[0001] The present invention concerns a method and a system for managing a flow of passenger
on a platform, in particular in the field of public transport, according to claims
1 and 11.
[0002] The present invention relates generally to changes of the number of passengers waiting
for a public transport and the management of these changes of number of passengers,
in particular to avoid overcrowded platforms.
[0003] A current problem related to public transport is the abrupt changes of the number
of people waiting for a public transport, notably during rush hours or in case of
a non predictable event. There is actually no method or system capable of managing
in real time these changes of the number of passengers waiting for public transports,
and as it is well known by people skilled in the art, overcrowded platforms are potential
sources for risk security.
[0004] Consequently, there is a standing need for improving the management of the flow of
passenger at platforms. It is therefore an objective of the present invention to provide
a method and a system for managing in real time a flow of passenger on a platform.
[0005] This objective is achieved according to the present invention with respect to the
method by a method for managing a flow of passengers at a platform, the method comprising
the steps of:
- determining in real time a number of passengers waiting at the platform in function
of the time;
- automatically adjusting a real serving frequency and a real dwell time of incoming
public transports at said platform in function of said number of passengers.
[0006] This objective is achieved according to the present invention with respect to the
system by a system for managing a flow of passengers at a platform, the system comprising:
- a detector capable of counting a number of passengers waiting at a platform;
- a supervision system capable of adjusting a real serving frequency and a real dwell
time of the incoming public transports at said platform in function of said number
of passengers.
[0007] The present invention is thus capable of adapting in real time a flow of public transports
arrivals and departures at a platform to changes of a number of passengers waiting
for said public transport at said platform. In other words, the transport capacity
of a public transport network might be adapted in real time to local variations on
said network of a number of passengers waiting for a public transport of said public
transport network and planning to travel with said public transport on said public
transport network. By "passenger" we refer to a person who has the intention to travel
with said public transport, and is thus waiting, e.g. on a platform at a station or
on a special place dedicated to the waiting, for said public transport. Preferentially,
the public transport according to the present invention may comprise several units.
Each unit may in particular be coupled at each of its extremities to another unit.
For example, such a public transport is a train, and said units are coaches designed
for transporting passengers.
[0008] In particular, the step of automatically adjusting the real serving frequency and
the real dwell time comprises:
- automatically calculating a theoretical serving frequency and a theoretical dwell
time for the incoming public transports at said platform in function of said number
of passengers and of a transport capacity of said public transport, i.e. calculating
a theoretical incoming frequency of incoming public transport serving said platform
and a theoretical dwell time for each incoming public transport separating its time
of arrival from its departure time, in order to avoid an overcrowded platform; and
- automatically adjusting the real serving frequency and the real dwell time of incoming
public transports at said platform to said number of passengers to the theoretical
serving frequency and dwell time that have been calculated.
[0009] Indeed, the supervision system may in particular comprise a controller capable of
determining a theoretical serving frequency and a theoretical dwell time of incoming
public transports at said platform in function of said number of passengers and of
a capacity transport of said public transport. In particular, the theoretical serving
frequency is proportionally correlated to the number of passengers waiting at the
platform. The supervision system is then in particular capable of adjusting the real
serving frequency and the real dwell time of the incoming public transports at said
platform to the theoretical serving frequency and theoretical dwell time that have
been calculated by its controller.
[0010] The theoretical, respectively real, serving frequency refers to a theoretical, respectively
real, number of public transports arriving at and then departing from the platform
by units of time, the time of arrival and the time of departure being separated by
said dwell time. The supervision system is for example an ATS (Automatic Train Supervision
System) that is able to manage public transport traffic.
[0011] Preferentially, said detector according to the present invention is capable of counting
in real time the number of passengers accessing said platform. Said number might be
the exact number of passengers waiting at said platform, or an estimate of the exact
number of passengers waiting at said platform. For example, it is able to count a
number of passengers crossing a platform access door (for entering the platform area
or leaving it), said access door being the single access to the platform. In this
case, the difference between the number of passengers crossing the platform access
door for entering and leaving the platform corresponds to the exact number of passengers
waiting for the public transport at said platform accessible via said access door.
