WO2022033677A1

WO2022033677A1 - System and method for adaptive traffic signal planning and control

Info

Publication number: WO2022033677A1
Application number: PCT/EP2020/072641
Authority: WO
Inventors: Saadhana B VENKATARAMAN; Vijaya Sarathi Indla; Saikat Mukherjee; Asit DEVA; Bony Mathew; Ram PADHY; Sagar PATHRUDKAR; Bristi SINGH; Alok KAJLA
Original assignee: Siemens Aktiengesellschaft
Priority date: 2020-08-12
Filing date: 2020-08-12
Publication date: 2022-02-17

Abstract

An adaptive traffic control system (100) and method (200) are provided, having a data collection system (101), a data processing system (102) and an adaptive traffic signal plan controller (ATC) (110) operably connected to traffic light(s) (110A). The data collection system (101) includes a video data collection module (104) and a crowdsource data collection module (105) for obtaining traffic management data. The data processing system (102) has a dynamic mode selection module (106) dynamically selecting the video data collection module (104) and/or the crowdsource data collection module (105) for obtaining the traffic management data and transmitting to the ATC (110), based on predefined parameters. The ATC (110) generates a traffic signal plan for the traffic light(s) (110A) based on the traffic management data.

Description

System and method for adaptive traffic signal planning and control

TECHNICAL FIELD

Various embodiments of the disclosure relate to on-road traffic management in general and more particularly, to a system and method for improving traffic signal planning and control.

BACKGROUND

Automobiles are a part of everyday life in urban and suburban areas of any country. Traffic Signals are known to be an integral part of the urban landscape and the surrounding communities, and are essential for controlling the flow of traffic on roads, large and small. Drivers must pay attention to traffic signals and failure to heed them results in increased traffic congestion and accidents.

The primary role of any traffic signal is to avoid congestion in an effective and efficient manner by trying to keep the traffic flow moving as much as possible. It is well known, for example, to “time” the traffic lights along any road having traffic lights, for example, a stretch of a highway so that vehicles progressing along the highway at the legal speed limit will encounter a reduced number of red lights causing them to have to stop. Such pre-defined timing of the traffic signals is adequate as long as automobile drivers are observing the speed limit and the traffic is not impeding their progress. However, it is fairly common for drivers on highways and internal city roads to exceed the speed limit without considering the timing of the lights. Moreover, with growing population, traffic congestion has become one of the biggest problems in almost all countries.

Conventional traffic controlling techniques, such as aforementioned timing technique, are typically simple and rule based thereby, rendering them to be inefficient to cope with constantly changing traffic profiles across various geographies. Traffic intersections are one of the most prevalent bottlenecks in urban environments and thus, traffic signal control plays a vital role in urban traffic management.

Presently, adaptive traffic signal plan generators also referred to as adaptive traffic controllers (ATCs) are deployed to cope with said multi-dimensional problem posed by growing dynamics associated with the automobiles, their driver behaviors, the demographics and the road conditions. ATCs continuously monitor the traffic conditions and change the signal plan accordingly. Hence, ATCs are proven to be more efficient as they are responsive in real time to the fluctuations of traffic demand. However, continuous collection of the traffic data for making accurate and speedy control decisions is a crucial part of ATC. ATC involves deploying various sensors at traffic intersections to monitor the traffic conditions. These sensors generally include cameras, load cells, proximity sensors etc., which lead to issues of high infrastructure and capital costs, maintainability issues, safety of the hardware as the sensors are prone to theft, as well as performance degradation probability with time and weather.

SUMMARY

Therefore, it is an object of the present disclosure to provide a system and a method for optimizing data processing associated with adaptive traffic control (ATC) such that reliability of ATC is increased, dependence on variety of sensors is decreased and therefore costs associated therewith are decreased.

The adaptive traffic control system disclosed herein comprises a data collection system obtaining traffic management data and an adaptive traffic signal plan controller, hereby referred to as ATC, operably connected to one or more traffic lights. The ATC generates a traffic signal plan for the traffic light(s) based on the traffic management data.

