadequate monitoring of the transition area.

According to one embodiment, the passenger transport system may further comprise an evaluation device configured to determine a movement speed of an object observed within the monitored surveillance volume relative to the comb plate based on detection signals of the object recognition device.

In other words, the evaluation device can infer a speed at which the detected object moves relative to the stationary comb plate based on data and signals which it receives from the object detection device and which represent the position and extent of an object located in the monitoring volume.

From the knowledge of this speed, as explained in more detail below, it can be deduced, among other things, whether the object is an undesirable object that is shifting relative to the moving conveyor belt or a presumably non-critical object that is moving along with the conveyor belt.

The evaluation device can be integrated into the object recognition device or be arranged spatially separate from the latter. The evaluation device can communicate with the object recognition device, i.e., exchange data. The evaluation device can comprise a computing unit, a data storage medium, an input interface, and/or an output interface. In particular, the evaluation device can be programmable and operated with evaluation software.

According to one embodiment, the evaluation device can be configured to execute or control a method according to embodiments of the second aspect of the invention, as described in particular herein in the following passages.

In this preferably computer-implemented method, the monitored volume is monitored contactlessly and three-dimensionally with respect to the position and extent of at least one object located therein, with neighboring volumes outside the monitored volume being excluded from monitoring. Monitoring can be understood here as continuously or cyclically observing the monitored volume, for example, at periodic intervals, in order to detect objects located therein.

According to one embodiment, the method may comprise determining a movement speed of an object observed within the monitored surveillance volume based on detection signals of the object recognition device.

In other words, the detection signals provided by the object detection device can be evaluated not only to detect an object located within the monitoring volume and, if necessary, to identify its dimensions or contour, but also to detect the current speed at which the object is moving relative to the object detection device. As explained in more detail below, information about the object's speed can provide an indication as to whether the object is an unwanted object.

According to a more specific embodiment, when carrying out the procedure The observed object is detected in first detection signals of the object detection device based on signal data that correlate with a size and/or contour of the object, and is recognized again in second detection signals of the object detection device, recorded at a later time interval, based on signal data that correlate with a size and/or contour of the object. To determine the movement speed, a spatial displacement of the detected and recognized object can be determined based on the first and second detection signals and related to the time interval.

In other words, the object detection device and the method performed by the evaluation device can be configured such that the system functions even if an observed object is briefly obscured, for example, by a person and thus cannot be detected. A "disappearance" of the object, i.e., the fact that the object is briefly not detected by the object detection device, and a "reappearance" of the object, i.e., the fact that the object is detected again by the object detection device after a short pause, should be processed in such a way that it is recognized that they are one and the same object. The object can be initially recognized in the acquired detection signals based on its size and/or contour, or recognized again after the brief "disappearance." By measuring the duration between the "disappearance" and the "reappearance" of the observed object, the speed at which the object was moving during its "disappearance" can be deduced.

Overall, the described approach makes it possible to observe objects moving relative to the object recognition device even if they are briefly obscured, i.e., become "invisible" to the object recognition device. An object can be uniquely identified by comparing it with a previously detected visual property, such as its size and contour, and recognized again after a brief "disappearance." In particular, the speed of the observed objects can also be determined in such cases.

According to one embodiment, the method may comprise outputting a trigger signal if the determined movement speed of the observed object is less than a current speed of the conveyor belt. In other words, if an object observed within the monitoring volume is detected to be moving slower than the conveyor belt, it can be concluded that the object is undesirable or even poses a threat to the operation of the passenger transport system. Upon detection of such a potentially critical object, a trigger signal can be issued, for example, in the form of an electrical or optical signal. This trigger signal can trigger various reactions.

In contrast, it can be assumed that objects that move with the conveyor belt and thus have the same speed as the conveyor belt or objects that move even faster than the conveyor belt can typically be considered uncritical, so that no trigger signal needs to be issued.

In this context, information about the current speed of the conveyor belt can be queried, for example, by a control system of the passenger transport system.

The decision as to whether or not to output the trigger signal can take a specified speed tolerance into account. For example, a trigger signal can only be output if the current movement speed of the observed object is at least 2%, at least 5%, at least 10%, or even at least 20% lower than the current speed of the conveyor belt. In particular, a trigger signal can be output if it is detected that the observed object is not moving, i.e., is stationary relative to the moving conveyor belt, or if the observed object is periodically moving back and forth.

