WO2024227778A1 - Method and device for outputting a predicted occupancy distribution in a public means of transportation - Google Patents

Abstract

The invention relates to a method for outputting a predicted occupancy distribution in a public means of transportation, having the steps of detecting the actual state of the occupancy distribution in the public means of transportation in the public area between two stops, ascertaining first prediction data relating to the occupancy distribution at a first point in time, generating a first prediction of the occupancy distribution from the detected actual state and the first ascertained prediction data, outputting first data derived from the predicted occupancy distribution on an output device at a first output point in time, ascertaining second prediction data relating to the occupancy distribution at a second point in time, generating a second prediction of the occupancy distribution from the detected actual state and the second ascertained prediction data, and outputting second data derived from the second predicted occupancy distribution on an output device at a second output point in time, wherein the output first data differs from the output second data. The invention also relates to an occupancy distribution prediction system.

Description

METHOD AND DEVICE FOR OUTPUTTING A FORECAST OCCUPANCY DISTRIBUTION IN A PUBLIC TRANSPORTATION

The invention describes a method for outputting a predicted occupancy distribution in a public transport means with the method steps of recording an actual state of the occupancy distribution in the public transport means between two stops, determining first forecast data for the occupancy distribution at a first point in time, creating a first forecast of the occupancy distribution from the recorded actual state and the first determined forecast data, outputting first data derived from the predicted occupancy distribution on an output device at a first output time, determining second forecast data for the occupancy distribution at a second point in time, creating a second forecast of the occupancy distribution from the recorded actual state and the second determined forecast data, outputting second data derived from the second predicted occupancy distribution on an output device at a second output time, wherein the output first data is different from the output second data, and an occupancy distribution forecast system.

State of the art

In buses, trams and other public transport vehicles, the situation regularly occurs at peak times that large groups of passengers (e.g. schoolchildren) get stuck in the aisles or passageways shortly after boarding the vehicle, thus not properly filling other areas of the vehicle and making it difficult or impossible for other passengers to board.

This results in delays and vehicle delays for the public transport company (transport company) and inconveniences for passengers, including missed connections, because they have to wait for the next scheduled vehicle. In such phases, the transport company may also be forced to to use additional vehicles, although in terms of sheer numbers all passengers could be transported with the scheduled vehicle deployment.

If the transport company wants to influence the passenger distribution in the vehicle, the driver must make the appropriate announcements after having gained an overview of whether passenger redistribution is even an option, i.e. whether there are any areas of the vehicle that are less busy. This is very tedious for the driver and one-off announcements or instructions are usually overheard or ignored by the passengers concerned. A transport company therefore has an interest in achieving a homogeneous passenger distribution in its vehicles and using as little material and personnel as possible.

A simple, inexpensive and largely automated device that can be installed in vehicles and that detects an uneven passenger distribution and helps to resolve this represents a desirable investment for transport companies. In this case, only those vehicles that are also used on the affected routes during the critical phases need to be equipped accordingly.

A system for recording passenger distribution is known, for example, from document DE 10 2004 040 057 A1, which provides for the identification of passengers by equipping tickets with transmitters, preferably RFID chips. Using suitable receivers, the distribution of tickets and thus the distribution of passengers in the area can be recorded. This data can then be made available to a traffic control system.

It is therefore an object of the present invention to provide an improved method for outputting a predicted occupancy distribution.

It is also an object of the invention to provide a occupancy distribution forecasting system.

Description of the Invention The object is achieved by means of the method according to the invention for outputting a predicted occupancy distribution in a public transport system between two stops. Advantageous embodiments of the invention are set out in the subclaims.

The method according to the invention for outputting a predicted occupancy distribution in a public transport vehicle has seven method steps: In the first method step, an actual state of the occupancy distribution in the public transport vehicle between two stops is recorded. An occupancy distribution in the sense of the invention is the number of passengers per unit area as a function of the length of the public transport vehicle. The occupancy distribution therefore indicates how many passengers are where over the length of the public transport vehicle. Furthermore, the occupancy distribution is the number of passengers per unit area as a function of the length of the boarding/alighting area of a stop. The actual state of the occupancy distribution also indicates how many passengers are where over the length of the public transport vehicle and/or the boarding/alighting area of a stop at a given time. The actual state of the occupancy distribution is recorded and determined, for example, by suitable sensors.

In the second step of the process, initial forecast data on the staffing distribution is determined at a first point in time. The first forecast data is optionally empirical data that was obtained from a number of determined staffing distributions under reproducible conditions.

In the third step, an initial forecast of the occupancy distribution is created from the recorded actual state and the first forecast data determined. Using knowledge of the actual state of the occupancy distribution and the current passenger behavior at a given point in time, a forecast of the occupancy distribution at a later point in time is created. This later point in time is, for example, after passengers have disembarked from public transport at the next stop. In the fourth method step, first data derived from the predicted occupancy distribution is output on an output device at a first output time. In particular, the passengers are signaled a direction of decreasing occupancy distribution and the direction to an area with low occupancy distribution. The output device can output such information optically and/or acoustically, for example.

In the fifth method step, second forecast data on the occupancy distribution at a second point in time is determined. For this purpose, the current passenger behavior is optionally recorded and used to determine the second forecast data. Passenger behavior in the sense of the invention is the behavior of a passenger on the one hand during the journey in the public transport between two stops and on the other hand on a platform of the next stop where the public transport stops to let passengers on and off. The passenger behavior includes in particular movements of the passenger, which changes the occupancy distribution in the public transport or at the next stop. The current passenger behavior at a given point in time is recorded and determined, for example, by suitable sensors.

