Method and device for computer-implemented determination of a means of transport used by a person during a journey

Abstract

A method for computer-implemented automatic determination of one or more means of transport used by a person during a journey, comprising at least one drive component which rotates during the journey and which comprises or consists of metal, wherein the following steps are carried out at a time after the person has ended his journey: i) Obtaining a magnetometer signal (MMS) comprising the temporal progression of a frequency-dependent signal energy of a distortion of an earth's magnetic field caused by the movement of the means of transport(s) while the person is using the means of transport(s), wherein the magnetometer signal (MMS) comprises a magnetic flux density detected by a magnetometer (MM) caused by the movement of the respective means of transport, wherein the magnetometer (MM) is worn on the body by the person during the journey; characterized by the step: ii) determining one or more means of transport used by the person during the journey by exclusively processing the magnetometer signal (MMS) by a trained data-driven model (MO), wherein the magnetometer signal (MMS) is fed as a digital input to the trained data-driven model (MO) and the trained data-driven model (MO) provides the means of transport (FM) used as a digital output.

Description

[0001] The invention relates to a method and a device for the computer-implemented automatic determination of one or more means of transport used by a person during a journey, which comprises at least one drive component which rotates during the movement and which comprises or consists of metal.

[0002] A means of transport within the scope of the present invention is understood to mean any means of transport suitable for transporting a person. The means of transport can be motorized or non-motorized. Means of transport within the meaning of the present invention include rail-based means of transport (trains, commuter trains, subways, trams) as well as non-rail-based means of transport, such as buses, passenger cars, trucks, motorcycles, e-scooters, bicycles, pedelecs, e-bikes, boats (ferries, and the like), etc. Elevators and escalators are also understood to be means of transport.

[0003] A means of locomotion of this type can have any number and type of drive mechanism, such as wheels, propellers in boats, gears in elevators or escalators. These represent rotating drive components that are caused to rotate by a drive during the movement of the means of locomotion, or they are passively rotating components.

[0004] For general transport research, but also for the provision of an automated, mobile ticketing system, it would be desirable if the means of transport used by a person during a journey could be determined automatically.

[0005] A mobile ticketing system known as FAIRTIQ is based on the use of information from a global navigation satellite system (GNSS) to automatically charge a user for a journey undertaken. Using GNSS, it is possible to determine the position and speed of the means of transport and thus its ground speed, i.e. the speed at which the means of transport moves over the ground. This requires manual user interaction to check in and out at the start and end of the journey. Check-in and check-out are carried out by the user via an application (so-called app) installed on a personal user device, e.g. a smartphone. Based on GNSS data from the personal user device, the lowest possible price is then calculated for the means of transport used.

[0006] A disadvantage of FAIRTIQ is that only above-ground means of transport, such as trains, buses, trams, and boats, can be taken into account, as the detection of a means of transport used is based on determining the position from GNSS data. Using the GNSS data recorded during the journey, the system is able to distinguish between different above-ground means of transport. If the person is not moving on a known trajectory that can be assigned to a means of transport from the GNSS data, a car or bicycle is assumed as the means of transport and is not counted. Underground means of transport, such as subways, cannot be taken into account, as it is impossible to determine the position and / or speed of a vehicle inside tunnels using GNSS.

[0007] In the Moving Lab project, mobility data collected by personal user devices, such as smartphones or wearables, using position and motion sensors, is processed for traffic research purposes. A person's exact position is determined several times per minute using GNSS. When a change in position is detected, acceleration profiles are generated that indicate the use of a specific mode of transport. The system calculates the route traveled by a person, including all modes of transport used, transfer points, and stops. This system requires the user to carry their personal user device with them and actively activate data collection. Interesting route features, such as a start time, an end time, route progress, or the mode of transport used, are then automatically determined based on GNSS information and information from an acceleration sensor.To improve the reliability of the collected information, the user is asked after each trip to select the mode of transport used via an interactive tool, e.g., an app on their personal device. This information is then used to analyze the collected route data and observed mobility behavior.