As used herein, “traffic management data” refers to data pertaining to automobile traffic and to ambient surroundings of the automobile traffic, captured in real-time or near realtime. The data collection system includes a video data collection module and a crowdsource data collection module. The video data collection module employs sensor(s) recording the traffic management data in real-time. As used herein, the term „sensor(s)“ mainly refers to camera(s) such as closed circuit television cameras recording images and videos of traffic present in surroundings thereof. According to one aspect of the present disclosure, the sensor(s) may include laser based sensors such as LiDARs and LADARs, ambient conditions sensors sensing temperature, humidity, pressure, particulate matter, etc. The crowdsource data collection module obtains the traffic management data from multiple infrastructure-devoid sources, for example, automobiles and portable electronic devices travelling along with the automobiles. The crowdsourced data mainly includes travelling speed, travelling time, travelling routes, etc., recorded by the automobiles and/or navigation applications installed on mobile phones or devices installed in or travelling along with the automobiles. The crowdsources data may be obtained from a mix of consumer devices and commerical fleets such as Taxis and delivery trucks regularly plying the roads. The consumer devices have the advantage of plying areas and roads deeper within an urban city thereby, providing access to narrower and less trodden routes.

The data collection system stores the traffic management data in a traffic management database in form of static data and dynamic data. The static data comprises, for example, data associated with the sensors such as a number of cameras installed in a particular geographical area, focal length of each camera, field of view of each camera, positional coordinates of each camera, shutter speed of each camera, network or road geometry, that is orientation of the road whether it curves or is a straight road, length of a road with respect to the camera installed thereon, number of traffic lights installed on every junction on each road, etc. According to one aspect of the present disclosure, the static data also includes traffic signal plans that the traffic lights follow at each junction, a log of events such as accidents, traffic congestions anomalies, etc., that may have occured. The dynamic data comprises, for example, data subject to change on a regular basis such as ambient conditions pertaining to weather, data recorded by the sensors at various time instances pertaining to traffic, data received from crowdsources including automobiles and respective portable electronic devices, etc.

The adaptive traffic control system disclosed herein, achieves the aforementioned object, in that a data processing system processes comprising a dynamic mode selection module dynamically selects the video data collection module and/or the crowdsource data collection module for obtaining the traffic management data, based on one or more predefined parameters. The predefined parameters comprise characteristic(s) associated with sensors) that the video data collection module employs and the real-time traffic management data recorded by the sensor(s). Advantaegously, the predefines parameters also comprise network geometry that is, characteristics associated with the terrain or road along which automobiles are plying, sections of each path on the terrain, map of the terrain, etc.

According to one aspect of the present disclosure, the predefined parameters include a set of rules that the dynamic mode selection module applies based on a status of the sensors, that is the cameras and/or the ambient sensors, and the traffic management data recorded by the sensor(s). According to this aspect the rules include a set of conditions pertaining to the sensor(s) that allow effective and efficient capture of the traffic management data. The conditions include no physical damage to the camera(s) and no physical occlusion present in field of view of the camera(s) either due to an object such as a heavy duty vehicle or a branch of tree blocking the field of view or due to weather such as fog and/or rainstorms blocking the field of view. When the above conditions are met, the dynamic mode selection module selects the video data collection module as an input mode. Moreover, when either of the above conducive conditions is not met, that is there exists physical damage or physical occlusion that renders the camera(s) non-functional until the root cause is cleared, the dynamic mode selection module selects crowdsource data collection module as an input mode. Furthermore, when there is no physical damage or physical occlusion present due to an external body, but there exists piling up of traffic beyond field of view of the camera as a result of road geometry or lack of camera(s) the dynamic mode selection module selects a hybrid input mode wherein both the vide data collection module and the crowdsource data collection module are selected as input mode.

Advantageously, when there is free flowing traffic where majority of the traffic is within the field of view of the camera(s), the data streams from video data collection module are solely considered as an input. Analyzing the number and nature of the flow of vehicles is performed periodically, and the moment an estimated traffic queue length exceeds the field of view of camera(s) without any physical damage or physical occlusion of the camera^), the data streams from both the video data collection module and the crowdsource data collection module are considered as an input. However, when there exists a physical damage or physical occlusion of the camera(s) due to an external object then the data streams from the crowdsource data collection module are solely considered as an input. The traffic flow is monitored parallelly at all times and if the queue lengths shrink beyond a certain predefined threshold, then the input mode is switched back to default mode of operation which is video data collection module. Advantageously, this switching enables enhanced reliability of incoming data along with automatic and real-time detection of traffic situation.

According to another aspect of the present disclosure, a recommendation may be provided by the adaptive traffic control system to a user of the adaptive traffic control system for selection of an input mode based on the predefined parameters. The user is, for example, an operator at a traffic control room responsible for managing traffic lights.