In the sixth step, a second forecast of the occupancy distribution is created from the recorded actual state and the second determined forecast data. In the seventh step, second data derived from the forecast occupancy distribution is output on an output device at a second output time.

According to the invention, the output first data are different from the output second data. In particular, the output second data are created based on an updated actual state of the occupancy distribution. The output second data are therefore also created based on an updated actual state of the occupancy distribution.

By means of the method according to the invention, the occupancy distribution in all areas of the public transport and the platform of the stop is taken into account in order to to calculate a possible movement of people to balance the occupancy distribution and to communicate this to the passengers in a suitable form as a visual and/or acoustic signal. This can shorten the boarding and alighting times at the stops and at the same time achieve an even utilization of the public transport.

For the purposes of this document, public transport is understood to mean any means of transport for the transport of people and/or goods that is generally accessible to everyone. This includes means of transport operated by specially licensed transport companies (for example railways, aviation, cable cars, shipping, elevators) or means of transport on licensed lines or routes or means of transport operated by licensed service providers, as well as any commercially operated means of transport for people and/or goods in the private sector.

For the purposes of this document, stops are all locations where public transport stops, where people get on and/or off and/or goods are loaded and/or unloaded. Stops can be fixed stopping points according to a timetable. However, individually agreed stopping points are also considered stops for the purposes of this document.

The sensors suitable for carrying out the process are, for example, optical sensors for image recognition (cameras) and sensors that work with emitted and received light. Furthermore, sensors can also be based on the frequency modulated or FMCW (Frequency Modulated Continuous Wave) method, in which the distance is determined from the frequency shift of the reflected light beam based on a frequency modulated light beam. Also known as the Doppler effect. FMCW is considered the next generation LiDAR technology. TOF, structured light, 905nm LiDAR are the classic technologies.

In a further development of the invention, the first time and the second time are different from one another and/or the first output time and the second output time are different from one another. Optionally, the first output time is located before the second output time. The The second data output derived from the forecast occupancy distribution therefore reflects an updated forecast occupancy distribution.

In a further embodiment of the invention, the first forecast data is determined from a first source and the second forecast data is determined from a second source. The first source is optionally empirical data that was obtained from a plurality of determined occupancy distributions under reproducible conditions. The second source is optionally measurement data that is recorded by suitable sensors and with which an actual state of the occupancy distribution is determined. In a further development of the invention, the source of the first forecast data is different from the source of the second forecast data.

In a further embodiment of the invention, the source of the first forecast data comprises empirical data. The first source is empirical data that was obtained from a plurality of determined occupancy distributions under reproducible conditions. The empirical data are therefore occupancy distributions from previous trips of the public transport that stop at the same stops.

In a further embodiment of the invention, the source of the second forecast data comprises data determined from current measurement data. The occupancy distribution indicates how many passengers are located where over the length of the public transport. Furthermore, the occupancy distribution is the number of passengers per unit area as a function of the length of the boarding/alighting area of a stop. The actual state of the occupancy distribution also indicates how many passengers are located where over the length of the public transport and/or the platform of a stop at a given time. The actual state of the occupancy distribution is recorded and determined, for example, by suitable sensors.

In a further embodiment of the invention, the second forecast data is recorded from measurement data that reflect the current passenger behavior. The passenger behavior includes in particular movements of the passenger, which changes the occupancy distribution in the public transport and/or at the next stop. The current passenger behavior at a given time is recorded, for example, by suitable sensors. and determined. In a further development of the invention, current passenger behavior is recorded in public transport or at the next stop.

In a further development of the invention, the actual state of the occupancy distribution is recorded in a first time interval, wherein the first time interval is on the route between two consecutive stops. In the first time interval, the public transport is usually in motion, all passengers have boarded the public transport at a first stop and are usually moving within the public transport in order to find and take their standing and/or sitting places or have already taken their standing and/or sitting places.

In a further embodiment of the invention, the analysis of passenger behavior takes place in a second time interval. The second time interval is usually in a period of time shortly before reaching the second stop. In the second time interval, the passengers have stowed their personal items (e.g. luggage), have left their seats and are moving towards the exits of the public transport or are in the immediate vicinity of them.

In a further development of the invention, the second time interval is different from the first time interval. While in the first time interval the passenger movement in the public transport is low, the passenger movement in the second time interval is high because the passengers, for example, grab their bags, stow their personal items in their bags and/or move towards the exits of the public transport.

In a further embodiment of the invention, the second time interval is before reaching the next stop. In the second time interval shortly before reaching the second stop, the passengers have stowed their personal items (e.g. luggage), left their seats and are moving towards the exits of the public transport or are in the immediate vicinity of it. In a further development of the invention, the second time interval is 2 minutes, preferably 1 minute, particularly preferably 30 seconds and especially preferably 15 seconds before reaching the next stop. In a further embodiment of the invention, the second time interval is before reaching the next stop. In the second time interval shortly before reaching the second stop, the passengers have stowed their personal items such as cell phones, books or similar in their bags, have left their seats and are moving towards the exits of the public transport or are in the immediate vicinity of it. In a further development of the invention, the second time interval is 2 minutes, preferably 1 minute, particularly preferably 30 seconds and especially preferably 15 seconds before reaching the next stop.