[0008] The methods known from the state of the art therefore require, on the one hand, interaction with the person undertaking a journey using one or more means of transport and / or are not suitable for all means of transport available for passenger transport, in particular underground means of transport.

[0009] DE 10 2018 204 863 A1 discloses a method for automatically determining the means of transport used by a person by evaluating distortions of the Earth's magnetic field detected by a magnetometer during use of the means of transport(s). After extracting magnetic and acceleration characteristics, a classification is performed. This serves to recognize and analyze human activities in order to use them specifically in personal healthcare applications, life coaching applications, or recommendation systems. For example, it is proposed to use the recognition of a mode of transport to make a suggestion, such as getting off the bus one stop earlier to increase physical activity.

[0010] US 2016 / 0 379 141 A1 discloses a method for automatically determining travel sections using a magnetometer and other sensors, wherein user inputs are processed.

[0011] DE 10 2017 222 924 A1 discloses a method for determining the speed of a means of transport having at least one metal wheel. During the movement of the means of transport, the wheel rotates. A signal comprising a frequency-dependent signal energy of a distortion of the Earth's magnetic field caused by the movement of the means of transport is evaluated. The speed of the distortion is determined from the frequency of occurrence of a distortion of the Earth's magnetic field caused by the rotation of the at least one wheel.

[0012] EP 3 614 158 A1 discloses a method for determining a transport mode that collects magnetometer and speed data from a mobile device, correlates the magnetometer with the speed data in groupings, and performs a spectral analysis of the groups of magnetometer data. The energy calculated for each of the frequency components determined from the spectral analysis is compared with a base value to generate a difference. Based on this difference, the vehicle is assigned a transport mode type.

[0013] It is an object of the invention to provide a method and a device which enable an automatic, computer-implemented determination of one or more means of transport used by a person during a journey.

[0014] This object is achieved by a method according to the features of claim 1, a device according to the features of claim 12, a computer program product according to the features of claim 14 and a computer program according to the features of claim 15. Advantageous embodiments emerge from the dependent claims.

[0015] According to a first aspect of the invention, a method for the computer-implemented automatic determination of one or more means of transportation used by a person during a journey is proposed. The means of transportation comprises at least one drive component that rotates during movement and comprises or consists of metal.

[0016] The means of transport is preferably, but not necessarily, a motorized means of transport. The means of transport can be rail-based, such as a train, a commuter train, a subway, a tram, or a non-rail-based, such as a bus, a passenger car, an e-scooter, a bicycle, a pedelec, an e-bike, or a watercraft. While in rail-based and non-rail-based means of transport the rotating drive component is a wheel, in a watercraft the rotating drive component is, for example, a drive screw. Means of transport within the meaning of the invention also include elevators or escalators, in which rotating deflection pulleys and / or drive wheels represent a rotating drive component within the meaning of the invention.

[0017] A journey within the meaning of the invention can, for example, be the successive use of different means of transport to travel from a starting point of the journey (location A) to a destination point of the journey (location B). Individual or all means of transport may be subject to a fee, with the amount of the transport fee depending on the type of transport and / or the route and / or the length of the journey. Individual or all means of transport may also be free of charge.

[0018] According to the invention, the following steps are carried out at a time after the person has completed his or her journey, i.e. after reaching the destination point of the journey:

[0019] In step i), a magnetometer signal is obtained. In the following, the term "magnetometer signal" refers to a digital signal. The digitization of the magnetometer signal can be carried out after receipt from a signal source. The term "obtaining a magnetometer signal" means that the magnetometer signal is received by a processor that carries out the method of the present invention. The magnetometer signal comprises the temporal profile of a frequency-dependent signal energy of a distortion of an earth's magnetic field caused by the movement of the means of transport(s) while the person is using the means of transport(s). The magnetometer signal comprises a magnetic flux density detected by a magnetometer, which is caused by the movement of the respective means of transport. The magnetometer is worn on the body by the person during the (entire) journey.The term “during the journey” means that the magnetometer is carried by the person at least from the starting point of the journey (location A) to the destination point of the journey (location B).

