Method and computer program for ensuring a correct maintenance of a personal transport system, and computer-readable medium

Abstract

A method for ensuring a correct maintenance of a personal transport system (20) is described. The method comprises: receiving a sensor signal from at least one sensor (42, 44, 46) worn by a maintenance worker (40) carrying out a maintenance task at the personal transport system (20); determining whether the determined maintenance task is carried out correctly from the received sensor signal; and generating a warning signal for the maintenance worker (40) when the determined maintenance task is not carried out correctly.

Description

[0001] Method and computer program for ensuring a correct maintenance of a personal transport system, and computer-readable medium

[0002] The technology described herein generally relates to monitoring a maintenance work in or at a personal transport system. More particularly, the technology relates to a method and a computer program for ensuring a correct maintenance of a personal transport system. The technology described herein further relates to a computer readable medium on which the computer program is stored.

[0003] A personal transport system may be an elevator, an escalator, or a moving pavement. Such personal transport systems normally serve for transporting passengers and / or goods throughout a building, e.g., over two or more floors of the building. For example, in case of an elevator, a multiplicity of passengers may be accommodated within a car of the elevator and may then be transported for example from a first floor to a second floor in the building. In this description, the terms “first” and “second” are used in this context only to differentiate one floor from another and not for stating that e.g. the first floor is in fact the first floor of the building. Therefore, the first and second floor may be any two different floors of a building. During such travel, further passengers may enter or exit the car during intermediate stops at other floors. The car moves vertically from one floor to the other through a hoistway of the building.

[0004] A personal transport system has to be maintained by a maintenance worker on a regular basis, e.g. in accordance with a regular maintenance interval or because of a fault of the elevator. In order to maintain the elevator safely, the worker has to wear safety devices corresponding to the risk area he has to access. The safety devices each may also be referred to as Personal Protection Equipment (PPE). For example, if the worker wants to access the pit of the hoistway, the worker has to wear a helmet. If the worker wants to access the top of the car, the worker has to wear the helmet and a harness, and so forth. It is already known that that the PPE may be equipped with one or more sensors for monitoring whether the maintenance worker carries out a maintenance task safely, i.e., that the maintenance worker considers all necessary safety measures.

[0005] The worker has to carry out at least one maintenance task when maintaining the personal transport system. Each maintenance task which may be carried out at the personal transport system may comprise two or more steps. These steps have to be performed correctly, at the right time, and in particular in the right order. If a step is not performed correctly or not performed at the correct time, the maintenance task may be not fulfilled. This may cause danger for the maintenance worker or afterwards for the passengers.

[0006] US 10,249, 163 B 1 discloses a method for ensuring a correct maintenance of a personal transport system using a sensor signal from a sensor located in the maintenance worker's work area.

[0007] WO 2023 / 274684 Al discloses a method for ensuring a safe maintenance of a personal transport system in form of an elevator using a sensor signal from a sensor worn by the maintenance worker.

[0008] Accordingly, there is a need for a method for ensuring a correct maintenance of a personal transport system.

[0009] Such need is met by the subject-matter of the independent claims. Advantageous embodiments are defined in the dependent claims as well as in the following specification and the associated figures.

[0010] According to a first aspect of the technology described herein, a method for ensuring a correct maintenance of a personal transport system is described. The method comprises: receiving a sensor signal from at least one sensor worn by a maintenance worker carrying out a maintenance task at the personal transport system; determining which maintenance task the maintenance worker is actually carrying out from the received sensor signal; determining whether the determined maintenance task is carried out correctly from the received sensor signal; and generating a warning signal for the maintenance worker when the determined maintenance task is not carried out correctly. Additionally, the method comprises: determining which steps have to be performed in which order to carry out the determined maintenance task; determining which step is currently performed by the maintenance worker from the received sensor signal; determining whether the step currently performed by the maintenance worker is the correct step at the right time for the determined maintenance task; and determining that the determined maintenance task is not being carried out correctly when the step currently carried out by the maintenance worker is not the correct step at the right time for the determined maintenance task. Embodiments of the method described herein may be implemented in hardware, software or a combination thereof.

[0011] According to a second aspect of the technology described herein, a computer program is described. The computer program comprises computer-readable instructions, which when being carried out by a processor of a computer cause the computer to carry out the method as described in the foregoing and below. The computer program may be provided in any computer readable language.