According to yet another aspect of the present disclosure, the adaptive traffic control system enables its user(s) to provide and update the predefined parameters. For example, the user may manually decide rules on when to opt for video data or crowdsourced data or both based on his/her experience. The data processing system comprises a traffic density estimation module that determines a macroscopic traffic density in a geographical area, based on the traffic management data obtained by the crowdsource data collection module when selected as an input mode. As used herein, the term ..macroscopic" refers to traffic density computed for various segments per traffic lane of a road. Advatageously, the macroscopic traffic density represents distribution of congestive occupancy of a road from crowdsourced travel time data. The traffic density estimation module uses crowdsourced data to find the travel time duration in traffic from one geographical coordinate to another. Likewise, the travel time duration for a trip associated with all the lanes in a junction. The traffic density estimation module comprises a traffic queue length estimation module that determines a peak traffic queue length in the geographical area for determining the macroscopic traffic density. The traffic queue length estimation module advantageously, employs segmentation techniques based on the road geometry that is, curved or straight, to allow discretization of the lane into measurable geographical coordinates.

The traffic density estimation module based on the travel time duration data determines the time spent in congestion for each lane in the junction. The traffic density estimation module determines a minimum travel time duration for the lane from the data collected, then it determines a current travel time duration for the lane from the same data collected, then it computes a difference between the above which is time spent in congestion for the lane. The above steps are repeated for each lane in the junction. The traffic queue length estimation module compares all the congestion times calculated for the lane, to determine a geographical co-ordinate in the lane from where the congestion queue begins. Advantageously, the traffic queue length estimation module also determined a width of the traffic queue based on the network geomtery data and thereby, detects granular information such as number of cars and buses present in the traffic queue. Advantageously, the traffic queue length estimation module enables the data processing system to learn traffic congestion patterns along with granulaties detected, over a period of time which are stored in the traffic management database and are applied by the ATC for efficient adaptive traffic signal planning and control.

The traffic density estimation module computes a velocity of automobiles in congestion, hereby referred to as crawl velocity. The crawl velocity is calculated by dividing the length of the lane by the maximum travel time duration for the lane. The maximum travel time duration is derived from the crowd sourced data. The traffic density estimation module, based on the above calculated parameters, computes length of congestion at each lane in the junction. The length of congestion is a product of crawl velocity and the time spent in congestion. The length of congestion essentially encodes the macroscopic traffic density in each lane at a junction.

The estimated macroscopic traffic density is used as an input to the ATC. Advantageously, the ATC includes a mapping function that maps traffic densities in each lane at a junction to the traffic signal plans generated. This mapping function is generated by formulating the problem of traffic signal control as that of a partially observable Markov Decision Process and trained by exposing to various scenarios with intermittent adversarial inputs. The mapping function thus, continuously learns to adapt to varying traffic conditions and becomes robust to adversarial conditions such as extreme traffic scenarios over a period of time.

The data processing system comprises a time division multiplexing module that synchronizes acquisition of the traffic management data via one or more sensors of the video data collection module with the traffic signal plan generated by the ATC. The time division multiplexing module enables processing of the traffic management data that the video data collection module obtains for various directions at a junction, to be time synchronized such that any analytics output derived by processing the traffic management data by the ATC is most relevant for the task being performed. For example, for the adaptive traffic signal plan generation task, the relevant analytics output is the maximum traffic queue length in each section or a lane leading to the junction. This peak queue length would typically occur at a time just before the corresponding traffic signal si scheduled to turn green. Thus, synchronizing the video data acquisition and processing timing with the available signal plan timing would save substantial computational effort and allow the same adaptive signal plan generation to be done with much lesser hardware requirements. This reduction in acquisition and computation redundancy results in effectively a large number of applications being deployable in limited computational resources.

Advantageously, The adaptive traffic control system can be generalized and scaled to a zone containing multitude of junctions and lanes, thereby, enabling a coordinated behavior between different junctions and optimization of traffic flow throughout the zone.

According to the present disclosure, the adaptive traffic control system includes a memory unit, and at least one processor communicatively coupled to the memory unit. The memory unit refers to all computer readable media, for example, non-volatile media, volatile media, and transmission media except for a transitory, propagating signal. The memory unit stores computer program instructions defined by the modules of the adaptive traffic signal control system, and the processor executes these computer program instructions respectively. The processor refers to any one of microprocessors, central processing unit (CPU) devices, finite state machines, microcontrollers, digital signal processors, an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA), etc., or any combination thereof, capable of executing computer programs or a series of commands, instructions, or state transitions. The processor may also be implemented as a processor set comprising, for example, a general-purpose microprocessor and a math or graphics co-processor.

Also disclosed herein, is a data processing system for adaptive traffic signal planning and control operably communicating with a data collection system comprising a video data collection module and a crowdsource data collection module, and an adaptive traffic signal plan controller (ATC). The the data processing system comprises the dynamic mode selection module, the traffic density estimation module having the traffic queue length estimation module, and the time division multiplexing module. The dynamic mode selection module dynamically selects the video data collection module and/or the crowdsource data collection module of the data collection system, for obtaining the traffic management data, based on one or more predefined parameters, as disclosed above.