[0020] The magnetometer signal is received by a sensor unit containing a magnetometer. A magnetometer is a sensor device for measuring magnetic flux densities. A Hall sensor, for example, can be used as a magnetometer. In principle, all known types of magnetometers can be used, provided they are suitable for being worn on the body. The sensor unit transmits the magnetometer signal, which includes the frequency-dependent signal energy of the distortion of the Earth's magnetic field caused by the movement of the vehicle, and the magnetometer signal is received at the interface for further processing by the processor to implement the method.

[0021] In step ii), one or more means of transport used by the person during the trip is determined by exclusively processing the magnetometer signal with a trained data-driven model. The magnetometer signal is fed into the trained data-driven model as a digital input, and the trained data-driven model provides the means of transport used as a digital output.

[0022] The method according to the present invention provides a simple method for automatically determining one or more means of transport used by a person during a journey. A trained data-driven model is used for this purpose. The model has been trained using training data, whereby either a supervised or unsupervised machine learning method can be used. In a supervised machine learning method, a plurality of magnetometer signals are processed as training data, together with information regarding the type of transport to which the temporal profile of the frequency-dependent signal energy of the distortion caused by the movement of the transport means belongs.In unsupervised machine learning, a large number of magnetometer signals from different means of transport are processed as training data, which are then automatically divided into clusters that can then be assigned to specific means of transport.

[0023] In principle, any known data-driven model trained with a machine learning method can be used in this procedure, such as neural networks, support vector machines, decision trees.

[0024] The method according to the invention utilizes the presence of the Earth's magnetic field, which causes a detectable distortion due to the movement of the means of transport, in particular due to the rotation of one or more drive components, and leads to a magnetic signature characteristic of a particular means of transport. This magnetometer signal is evaluated using the data-driven model to determine the type of transport means(s) used. In particular, a frequency curve of the occurrence of the distortion caused by the rotation of at least one drive component is determined. In other words, the temporal change in the frequency is taken into account.This can be used by the data-driven model to identify, based on logical criteria, the frequency in the signal that can be clearly assigned to the occurrence of the distortion caused by the rotation of at least one drive component.

[0025] One advantage of this approach is that the proposed method eliminates the need to process GNSS signals to determine the means of transport used. This allows the method to be applied in various environments, particularly urban areas and tunnels, without any degradation in quality. The method is therefore very robust with regard to the environment in which the person using the means of transport moves.

[0026] According to an expedient embodiment, the steps described above can be performed not only at a time after the person's journey has ended, but also at one or more times during the journey, in particular after receiving user information that a journey segment has ended. The user information can be generated, for example, as a result of user interaction with a personal user device, e.g., a smartphone or portable computer.

[0027] According to a further expedient embodiment, information based on the means of transport used is output via an interface. This information can, for example, be an identifier or classifier representing the means of transport used. Alternatively, the information about the means of transport used can be output directly. The information can then be further processed, e.g., in an automated ticketing system that does not require user interaction from the person using the means of transport, or for general mobility research.

[0028] It is also expedient for the magnetometer to be worn by the person close to the ground, in particular on or in a shoe or on the ankle. Generally, the magnetometer is intended to be arranged in spatial proximity to one of the rotating components to detect the distortion of the Earth's magnetic field caused by the movement of the means of locomotion. This embodiment is based on the finding that the closer the sensor unit is arranged to the at least one rotating drive component, the easier it is to determine the frequency of occurrence of the distortion of the Earth's magnetic field caused by the rotation of the rotating drive component from the magnetometer signal.

[0029] A further advantageous embodiment provides for the determination of the means of transport used during the trip to be carried out by a processing unit worn by the person. The processing unit can be a standalone, wearable device separate from the magnetometer. The processing unit can be implemented in a personal user device, such as a smartwatch, a fitness tracker, a smartphone, and the like.