In a further embodiment of the invention, the actual state of the occupancy distribution is recorded when passenger movement in public transport is low. Passenger movement in public transport is not low immediately after departure from a stop, but at a later point in time when passengers have taken their standing and/or sitting places and possibly stowed their luggage. Passenger movement at this later point in time usually only occurs, for example, to use the toilets and/or to consume drinks and/or food. The proportion of these passengers is usually small, however. When passenger movement is low, occupied and/or complementary unoccupied standing and/or sitting places in public transport can be recorded and determined more reliably than when passenger movement is high.

In a further embodiment of the invention, the actual state of the occupancy distribution is recorded by a first sensor, the first sensor being a sensor for counting passengers at doors and/or ceilings and/or passageways and/or a sensor for recording image material, and the data recorded by the first sensor being analyzed with regard to the actual state of the occupancy distribution. A first sensor is optionally advantageously arranged at entrance doors or portals into the public transport and/or doors in the interior, e.g. passenger compartments with standing and/or seating areas, and records the number of passengers there. Optionally, the first sensor is a depth image camera. The first sensor can record a depth image of the transport space, with which an occupancy distribution within the transport space is determined based on the depth image. In a further In accordance with the invention, the first sensor is a sensor for counting passengers at doors and/or passageways and/or a sensor for capturing image material.

In a further embodiment of the invention, the analysis of passenger behavior includes the analysis of movement flows of people and/or a distribution of people at the next stop. By recording the distribution of people at the next stop, it is possible to predict in which area of the stop passengers will board the public transport and to calculate a corresponding passenger distribution in the public transport.

In a further embodiment of the invention, the passenger behavior is analyzed with regard to the preparatory actions for passengers to get off. In a further embodiment of the invention, the preparatory actions for passengers to get off the public transport vehicle include the passengers getting up from their seats, moving towards the exit of the public transport vehicle and/or stowing personal items. By recording the preparatory actions of the passengers to get off, a predicted occupancy distribution after the passengers get off when the public transport vehicle stops at a stop is possible.

The object is further achieved with the occupancy distribution forecasting system according to the invention. Advantageous embodiments of the invention are also set out in the subclaims.

The occupancy distribution forecasting system for a public transport vehicle according to the invention has a first sensor unit with a first sensor, the first sensor unit being intended and suitable for counting passengers. A first sensor is optionally advantageously arranged at entrance doors to the public transport vehicle and/or doors in the interior, e.g. passenger compartments with standing and/or seating areas, and records the number of passengers present there. Optionally, the first sensor is a depth image camera. The first sensor can record a depth image of the transport space, with which an occupancy distribution of the transport space is determined based on the depth image. In a further embodiment of the invention, the first Sensor a sensor for counting passengers at doors and/or passageways and/or a sensor for capturing images.

The occupancy distribution forecasting system according to the invention further comprises a second sensor unit with a second sensor, wherein the second sensor unit is intended and suitable for detecting passenger behavior in public transport and/or at a stop. The second sensor is optionally a sensor for image recognition, with which people can optionally be distinguished from objects.

Furthermore, the occupancy distribution forecasting system according to the invention has a control unit that is intended and suitable for receiving the data from the first and second sensor units and for creating a forecast occupancy distribution from the received data, as well as an output device for outputting data derived from the forecast occupancy distribution. The control unit has a memory with a suitable software program. The control unit can be arranged locally or decentrally, i.e. in the means of transport or at the stop. Optionally, it is also possible to provide a central control that communicates, for example, via cloud computing and/or fog computing. Mixed forms of central and decentralized control are also possible.

The output device has a plurality of optical displays and acoustic display devices (loudspeakers). Displays and loudspeakers are arranged, for example, in the ceiling area of the passenger compartment. In areas with a high predicted passenger distribution, the displays are intended to light up red, for example, while in areas with a low predicted passenger distribution, the displays light up green. Due to the color signaling, a passenger is motivated to move from an area with a high predicted passenger distribution to an area with a low predicted passenger distribution. A preferred direction can also be indicated, for example, by a running direction of an LED running light arranged in the floor of a carriage and/or a platform at a stop. It is also conceivable to signal a direction in which carriages with a low predicted passenger distribution are to be expected. In addition, the loudspeakers announcements are played to inform passengers of an area in the carriages with a low predicted passenger distribution.

In a further development of the invention, the functionality of the first sensor is different from that of the second sensor. While the first sensor enables passenger counting and localization, the second sensor records the passenger behavior of individual passengers.

In a further embodiment of the invention, the control unit is coupled to the first sensor unit, to the second sensor unit and/or to the output unit. A method for outputting a predicted occupancy distribution in a public transport vehicle can be carried out by means of the control unit.

In a further aspect of the invention, the second sensor unit comprises a plurality of second sensors. The number and arrangement of the second sensors is selected such that all passengers in a public transport vehicle and/or on the platform of a stop can be detected.

In a further embodiment of the invention, the plurality of second sensors of the second sensor unit monitor passenger behavior in public transport and/or at the next stop. The second sensors are image recognition sensors that can optionally be used to distinguish people from objects. In a further development of the invention, the plurality of second sensors of the second sensor unit differ in their functionality.

In a further embodiment of the invention, the data recorded by the second sensor is evaluated using AI/ML or AI/DL support. The control unit uses machine or deep learning methods that are applied to the data recorded by the sensor. This allows the passengers' preparatory actions for getting off to be significantly differentiated from other actions.

In a further embodiment of the invention, the analysis of passenger behaviour comprises the analysis of movement flows of persons and/or a distribution of persons at the next stop. By recording the distribution of people at the next stop, it is possible to predict in which area of the platform at the stop passengers will board public transport and to calculate the corresponding passenger distribution in public transport.

