U.S. patent application number 12/054530 was filed with the patent office on 2008-07-10 for proactive mechanism for supporting the global managment of vehicle traffic flow.
This patent application is currently assigned to International Business Machines Corporation. Invention is credited to Chatschik Bisdikian, Joel W. Branch, Norman H. Cohen, John S. Davis, Maria R. Ebling, Daby M. Sow.
Application Number | 20080167795 12/054530 |
Document ID | / |
Family ID | 38971337 |
Filed Date | 2008-07-10 |
United States Patent
Application |
20080167795 |
Kind Code |
A1 |
Bisdikian; Chatschik ; et
al. |
July 10, 2008 |
Proactive Mechanism for Supporting the Global Managment of Vehicle
Traffic Flow
Abstract
The invention addresses the disadvantage of automated vehicle
traffic management systems by providing a solution for predicting
and avoiding traffic congestion before the congestion is
experienced. The invention includes a decentralized mechanism for
predicting congestion. An exemplary embodiment of the invention
uses a system of sensors to determine current traffic flow; the
sensors communicate their state to a network of servers and then an
algorithm is applied to the collected data to predict traffic
congestion; upon detection of congestion, signals are communicated
to the system controllers to prevent the traffic congestion.
Inventors: |
Bisdikian; Chatschik;
(Chappaqua, NY) ; Branch; Joel W.; (Troy, NY)
; Cohen; Norman H.; (Suffern, NY) ; Davis; John
S.; (New York, NY) ; Ebling; Maria R.; (White
Plains, NY) ; Sow; Daby M.; (North White Plains,
NY) |
Correspondence
Address: |
FREDERICK W. GIBB, III;Gibb & Rahman, LLC
2568-A RIVA ROAD, SUITE 304
ANNAPOLIS
MD
21401
US
|
Assignee: |
International Business Machines
Corporation
Armonk
NY
|
Family ID: |
38971337 |
Appl. No.: |
12/054530 |
Filed: |
March 25, 2008 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
11488939 |
Jul 18, 2006 |
|
|
|
12054530 |
|
|
|
|
Current U.S.
Class: |
701/118 |
Current CPC
Class: |
G08G 1/096844 20130101;
G08G 1/0116 20130101; H04L 67/12 20130101; G08G 1/096811 20130101;
G08G 1/0133 20130101; G08G 1/0145 20130101; G08G 1/0112
20130101 |
Class at
Publication: |
701/118 |
International
Class: |
G08G 1/127 20060101
G08G001/127 |
Claims
1. A method of managing traffic flow in a transportation system,
said method comprising: creating an abstract representation of said
transportation system that includes decision nodes representing
intersections of paths within said transportation system and that
includes links representing paths between said decision nodes,
receiving current traffic flow rates for each of said links from a
plurality of sensors located within said links; predicting future
traffic flow rates for each outgoing link of each node based on
current traffic flow rates for each incoming link of each said node
and based on current traffic flow rates of each said outgoing link;
comparing predicted traffic flow rates for each said outgoing link
of each node to determine which of said outgoing links have the
lowest traffic flow rates; routing users of said transportation
system to outgoing links that have said lowest traffic flow rates;
and after routing said users, adding users routed to a link to a
predicted traffic flow rate for said link.
2. The method according to claim 1, further comprising: identifying
a link as being congested if either of a current traffic flow level
or a predicted traffic flow level exceeds a predetermined maximum;
and routing users of said transportation system away from congested
links.
3. The method according to claim 1, further comprising receiving a
final destination from a user of said transportation system,
wherein said routing only considers links that will allow said user
to reach said final destination.
4. A method of managing vehicle traffic flow in a vehicle
transportation system, said method comprising: creating an abstract
representation of said transportation system that includes decision
nodes representing intersections of roads within said
transportation system and that includes links representing roads
between said decision nodes, receiving current vehicle traffic flow
rates for each of said links from a plurality of sensors located
within said links; predicting future vehicle traffic flow rates for
each outgoing link of each node based on current vehicle traffic
flow rates for each incoming link of each said node and based on
current vehicle traffic flow rates of each said outgoing link;
comparing predicted vehicle traffic flow rates for each said
outgoing link of each node to determine which of said outgoing
links have the lowest vehicle traffic flow rates; routing vehicles
of said transportation system to outgoing links that have said
lowest vehicle traffic flow rates; and after routing said vehicles,
adding vehicles routed to a link to a predicted vehicle traffic
flow rate for said link.