[0012] According to a third aspect of the technology described herein, a computer readable medium is described. The computer readable medium comprises the above-mentioned computer program. For example, the computer readable medium may have stored thereon the above computer program. For example, the computer readable medium may be a volatile or non-volatile data memory. For example, the computer readable medium may be a flash memory, a DVD, a CD, a ROM, a RAM, an EPROM or similar devices. Alternatively, the computer readable medium may be part of another computer or server or of a data cloud from which the computer program product may be downloaded for example via a network such as the Internet.

[0013] It shall be noted that the features, advantages and / or effects of the embodiments of the technology described herein are described basically with respect to the above method. One skilled in the art will recognize that these features, advantages and / or effects may be suitably transferred from one embodiment to another and / or from the first aspect to the second and / or third aspects. Therefore, with due regard to a concise description of the present invention, a repetitive explanation of these features, advantages and / or effects is omitted. Further, features of different embodiments may be modified, adapted, combined and / or replaced, etc. in order to come to further embodiments of the technology within the scope of the appended claim described herein.

[0014] The sensor signal may be generated by the sensor and may be sent to the computer carrying out the method. The sensor may be arranged at an equipment of the maintenance worker, e.g., at a safety equipment of the maintenance worker or in a mobile device of the maintenance worker. The safety equipment may be a helmet, a harness, a wristband, etc. The harness may hold the worker, if the worker falls, e.g., into a hoistway of an elevator. The wristband may comprise an electronic circuit, for example an RFID chip or a sender, which may be detected by a sender of the elevator or, respectively, which may send a proximity signal to a corresponding receiver of the elevator. The mobile device may be a mobile phone or a tablet computer. The sensor may be a sensor out of a group of sensors, the group of sensors comprising: a gyroscope, a magnetometer, an acceleration sensor, a barometer, a microphone, and a camera. Optionally, sensor signals of two or more of the sensors may be received and it may be determined from two or more of the received sensor signals which motions the maintenance worker is actually carrying out.

[0015] The computer may be a part of the personal transport system, e.g., a controller of the personal transport system. Alternatively, the computer may be independent from the personal transport system. The method may also be carried out by a system comprising several entities. For example, the sensor signal may be generated by the sensor, the sensor signal may be received by the mobile device of the maintenance worker or by the controller of the personal transport system and then may be forwarded to a remote computer, e.g., a server, in which it is determined whether the maintenance work is carried out correctly or not.

[0016] The determining of which maintenance task the maintenance worker is actually carrying out may also be done by an artificial intelligence or machine learning algorithm, in short summarized as “second Al” in the following. The second Al may be the same or another Al as the first Al. In case of the second Al being the first Al, the corresponding Al may be trained to determine which maintenance task the maintenance worker is actually carrying out and whether the determined maintenance task is carried out correctly. In case of the second Al being different from the first Al, the second Al may be trained in advanced, e.g., by supervised learning. In this context, the second Al may be trained by a second learning data set comprising a number of recorded sensor signals of the sensors worn by maintenance workers when carrying out different maintenance tasks at earlier times, wherein the maintenance tasks of the second learning data set may be labelled before starting the supervised learning. The second Al may be a classifier which has been trained to classify certain data patterns in the sensor signal as being the result of corresponding maintenance tasks carried out by the maintenance worker. The trained ready to use second Al may be stored on the memory of the entity carrying out the method or may be stored in the remote device. Alternatively, instead of determining which maintenance task the maintenance worker is actually carrying out, it may be known in advance which maintenance task the maintenance worker is currently carrying out. Alternatively, the maintenance worker may input in the mobile device which maintenance task will be carried out before carrying out the corresponding maintenance task.

[0017] The determining of whether the determined maintenance task is carried out correctly may be done by an artificial intelligence or machine learning algorithm, in short summarized as “first Al” in the following. The first Al may be trained in advanced, e.g., by supervised learning. In this context, the first Al may be trained by a first learning data set comprising a number of recorded sensor signals of the sensors worn by maintenance workers when carrying out the maintenance task correctly at earlier times, wherein those parts of the first learning data set referring to the maintenance task may be labelled as such before starting the supervised learning. The first Al may be a classifier which has been trained to classify certain data patterns in the sensor signal as being the result of corresponding maintenance tasks being carried out correctly by the maintenance worker. The trained ready to use first Al may be stored on a memory of the entity carrying out the method, or may be stored in the remote device, wherein the remote device may communicate with the computer for sending the classification result to the maintenance worker.