Passenger transport systems of the aforementioned type have two transition areas, which is why an object detection device is preferably installed in both. Only the transition area where the conveyor belt moves toward the comb plate poses a risk. Since older users often wait briefly on the comb plate when entering the conveyor belt, the object detection device could misinterpret this as a lying object. To avoid such "false signals," the trigger signal is preferably issued depending on the direction of travel. In particular, according to one embodiment, an operating mode of the passenger transport system can be changed in response to the trigger signal.

The term "operating mode" can be interpreted broadly. In particular, an acoustic or visual alarm can be issued in response to the emitted trigger signal. The alarm can be issued directly in the area of the passenger transport system, for example, so that it can be perceived by passengers there. Alternatively or additionally, the alarm can also be issued in a remote monitoring center. Alternatively or additionally, the operating mode of the escalator or moving walkway can be changed. For example, its speed can be reduced to creep speed or the conveyor belt can even be brought to a complete stop.

The following example illustrates the essential functions of the object detection device. With the aforementioned object detection device, the monitoring volume can be limited by filter settings in the software of the recording device so that its width almost corresponds to the distance between the two base plates of the passenger transport system. The length of the monitoring volume can be limited to 50 mm and extend from the comb intersection line across the conveyor belt. A defined vertical distance of 10 mm can be set between the tread surface level and the monitoring volume, and the height of the monitoring volume can be limited to 50 mm, so that an upper horizontal surface of the monitoring volume extends 60 mm above the tread surface level. Objects such as a bottle, a suitcase, or a child's arm that are located within the monitoring volume are detected by the object detection device. Small, flat objects such as chewing gum, a ballpoint pen, or a scrap of paper are not detected due to the defined vertical distance of 10 mm. Furthermore, only detected objects whose speed is lower than the speed of the conveyor belt are taken into account. A sufficiently large object that has run into the comb cutting line has a speed of 0 m/s and is therefore taken into account. If the presence of an object is continuously detected over a period of 0.5 to 2 seconds, the object detection device releases a first trigger signal, which activates an acoustic warning signal. If the presence of an object is continuously detected for a If the object is detected for a period of more than 2 seconds, the object detection device triggers a second trigger signal to interrupt the power supply to the motor and brake of the passenger transport system's drive unit. The warning signal remains active, and the start switch of the passenger transport system must not be activated until the object detected by the object detection device has been removed.

According to a further embodiment, the monitoring volume is monitored by means of an evaluation device which is configured for machine learning and which has been previously trained using training data, wherein the training data represent detection signals of the object recognition device with regard to possible operating situations when monitoring the monitoring volume.

In other words, the evaluation device may have an artificial intelligence capable of machine learning, which has been previously trained to recognize objects located within the monitoring volume and which are to be classified as potentially critical in the detection signals provided by the object detection device.

The artificial intelligence can be trained in advance with training data corresponding to the detection signals supplied by the object detection device. Such training data can, for example, be real measurement data and can, for example, have been recorded using a real passenger transport system, whereby as many different operating situations as possible are to be recorded for training purposes. Specifically, a large number of objects such as cans, bottles, bricks, bags of various sizes, body parts of mannequins and the like can be placed individually or together in different positions on the moving conveyor belt. Each time an object is detected (i.e., when a detection signal or signal pattern corresponding to the detected object deviates from the signal pattern without a detected object and specified criteria such as the speed of the object relative to the travel speed of the passenger transport system are met), it can be confirmed, for example by means of a manual input, that this must be recognized by the artificial intelligence as an unwanted object. Alternatively or additionally, training data can be used that has been artificially generated, for example by generating signal patterns to be recognized through simulation or modeling on a 3D CAD system, which can also be used to simulate movements. Overall, the training data can represent a wide variety of scenarios that may occur or be conceivable during the operation of the passenger transport system. The training data can be identified and labeled as such. Accordingly, the artificial intelligence can perform supervised machine learning. Alternatively, the system can also be configured for unsupervised learning, whereby objects detected during daily operation, for example, along with their significant outlines, position, and extent, are continuously saved as training data.

The use of an evaluation device capable of machine learning, which has been suitably trained in advance, can enable particularly reliable detection of unwanted objects in the transition area of the passenger transport system.

Embodiments of the invention are described below with reference to the accompanying drawings, wherein neither the drawings nor the description are to be construed as limiting the invention.