5. The method according to claim 4, further comprising: identifying
a link as being congested if either of a current vehicle traffic
flow level or a predicted vehicle traffic flow level exceeds a
predetermined maximum; and routing vehicles of said transportation
system away from congested links.
6. The method according to claim 4, further comprising receiving a
final destination from a vehicle of said transportation system,
wherein said routing only considers links that will allow said
vehicle to reach said final destination.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application is a Continuation of U.S. application Ser.
No. 11/488,939 filed Jul. 18, 2006, the complete disclosure of
which, in its entirety, is herein incorporated by reference.
FIELD OF THE INVENTION
[0002] The present invention generally relates to a method and
system to manage vehicle traffic flow in a decentralized manner;
and, more particularly, to proactive prediction of congestion in a
traffic system with a technique for prevention of traffic
congestion.
BACKGROUND OF THE INVENTION
[0003] Vehicle traffic management involves the policies and
infrastructure used to manage vehicular traffic flow in a
large-scale road system such as a roadway system associated with a
municipality or a government controlled highway system. The goal of
a vehicle traffic management system is to maximize traffic flow
through the system while maintaining a safe environment.
[0004] Vehicle traffic management systems consist of roads that
intersect and such intersections offer vehicles the opportunity to
choose between several routes to their final destination.
Throughout a roadway system, controllers may exist for controlling
flow between routes; an exemplary embodiment of a controller may be
a traffic light which repeatedly modifies its state to determine
how vehicles should proceed.
[0005] Vehicle traffic management may involve manual as well as
automated techniques. An exemplary embodiment of a manual vehicle
traffic management system may involve human administrators that
monitor traffic flow from a central location and manipulate the
system's controllers in response to traffic congestion. An
exemplary embodiment of an automated vehicle traffic system may
involve a computer system that monitors traffic flow and uses
algorithms to automatically manipulate the system's controllers in
response to traffic congestion.
[0006] As populations expand and the number of vehicles on roadways
continues to increase, the process of managing traffic becomes an
increasingly difficult task. The disadvantage of manual vehicle
management systems is that they are limited in response time by the
human operators. As roadway systems grow in complexity, the human
response time becomes a formidable limitation. Automated vehicle
traffic management systems are not constrained by the speed
limitations that manual systems face. One advantage of automated
vehicle traffic management systems is that they can be implemented
in a distributed manner. A distributed implementation of an
automated vehicle management system can enable very fast
operation.
[0007] A plethora of controllers are available for use by vehicle
traffic management systems. An exemplary traffic controller is a
direction routing system such as Yahoo!Maps, MapQuest, Google Maps
and other online tools. The current realization of these tools is
to formulate the complete route prior to the beginning of a
vehicle's trip. A modification of these tools may involve
incorporation of real time traffic so that the route can be
iteratively updated as the trip progresses. An automated system in
which real time traffic allows for iteratively updated routes
offers many advantages such as quick response to encountered
congestion. One key feature that such a system would not offer is a
technique for preventing traffic congestion.
SUMMARY OF THE INVENTION
[0008] The invention addresses the disadvantage of automated
vehicle traffic management systems by providing a solution for
predicting and preventing traffic congestion before the congestion
is experienced. The invention includes a decentralized mechanism
for predicting congestion. An exemplary embodiment of our invention
uses a system of sensors to determine current traffic flow; the
sensors communicate their state to a network of servers and then an
algorithm is applied to the collected data to predict traffic
congestion; upon detection of congestion, signals are communicated
to the system controllers to avoid the traffic congestion.
[0009] For example, in one aspect of the invention, an exemplary
sensor detects the average speed of vehicles (flow rates) on each
roadway path and this information is aggregated for predicting
congestion. In another aspect of the invention, the system
controllers are actuators within the vehicles that automatically
turn at roadway intersections in response to instructions. In
another aspect of the invention, an exemplary sensor detects the
average speed of vehicles on each roadway path and this information
is aggregated for predicting congestion.