[0012] In particular, said detector comprises a camera system capable of imaging said platform
or at least a part of the platform, e.g. each access door of said platform, by means
of at least one camera. The camera of the camera system may be a visible light camera
and/or an infrared camera, which advantageously allows to differentiate cold objects
from warm objects (e.g. passengers). In particular, said camera system is realized
according to a first design which enables said camera system to image the whole surface
of said platform, and/or a second design which enables said camera system to image
each access door of said platform, by means of one or several cameras. By imaging
the whole surface of the platform, and/or each access door, it is advantageously possible
to count the number of passengers that are going to travel by said public transport
serving the platform, as well as to recognize a passenger needing a special dwell
time. For this purpose, the detector comprises in particular an image treatment unit
that is capable of counting passengers in an image taken by said camera system, preferentially
in each image taken by said camera system. Furthermore, said image treatment unit
is preferentially also capable of detecting or recognizing in an image taken by said
camera system a passenger having a disability or unconventional behavior. In other
words, said image treatment unit is capable of analyzing each image taken by said
camera system, and from each image, to determine the number of passengers appearing
in the image, and preferentially, if a passenger has a disability or a special behavior.
[0013] Preferentially, said camera system comprises also at least one camera designed for
imaging an entrance door of the public transport. In particular, each entrance door
of the public transport might be imaged by at least one camera of said camera system.
For example, said camera system may comprise at least one on board camera that is
mountable at a fixed position in said public transport and designed for imaging an
entrance door of said public transport, and/or a platform camera at a position from
which it is capable of imaging said entrance door from the platform. By imaging an
entrance door of the public transport, the detector is capable to detect a passenger
having a problem for jumping into said public transport or a behavior that needs to
be communicated to a control center which may command a personal intervention.
[0014] Indeed, the image treatment unit comprises in particular a passengers counting algorithm
for determining the number of passengers in an image, and preferentially a profiling
algorithm. Said passengers counting algorithm is in particular capable of facial recognition
and/or person recognition. The techniques at the basis of the algorithm of the present
invention might be for example background plane subtraction and/or boosting techniques.
By means of the profiling algorithm, the image treatment unit is in particular capable
of profiling the passengers, which allows a classification of the passengers, notably
into three profiles: normal passenger, passenger having a disability, suspect passenger
(passenger having a behavior susceptible to result in a personal intervention). The
profiling algorithm is preferentially capable of determining the behavior of a person
on the platform, for example if said person is waiting for the public transport, or
wants to enter into the public transport, or is blocking a door. Preferentially, the
image treatment unit is also capable of fusing an image taken by an IR camera with
an image taken by a visible light camera in order to enhance the detection of passengers
on the platform.
[0015] Preferentially, said passenger counting algorithm is capable of calculating an estimate
of the number of passengers occupying the platform from a calculation of a platform
surface availability, i.e. an estimation of the surface of the platform that is still
available for the passengers (i.e. not occupied by passengers). For this purpose,
the surface of the platform might in particular have a special painting or feature
that facilitates the estimate of passengers occupying said platform.
[0016] The detector is in particular capable to communicate to the supervision system, notably
to its controller and by means of a wire or wireless connection of said detector with
said supervision system, a data comprising the number of passengers that has been
previously determined by the detector, and preferentially a time stamp (i.e. a date)
specifying at what time this number of passengers has been determined by the detector.
Said time stamp may be provided by an internal clock of the detector, or an external
clock. Preferentially, the detector is also capable of signaling to the controller
a presence of a passenger having a disability or a suspect passenger, e.g. who has
a special behavior (for example, passengers blocking a door of the public transport),
so that said supervision system, notably its controller, may calculate a theoretical
dwell time adapted to the disability of said passenger having the disability, or respectively
signal to the control center said special behavior. Advantageously, the present invention
allows thus to automatically adapt the dwell time of the public transport to a disability
of a passenger, in particular by increasing the dwell time if at least one passenger
who is going to travel by said public transport is susceptible of having a disability,
e.g. a problem for jumping in said public transport. Advantageously, for an on site
quick intervention, the supervision system, notably its controller, or the image treatment
unit, may be directly connected to a voice-based notification system of the platform
and/or of the public transport for automatically giving notification to passengers
(e.g. if passengers are blocking doors). The present invention allows thus targeted
instructions to be notified to passengers at doors of the public transport or at the
platform, and a targeted management of the dwell time depending on results of the
profiling algorithm.