Figure 1 shows a perspective view of a passenger transport system according to an embodiment of the present invention.

Figure 2 shows a longitudinal sectional view of the passenger transport system from Figure 1.

Figure 3 shows a top view of the passenger transport system from Figure 1.

The figures are merely schematic and not to scale. The same reference numerals designate the same or equivalent features in the various figures.

Figure 1 shows a perspective view of a passenger transport system 1 in the form of an escalator 3. Figures 2 and 3 show a longitudinal sectional view and a top view, respectively, of a transition area 5 of the passenger transport system 1.

The passenger transport system 1 comprises a movable conveyor belt 7. The conveyor belt 7 is composed of a plurality of successively arranged and interconnected step units 8 in the form of steps. The conveyor belt 7 can, for example, by a drive wheel 51 (see Figure 2) in a rotating manner along a travel path. At the transition region 5, the moving conveyor belt 7 borders a stationary comb plate 9 and a base plate 11 longitudinally adjacent to the latter and also stationary. A boundary line at which the conveyor belt 7 borders the comb plate 9 is referred to as the comb cutting line 41.

In the transition region 5, as can be seen in particular in Figure 2, the tread units 8 of the conveyor belt 7 move in such a way that a surface 21 of a tread surface 23 of a tread unit 8 adjacent to the comb plate 9 extends in a plane which is referred to herein as the tread surface plane 19. The comb plate 9 and the base plate 11 extend at least partially above this tread surface plane 19, so that a plane referred to herein as the base plate plane 37 runs along a surface 39 of the tread surface of the base plate 11 above and at a distance DBT measured in the vertical direction of a few millimeters to a few centimeters, for example between 1 mm and 2 cm, from the tread surface plane 19.

As can also be seen in Figures 1 and 3, a balustrade 47, which supports a handrail 49, extends on each side of the conveyor belt 7. At the base of the balustrade 47, a balustrade base 33 extends, which is directly adjacent to the conveyor belt 7 or at most slightly spaced from it and is referred to herein as a side boundary element 31. A surface 36 of the side boundary element 31 directed toward the conveyor belt extends in a side boundary plane 35.

During operation of the passenger transport system 1, situations may arise in which unwanted objects 15 are located within the transition area 5. For example, waste that a passenger has dropped during their journey on the escalator 3 may be transported along with the conveyor belt 7 and then deposited on the comb plate 9. In the figures, a bottle is shown as an example of such an unwanted object 15. Since the comb plate 9 projects upwards above or protrudes beyond the tread plane 19 of the conveyor belt 7, such unwanted objects 15 cannot be transported out of the passenger transport system 1 by the conveyor belt 7, but rather settle in the transition area 5 and in particular adjacent to the comb cutting line 41. In an alternative situation, for example, a piece of clothing or a shoe may become lodged in the transition area 5 between the stationary comb plate 9 and the moving Conveyor belt 7 may become trapped. In both cases, undesirable or even dangerous situations may occur for passengers of escalator 3 and/or damage to escalator 3.

In order to be able to detect unwanted objects in the transition area 5 both reliably and with as little technical effort as possible, the passenger transport system 1 has an object detection device 13. In the example shown, the object detection device 13 has a 3D camera 29, for example in the form of a TOF camera or a stereo camera. Such a 3D camera 29 is capable of recording three-dimensional images of its surroundings and, in addition to a two-dimensional lateral image resolution, also providing information about the distances of surfaces or objects recorded in this image. In other words, it is possible, among other things, to configure this camera 29 such that it only records images within a predetermined volume.

For the passenger transport system 1 described herein, the object detection device 13 is specifically configured such that it exclusively monitors a spatially limited monitoring volume 17 adjacent to the comb plate 9 in the transition area 5 for unwanted objects 15, but does not monitor adjacent neighboring volumes 27 outside the monitoring volume 17, i.e., ignores them or does not detect them at all. The monitoring volume 17 can, in principle, have any desired shape and contour and can have vertical dimensions H, longitudinal horizontal dimensions L in a direction parallel to a direction of movement of the conveyor belt 7, and transverse horizontal dimensions B in a direction transverse to the direction of movement of the conveyor belt 7 and transverse to the vertical. Preferably, the monitoring volume 17 can be cuboid-shaped, i.e., it can have a uniform height as vertical dimensions H, a uniform length as longitudinal horizontal dimensions L, and a uniform width as transverse horizontal dimensions B. For example, the camera 29 used as object detection device 13 can continuously record three-dimensional images or video sequences within a field of view. The detection signals recorded in this way can represent a type of point cloud. The point cloud can be filtered within a recording time such that only detection signals remain that represent objects 15 and passengers' feet only within the monitoring volume 13. The object recognition device 13 is specifically configured to detect objects 15 located within the monitoring volume 17 in a contactless and three-dimensional manner with regard to their position and extent. In other words, the camera 29 of the object recognition device 13 can, for example, provide three-dimensional image data, which can be used to analyze whether or where an object 15 is located within the monitoring volume 17 and what dimensions this object 15 has. Preferably, the object recognition device 13 can also be used to determine the precise shape and/or contour of the detected object 15.