DESCRIPTION OF THE RELATED ART
[0010] Prior to the present invention, there existed no
decentralized technique for predicting vehicular traffic congestion
in a vehicle traffic management system. All the following are
incorporated herein by reference. In particular, US Patent
Application US20050222751A1 includes no predictive components. In
particular, US Patent Application US2005065711A1 includes no
predictive components. In particular, U.S. Pat. No. 5,696,503
predicts location of individual vehicles but does not provide an
algorithm for predicting congestion. In particular, US Patent
Application US2004246147A1 includes no predictive components. In
particular, US Patent Application US20050164673A1 includes no
predictive components. In particular, US Patent Application
US20050003802A1 includes no predictive components. In particular,
US Patent Application US20040140909A1 includes no predictive
components. In particular, U.S. Pat. No. 6,351,709 includes a route
updating mechanism but no predictive components. In particular,
U.S. Pat. No. 6,853,915 receives route information from a
centralized location (as opposed to a local database) but does not
include prediction. In particular, U.S. Pat. No. 6,480,783 provides
multiple alternative routes but does not include prediction. In
particular, US Patent Application US20010001848A1 uses historical
traffic data but does not predict congestion based on current data.
In particular, U.S. Pat. No. 5,801,943 includes no predictive
components.
BRIEF DESCRIPTION OF THE DRAWINGS
[0011] To describe the foregoing and other exemplary purposes,
aspects, and advantages, we use the following detailed description
of an exemplary embodiment of the invention with reference to the
drawings, in which:
[0012] FIG. 1 is a block diagram of an exemplary sensor and
actuator (SANet) architecture;
[0013] FIG. 2 is an illustration of an exemplary highway
topology;
[0014] FIG. 3 is an illustration of an exemplary description of the
link/node terminology for abstractly describing the components of a
highway topology;
[0015] FIG. 4 is an illustration of an exemplary abstract sensor
and actuator layout for an exemplary highway topology;
[0016] FIG. 5 is a flow diagram illustrating the methodology for
using local information to route a vehicle through a highway
system;
[0017] FIG. 6 is a flow diagram illustrating the methodology for
predicting traffic congestion of a highway system; and
[0018] FIG. 7 is a flow diagram illustrating the methodology for
generating projected link flow values of a highway system.
DETAILED DESCRIPTION OF EXEMPLARY EMBODIMENTS OF THE INVENTION
[0019] Referring now to the drawings, and more particularly to
FIGS. 1-7, we describe exemplary embodiments of the method and
structures according to the present invention.
[0020] FIG. 1 100 displays an exemplary embodiment of a sensor and
actuator network (SANet) architecture. As displayed, a set of
sensors 115, 116, 125, 126 and a set of actuators 111, 112, 121,
122, 123 are deployed for sensing data and actuating (enacting
commands), respectively. Two servers 110, 120 are used to monitor
data collected from sensors and to send commands to actuators. The
layout is organized in a distributed fashion so that one server 110
displayed in the exemplary architecture 100 communicates to two
sensors 115, 116 and two actuators 111, 112; similarly, the other
server 120 displayed in the exemplary architecture 100 communicates
with two sensors 125, 126 and three actuators 121, 122, 123. The
two servers 110, 120 are coordinated by a server manager 130.
[0021] FIG. 2 200 illustrates an exemplary highway topology. The
topology consists of a system of roadways that intersect, merge and
diverge. Four entrance points 205, 225, 245, 270 are shown where
vehicles can enter the system and all vehicle flow moves rightward.
There are two types of inflection points: a merge point and a
decision point. A merge inflection point involves the merging of
two or more road ways into a smaller number of road ways. A
decision inflection point involves the diverging of a road way into
a larger set of road ways. Seven inflection points 210, 215, 230,
235, 250, 255, 265 are shown in FIG. 2. FIG. 2 shows four decision
points 210, 215, 250, 255 with traffic lights depicted to serve as
controllers of traffic flow. FIG. 2 shows three merge points 230,
235, 265.