[0030] Alternatively, the means of transport used during the journey can be determined by a central processing unit, particularly one located in the cloud. For this purpose, the magnetometer signal is transmitted directly to the central processing unit or via the aforementioned personal user device. The magnetometer signal can be transmitted immediately after receipt, provided a communication connection exists, or at a later time when a communication connection can be established.

[0031] Another useful design provides that the data-driven model also provides the duration of use of a respective means of transport as a digital output.

[0032] According to a further expedient design, the following steps are carried out: In step iii), several pieces of location information about the person during the trip are obtained. Each piece of location information comprises location information determined by a positioning unit and associated time information. According to a suitable embodiment, the position information is determined by a personal item, in particular a mobile device or a pocket computer, or a wearable device belonging to the person. It can also be determined by a different explicit positioning unit, which includes, for example, a GNSS receiver.

[0033] In step iv), locations visited during the journey are determined. The locations include a first location where the person began using one of the means of transport and a second location where the person ended using that means of transport. The visited locations are determined by linking the location information and the digital output of the means of transport(s). This makes it possible to determine locations visited, in particular stops or transfer stations, in addition to the means of transport used. This enables improved analysis of mobility behavior or automated use in a ticketing system.

[0034] In particular, the processing comprises the following steps: determining a first time stamp from the magnetometer signal at which the start of use of the means of transport is determined; determining a second time stamp from the magnetometer signal at which the end of this means of transport is determined; determining the location information as the first location whose time information corresponds to the first time stamp; and determining the location information as the second location whose time information corresponds to the second time stamp.

[0035] In addition to the method described above, the invention relates, according to a second aspect, to a device for the computer-implemented automatic determination of one or more means of transport used by a person during a journey, wherein the device comprises a processor as a processing unit which is configured to carry out the invention according to one or more preferred embodiments of the method.

[0036] Furthermore, the invention relates to a computer program product with program code stored on a non-volatile machine-readable medium and configured to carry out a method according to one or more preferred embodiments when the program code is executed on a computer.

[0037] Finally, the invention relates to a computer program with a program code for carrying out a method according to one or more preferred embodiments when the program code is executed on a computer.

[0038] The invention is explained in more detail below using exemplary embodiments. They show: Fig. 1 shows a first embodiment of a device according to the invention with an internal magnetometer for determining one or more means of transport used during a journey; Fig. 2 shows a second embodiment of a device according to the invention with an external magnetometer for determining the means of transport used during a journey; Fig. 3 a frequency-time diagram showing, as an example, a frequency-dependent signal energy detected by the magnetometer during movement on foot and by bicycle; Fig. 4 a frequency-time diagram showing, by way of example, a frequency-dependent signal energy detected by the magnetometer when moving by tram, on foot and by subway; Fig. 5 a frequency-time diagram showing, by way of example, a frequency-dependent signal energy detected by the magnetometer during movement by a bus; Fig. 6 a frequency-time diagram showing, by way of example, a frequency-dependent signal energy detected by the magnetometer when moving on foot in the vicinity of a railway station; Fig. 7 a frequency-time diagram showing, by way of example, a frequency-dependent signal energy detected by the magnetometer during movement on a motorcycle; Fig. 8 is a frequency-time diagram showing, by way of example, a frequency-dependent signal energy detected by the magnetometer during movement in a motor vehicle; and Fig. 9 a schematic representation of a flow diagram of the method according to the invention for automatically determining one or more means of transport used by a person during a journey.

[0039] Fig. Figure 1 shows an exemplary embodiment of a device 10 according to the invention for the computer-implemented automatic determination of one or more means of transport used by a person during a journey, which means or means are not shown in detail in the figures. The means of transport has at least one drive component made of or including metal, e.g., a wheel, a propeller, or a pulley, that rotates during travel.