[0017] Said number of passengers waiting for the public transport, and in particular said
time stamp, are preferentially used by as input parameters by the supervision system,
notably by its controller, for adjusting said real serving frequency and real dwell
time. In particular, if the number of passengers per unit of time overcomes a threshold
value at a platform, each platform being in particular characterized by its own threshold
value, then said supervision system is able to manage the public transport traffic
for increasing the frequency of the public transports at said platform. The public
transport traffic management is in particular done in function of a number of running
public transports and their associated dwell time, and running speeds. Preferentially,
in order to increase a real serving frequency, the supervision system is able to modify
an upstream traffic of public transport by reducing dwell times of coming public transports
in upstream stations and/or increasing the running speed of coming public transports,
while respecting security parameters such as inter public transport distance, speed
limits, etc. Preferentially, if a platform overcrowdedness overcomes a limit value,
or if the threshold value of several platforms is overcome, then the supervision system
according to the invention is able to inject at least one additional public transport
unit in the traffic.
[0018] Preferentially, the controller is able to use said number of passengers waiting for
the public transport, and in particular said time stamp, as input parameters for calculating
the theoretical serving frequency and the theoretical dwell time of incoming public
transports at the platform. Preferentially, the controller is capable of quantifying
the overcrowdedness of the platform in function of time, wherein said overcrowdedness
might be defined as the ratio of the number of passengers counted by the detector
in function of time over a platform passenger capacity. In particular, from said time
stamp, the controller is also capable of calculating a short time trend of an evolution
of the number of passengers in function of time in order to predict an increase or
decrease of the real serving frequency of the public transports at the platform. Preferentially,
the supervision system comprises a database and is capable of storing in said database
said number of passengers in function of time and of making probabilistic calculation,
notably by means of its controller, of the number of passengers in function of time
from the stored numbers of passengers in function of time, notably to make long time
trends, which may serve for example to determine a rush hour number of passengers.
In particular, the controller determines said theoretical serving frequency and said
theoretical dwell time only if the overcrowdedness exceeds the threshold value (e.g.
when the ratio of the number of passengers counted by the detector over the platform
passenger capacity equals 0.7), and if the trend predicts an increase of the number
of passengers in function of time. Said number of passengers might be thus used for
triggering the adaptation of the real serving frequency and the real dwell time by
the supervision system.
[0019] Preferentially, the supervision system is capable of adjusting the real serving frequency
of the incoming public transports at said platform, and the real dwell time of said
incoming public transports, to the theoretical serving frequency and theoretical dwell
time, for example by speeding up public transports on the public transport network
and/or by decreasing the dwell time at other platforms and/or at said platform and/or
by injecting at least another public transport unit in said public transport network.
[0020] The invention will now be described in a preferred but not exclusive embodiment with
reference to the accompanying drawing, wherein:
- Figure 1
- is a schematic illustration of one embodiment of the system according to the invention;
[0021] Figure 1 shows a schematic illustration of one embodiment of a system for managing
a flow of passengers at a platform 1, the system comprising:
- a detector capable of counting a number of passengers waiting at the platform 1, said
detector comprising in particular a camera system comprising itself at least one camera
21 and an image treatment unit 22;
- a supervision system 3 capable of adjusting a real serving frequency and a real dwell
time of an incoming public transport 11 at said platform 1 to said number of passengers,
wherein said supervision system 3 may be connected to said detector, in particular
to the image treatment unit 22, so that the image treatment unit 22 is able to communicate
to the supervision system 3 the number of passengers and a time stamp as input parameters.