[0022] FIG. 3 300 illustrates an exemplary abstract representation
of the components of a traffic system. Road ways are represented by
the black arcs and inflection points are represented by the
circular nodes. FIG. 3 displays three inflection points 310, 340
and 350. Point A 310 is a decision point, point D 340 is a merge
point and point E is a decision point. As indicated by the arrows,
traffic in the exemplary illustration 300 moves rightward. The
decision point at A 310 offers a choice for a vehicle to be routed
along link AB between points 310 and 320. The merge point D 340
displays the merging of the route AD between points 310 and 340
with the route CD between points 330 and 340. The decision point E
offers selection between route DF (between points 350 and 360),
route EG (between points 350 and 370) and route EH (between points
350 and 380). Each link has a maximum traffic flow rate as well as
an actual traffic flow rate. If the actual traffic flow rate of a
given link exceeds (or is projected to exceed) the link's maximum
traffic flow rate, then that link is experiencing (or is projected
to experience) traffic congestion and traffic is routed away from
the congested link(s). For a given inflection point, the total flow
rate of all incoming links will equal the total flow rate of all
outgoing links. A top inflection point is defined as an inflection
point that has no incoming link that is the output of another
inflection point. Point A 310 in FIG. 3 is a top inflection point
while point D 340 and point E 350 are not top inflection points in
the exemplary illustration.
[0023] FIG. 4 400 illustrates the application of the exemplary
abstract representation discussed in FIG. 3 300 as applied to the
exemplary traffic topology of FIG. 2 200. Six inflection points
410, 415, 440, 445, 450, 455 are shown. Four decision points 410,
415, 450, 455 are shown. Two merge points 440, 445 are shown. The
entrance points to the highway system are abstractly represented
via points A 405, D 435, H 445 and L 465. The exit points are
abstractly represented via points C 420, G 440 and K 460. In this
exemplary illustration, inflection points B 410 and I450 are top
inflection points.
[0024] FIG. 5 shows one possible embodiment of our localized
direction-routing scheme. The computation starts in step 510 with
the detection of a vehicle at a decision node (DN). When this
happens, the final destination of the vehicle or object is sent to
the DN, 520. The final destination is selected by the user of the
vehicle or by some other user/manager of the transportation system.
The DN then determines the set of possible outgoing links, 530.
Each of these links corresponds to a path from the current position
of the vehicle to its destination. The DN tests if there is more
than one outgoing link for this vehicle, 540. If there is more than
one outgoing link, the DN evaluates the cost (e.g., in travel time
or delay) associated with each of these links by making requests to
downstream DNs, asking for input and output flow rates, 550. The DN
then selects the link with the lowest cost (e.g., lowest flow rate)
560 and actuates the decision to drive down this link 570. If there
is only one link detected in step 540, the execution moves to step
570, where the DN actuates the decision to drive down this link.
The execution ends on step 580 with the DN updating and recording
the outgoing link flow rate, 580, for future decision that it may
have to compute. Thus, after a DN instructs an object or vehicle to
use a link, the prediction of vehicle flow rate on that link can be
modified to include the new object even before the object begins to
traverse the link.
[0025] FIG. 6 shows one possible embodiment of our method to
predict traffic congestion. The computation starts in step 610 with
a measurement of the current flow on each link in the system.
Predictions of all these rates are obtained in step 630. The
algorithm then retrieves L.sub.c, the list of all links with
unknown congestion status, 640, and tests if the list is empty,
650. If L.sub.c is not empty, the algorithm picks one link from the
list and uses predefined maximum tolerable flows, l.sub.max, for
each link to determine congestion on this link, 660. To determine
this congestion, the algorithm compares the projected link flow
l.sub.p (obtained in step 630) with the maximum tolerable flow for
that link, l.sub.max. If l.sub.p is greater than l.sub.max, the
algorithm declares the corresponding link to be congested 670. It
then removes this link from L.sub.c. If l.sub.p is less or equal to
l.sub.max, the computation continues to step 670 with the removal
of the link from L.sub.c. Once the list becomes empty, the
computation ends 690.