[0040] A means of transport is any means of transport suitable for the transport of at least one person, whether non-motorized or motorized. This includes rail-based means of transport, such as a train, a commuter train, a subway, a tram, as well as non-rail-based means of transport, such as a bus, a passenger car, a motorcycle, a bicycle, a pedelec, an e-bike, a truck, but also boats, such as ferries, as well as elevators and escalators.

[0041] In the Fig. The device 10 shown in Figure 1 comprises an interface 11 for receiving a magnetometer signal MMS from a sensor unit 20, a processing unit 12 for processing the information received in the magnetometer signal MMS, and an optional transmission interface 13 in a common housing. The sensor unit 20 comprises a magnetometer MM and, if appropriate, additional sensors.

[0042] The device 10 represents, for example, a so-called wearable, i.e., a portable computer system that is attached to the body of the person (also referred to as the user in this description) during use, as is known, for example, from smartwatches or activity trackers. The device 10 can also be a personal item of use or be integrated into it. The personal item of use is, for example, a mobile device or a pocket computer.

[0043] The Fig. The device shown in Figure 1 is designed to be attached to the body, preferably to a leg, or to or in a shoe of the person. For this purpose, the device 10 comprises corresponding fastening means, e.g., in the form of one or more loops, clips, hook-and-loop fasteners, and the like. If the device 10 is to be attached to or in the person's shoe, it can be attached, for example, to a shoelace or a special cavity provided on or in the shoe. Such cavities are known, for example, from sports shoes for accommodating step sensors.

[0044] The attachment of the device 10 to a leg or shoe of the person enables a simplified evaluation of the magnetometer signal MMS received by the interface 11, since the information contained in the magnetometer signal MMS, from which the means of transport used during a journey is determined, is contained in greater detail the closer the sensor unit 20 of the device 10 is arranged to the area of ​​the rotating component of the means of transport.

[0045] The Fig. The variant shown in 2 differs from the one in Fig. 1 in that the sensor unit 20, which comprises the magnetometer MM, is designed as a separate unit from the device 10 for determining the means of transport used by the person during a journey. Fig. 2, it is therefore sufficient if only the sensor unit 20 is attached to a leg or to or in the shoe of the person. For this purpose, only the sensor unit 20 has the necessary fastening means. The device 10, which is separate from the sensor unit 20, can, in contrast, be present as a separate component or integrated into a wearable or personal item of use, such as a mobile device or a pocket computer, of the person. The device 10 and the sensor unit 20 are designed such that the signal emitted by the sensor unit can be received at the interface 11 of the device 10. The signal can be transmitted via a short-range wireless communication channel, e.g., Bluetooth, Zigbee, and so on. The interfaces of the device 10 and the sensor unit 20 required for this purpose are shown in Fig. 2 is not explicitly shown.

[0046] The optional transmission interface 13 is used to output the preprocessed or unprocessed magnetometer signal MMS. For this purpose, the magnetometer signal MMS can be transmitted to a processor, e.g., a central processing unit in the cloud, which executes the method described in detail below for the computer-implemented automatic determination of one or more means of transport used by the person during a trip. Alternatively, the method for determining the means of transport used during a trip can be carried out by the processing unit 12 of the device 10.

[0047] The following explanations, which explain the method for determining the means of transport used by the person wearing the device 10 or the sensor unit 20 during a journey, apply to both, in the Fig. 1 and Fig. 2, embodiments of the device 10. These statements also apply regardless of whether the method is carried out by the processing unit 12 of the device 10 or a central processing unit, e.g., a computer located in the cloud. Therefore, reference is made below to a processor PR, which can optionally be represented by the processing unit 12 of the device 10 or a central processing unit.

[0048] The processor PR (see Fig. 9) The magnetometer signal MMS received by the sensor unit 20 or its magnetometer MM comprises the temporal progression of a frequency-dependent signal energy of a distortion of the Earth's magnetic field caused by the movement of the means of transport(s) during the journey while the person is using the means of transport(s). The magnetometer signal MMS comprises a magnetic flux density detected by the magnetometer MM, which is caused by the movement of the respective means of transport. In order to be able to determine such a distortion of the Earth's magnetic field with the aid of the magnetometer MM, it is necessary that at least one rotating component of the respective means of transport comprises or consists of a metal.