In a further embodiment of the invention, the passenger behavior is analyzed by a second sensor. The second sensor is optionally an image recognition sensor, which can optionally be used to distinguish people from objects.

In a further embodiment of the invention, the second sensor is different from the first sensor. While passenger counting and localization is possible using the first sensor, the second sensor records the passenger behavior of individual passengers.

In a further aspect of the invention, the second sensor is a camera for image capture. The second sensor is a sensor for image recognition, which can optionally be used to distinguish people from objects.

In an advantageous embodiment of the invention, the forecast of the occupancy distribution is created for a point in time immediately after passengers get off the public transport at the next stop. By recording the distribution of people at the next stop, it is possible to predict in which area of the platform at the stop passengers will board the public transport and to calculate a corresponding passenger distribution in the public transport. In addition, occupied and/or complementary unoccupied standing and/or seated places in the public transport are recorded and determined. By displaying unoccupied standing and/or seated places in the public transport, the passenger flow can be directed, passengers can board and disembark more quickly and the utilization of the public transport is improved.

In a further development of the invention, the time is not in the first time interval and/or the second time interval. The first time interval is on the way between two consecutive stops. In the first time interval, the public Means of transport are usually in motion, all passengers have boarded the public transport at a first stop and are usually moving within the public transport to find and take their standing and/or seated places or have already taken their standing and/or seated places. The second time interval is usually a period of time shortly before reaching the second stop. In the second time interval, passengers have stowed their personal items in their bags (e.g. luggage), have left their seats and are moving towards the exits of the public transport or are in the immediate vicinity of them. The time of the forecast of the occupancy distribution is immediately after passengers have disembarked from the public transport at the next stop.

Fig. 1 a: Occupancy distribution forecasting system in a public transport system

Fig. 1 b: Occupancy distribution forecast system on a railway platform

Fig. 2 a: Current state of the occupancy distribution in a public transport vehicle

Fig. 2 b: Current state of the occupancy distribution on a platform

Fig. 3 a: Occupancy distribution forecasting system in a public transport vehicle, first and second sensors

Fig. 3 b: Occupancy distribution forecasting system on a railway platform, first and second sensors

Fig. 4 a: Actual state and predicted occupancy distribution in a public transport system

Fig. 4 b: Actual state and predicted occupancy distribution on a platform

Fig. 5 a: Occupancy distribution forecasting system in a public transport vehicle, first and second sensors, output device

Fig. 5 b: Occupancy distribution forecasting system in a railway platform, first and second sensors, output device

Fig. 6 a: Flowchart of the process for outputting a forecast

Occupancy distribution in a public transport vehicle

Fig. 6 b: Flowchart of the process for outputting a forecast

Occupancy distribution in a public transport system, permanent output of the actual state and the predicted occupancy distribution Fig. 6 c: Flowchart of the procedure for outputting a predicted occupancy distribution in a public transport system, permanent output of both the actual state and the predicted occupancy distribution

Fig. 7 a: Connections of first sensors, second sensors with the control unit and output devices as well as data flow, sensors in public transport and at the second stop

Fig. 7 b: Connections of first sensors, second sensors with the control unit and output devices as well as data flow, sensors in public transport and at second stop, output devices in public transport and at second stop

Fig. 7 c: Connections of first sensors, second sensors with the control unit and output devices as well as data flow, sensors in public transport and at stops, output devices in public transport and at stops

Fig. 1 shows an embodiment of the occupancy distribution forecast system BVPS according to the invention arranged in a public transport OV (Fig. 1 a) and on a platform B of a stop Hn (Fig. 1 b). The public transport OV in this and all other embodiments is a light rail which travels on tracks G at regular intervals to a plurality of stops H1, H2 and stops at the stops H1, H2. The public transport OV has a plurality of carriages W coupled to one another which are continuously connected. The carriages W have passenger seats P. By means of the doors T, a passenger F can enter and leave the carriages W and the passenger compartments A. Passengers F get on and off at the stops H1, H2, H3, with the stop H3 being reached after the stop H2 and the stop H2 after the stop H1.

The stops H1 and H2 each have a platform B. The platform B has a number of doors T and lifts E for passengers F to enter and exit from the platform B. The platform B itself is separated from the track G, on which the public transport OV runs, by a platform edge BK. In this embodiment, the actual state of the occupancy distribution with passengers F of the carriages W of the passengers F of the public transport OV during the journey of the public transport OV (Fig. 1 a) and the passengers on the platform B (Fig. 1 b) is recorded at a given point in time. Passengers F and their positions in the carriages W are recorded by means of the first sensors S1, and passengers F and their positions on a platform B of a stop Hn are recorded by means of the second sensors S2.

The occupancy distribution forecast system BVPS has the first sensor unit SE1 with a plurality of first sensors S1. The first sensors S1 are arranged in the passenger compartments A of the carriages W and in the areas of the doors T so that they have an unobstructed line of sight (LoS = Line of Sight) of the passengers in the passenger compartments A and in the areas between the carriages W.

Furthermore, in this embodiment, the actual state of the occupancy distribution with passengers F of platform B of a stop Hn is recorded at a given point in time. For this purpose, platform B has the occupancy distribution forecast system BVPS (Fig. 1 b). The occupancy distribution forecast system BVPS has the second sensor unit SE2 with a plurality of second sensors S2, which are connected to the control unit C. The control unit C determines the actual state of the occupancy distribution based on the data provided by the second sensors S2, the occupancy distribution being the function of the number of passengers F over the length L(V) of platform B (Fig. 2 b). The actual state of the occupancy distribution therefore indicates how many passengers F are where on platform B at a given point in time.