To simplify the analysis of detection signals provided by the object detection device 13, the object detection device 13 is configured and arranged such that the monitoring volume 17 observed by it is located at least slightly above the tread plane 19. A defined vertical distance DH exists between the tread plane 19 and a lower boundary surface of the monitoring volume 17 directed vertically toward it. The defined vertical distance DH should, for example, be between 1 mm and 100 mm, preferably between 3 mm and 50 mm.

In particular, it may be preferable to select the defined vertical distance DH sufficiently large that it is greater than the distance DBT between the floor plate plane 37 and the tread plane 19, so that the monitoring volume 17 is also vertically spaced from the floor plate plane 37.

By selecting or arranging the monitoring volume 17 in this way, it can be ensured that the object detection device 13 does not also record the moving conveyor belt 7 and, if applicable, the comb plate 9 and the base plate 11, which would otherwise make analysis of the recorded detection signals more difficult. In other words, it is deliberately avoided that the object detection device 13 generates detection signals that represent information regarding the moving conveyor belt 7 and/or the comb plate 9 and the base plate 11, since it has been recognized that such information is not helpful for the desired monitoring of the transition area 5 for unwanted objects 15 and makes evaluation of the detection signals more difficult. Instead, the object detection device 13 is limited to only detecting detection signals relating to the area above the conveyor belt 7 and preferably also to record the monitoring volume 17 located above the comb plate 9 and the base plate 11.

Similarly, the object detection device 13 is further configured and arranged such that the monitoring volume 17 observed by it is located at least slightly laterally spaced from the lateral boundary planes 35, in which the surfaces 36 of the lateral boundary elements 31 extend, said surfaces 36 being directed toward the conveyor belt 7. A transverse horizontal distance DB exists between the respective lateral boundary plane 35 and a lateral boundary surface of the monitoring volume 17 directed laterally toward this plane. The transverse horizontal distance DB can be between 1 mm and 100 mm, preferably between 3 mm and 50 mm. This lateral spacing toward the lateral boundary elements 31 also makes it easier to evaluate or analyze the detection signals supplied by the object detection device 13, since they are not disrupted by the generally superfluous recording of data representing the lateral boundary elements 31.

The object recognition device 13 is further configured and arranged such that the monitoring volume 17 extends vertically above the comb cutting line 41 and preferably in a region 43 vertically above a partial region of the conveyor belt which adjoins the comb cutting line 41 in the longitudinal horizontal direction.

The monitoring volume 17 can have a limited height H of less than or equal to 200 mm, but preferably greater than 5 mm. This can, among other things, ensure that the monitoring volume 17 is as small as possible and only covers height ranges in which unwanted objects, such as waste carried on the conveyor belt 7, are typically expected. A length L of the monitoring volume 17 can be selected to be less than or equal to 400 mm, but is preferably greater than 50 mm. This also allows the monitoring volume 17 to be limited to an area sufficient to detect unwanted objects 15, thus keeping the amount of data signals provided by the object detection device 13 to a minimum.

The object recognition device 13 can be arranged on or in one of the side boundary elements 31. For example, the object recognition device 13 can be arranged on or in the - TI -

Balustrade base 33 must be mounted.

The passenger transport system 1 additionally has an evaluation device 45. The evaluation device 45 can, for example, be integrated into the object detection device 13, but can also be located spatially remote from it. The evaluation device 45 serves to analyze the detection signals supplied by the object detection device 13 in such a way that a movement speed of an object 15, which was detected within the monitoring volume 17, relative to the comb plate 9 can be deduced from them. In other words, the detection signals supplied by the object detection device 13 should not only be used to detect an object 15 within the monitoring volume 13 in general, but should preferably also be used to determine how fast this object 15 is moving relative to stationary components of the passenger transport system 1.