[0026] FIG. 7 illustrates an exemplary flow diagram for generating
projected flow values of links in a traffic system. The algorithm
starts 710 by selecting all top inflection points of a given
abstract SANet layout and placing them in a list L.sub.T. If there
are no top points in L.sub.T then the algorithm ends 780. If there
are top inflection points, then one top inflection point (P.sub.T)
is arbitrarily removed 720 from the list. For the selected
inflection point P.sub.T, the flow rates of all incoming links are
summed 730. Then the outgoing links of P.sub.T are placed 740 in
the list. The size of the list L.sub.i is checked 750 and if the
list is not empty then one outgoing link L.sub.i is removed from
the list. The projected flow rate of link L.sub.i is determined
from the input flow of P.sub.T. Thus, the flow from all incoming
links to a node is assigned to all outgoing links using a
methodology that most evenly balances traffic flow on the outgoing
links. During this balancing process the traffic that is restricted
to certain links (cannot reach its desired destination on all
outgoing links) is assigned to the restricted links first. Then,
the remaining traffic flowing out of the node in question is
assigned to the outgoing links. The least used (least congested,
lowest flow) link is selected when making each traffic element
assignment to balance the use of the outgoing links. After each
traffic element is assigned, the flow on each link is recalculated
and the next assignment is made to the then least used link. Once
all output link flow rates have been determined, then the algorithm
is repeated at step 710 by selecting a new top inflection point
from the abstract SANet layout. The algorithm recursively
propagates flow rates until projected flow values for all links
have been determined.
[0027] Therefore, as shown above, the invention provides a method
of managing traffic flow in a transportation system by creating an
abstract representation of the transportation system that includes
decision nodes representing intersections of paths within the
transportation system and that includes links representing paths
between the decision nodes. The method receives current traffic
flow rates for each of the links from a plurality of sensors
located within the links. From this information the invention
predicts future traffic flow rates for each outgoing link of each
node based on current traffic flow rates for each incoming link of
each the node and based on current traffic flow rates of each the
outgoing link.
[0028] The invention compares predicted traffic flow rates for each
outgoing link of each node to determine which of the outgoing links
have the lowest traffic flow rates. This allows the method to route
users of the transportation system to outgoing links that have the
lowest traffic flow rates. In addition, after a user is routed to a
certain link, the method, adds users routed to a link to a
predicted traffic flow rate for the link. The method identifies a
link as being congested if either of a current traffic flow level
or a predicted traffic flow level exceeds a predetermined maximum.
This allows the invention to route users of the transportation
system away from congested links. Thus, the invention can receive a
final destination from a user (an indication of where the user
desires the transportation system to take them) and the routing
only considers links that will allow the user to reach the final
destination.
[0029] The traffic management system has the ability to monitor
traffic conditions in real-time via an attached group of sensor
devices such as roadside sensors, traffic cameras, and in-car speed
and position sensors. Other information describing events which may
affect traffic (e.g., major sporting events) can be manually
entered into the system. The system also has the ability to
transmit traffic command instructions (indicating what route or
route segment to follow) or decision rule sets (e.g., choose route
#1 with probability p.sub.1, route #2 with probability p.sub.2,
etc.) to drivers via in-car navigation devices.
[0030] In one embodiment a user will enter a travel destination
into a vehicle navigation device. The device will send the source
and destination locations to the travel management system. The
travel management system will calculate a route for the user based
on a group of factors. First, it will consider any current traffic
congestion affecting the user's commute from source and
destination. Secondly, it will consider any impending traffic flows
that will affect the user's commute between the source and
destination. The invention performs this step primarily by
considering any decisions the system has made or is making to route
other vehicles in such a way that could affect the user's impending
commute. This is an example of the coordinated decision-making
process provided by embodiments herein. Third, the system also
incorporates knowledge of other events and predicts any affects on
the user's commute (e.g., major sporting and entertainment events).
The system calculates a route, or produces rules for locally
deciding a route, attempting to avoid the effects of all the
previous factors and sends this information back to the user's
navigation system. The system will display either the route chosen
by the system or a set of routes from which the user may choose.
This entire process may repeat as the user commutes along his or
her path so that updated decisions are made available.
[0031] While the invention has been described in terms of several
exemplary embodiments, those skilled in the art will recognize that
the invention can be practiced with modification within the spirit
and scope of the appended claims. For example, the present
invention applies to vehicle highway management systems in which a
system of sensors are used to monitor highway traffic, an algorithm
is applied to the collected data and corresponding signals are sent
to a system of controllers such that each controller resides in an
individual car and controls that car's routing decisions. In
another embodiment of the present invention, an indoor moving
sidewalk used to move pedestrians between indoor locations could be
deployed with a network of sensors for monitoring the moving
pedestrian traffic and routing decisions could be made using a
system of controllers throughout the system. Further, it is noted
that, Applicants' intent is to encompass equivalents of all claim
elements, even if amended later during prosecution.
* * * * *