[0018] The warning signal may be sent to an output unit. The output unit may be configured for outputting a feedback or a warning to the maintenance worker upon receiving the warning signal. So, the warning signal may cause the output unit to output a corresponding warning for the maintenance worker. The output unit may comprise a display and the warning may be output in written or drawn form. Alternatively or additionally, the output unit may comprise a speaker and the warning may be output in acoustic form, e.g., in form of speech or sound. The output unit may be arranged at the equipment of the maintenance worker or may be a part of the mobile device of the maintenance worker.

[0019] As a warning, a chatbot can also be launched, which attempts to determine the reasons for the deviations from the correct way of performing the maintenance task in a conversation with the worker and suggests possible solutions. Optionally, a conformation signal may be generated when the determined maintenance task is carried out correctly such that the work of the maintenance worker is confirmed during the maintenance task is carried out correctly.

[0020] The maintenance task may be a maintenance task out of a group of maintenance tasks, the group of maintenance tasks comprising: cleaning the rails and verifying the correct functioning of doors, diagnosing and resolving short circuit issues, and mechanical fixes.

[0021] Every maintenance task may be characterized by a set of steps to be carried out one after the other to fulfil the corresponding maintenance task. For example, a data base may be provided, the data base containing a number of maintenance works and the corresponding steps to be performed for carrying out the maintenance works. The steps may be steps out of a group of steps, the group of steps comprising unlocking a door, e.g., of an elevator, a hoistway or a switchboard of the personal transport system, pressing a button, turning a screwdriver, locking a brake, etc. For example, for the maintenance task of cleaning and rails and verifying the door mechanism the steps 1) switching the elevator to maintenance mode 2) opening the door 3) locking the door 4) cleaning the door 5) pressing the buttons to move the elevator 6) moving the elevator up and down 7) switch the safety button on top of the elevator 8) moving out of the shaft 9) closing the elevator door, may be performed. Alternatively, for the maintenance task of finding short circuit, the steps 1) switch the elevator to maintenance mode 2) open the panel 3) use a jumper cable to diagnose the issue 4) close the panel, may be performed.

[0022] When it is known which maintenance task is actually performed, it may be determined which step has to be performed next by the maintenance worker to proceed with fulfilling the maintenance task. Then, it can be checked whether the step currently carried out by the maintenance worker is the correct step at the correct time for fulfilling the maintenance task.

[0023] That the steps are carried out correctly may refer to a time when the corresponding steps are carried out. For example, opening a door or pressing a button may be correct at one time, but may be not correct at another. Alternatively or additionally, that the steps are carried out correctly may refer to an order in which the motions are to be carried out. For example, it might be correct to open the door after pressing a certain button, but not before pressing the button. Alternatively or additionally, that the steps are carried out correctly may refer to the step itself. For example, for one maintenance task it might be correct to use a screwdriver, whereas for another maintenance task it might be correct to open a maintenance door of the personal transport system.

[0024] The maintenance task may be determined as being carried out correctly when the step which is in fact carried out does correspond to the step which should be carried out. In this case, the confirmation signal may be sent to the maintenance worker.

[0025] A third Al may be used to determining which step is currently carried out by the maintenance worker from the received sensor. The third Al may be the same Al as the first and / or second AIs. The third Al may be trained in advanced, e.g., by supervised learning. In this context, the third Al may be trained by the first, second, or a third learning data set comprising a number of recorded sensor signals of the sensors worn by maintenance workers when carrying out the maintenance tasks and the corresponding steps correctly at earlier times, wherein the maintenance tasks and / or steps may be labelled before starting the supervised learning. The third Al may be a classifier which has been trained to classify certain data patterns in the sensor signal as being the result of corresponding steps carried out in the right order to fulfil the maintenance tasks carried out by the maintenance worker.