The information about the movement speed of the object 15 can then be used to decide whether the object 15 is a potentially critical and thus undesirable object 15, such as left-over waste, or whether currently only an object is detected in the monitoring volume 13 which is regularly transported through the monitoring volume 13 during operation of the escalator 3, such as a passenger's foot.

In particular, the detection signals of the object recognition device 13 can be evaluated by the evaluation device 45 in such a way that only if the determined movement speed of the observed object 15 is lower than a current speed of the conveyor belt 7, it is concluded that it is potentially an undesirable object 15. In this case, a trigger signal can then be output.

In response to the trigger signal, an operating mode of the passenger transport system 1 can then be changed. For example, the trigger signal can be forwarded from the evaluation device 45 to a control system of the passenger transport system, which, upon receipt of the trigger signal, causes, for example, a drive of the passenger transport system to only slow down the conveyor belt 7, i.e., to continue moving it in a crawling motion. or even to completely stop the movement of the conveyor belt 7. In addition or alternatively, the trigger signal can trigger the output of a warning signal. The warning signal can be issued visually, acoustically, or in another way. The warning signal can be used to warn passengers of the passenger transport system. In addition or alternatively, a remote monitoring center can be warned using the warning signal.

Since the approach described herein for monitoring the transition area 5 of the passenger transport system 1 is also intended to function when objects carried by the conveyor belt 7 are briefly obscured, for example, by persons and are therefore temporarily undetectable by the object recognition device 13, it may be advantageous to design the object recognition device 13 or an evaluation of its detection signals carried out by the evaluation device 45 in such a way that an object 15 can briefly "disappear" and be detected again after its reappearance.

For this purpose, the observed object 15 can initially be detected in the first detection signals of the object detection device based on signals that correlate with a size and/or contour of the object 15. If the object reappears after a time interval after briefly disappearing, it can be detected based on signal data that in turn correlate with its size and/or contour. To detect or re-detect the size and/or contour of the object, corresponding image data can, for example, be subjected to image analysis. The intrinsic speed of the object 15 detected and re-detected in this way can then be determined by relating the spatial displacement of the object during its "disappearance" to the time interval within which it "disappeared."

Detection of unwanted objects 15 within the monitoring volume 13 can be further improved, among other things, by using a machine capable of machine learning as the evaluation device 45 and by training this learning-capable evaluation device 45 in advance with training data before commissioning of the passenger transport system 1, which training data correspond to the detection signals of the object detection device 13 to be evaluated later during operation. The training data can, for example, be recorded during a test phase in which the passenger transport system 1 is deliberately operated in such a way that undesirable objects 15 are present in the monitoring volume 17. Alternatively or additionally, training data can also be generated artificially, for example through computer simulations.

The training data can be designed in such a way and/or the adaptive evaluation device can learn with the aid of the training data in such a way that later, during actual operation of the passenger transport system 1, for example, a movement analysis and/or an analysis of geometric properties of observed objects 15 can be carried out in such a way that undesirable objects 15 can be distinguished from non-critical objects with a high degree of probability.

In addition, by training the evaluation device 45, for example, situations can be learned in which observed objects are temporarily obscured and thus appear to disappear for the object recognition device 13, but reappear after a time interval.

Furthermore, the evaluation device 45 can learn through training at which movement speeds of observed objects 15 there is a high probability that they are unwanted objects.

Finally, it should be noted that terms such as "having," "comprising," etc., do not exclude other elements or steps, and terms such as "a" or "an" do not exclude a plurality. Furthermore, it should be noted that features or steps described with reference to one of the above embodiments may also be used in combination with other features or steps of other embodiments described above, as long as they are included within the scope of the appended claims.