[0049] The means of transport used by the person during the journey are determined by the processor PR by processing the magnetometer signal MMS with a trained data-driven model MO. The magnetometer signal MMS is fed into the trained data-driven model MO as a digital input. The trained data-driven model MO then provides the means of transport used FM as a digital output, see Fig. 9.

[0050] The data-driven model MO was trained using training data. Training can be performed using either a supervised or an unsupervised machine learning method. In a supervised machine learning method, a large number of magnetometer signals (MMS) are used as training data, along with information about the type of vehicle to which the temporal profile of the frequency-dependent signal energy of the distortion caused by the vehicle's movement belongs. In an unsupervised machine learning method, a large number of training data sets, each comprising magnetometer signals from one or more different vehicles, are divided into clusters, which can then be assigned to specific vehicles.Since the training of a data-driven model using a supervised or unsupervised machine learning method is known in principle to the expert, a detailed description is omitted here.

[0051] In principle, any known data-driven model MO trained with a machine learning method can be used in this procedure.

[0052] The steps described above are performed at least after the person has completed their journey in order to be able to determine all means of transport used during the journey. Additionally, the steps described above can be performed at one or more times during the journey. This is particularly useful when the user enters user information, e.g., into a personal user device, to signal that a journey segment has ended.

[0053] In order to enable further processing of the digital output, which includes the means of transport FM(s) used during a journey, it may be expedient to output information based on the means of transport FM(s) used via a UI interface. The information may specify the means of transport used or, for example, be an identifier or classifier representing the means of transport. The information can be used, on the one hand, to verify the correct identification of the means of transport used by the user and, on the other hand, for further processing, e.g., within the framework of an automated ticketing system.

[0054] In the following described Fig. Figures 3 to 8 show various frequency-time diagrams, each showing a frequency-dependent signal energy recorded by the magnetometer MM during locomotion with different means of locomotion as a respective magnetometer signal MMS. The frequency-time diagrams were recorded with an MTw sensor from Xsens, which was attached to the subject's shoe, at a frequency of 100 Hz.

[0055] The total magnetometer signal MMS is given for each time point t by a specific frequency-dependent signal energy SE in the range from 0 to -80. The frequency-dependent signal energy SE can be visualized, for example, by colors, as shown by the orientation bar displayed alongside the diagrams and backed by a color gradient.

[0056] Fig. Figure 3 shows the frequency-time curve (ft curve), referred to as the magnetic signature, of the two active modes "running" and "cycling." The magnetic signature for "running" is visible in the time range between t = 300s and t = 440s. The person then continued to move on a bicycle. This can be seen in the magnetic signature in the time ranges t = 480s and t = 700s. The frequency-time curve allows the energy of the various "means of transport" to be determined from the measured magnetometer signal. The frequencies displayed at the frequency f ≈ 0 Hz represent the Earth's magnetic field. The components whose signal energy SE lies in the range from -25 to -80 represent changes in the measured Earth's magnetic field.

[0057] Changes in the measured magnetic field during "walking" are linked to the movement of the person's leg. Therefore, the corresponding frequencies correspond to a walking frequency. The changes in the Earth's magnetic field detected by the MM magnetometer during locomotion by the bicycle have two different origins. On the one hand, they are caused by the person's leg approaching and moving away from the metal parts of the bicycle frame. This corresponds to the cadence when cycling. On the other hand, variations in the measured Earth's magnetic field are detected by the metal parts of the bicycle. These components correspond to the frequency of the bicycle's rotational speed.