[0026] According to one embodiment, it is determined whether the step currently performed by the maintenance worker is the correct step to be performed at the right time for the determined maintenance task or not by: determining a set of probabilities comprising for each step of the determined maintenance task a corresponding probability, wherein the probabilities are representative of how probable it is that the corresponding step has to be carried out currently; determining a probability with which the step currently performed is performed from the set of probabilities; and determining that the step currently performed is not the correct step at the right time for the determined maintenance task when the determined probability for this step is smaller than a predetermined probability threshold.

[0027] According to one embodiment, it is determined whether the step currently performed by the maintenance worker is the correct step to be performed at the right time for the determined maintenance task or not by: determining a set of probabilities comprising for each step of the determined maintenance task a corresponding probability, wherein the probabilities are representative of how probable it is that the corresponding step has to be carried out currently; determining the probability with which the step currently performed is performed from the set of probabilities; and determining that the step currently performed is not the correct step at the right time for the determined maintenance task when the determined probability for this step is not the highest probability of the determined set of probabilities.

[0028] According to one embodiment, the method comprises, when the determined maintenance task is not carried out correctly: determining a severity of a mistake representing the not correctly carried out maintenance task; and generating the warning signal depending on the determined severity. For example, the warning signal may be generated such that it causes a long vibration with a sound for something urgent, e.g., an elevator is not locked while the maintenance worker is doing the maintenance work. In contrast, the warning signal may be generated such that it causes a short vibration for something not urgent, e.g., that the elevator is not locked but the maintenance worker hasn’t started the maintenance work yet.

[0029] In the following, advantageous embodiments of the technology described herein will be described with reference to the enclosed drawings. However, neither the drawings nor the description shall be interpreted as limiting the technology described herein.

[0030] Fig. 1 shows examples of a maintenance worker and a personal transport system.

[0031] Fig. 2 shows a flow chart of a method for ensuring a correct maintenance of the personal transport system, according to an embodiment of the technology described herein.

[0032] Fig. 3 shows a flow chart of a method for determining whether a maintenance task is carried out correctly, according to an embodiment of the technology described herein.

[0033] Fig. 4 shows a flow chart of a method for determining whether a step currently performed by the maintenance worker is the correct step to be performed at the right time for the determined maintenance task or not, according to an embodiment of the technology described herein.

[0034] Fig. 5 shows a flow chart of a method for determining whether a step currently performed by the maintenance worker is the correct step to be performed at the right time for the determined maintenance task or not, according to an embodiment of the technology described herein.

[0035] The figures are only schematic and not to scale. Same reference signs refer to same or similar features.

[0036] Fig. 1 shows examples of a maintenance worker 40 and a personal transport system 20. In the described embodiment, the personal transport system 20 is embodied as an elevator. Alternatively, the personal transport system 20 may be embodied as an escalator or a moving pavement, for example. The elevator comprises a car 22, a hoistway 24, a car holder 30, a cable 32, a motor 34, a controller 35, a first door 36 and a second door 38.

[0037] The personal transport system 20 is arranged in a building (not shown) having several floors. In particular, the building may have a first floor 26 and a second floor 28, wherein in this context the terms “first” and “second” are only used to differentiate one floor from another and may refer to any of the floors of the building. The personal transport system 20 is configured for transporting a load, e.g. one or more people and / or any another load, from one floor to the other, for example from the first floor 26 to the second floor 28 or vice versa.

[0038] The car 22 is held and may be vertically transported within the hoistway by the car holder 30 and the cable 32. The cable 32 may be moved by the motor 34. The motor 34 may be controlled by the controller 35. The car 22 is configured for accommodating the load, i.e. the one or more people and / or the other load. The load and / or the maintenance worker 40 may enter the car 22 through the first door 36, when the car 22 is on the first floor 26, or the second door 38, when the car 22 is on the second floor 28. The maintenance worker 40 may also step on top of the car 22, e.g., during maintenance. For example, the maintenance worker 40 may enter the hoistway 24 through the second door 38 and step on top of the car 22, when the car 22 is on the first floor 26. The maintenance worker 40 may also enter a pit 54 of the hoistway 24 under the lowest one of the floors 36, 38 of the building.

[0039] The maintenance worker 40 wears at least one, preferably several safety devices, in particular a helmet 42, a harness 44, and / or a wristband 46. In addition, the maintenance worker 40 may carry a mobile device (not shown), such as a mobile phone, a smart watch, a notebook, and / or a tablet computer.