Each first sensor S1 and each second sensor S2 is a depth image camera. The evaluation of three-dimensional depth images captured by the first sensors S1 and second sensors S2 has the advantage that, in contrast to a pure people count, a space actually occupied by the people and/or pieces of luggage can be recorded. For example, it is easily possible, particularly in rail transport, to detect excessive occupancy of an area of the transport space with pieces of luggage or, in particular, bicycles. Fig. 2 shows the actual state of the occupancy distribution in a public transport vehicle (Fig. 2 a) as well as the actual state of the occupancy distribution on a platform B (Fig. 2 b) at a point in time. At this point in time, the passenger seats P of the carriages W are not fully occupied, the carriages W have free passenger seats FP. The total number of the first sensors S1 detect the position of each passenger F in the carriages W at this given point in time. The control unit C determines the actual state of the occupancy distribution based on the data provided by the first sensors S1, whereby the occupancy distribution is the function of the number of passengers F over the length L(OV) of the public transport vehicle OV. The actual state of the occupancy distribution therefore indicates how many passengers F are where in the public transport vehicle OV at a given point in time. Complementary to the occupancy distribution is the distribution of free passenger seats FP over the length L(OV) of the public transport OV (Fig. 2 a). This distribution of free passenger seats FP over the length L(OV) of the public transport OV is particularly interesting for the passengers F and the transport company itself in order to ensure an even and as complete as possible occupancy of the public transport OV.

Based on the data provided by the second sensors S1, the control unit C determines the actual state of the occupancy distribution on platform B of stop Hn, where the occupancy distribution is the function of the number of passengers F over the length L(V) of platform B (Fig. 2 b). The actual state of the occupancy distribution therefore indicates how many passengers F are where on platform B at a given time. The first sensors S1 and the second sensors S2 are connected to the control unit C, which in this embodiment is arranged on platform B. However, the control unit C can also be arranged in the public transport OV.

Fig. 3 shows embodiments of the occupancy distribution forecasting system BVPS according to the invention arranged in a public transport OV (Fig. 3 a) and on a platform B (Fig. 3 b). First S1 and second sensors S2 are arranged in such a way that the entire transport area A and the entire platform B can be recorded three-dimensionally over the sum of all detection areas. In addition to the embodiments shown so far (see Fig. 1), a first sensor S1 and second sensor S2 create, in addition to the actual state of the occupancy distribution Images of passengers F. The first sensors S1 and second sensors S2 continuously update the images, digitize the images and forward them to the control unit C.

The first sensors S1 are arranged in the passenger compartments A of the carriages W and in the areas of the doors T in such a way that they have an unobstructed line of sight (LoS = Line of Sight) of the passengers in the passenger compartments A and in the areas between the carriages W (Fig. 3 a). The second sensors S2 are also arranged on the platform B (Fig. 3 b) in such a way that they have an unobstructed line of sight of passengers F (Fig. 3 b).

The control unit C analyzes the passenger behavior FV from these images created by the first sensors S1 and the second sensors S2. The control unit C has a memory and a suitable software program for analyzing the images.

The analysis of passenger behavior includes the analysis of movement flows of passengers F in the public transport OV (Fig. 3 a). In the public transport OV, the analysis of movement flows of passengers F includes the predicted passenger distribution FP(PRO) over the length L(OV) of the public transport OV at the time when the public transport OV reaches the stop H2 (see Fig. 4 a). The predicted passenger distribution FP(PRO) therefore indicates how many passengers F will be where in the public transport OV at the time when the public transport OV reaches the stop H2. On platform B, the analysis of movement flows of passengers F includes the predicted passenger distribution FP(PRO) over the length L(B) of the platform B at the time when the public transport OV reaches the stop H2 (Fig. 4 b).

To analyse the predicted passenger distribution FP(PRO), the control unit C uses the determined actual state of the occupancy distribution at a given point in time, which was determined using the first sensors S1 and second sensors S2 (see Fig. 1, Fig. 2). In addition, the preparatory actions of the passengers F for getting off are analysed using the first S1 and second sensors S2. Preparatory actions include, for example, passengers F getting up from their seats, stowing personal items, putting on clothes and, in particular, the movement of passengers towards the exit T of the public transport OV. The control unit C uses machine or deep learning methods that are applied to the data recorded by the sensor. This significantly distinguishes the preparatory actions of passengers F to get off from other actions.

Fig. 5 shows further embodiments of the occupancy distribution forecast system BVPS according to the invention arranged in a public transport OV (Fig. 5 a) and on a platform B of a stop H2 (Fig. 5 b). The arrangement and functioning of the occupancy distribution forecast system BVPS has already been described in the previous embodiment (see Fig. 3). The analysis of the predicted passenger distribution FP(PRO) is also described in the previous embodiment (see Fig. 3, Fig. 4).