Advantageously, one or more components of the data processing system are deployed in a cloud-computing environment. As used herein, “cloud computing environment” refers to a processing environment comprising configurable computing physical and logical resources, for example, networks, servers, storage, applications, services, etc., and data distributed over the cloud platform. The cloud computing environment provides on-de- mand network access to a shared pool of the configurable computing physical and logical resources. The industrial database, according to another embodiment of the present invention, is a location on a file system directly accessible by the information retrieval system. Advantageously, the traffic density estimation module, the traffic queue length estimation module and the dynamic mode selection module are deployed in a cloud computing environment whereas the time division multiplexing module is deployed as an edge deivce in proximity of the video data collection module and the traffic signals at a junction.

Also disclosed herein, is a method for adaptive traffic signal planning and control. The method employs aforementioned adaptive traffic control system. The method comprises obtaining traffic management data by the video data collection module and/or the crowdsource data collection module of the data collection system, dynamically determining an input mode for obtaining the traffic management data by the dynamic mode selection module of the data processing system, wherein the input mode is the video data collection module and/or the crowdsource data collection module, transmitting the traffic management data from the dynamically selected input mode to the adaptive traffic signal plan controller (ATC), by the data processing system, and geenerating, by the ATC, a traffic signal plan for traffic light(s) operably connected to the ATC based on the traffic management data received from the data processing system.

Also disclosed herein, is a computer program product comprising a machine-readable instructions stored therein, which when executed by at least one server/processor perform the aforementioned method for adaptive traffic signal planning and control.

The above summary is merely intended to give a short overview over some features of some embodiments and implementations and is not to be construed as limiting. Other embodiments may comprise other features than the ones explained above.

BRIEF DESCRIPTION OF THE DRAWINGS

The above and other elements, features, steps and characteristics of the present disclosure will be more apparent from the following detailed description of embodiments with reference to the following figures:

FIG 1 A illustrates schematic representation of an adaptive traffic control system, according to an embodiment of present disclosure.

FIG 1 B is a schematic representation of components of a cloud-computing environment in which the data processing system of the adaptive traffic control system shown in FIG 1A is deployed, according to an embodiment of present disclosure.

FIG 2A is a process flowchart representing a method for processing data for adaptive traffic control, according to an embodiment of present disclosure. FIG 2B is a process flowchart representing a method for dynamically controlling a mode of input for adaptive traffic control, according to an embodiment of present disclosure.

FIG 2C is a process flowchart representing a method for time division multiplexing of video data for adaptive traffic control, according to an embodiment of present disclosure.

FIG 2D is a process flowchart representing a method for determining traffic density based on crowdsourced data for adaptive traffic control, according to an embodiment of present disclosure.

FIGS 3A-3B illustrate schematic representations showing traffic queue lengths required for determining the traffic density for adaptive traffic control, according to an embodiment of present disclosure.

FIG 4 illustrates a schematic representation of adaptive traffic control based on time division multiplexed video data, according to an embodiment of present disclosure.

DETAILED DESCRIPTION OF EMBODIMENTS

In the following, embodiments of the disclosure will be described in detail with reference to the accompanying drawings. It is to be understood that the following description of embodiments is not to be taken in a limiting sense.

The drawings are to be regarded as being schematic representations and elements illustrated in the drawings, which are not necessarily shown to scale. Rather, the various elements are represented such that their function and general purpose become apparent to a person skilled in the art. Any connection or coupling between functional blocks, devices, components, or other physical or functional units shown in the drawings or described herein may also be implemented by an indirect connection or coupling. A coupling between components may also be established over a wireless connection. Functional blocks may be implemented in hardware, firmware, software, or a combination thereof.

FIG 1A illustrates schematic representation of an adaptive traffic control system 100, according to an embodiment of present disclosure. The adaptive traffic control system 100 comprises a data collection system 101 , a data processing system 102, and an adaptive traffic controller 110 hereby referred to as ATC, operably communicating therebetween. The ATC 110 is physically connectable to a traffic signal 110A. The ATC 110 is known to be deployable as an edge device installable at and connectable to a traffic signal 110A for dynamically generating traffic signal plans which are followed by the traffic signal 110A for smooth flow of traffic at a junction.