[0040] Each of the safety devices 42, 44, 46 may comprise one or more sensors. The sensors of the safety devices 42, 44, 46 may communicate with the controller 35 of the personal transport system and / or with the controller of the mobile device. In addition, the mobile device may also comprise one or more sensors. Further, the mobile device may comprise a controller, an input unit, such as a microphone, keyboard, and / or a touchpad, and / or an output unit, such as a display and / or a speaker. The mobile device may be configured to enter the internet to communicate with a remote server. The mobile device may be configured to communicate with the controller 35 of the personal transport system.

[0041] The sensors may be arranged and configured such that signals of the sensors may be representative for that whether a maintenance task currently carried out by the maintenance worker 40 is carried out correctly, and optionally which maintenance task is currently carried out by the maintenance worker 40. The sensor may be a sensor out of a group of sensors, the group of sensors comprising: a gyroscope, a magnetometer, an acceleration sensor, a barometer, a microphone and a camera.

[0042] Every maintenance task may be characterized by a set of steps to be carried out one after the other to fulfil the corresponding maintenance task. A data base may be provided, the data base containing a number of maintenance tasks and the corresponding steps to be performed for carrying out the maintenance works. The steps may be steps out of a group of steps, the group of steps comprising unlocking a door, e.g., of the elevator, the hoistway 24, or a switchboard of the personal transport system 20; pressing a button; turning a knob or a screwdriver; locking a brake; etc. For example, for the maintenance task of cleaning and rails and verifying the door mechanism the steps 1) switching the elevator to maintenance mode 2) opening the door 3) locking the door 4) cleaning the door 5) pressing the buttons to move the elevator 6) moving the elevator up and down 7) switch the safety button on top of the elevator 8) moving out of the shaft 9) closing the elevator door, may be performed. Alternatively, for the maintenance task of finding short circuit, the steps 1) switch the elevator to maintenance mode 2) open the panel 3) use a jumper cable to diagnose the issue 4) close the panel, may be performed.

[0043] Further, each step may be defined by one or more motions. For example, when using a screwdriver, a hand of the maintenance worker 40 rotates again and again while moving slowly along the axis of rotation. In contrast, when pressing a button, the hand moves in one direction and back only. When the motions are recognized, e.g., by a gyroscope of the smart watch or wrist band, it may be derived which step the maintenance worker 40 is currently performing.

[0044] When it is known which maintenance task is actually performed, it may be determined which step has to be performed next by the maintenance worker 40 to proceed with fulfilling the maintenance task. Then, it can be checked whether the step currently carried out by the maintenance worker 40 is the correct step at the correct time for fulfilling the maintenance task.

[0045] The methods described with respect to figures 2, 3, 4, and 5 in the following may be carried out by any computer communicating with the sensors. For example, the methods may be carried out at least in part by the mobile device of the maintenance worker 40, by the controller 35 of the personal transport system, and / or by a server communicating with the mobile device. Each of the methods may be implemented in hardware, software or a combination thereof.

[0046] Fig. 2 shows a flow chart of a method for ensuring a correct maintenance of the personal transport system 20, according to an embodiment of the technology described herein.

[0047] In a step S2, a sensor signal from at least one of the sensors worn by the maintenance worker 40 carrying out a maintenance task at the personal transport system 20 is received. The sensor signal may be generated by the corresponding sensor and may be sent to the entity carrying out the method. Optionally, sensor signals of two or more of the sensors may be received. In a step S4, it may be determined from the received sensor signal which maintenance task the maintenance worker 40 is actually carrying out. Alternatively, the maintenance task currently carried out is known in advance or may be input by the maintenance worker 40, e.g., via the mobile device, such that the entity carrying out the method already knows in advance which maintenance task is carried out by the maintenance worker currently.

[0048] The determining of which maintenance task the maintenance worker 40 is actually carrying out may be done by an artificial intelligence or machine learning algorithm. The Al may be trained in advanced, e.g., by supervised learning. In this context, the Al may be trained by a learning data set comprising a number of recorded sensor signals of the sensors worn by maintenance workers 40 when carrying out different maintenance tasks at earlier times, wherein the maintenance tasks of the learning data set may be labelled before starting the supervised learning. The Al may be a classifier which has been trained to classify certain data patterns in the sensor signal as being the result of corresponding maintenance tasks carried out by the maintenance worker 40. The trained ready to use Al may be stored on the memory of the entity carrying out the method or may be stored in a remote location, e.g., on the server.