In addition, the occupancy distribution forecast system BVPS has an output device A. The output device A has a plurality of optical displays D and acoustic display devices (loudspeakers) L. Displays D and loudspeakers L are arranged in the ceiling area of the passenger compartment A (Fig. 5 a). In the area of a high predicted passenger distribution FP(PRO), the displays D are intended to light up red, while in areas with a low predicted passenger distribution FP(PRO), the displays D light up green. Due to the color signaling, a passenger F is motivated to move from an area with a high predicted passenger distribution FP(PRO) to an area with a low predicted passenger distribution FP(PRO). A preferred direction can also be indicated, for example, by a running direction of an LED running light arranged in the floor of a carriage W. Particularly in the case of multi-unit trains, it is conceivable to signal a direction in which carriages W with a low predicted passenger distribution FP(PRO) are to be expected. In addition, corresponding announcements are played through the loudspeakers L, which inform passengers F of an area in the carriages W with a low predicted passenger distribution FP(PRO). In the same way, on platform B, passengers F are shown an area with a low predicted passenger distribution FP(PRO) in the incoming public transport OV visually by the displays D and acoustically by the loudspeakers L (Fig. 5 b). This also motivates passengers F to find an area with a low predicted passenger distribution FP(PRO) on platform B before boarding the public transport OV. This speeds up the boarding and alighting of passengers F from and to the public transport OV.

Fig. 6 shows embodiments of the method for outputting a-PROGn of predicted occupancy distributions. It shows which method step is carried out on which component of the occupancy distribution forecasting system BVPS at which time t. The public transport OV leaves the first stop H1 at time v-H1, reaches the second stop H2 at time a-H2 and requires the travel time FZ for this (Fig. 6 a). In the time interval Z1, at time Zp1, the control unit C creates a first forecast of the occupancy distribution e-PROG1. The source of this first forecast is empirical data from previously determined occupancy distributions, which are stored in a memory of the control unit C.

In the time interval Z1 at the time Zp1.1, the actual state of the occupancy distribution is recorded d-OV using the sensors S1 arranged in the public transport OV if the passenger movement in the public transport OV is low. In this time interval Z1 at the time Zp1.1, the passenger movement within the carriages W is usually low, the passengers F have taken their seats. In addition, at the same time Zp1.1, the current passenger behavior is recorded using the first sensors S1. The recorded sensor data d-OV from the first sensors S1 are sent to the control unit C, which determines the actual state of the occupancy distribution and the current passenger behavior, taking into account the first forecast of the occupancy distribution c-OV created using empirical data e-PROG1. The forecast passenger distribution is sent to the output unit A, which outputs the data of the forecast passenger distribution via loudspeakers L and displays D at the time Zp1.1 a-PROG.

Within the second time interval Z2 15 s before reaching a-H2 of the next stop H2, the second sensors S2 on platform B of the second stop H2 are used to At time Zp2, the actual state of the occupancy distribution and the passenger behavior FV of the passengers F are recorded dH. In further embodiments of the method, the second time interval Z2 can be 2 minutes, preferably 1 minute, and particularly preferably 30 seconds. The recorded sensor data dH of the second sensors S2 are also sent to the control unit C, which determines a predicted passenger distribution C-PROG2. The predicted passenger distribution is determined C-PROG2 for the time ZP, which is immediately after passengers F get off the public transport OV at the next stop H2. The predicted passenger distribution is sent to the output unit A, which outputs data of the predicted passenger distribution via loudspeakers L and displays D at time Zp2 a-PROG2.

In a variant (Fig. 6 b) of the method, within the third time interval Z3, which lies within the first time interval Z1 and has a shorter duration than the first time interval Z1, multiple recordings d-OVn, determinations c-OVn and outputs a-OVn of actual states of the occupancy distribution take place at regular, adjustable time intervals at the times Zp1.1 to Zpl .n. The outputs a-OVn are made via loudspeakers L and displays D within the public transport OV. Passengers F thus receive updated information about free seats within the public transport OV during the travel time FZ.

In a further variant (Fig. 6 c) of the method, a plurality of detections d-Hn are additionally carried out within the second time interval Z2 by means of the second sensors S2 arranged on the platform B of the second stop H2: At time Zp2.1, a first detection d-H1 of the actual state of the occupancy distribution and the passenger behavior FV of the passengers F is carried out, a first determination C-PROG1 of a predicted passenger distribution and a first output a-PROG1 of data via loudspeakers L and displays D. At time Zp2.2, a second detection d-H2 of the actual state of the occupancy distribution and the passenger behavior FV of the passengers F is carried out by means of the second sensors S2 arranged on the platform B of the second stop H2, a second determination C-PROG2 of a predicted passenger distribution and a second output a-PROG2 of data via loudspeakers L and displays D. The outputs a-PROG1 and a-PROG2 are carried out on the Platform B of the second stop H2, passengers F receive updated information about the occupancy distribution in public transport OV and on platform B. Passengers F can board and alight more quickly and the stopping time of public transport OV at stop H2 is shortened.

Fig. 7 shows variants of the connections of first sensors S1, second sensors S2 with the control unit C and the output devices L, D as well as the data flow between the mentioned components both in the public transport OV and on the platform B of a stop H2, H3. The control unit C is arranged in such a way that a preferably wireless data exchange can take place between the components arranged in the moving public transport OV and in the stops H2, H3.

In a first variant (Fig. 7 a), the first sensors S1 are arranged in the public transport vehicle OV and the second sensors S2 on the platform of the stop H2 and send the recorded d-OVn, d-Hn data to the control unit C, which sends the determined actual states c-OVn, c-Hn and predicted states c-PROGn of the occupancy distribution to the output devices L, D of the stop H2, where they are output a-OVn, a- PROGn.

In a further variant (Fig. 7 b), first sensors S1 are arranged in the public transport OV and send the recorded d-OVn data to the control unit C. Second sensors S2 are arranged at the stop H2 and also send their recorded d-Hn data to the control unit C. The control unit C determines c-OVn, c-Hn actual states and predicted states c-PROGn of the occupancy distribution of both the public transport OV and the stop H2, sends the determined c-OVn, c-Hn actual states and predicted states c-PROGn of the occupancy distribution to the output devices L, D of the stop H2 and the public transport OV, where they are output a-OVn, a-PROGn respectively.