Preferentially, the supervision system 3 comprises a controller 31 capable of determining
a theoretical serving frequency and a theoretical dwell time of incoming public transports
11 at said platform 1 in function of said number of passengers and of a capacity transport
of said public transport 11. The supervision system 3 is in particular capable of
adjusting a real serving frequency and a real dwell time of the incoming public transports
11 to the theoretical serving frequency and theoretical dwell time that have been
calculated by said controller 31.
[0022] The supervision system 3 is in particular capable of directly communicating with
and commanding the public transport 11, for example via a wireless communication.
[0023] Preferentially, the cameras 21 are able to image the platform 1 and/or each door
entrance 111 of the public transport 11. In particular, other cameras 211 may each
image a platform entrance door 12, wherein the platform entrance doors 12 are the
only way to access said platform 1. Other counting systems or techniques might be
used by a person skilled in the art for counting a number of people entering or leaving
a platform area, like for example techniques based on a tickets counter, or a turnstile.
[0024] The images taken by the cameras 21, 211 are then analyzed in real time by the image
treatment unit 22, which is able, by means of a passengers counting algorithm to determine
the number of passengers in each image. Said passengers counting algorithm is for
example capable of counting in function of the time the number of passengers passing
each platform access door 12 for entering and/or leaving the area of the platform
1. In particular, from the difference of passengers passing the platform access door
12 for leaving and for entering the area of the platform 1, the image treatment unit
22 is able to determine in real time the platform occupancy, i.e. the number of passengers
waiting at said platform 1. In particular, the platform occupancy at a time t equals
the number of passengers having entered the area of the platform until time t minus
the number of passengers having left said area until time t, wherein a passenger may
enter the area either by passing said platform access door 12 (e.g. counted from images
taken by said other camera 211) or by deboarding an income public transport (e.g.
counted from images taken by said camera 21), and may leave said area by passing the
platform access door 12 (e.g. counted from images taken by said other camera 211)
or by boarding an income public transport (e.g. counted from images taken by said
camera 21). According to a preferential embodiment of the invention, the image treatment
unit 22 is able to determine the number of passengers waiting for a public transport
11 directly from images of the platform 1 taken by cameras 21 by means of said passengers
counting algorithm in a manner free of any device for counting passengers passing
the platform access door 12. Preferentially, said image treatment unit 22 comprises
a profiling algorithm capable of recognizing at least three different types or profiles
of passengers:
- normal passenger: i.e. passengers having a normal behavior;
- passenger having a disability: i.e. passengers susceptible of having a problem for
jumping into the public transport and needing more time or help for jumping into said
public transport;
- suspect passenger: passenger having a behavior susceptible to result in a personal
intervention, or leading to a risk security, or to damages.
[0025] In particular, the image treatment unit 22 is directly connected to a voice-based
notification system of the platform 1 and/or of the public transport 11 for providing
predefined notification to passengers, notably in function of an identified profile.
Information or data related to the number of passengers is then transmitted by the
detector to the supervision system 3. Said information or data comprises in particular
said number of passengers in function of time, and preferentially a time stamp. Said
information or data may also comprise a profile data if a passenger having a disability
has been detected or if a suspect passenger has been detected.
[0026] The supervision system 3 is then able to adjust a public transport traffic at said
platform in function of said number of passengers in function of time. In particular,
said supervision system 3 is able to increase the incoming frequency of public transports
11 at a platform 1 if said number of passengers overcomes a threshold value.
[0027] Preferentially, the controller 31 of the supervision system 3 is capable of calculating
in real time a theoretical serving frequency and a theoretical dwell time of incoming
public transports 11 at said platform 1 from said number of passengers while taking
into account the capacity transport of the public transport 11. Preferentially, the
supervision system 3 is able to store in a database said number of passengers in function
of time, and in particular, the corresponding calculated theoretical serving frequency
and theoretical dwell time. Advantageously, said controller 31 might be able of calculating
and predicting a probable theoretical serving frequency and a probable theoretical
dwell time in function of time from previously stored values of numbers of passengers
in function of time stored in said database.