[0049] In a step S6, it may be determined whether the determined maintenance task is carried out correctly from the received sensor signal. In case of two or mor sensor signals being received, it may be determined from the two or more received sensor signals whether the determined maintenance task is carried out correctly. The determining of whether the determined maintenance task is carried out correctly may be done by an artificial intelligence or machine learning algorithm, in short summarized as “first Al” in the following. In this context, the aforementioned Al for determining which maintenance task the maintenance worker 40 is actually carrying out may be referred to as “second Al”.

[0050] The first Al may be trained in advanced, e.g., by supervised learning. In this context, the first Al may be trained by a first learning data set comprising a number of recorded sensor signals of the sensors worn by maintenance workers when carrying out the maintenance task correctly at earlier times, wherein those parts of the first learning data set referring to the maintenance task may be labelled as such before starting the supervised learning. In this context, the learning data set for training the second Al may be referred to as “second learning data set”. The first Al may be a classifier which has been trained to classify certain data patterns in the sensor signal as being the result of corresponding maintenance tasks carried out correctly by the maintenance worker 40. The trained ready to use first Al may be stored on a memory of the entity carrying out the method, or may be stored in the remote location, wherein the corresponding remote server may communicate with the mobile device for sending the classification result to the maintenance worker 40.

[0051] When it is determined that the determined maintenance task is carried out correctly, the method may be repeatedly carried out starting in step S2 again until the maintenance task is completed. Optionally, a conformation signal may be generated when the determined maintenance task is carried out correctly such that the work of the maintenance worker 40 is confirmed during the maintenance task is carried out correctly. When it is determined that the determined maintenance task is not carried out correctly, the method may proceed in optional step S 8 or in step S10.

[0052] In optional step S8, a severity of a mistake representing the not correctly carried out maintenance task may be determined. The severity may be defined by a predetermined set of values representing different grades of the severity, e.g., from “not severe”, over “severe”, to “very dangerous” or “very serious”.

[0053] In a step S 10, a warning signal for the maintenance worker 40 may be generated. The warning signal may be sent to an output unit. The output unit may be configured for outputting a feedback or a warning to the maintenance worker 40 upon receiving the warning signal. So, the warning signal may cause the output unit to output a corresponding warning for the maintenance worker 40. The output unit may comprise a display and the warning may be output in written or drawn form. Alternatively or additionally, the output unit may comprise a speaker and the warning may be output in acoustic form, e.g., in form of speech or sound. The output unit may be arranged at the equipment of the maintenance worker 40 or may be a part of the mobile device of the maintenance worker 40.

[0054] In case of the optional step S8 have been carried out, the warning signal may be generated depending on the determined severity. For example, the warning signal may be generated such that it causes a long vibration with a sound for something urgent, e.g., the maintenance worker 40 forgot to lock the car 22 of the elevator but is already performing other steps of the corresponding maintenance task. In contrast, the warning signal may be generated such that it causes a short vibration for something not urgent, e.g., that the maintenance worker 40 forgot to turn a light off before leaving the hoistway 24.

[0055] The step S6, in which it is determined whether the determined maintenance task is carried out correctly from the received sensor signal, may comprise a sub-routine comprising two or more sub-steps. The sub-routine may be referred to as method for determining whether a maintenance task is carried out correctly and is described in the following.

[0056] Fig. 3 shows a flow chart of a method for determining whether a maintenance task is carried out correctly, according to an embodiment of the technology described herein.

[0057] In a step S12, it may be determined which steps have to be performed in which order to carry out the determined maintenance task. For example, a data base or lookup table may be provided in advance, the data base containing a number of maintenance tasks and the corresponding steps to be performed in the correct order for carrying out the corresponding maintenance task. The data base or, respectively, lookup table may be stored on the mobile device or on the remote server. Then, when carrying out step S12, it might be looked up in the data base which steps have to be performed in which order to carry out the determined maintenance task.