In a variant of the above embodiment (see Fig. 7 b), second sensors S2 are arranged at three different stops H1, H2, H3 and send their recorded d-Hn data to the control unit C (Fig. 7 c). First sensors S1 are also arranged in the public transport OV and send the recorded d-OVn data to the control unit C. The control unit C determines c-OVn, c-Hn actual states and predicted States c-PROGn of the occupancy distribution both of the public transport OV and of each individual stop H1, H2, H3, sends the determined c-OVn, c-Hn actual states and predicted states c-PROGn of the occupancy distribution to the output devices L, D of each stop H1, H2, H3 and of the public transport OV, where they are respectively output a-OVn, a-PROGn.

According to the invention, the method presented here and the arrangement of the connections of first sensors S1, second sensors S2 with the control unit C and the output devices L, D as well as the data flow between the said components can be carried out in the public transport OV with any number of stops Hn.

REFERENCE SYMBOL LIST

A passenger compartment

B platform

BK platform edge

BVPS occupancy distribution forecasting system

C control unit

D Display device/display

E elevator

F Passenger

FP Free Passenger PI rates

FV Current Passenger Behavior

vehicle driving time

G track

H1, H2, H3 stops

L speaker

L(OV) Length of public transport

L(B) length of the platform

P Unoccupied seat/standing place

PB Predicted Occupation Distribution

PF Occupied seat/standing space

OV Public Transport

SE1 First sensor unit

SE2 Second sensor unit

51 First Sensor

52 Second Sensor t Time axis

T door

Z1 First time interval Z2 Second time interval

Z3 Third time interval

ZP Time of the predicted staffing distribution

Zp1, Zp1.1, Zp1.2, First Time Point

Zpl.n

Zp2, Zp2.1, Zp2.2 Second time v-H1 Leaving the first stop a-H2 Arrival at the second stop e-PROG1 First forecast of the occupancy distribution based on

Experience data d-OV, d-OV1, d-OV2, recording of the current passenger behavior and the actual state d-OVn of the occupancy distribution in public transport d-H, d-H1, d-H2 recording of the current passenger behavior and the actual state of the occupancy distribution on the platform of the stop c-OV, C-OV1, c-OV2, determination of the current passenger behavior and the actual state c-OVn of the occupancy distribution of public transport a-OV, a-OV1, a-OV2, output of the actual state of the occupancy distribution of a-OVn public transport c-PROG, C-PROG1, calculation of a forecast occupancy distribution