Claims

Patent claims 1. A passenger transport system (1) in the form of an escalator (3) or a moving walkway, wherein the passenger transport system has a transition area (5) in which a movable conveyor belt (7) of the passenger transport system (1) adjoins a stationary base plate (11) of the passenger transport system (1) provided with a comb plate (9), and wherein the passenger transport system (1) has an object detection device (13) for detecting objects (15) in a monitoring volume (17), which monitoring volume (17) is defined by vertical dimensions (H), longitudinal horizontal dimensions (L) and transverse horizontal dimensions (B), characterized in that the monitoring volume (17) is spaced at a defined vertical distance (DH) from a tread plane (19) and is located above this in the transition area (5), and in which tread plane (19) an upwardly directed surface (21) of a tread (23) extends from a tread unit (25) of the conveyor belt (7) adjacent to the comb plate (9), wherein the object detection device (13) is configured and arranged on the passenger transport system (1) in such a way as to monitor the monitoring volume (17) in a contactless and three-dimensional manner with regard to a position and an extent of at least one object (15) located therein and to exclude surrounding neighboring volumes (27) outside the monitoring volume (17) from the monitoring. 2. Passenger transport system according to claim 1, wherein the object recognition device (13) comprises a 3D camera (29), a 3D lidar system and/or a 3D laser scanner for monitoring the monitoring volume (17). 3. Passenger transport system according to one of the preceding claims, wherein the monitoring volume (17) is located horizontally between and in each case spaced at a transverse horizontal distance (DB) from side boundary planes (35) in which mutually opposite surfaces (36) of side boundary elements (31) of the passenger transport system (1) extend on opposite sides of the transition region (5). 4. Passenger transport system according to one of the preceding claims, wherein an upwardly directed surface (39) of a tread surface of a floor plate (11) which adjoins the comb plate (9) extends in a floor plate plane (37), and wherein the defined vertical distance (DH) is greater than a distance (DBT) of the floor plate plane (37) to the tread surface plane (19). 5. Passenger transport system according to claim 3 or 4, wherein the defined vertical distance (DH) is between 1 mm and 100 mm, or between 3 mm and 50 mm. 6. Passenger transport system according to one of claims 3 to 5, wherein the transverse horizontal distance (DB) is between 1 mm and 100 mm, or between 3 mm and 50 mm. 7. Passenger transport system according to one of the preceding claims, wherein the monitoring volume (17) extends vertically above a comb cutting line (41) at which the conveyor belt (7) adjoins the comb plate (9) and/or extends vertically in a region (43) above a partial region of the conveyor belt (7) adjoining the comb cutting line (41) in the longitudinal horizontal direction, and wherein furthermore: a height (H) of the monitoring volume (17) in the direction of the vertical dimensions is less than or equal to 200 mm, and/or a length (L) of the monitoring volume (17) in the direction of the longitudinal horizontal dimensions is less than or equal to 400 mm. 8. Passenger transport system according to one of the preceding claims, wherein the object detection device (13) is arranged on and/or in a side boundary element (31) of the passenger transport system (1). 9. Passenger transport system according to one of the preceding claims, further comprising an evaluation device (45) which is configured to determine a movement speed of an object (15) observed within the monitored surveillance volume (17) relative to the comb plate (9) based on detection signals of the object recognition device (13). 10. A method for operating a passenger transport system (1) according to one of claims 1 to 8, wherein the method comprises: contactless and three-dimensional monitoring while maintaining a defined vertical distance (DH) of the monitoring volume (17) to a tread plane (19) with regard to a position and an extension of at least one object (15) located in the monitoring volume (17), wherein neighboring volumes (27) outside the monitoring volume (17) are excluded from the monitoring. 11. The method of claim 10, further comprising: Determining a movement speed of an object (15) observed within the monitored surveillance volume (17) based on detection signals of the object recognition device (13). 12. The method according to claim 11, wherein the observed object (15) is detected in first detection signals of the object detection device (13) on the basis of signal data which correlate with a size and/or contour of the object (15), and is detected again in second detection signals of the object detection device (13) recorded one time interval later on the basis of signal data which correlate with a size and/or contour of the object (15), and wherein, in order to determine the speed of movement, a spatial displacement of the detected and recognized object (15) is determined based on the first and second detection signals and is set in relation to the time interval. 13. The method according to claim 11 and 12, further comprising: Outputting a trigger signal when the determined movement speed of the observed object (15) is less than a current speed of the conveyor belt (7). 14. Method according to claim 13, further comprising: Changing an operating mode of the passenger transport system (1) in response to the trigger signal. 15. The method according to any one of claims 10 to 14, wherein the monitoring volume (17) is monitored using an evaluation device (45) which is configured for machine learning and which was previously trained using training data, wherein the training data represent detection signals of the object recognition device (13) with regard to possible operating situations when monitoring the monitoring volume (17).