[0058] In a step S14, it may be determined which step is currently performed by the maintenance worker 40 from the received sensor signal. A third Al may be used to determine which step is currently carried out by the maintenance worker 40 from the received sensor signal. The third Al may be the same Al as the first and / or second AIs, or another Al. The third Al may be trained in advanced, e.g., by supervised learning. In this context, the third Al may be trained by the first, second, or a third learning data set comprising a number of recorded sensor signals of the sensors worn by maintenance workers 40 when carrying out the maintenance tasks and the corresponding steps correctly at earlier times, wherein the maintenance tasks and / or steps may be labelled before starting the supervised learning. The third Al may be a classifier which has been trained to classify certain data paterns in the sensor signal as being the result of corresponding steps to be carried out in the right order to fulfil the maintenance tasks carried out by the maintenance worker 40.

[0059] In a step SI 6, it may be determined whether the step currently performed by the maintenance worker 40 is the correct step at the right time for the determined maintenance task, e.g., by comparing the determined step with the steps of the corresponding maintenance task in the data base. For example, it may be determined that the determined maintenance task is not being carried out correctly when the step currently carried out by the maintenance worker 40 is not the correct step at the right time for the determined maintenance task. In this case, the method may proceed in the parent-routine, in particular in steps S8 or S10 explained above. In contrast, the maintenance task may be determined as being carried out correctly when the step which is in fact carried out does correspond to the step which should be carried out. In this case, the method may proceed again in step S14 until the maintenance task in completed. In addition, the confirmation signal may be sent to the maintenance worker 40.

[0060] In general, that the steps are carried out correctly may refer to a time when the corresponding steps are carried out. For example, opening a door or pressing a buton may be correct at one time, but may be not correct at another. Alternatively or additionally, that the steps are carried out correctly may refer to an order in which the motions are to be carried out. For example, it might be correct to open the door after pressing a certain buton, but not before pressing the buton. Alternatively or additionally, that the steps are carried out correctly may refer to the step itself. For example, for one maintenance task it might be correct to use a screwdriver, whereas for another maintenance task it might be correct to open a maintenance door of the personal transport system.

[0061] The step S 16, in which it is determined whether the step currently performed by the maintenance worker 40 is the correct step at the right time for the determined maintenance task, may comprise a sub-routine comprising two or more sub-steps. The sub-routine may be referred to as method for determining whether a step currently performed by the maintenance worker is the correct step to be performed at the right time for the determined maintenance task or not, and is described in the following with respect to figure 4 or, as an alternative, with respect to figure 5. Fig. 4 shows a flow chart of a method for determining whether a step currently performed by the maintenance worker 40 is the correct step to be performed at the right time for the determined maintenance task or not, according to an embodiment of the technology described herein.

[0062] In a step S22, a set of probabilities comprising for each step of the determined maintenance task a corresponding probability may be determined. The probabilities may be representative of how probable it is that the corresponding step has to be carried out currently.

[0063] In a step S24, a probability with which the step currently performed is performed is determined from the set of probabilities.

[0064] In a step S26, it may be determined whether the probability with which the step currently performed is performed is smaller or higher than a predetermined probability threshold. In particular, it may be determined that the step currently performed is not the correct step at the right time for the determined maintenance task when the determined probability for this step is smaller than the predetermined probability threshold. In this case, the method may proceed in step S8 or S10 of the parent-routine explained above. In contrast, it may be determined that the step currently performed is the correct step at the right time for the determined maintenance task when the determined probability for this step is higher than the predetermined probability threshold. In this case, the method may proceed in step S34 until the maintenance task is completed.

[0065] Fig. 5 shows a flow chart of an alternative method for determining whether a step currently performed by the maintenance worker 40 is the correct step to be performed at the right time for the determined maintenance task or not, according to an embodiment of the technology described herein.

[0066] Step S32 may correspond to step S22.

[0067] Step S34 may correspond to step S24. In a step S36, it may be determined that the step currently performed is not the correct step at the right time for the determined maintenance task when the determined probability for this step is not the highest probability of the determined set of probabilities. In contrast, it may be determined that the step currently performed is the correct step at the right time for the determined maintenance task when the determined probability for this step is the highest probability of the determined set of probabilities.