C-PROG2 a-PROG, a-PROG1, output of a predicted occupancy distribution a-PROG2

Claims

PATENT CLAIMS 1. Method for outputting a predicted occupancy distribution in a public transport (OV) between two stops (H1, H2) with the following procedural steps: • Recording the current state of the occupancy distribution in public transport (OV) between two stops (H1, H2), • Determination (e-PROG1) of initial forecast data on the staffing distribution at a first point in time (Zp1), • Creating an initial forecast of the staffing distribution from the recorded actual status (d-IST) and the first determined (e-PROG1) forecast data, • Output (a-PROG1) of first data derived from the predicted occupancy distribution (PB1) on an output device (L, D) at a first output time (Zp1.1) • Determining second forecast data (P2) for the staffing distribution at a second point in time (Zp2), • Creating a second forecast of the staffing distribution (PB2) from the recorded actual status and the second determined forecast data (P1), • Output of second data derived from the second predicted occupancy distribution (PB2) on an output device (L, D) at a second output time (Zp2.1), characterized in that the output first data are different from the output second data. 2. Method for outputting a predicted occupancy distribution in a public transport (OV) between two stops (H1, H2) according to claim 1, characterized in that the first time (Zp1) and the second time (Zp2) are different from one another and/or the first output time (Zp1.1) and the second output time (Zp2.1) are different from one another. 3. Method for outputting a forecast occupancy distribution in a public transport means (OV) between two stops (H1, H2) according to claim 1 or 2, characterized in that the first forecast data are determined from a first source and the second forecast data are determined from a second source. 4. Method for outputting a forecast occupancy distribution in a public transport (OV) between two stops (H1, H2) according to claim 3, characterized in that the source of the first forecast data comprises empirical data. 5. Method for outputting a forecast occupancy distribution in a public transport (OV) between two stops (H1, H2) according to one or more of the preceding claims, characterized in that the source of the second forecast data comprises data determined from current measurement data. 6. Method for outputting a forecast occupancy distribution in a public transport means (OV) between two stops (H1, H2) according to claim 5, characterized in that the second forecast data are acquired from measurement data which reflect the current passenger behaviour (FV). 7. Method for outputting a predicted occupancy distribution in a public transport means (OV) between two stops (H1, H2) according to claim 6, characterized in that the recording of current passenger behaviour (FV) takes place in the public transport means (OV) or at the next stop (H2). 8. Method for outputting a forecast occupancy distribution in a public transport (OV) between two stops (H1, H2) according to one or more of claims 3 to 7, characterized in that the source of the first forecast data is different from the source of the second forecast data. 9. Method for outputting a predicted occupancy distribution in a public transport means (OV) between two stops (H1, H2) according to one or more of the preceding claims, characterized in that the actual state of the occupancy distribution is recorded in a first time interval (Z1), wherein the time interval (Z1) lies on the route between two consecutive stops (H1, H2). 10. Method for outputting a predicted occupancy distribution in a public transport (OV) between two stops (H1, H2) according to one or more of the preceding claims, characterized in that the analysis of the passenger behavior takes place in a second time interval (Z2). 11. Method for outputting a predicted occupancy distribution in a public transport means (OV) between two stops (H1, H2) according to claim 10, characterized in that the second time interval (Z2) is different from the first time interval (Z1). 12. Method for outputting a predicted occupancy distribution in a public transport means (OV) between two stops (H1, H2) according to claim 10 or 11, characterized in that the second time interval (Z2) is before reaching a next stop (H2). 13. Method for outputting a predicted occupancy distribution in a public transport (OV) between two stops (H1, H2) according to claim 12, characterized in that the second time interval (Z2) is 2 min, preferably 1 min, particularly preferably 30 s and especially preferably 15 s before reaching (a-H2) the next stop (H2). 14. Method for outputting a predicted occupancy distribution in a public transport (OV) between two stops (H1, H2) according to one or more of the preceding claims, characterized in that the actual state of the occupancy distribution is detected by a first sensor (S1), wherein the first sensor (S1) is a sensor for counting passengers at doors and/or passageways and/or a sensor for detecting image material, and wherein the data detected by the first sensor (S1) are analyzed with regard to the actual state of the occupancy distribution. 15. Method for outputting a predicted occupancy distribution in a public transport vehicle (OV) between two stops (H1, H2) according to claim 14, characterized in that the analysis of the passenger behavior (FV) is carried out with regard to the preparatory actions for the alighting of passengers (F), wherein the preparatory actions for the alighting of passengers (F) in the public transport vehicle (OV) include the passengers (F) getting up from their seats, the movement in the direction of the alighting of the public transport vehicle (OV) and/or the stowing of personal items. 16. Method for outputting a predicted occupancy distribution in a public transport (OV) between two stops (H1, H2) according to claim 14 or 15, characterized in that the analysis of the passenger behaviour (FV) is carried out with the aid of AI/ML/or AI/DL. 17. Method for outputting a predicted occupancy distribution in a public transport means (OV) between two stops (H1, H2) according to one or more of claims 14 to 16, characterized in that the analysis of the passenger behaviour (FV) comprises the analysis of movement flows of persons and/or a distribution of persons at the next stop (H2). 18. Method for outputting a predicted occupancy distribution in a public transport vehicle (OV) between two stops (H1, H2) according to one or more of claims 14 to 17, characterized in that the analysis of the passenger behavior (FV) is carried out by a second sensor (S2). 19. Method for outputting a predicted occupancy distribution in a public transport means (OV) between two stops (H1, H2) according to claim 18, characterized in that the second sensor (S2) is different from the first sensor (S1). 20. Method for outputting a predicted occupancy distribution in a public transport means (OV) between two stops (H1, H2) according to claim 19, characterized in that the second sensor (S2) is a camera for image capture. 21. Method for outputting a forecast occupancy distribution in a public transport means (OV) between two stops (H1, H2) according to one or more of the preceding claims, characterized in that the forecast of the occupancy distribution is created for a point in time (Zp) immediately after passengers (F) have disembarked from the public transport means (OV) at the next stop (H2). 22. Method for outputting a predicted occupancy distribution in a public transport (OV) between two stops (H1, H2) according to claim 21, characterized in that the time (Zp) is not in the first time interval (Z1) and/or second time interval (Z2). 23. Occupancy distribution forecasting system (OVPS) for a public transport (OV) with: • a first sensor unit (SE1) with a first sensor (S1), wherein the first sensor unit (SE1) is intended and suitable for counting passengers (F) in public transport (OV) and/or recording passenger behaviour (FV) in public transport (OV), • a second sensor unit (SE2) with a second sensor (S2), wherein the second sensor unit (SE2) is intended and suitable for counting passengers (F) at a stop (H1, H2) and/or recording passenger behavior (FV) at a stop (H1, H2), • a control unit (C) which is intended and suitable for receiving the data from the first and second sensor units (SE1, SE2) and for creating a predicted occupancy distribution (PB) from the received data, • an output device (A) for outputting data derived from the predicted occupancy distribution (PB). 24. Occupancy distribution forecasting system (BVPS) for a public transport (OV) according to claim 23, characterized in that the first sensor (S1) is different in its functionality from the second sensor (S2). 25. Occupancy distribution forecasting system (BVPS) for a public transport (OV) according to claim 23 or 24, characterized in that the control unit (C) is coupled to the first sensor unit (SE1), to the second sensor unit and/or to the output unit (A). 26. Occupancy distribution forecasting system (BVPS) for a public transport (OV) according to one or more of claims 23 to 25, characterized in that the second sensor unit (SE2) comprises a plurality of second sensors (S2.1, S2.2). 27. Occupancy Distribution Forecasting System (BVPS) for a public transport system (OV) according to claim 26, characterized in that the plurality of second sensors (S2.1, S2.2) of the second sensor unit (SE2) monitor the passenger behavior (FV) in the public transport (OV) and at the next stop (H2). 28. Occupancy distribution forecasting system (BVPS) for a public transport (OV) according to claim 26 or 27, characterized in that the plurality of second sensors (S2.1, S2.2) of the second sensor unit (SE2) differ in their functionality. 29. Occupancy distribution forecasting system (BVPS) for a public transport (OV) according to one or more of claims 23 to 28, characterized in that the evaluation of the data KI/ML or AI/DL supported.