[0058] Fig. Figure 4 shows the magnetic signature recorded for a person while using a tram, a walking track, and a subway. The magnetic signature for the tram can be seen in the frequency-time diagram from t = 1500s to t = 1800s. At the latter time, the person leaves the tram and walks towards the subway station. During this walking passage, two stops at signaling systems can be detected. The person then enters the subway station using two escalators in the time range t = 2480s and t = 2550s until reaching the subway platform. At time t = 2550s, the person enters the subway and exits at time t = 3100s.

[0059] The location at which a person changes their means of transport or takes a break, e.g., due to a red traffic light, can be determined, for example, by combining the processed magnetometer signal MMS with position information. Each position information item comprises location information determined by a positioning unit, e.g., the personal user terminal, as well as associated time information. Locations visited during the journey, such as a bus stop, traffic lights, and the like, can then be determined by linking the location information and the digital output of the means of transport used. For this purpose, it is assumed that the means of transport has a starting location and an end location (destination). The starting location is referred to as the first location at which the person began using the means of transport.The end location is referred to as the second location at which the person ended the use of this means of transport. The linking processing comprises the steps of determining a first timestamp from the magnetometer signal at which the start of the use of the means of transport is determined, determining a second timestamp from the magnetometer signal at which the end of the use of this means of transport is determined, determining the location information as the first location whose time information corresponds to the first timestamp, and determining the location information as the second location whose time information corresponds to the second timestamp.

[0060] Fig. Figure 5 shows a magnetic signature recorded by a person using a bus. The use of the bus begins at time t = 100 s and ends at time t = 350 s. The person traveled on foot for the remainder of the journey until time t = 840 s, which can be determined from the magnetic signature.

[0061] Fig. Figure 6 shows the magnetic signature of a person walking to a train station. At time t = 2450 s, the person reaches the train station. An indicator in the magnetic signature can be detected at a frequency of f = 16⅔ Hz, which is the frequency used for high-voltage lines in long-distance passenger transport in German-speaking countries. This frequency can also be detected in the magnetometer signal MMS when the person is standing on the platform and not moving, e.g., while waiting for the next train. Fig. 6 shows that the person leaves the station at time t = 3500s.

[0062] Fig. 7 shows the magnetic signature of a motorcycle. Fig. Figure 8 shows the magnetic signature while the person is driving a motor vehicle. In both versions, the person wears the magnetometer while driving.

[0063] The described method can be used for a variety of commercial applications. The method can be used for intelligent e-ticketing systems that automatically detect the means of transport and the distance traveled to charge the person the correct price.

[0064] The process can be used as an additional information source for personal user devices. User devices typically adapt user preferences to the environment. For example, calendar notifications, incoming calls, or emails are forwarded to the user depending on the mode of transportation the user is using. For example, forwarding can be made dependent on whether the user is traveling by bicycle (fewer forwardings due to the risk of distraction), on a tram, or in an airplane.

[0065] Fitness trackers can accurately estimate the user's daily movement by obtaining additional information about the modes of transport used by the user.

[0066] The invention can be integrated into electronic tracking devices used to monitor individuals on bail or parole. The method can provide information about the individual's movement patterns or help trigger an alarm, for example, when certain means of transport are used.

[0067] The method can also be used for publicly funded research projects, such as the “Moving Lab” project described in the introduction, to obtain reliable data and statistics on the means of transport preferred by citizens. List of reference symbols FM Means of transport MMS magnetometer signal MM Magnetometer MO data-driven model POS position information PR Processor SE signal energy UI interface 10 Device 11 Interface 12 processing unit 13 Transmission interface 20 sensors