[0068] One or more of the methods described in the foregoing may be embodied as a computer program. The computer program may comprise computer-readable instructions, which when being carried out by a processor of a computer, e.g., of the mobile device or the server, cause the corresponding entity to carry out the corresponding method. The computer program may be provided in any computer readable language.

[0069] The computer program may be stored on a computer readable medium. The computer readable medium may comprise the above-mentioned computer program. For example, the computer readable medium may have stored thereon the above computer program. For example, the computer readable medium may be a volatile or non-volatile data memory. For example, the computer readable medium may be a flash memory, a DVD, a CD, a ROM, a RAM, an EPROM or similar devices. Alternatively, the computer readable medium may be part of another computer or server or of a data cloud from which the computer program product may be downloaded for example via a network such as the Internet.

[0070] Finally, it should be noted that the term “comprising” does not exclude other elements or steps and the “a” or “an” does not exclude a plurality. Also elements described in association with different embodiments may be combined. For example, one, two, three or all of the embodiments of the elevator 20 shown in figures 1 to 4 may be combined in or may be parts of the same elevator 20. Further, each of these embodiments of the elevators 20 may be combined with one, two or all of the above methods within the scope of the appended claim. Further, two or all of the above methods may be combined within the scope of the appended claim. For example, the method explained with respect to figure 7 may be implemented into the method explained with respect to figure 6. Alternatively or additionally, the methods explained with respect to figures 6 and 7 may combined with the method explained with respect to figure 8 in so far that the methods embody a communication between the controller 35 of the elevator 20 and one or more of the controllers of the safety devices 42, 44, 46. It should also be noted that reference signs in the claims should not be construed as limiting the scope of the claims.

Claims

Claims:

1. A method for ensuring a correct maintenance of a personal transport system (20), the method comprising:- receiving a sensor signal from at least one sensor (42, 44, 46) (S2);- determining which maintenance task the maintenance worker (40) is actually carrying out from the received sensor signal (S4);- determining whether the determined maintenance task is carried out correctly from the received sensor signal (S6); and- generating a warning signal for the maintenance worker (40) when the determined maintenance task is not carried out correctly (S10), characterized in that the named at least one sensor (42, 44, 46) is worn by a maintenance worker (40) carrying out a maintenance task at the personal transport system (20) and the determining whether the determined maintenance task is carried out correctly (S6) comprises:- determining which steps have to be performed in which order to carry out the determined maintenance task (S12);- determining which step is currently performed by the maintenance worker (40) from the received sensor signal (S14);- determining whether the step currently performed by the maintenance worker (40) is the correct step at the right time for the determined maintenance task (S16); and- determining that the determined maintenance task is not being carried out correctly when the step currently carried out by the maintenance worker (40) is not the correct step at the right time for the determined maintenance task.

2. Method according to claim 1, wherein it is determined whether the step currently performed by the maintenance worker (40) is the correct step to be performed at the right time for the determined maintenance task or not by:- determining a set of probabilities comprising for each step of the determined maintenance task a corresponding probability (S22), wherein the probabilities are representative of how probable it is that the corresponding step has to be carried out currently;- determining a probability with which the step currently performed is performed from the set of probabilities (S24); and- determining that the step currently performed is not the correct step at the right time for the determined maintenance task when the determined probability for this step is smaller than a predetermined probability threshold.

3. Method according to claim 1, wherein it is determined whether the step currently performed by the maintenance worker (40) is the correct step to be performed at the right time for the determined maintenance task or not by:- determining a set of probabilities comprising for each step of the determined maintenance task a corresponding probability, wherein the probabilities are representative of how probable it is that the corresponding step has to be carried out currently (S32);- determining the probability with which the step currently performed is performed from the set of probabilities (S34); and- determining that the step currently performed is not the correct step at the right time for the determined maintenance task when the determined probability for this step is not the highest probability of the determined set of probabilities (S36).

4. Method in accordance with one of the preceding claims, when the determined maintenance task is not carried out correctly comprising:- determining a severity of a mistake representing the not correctly carried out maintenance task (S8); and- generating the warning signal depending on the determined severity (S 10).

5. A computer program comprising computer-readable instructions, which when being carried out by a processor of a computer (50) cause the computer (50) to carry out the method in accordance with one of the preceding claims.

6. A computer readable medium comprising the computer program according to claim 5.

Images