[0028] The supervision system 3 and/or the image treatment unit 22 might be also able to
signal to a control center any profile corresponding to the «suspect passenger» profile
and/or the «passenger having a disability» profile, so that said control center may
decide if a personal intervention on site is necessary. Preferentially, said controller
31 is also able to communicate to said the supervision system 3 a trend characterizing
a variation of said theoretical serving frequency, i.e. a prediction of a variation
of said theoretical serving frequency in function of time, so that said supervision
system 3 may smoothly adapt the serving frequency of the public transport to the theoretical
serving frequency.
[0029] The supervision system 3 is in particular capable of commanding in real time a change
of the serving frequency of incoming public transport at a platform and the dwell
time of said public transport at said platform. It is also capable of increasing or
decreasing the serving frequency at a platform in order to reach predicted theoretical
serving frequency calculated by its controller 31. For example, different types of
rush hours might be stored in a database of the supervision system 3, and for each
type of rush hour, a theoretical serving frequency might have been calculated. In
this case, the supervision system 3 is capable of increasing the serving frequency
of incoming public transports so that at said rush hours, the real serving frequency
equals to the theoretical serving frequency.
[0030] To summarize, the method and the system according the invention present the following
advantages:
- they allow to avoid dangerous situation of crowded platforms;
- they regulate the public transport traffic to the number of passengers waiting for
said public transport;
- they regulate the dwell time of public transport to the type of passenger, and allow
a more efficient boarding and deboarding;
- they adjust the traffic offer of public transport to the real time demand;
- they improve the comfort and security of the passengers by avoiding overcrowded coaches
and platforms;
- they are able to automatically adapt in real time the public transport schedule to
the number of passengers waiting for said public transport at the platform.
1. System for managing a flow of passengers at a platform (1), the system comprising:
- a detector capable of counting a number of passengers waiting at the platform (1);
- a supervision system (3) capable of adjusting a real serving frequency and a real
dwell time of the incoming public transports (11) in function of said number of passengers.
2. System according to claim 1, wherein said detector is capable of counting in real
time the number of passengers crossing a platform access door (12) of said platform
(1) .
3. System according to one of the claims 1 or 2, wherein the detector comprises a camera
system capable of imaging at least a part of the platform (1) by means of at least
one camera (21, 211).
4. System according to claim 3, wherein the camera system comprises at least one camera
(21) designed for imaging an entrance door (111) of the public transport (11).
5. System according to one of the claims 1 to 4, wherein the detector comprises an image
treatment unit (22) capable of counting passengers in an image taken by said camera
system by means of a passengers counting algorithm.
6. System according to claim 5, wherein the image treatment unit (22) comprises a profiling
algorithm for recognizing in said image a passenger having a disability or unconventional
behavior.
7. System according to claim 6, wherein the image treatment unit (22) is capable of profiling
the passengers according to three profiles: normal passenger, passenger having a disability,
suspect passenger.
8. System according to one of the previous claims, wherein the image treatment unit (22)
is connected to a voice-based notification system of the platform (1) and/or of the
public transport (11).
9. System according to one of the previous claims, wherein the detector is able to communicate
to the supervision system (3) a data comprising the number of passengers that has
been previously determined and a time stamp.
10. System according to one of the previous claims, wherein the supervision system (3)
comprises a database and is capable of storing in said database said number of passengers
in function of time and of making a probabilistic calculation of the number of passengers
in function of time from data of said database.
11. Method for managing a flow of passengers at a platform (1), the method comprising
the steps of:
- determining in real time a number of passengers waiting at the platform (1) in function
of time;
- automatically adjusting a real serving frequency and a real dwell time of incoming
public transports (11) at said platform (1) in function of said number of passengers.
12. Method according to claim 11, wherein determining in real time a number of passengers
comprises counting in real time the number of passengers crossing a platform access
door (12).
13. Method according to one of the claims 11 or 12, comprising the step of detecting a
passenger having a disability or an unconventional behavior by means of a profiling
algorithm.
14. Method according to one of the previous claims, characterized by increasing the real dwell time if at least one passenger who is going to travel by
said public transport (11) is susceptible of having a disability.
15. Method according to one of the previous claims, characterized by calculating a theoretical serving frequency and a theoretical dwell time only if
an overcrowdedness exceeds a threshold value.