Claims

[1] A method for the computer-implemented automatic determination of one or more means of transport used by a person during a journey, comprising at least one drive component which rotates during the journey and which comprises or consists of metal, wherein the following steps are carried out at a time after the end of the person's journey: i) Obtaining a magnetometer signal (MMS) comprising the temporal course of a frequency-dependent signal energy of a distortion of an earth's magnetic field caused by the movement of the means of transport(s) while the person is using the means of transport(s), wherein the magnetometer signal (MMS) comprises a magnetic flux density detected by a magnetometer (MM) caused by the movement of the respective means of transport, wherein the magnetometer (MM) is worn on the body by the person during the journey; characterized by the step: ii) determining one or more means of transport used by the person during the journey by exclusively processing the magnetometer signal (MMS) by a trained data-driven model (MO), wherein the magnetometer signal (MMS) is fed as a digital input to the trained data-driven model (MO) and the trained data-driven model (MO) provides the means of transport (FM) used as a digital output. [2] Method according to claim 1, characterized by that steps i) and ii) are carried out at one or more times during the journey, in particular after the user has received information that a journey section has ended. [3] Method according to claim 1 or 2, characterized by that the trained data-driven model (MO) is a supervised or unsupervised machine learning method. [4] Method according to one of the preceding claims, characterized bythat information based on the means of transport (FM) used is output via an interface. [5] Method according to one of the preceding claims, characterized by that the magnetometer (MM) is worn by the person close to the ground, particularly on or in a shoe or on the ankle. [6] Method according to one of claims 1 to 5, characterized by that the determination of the means of transport (FM) used during the journey is carried out by a processing unit worn by the person on his or her body. [7] Method according to one of claims 1 to 5, characterized by that the determination of the means of transport (FM) used during the journey is carried out by a central computing unit, in particular one located in the cloud. [8] Method according to one of the preceding claims, characterized bythat the data-driven model (MO) also provides the duration of use of a respective means of transport as a digital output. [9] Method according to one of the preceding claims, characterized by that the next steps are carried out: iii) obtaining a plurality of location information items (POS) of the person during the journey, wherein each location information item (POS) comprises at least one item of location information determined by a positioning unit and an associated time information item; iv) determining locations visited during the journey, comprising a first location where the person started using one of the means of transport (FM) and a second location where the person ended using that means of transport, by linking the location information (POS) and the digital output of the means of transport used. [10] Method according to claim 9, characterized by that the processing includes the following steps: - the determination of a first time stamp from the magnetometer signal (MMS) at which the start of the use of the means of transport (FM) is determined, - the determination of a second time stamp from the magnetometer signal (MMS) at which the end of the use of this means of transport (FM) is determined, - the determination of the location information as the first location whose time information corresponds to the first timestamp, and - the determination of the location information as a second location, whose time information corresponds to the second timestamp. [11] Method according to claim 9 or 10, characterized by that the position information (POS) is determined by a personal item, in particular a mobile device or a pocket computer, or a wearable device of the person. [12] A device for the computer-implemented automatic determination of one or more means of transport used by a person during a journey, comprising at least one drive component which rotates during the journey and which comprises or consists of metal, the device comprising as a processing unit a processor (PR) which is configured to carry out the following steps at a time after the end of the person's journey: i) Obtaining a magnetometer signal (MMS) comprising the temporal course of a frequency-dependent signal energy of a distortion of an earth's magnetic field caused by the movement of the means of transport(s) while the person is using the means of transport(s), wherein the magnetometer signal (MMS) comprises a magnetic flux density detected by a magnetometer (MM) caused by the movement of the respective means of transport, wherein the magnetometer (MM) is worn on the body by the person during the journey; characterized by the step: ii) determining one or more means of transport used by the person during the journey by exclusively processing the magnetometer signal (MMS) by a trained data-driven model (MO), wherein the magnetometer signal (MMS) is fed as a digital input to the trained data-driven model (MO) and the trained data-driven model (MO) provides the means of transport (FM) used as a digital output. [13] Device according to claim 12, characterized by that it is designed to carry out a method according to one of claims 2 to 4, 6 and 8 to 11. [14] Computer program product comprising program code stored on a non-volatile machine-readable medium, arranged to carry out a method according to one of claims 1 to 11 when the program code is executed on a computer. [15] Computer program with program code for carrying out a method according to one of claims 1 to 11, when the program code is executed on a computer.

Images