The data collection system 101 of the adaptive traffic control system 100 comprises a video data collection module 104 and a crowdsource data collection module 105. The video data collection module 104 primarily receives inputs from one or more sensors such as cameras for example closed circuit television (CCTV) cameras. This input mainly comprises video streams captured by the cameras of their surroundings. The sensors may also include ambient condition monitoring sensors including temperature sensors, humidity sensors, wind speed sensors, particulate matter sensors, etc., that provide weather related information and/or LIDAR sensors, RADAR sensors, etc., that provide network geometry related information of the surroundings. The crowdsource data collection module 105 receives data pertaining to a particular area primarily from multiple automobiles plying in that area. Such data mainly includes a speed at which an automobile is moving, a navigation path of the automobile, and travel times required by the automobile to cover various sections of the path on which it is plying. The data may also include reports or feedback provided by the automobile drivers regarding road conditions, accidents, etc.

The data collection system 101 periodically transmits the data collected by the video data collection module 104 and the crowdsource data collection module 105 to a traffic management database 103. The traffic management database 103 categorically stores the data received, into majorly static data 103A and dynamic data 103B. The static data 103A includes the network geometry that is characteristics associated with the terrain along which automobiles are plying, sections of each path on the terrain, parameters associated with sensors, for example, positional co-ordinates of cameras, focal lengths of the cameras, etc. The dynamic data 103B includes weather data such as ambient weather conditions including temperature, humidity, wind, particulate matter levels, etc. The dynamic data may also include the inputs received from the video data collection module 104, that is, video streams and the crowdsource data collection module 105. The traffic management database 103 may also receive data from one or more external sources (not shown) including traffic control rooms, navigation satellites, manned or unmanned vehicles patrolling entities, etc. The data collection system 101 also transmits the data collected by the video data collection module 104 and the crowdsource data collection module 105 to the data processing system 102. The data processing system 102 comprises a dynamic mode selection module 106, a time division multiplexing module 107, and a traffic density estimation module 108. The traffic density estimation module 108 comprises a traffic queue length estimation module 109 therewithin. The time division multiplexing module 107 receives data from the video data collection module 104. The traffic density estimation module 108 receives data from the crowdsource data collection module 105. The dynamic mode selection module 106 dynamically selects an input for the ATC 110. The input comprises a video data and/or a crowdsourced data based on pre-defined parameters.

The adaptive traffic control system 100 shown in FIG 1A is a combination of a cloudbased system and an edge system. The sensor module 104, the time division multiplexing module 107 and the ATC 110 are deployable as an edge device installed on or in proximity of the traffic signal 110A. The dynamic mode selection module 106 and the traffic density estimation module 108 are deployable in a cloud-based environment. The traffic management database 103 is deployable either in a cloud-based environment or as an edge device operably communicating with rest of the components of the adaptive traffic control system 100.

FIG 1 B is a schematic representation of components of a cloud-computing environment 111 in which the data processing system 102 of the adaptive traffic control system 100 shown in FIG 1A is deployed, according to an embodiment of present disclosure. The adaptive traffic control system 100 employs the data processing system 102 and an application programming interface (API) 112. The API 112 employs functions 112A-112N each of which enable the data processing system 102 to retrieve and/or receive data stored in the traffic management database 103 and/or from the data collection system 101 having the video data collection module 104, and the crowdsource data collection module 105. The traffic management database 103 comprises data models 113A-113N which store data received from the video data collection module 104, the crowdsource data collection module 105 and/or the external data sources (not shown). It may be noted that each of the data models 113A-113N can store data in a compartmentalized manner pertaining to a category to which the data is relevant. For example, aforementioned categories of static data and dynamic data. Also, each of the functions 112A-112N is configured to access one or more data models 113A-113N in the traffic management database 103. The data processing system 102 works autonomously. However, there may be a provision that ena- bles a user of the adaptive traffic control system 100, to access the data processing system 102 via an interactive graphical user interface (not shown) to configure or selectively manage processing of the data received from the video data collection module 104 and the crowdsource data collection module 105.

The data processing system 102 dynamically selects desired mode of input with help of the dynamic mode selection module 106 and transforms the input into an API call. The data processing system 102 forwards this API call to the API 112 which in turn invokes one or more appropriate API functions 112A-112N responsible for retrieving/storing the data provided. Then, the API 112 determines one or more data models 113A-113N within the traffic management database 103 for performing said operation of retrieval/storage of data. The API 112 returns the retrieved data, or an acknowledgement of data stored to the data processing system 102 which in turn may forward the same to the user of the adaptive traffic control system 100, via the graphical user interface. The data that the user may want to retrieve may include, for example, reports of traffic signal plans generated for the ATC 110, traffic congestions, if any, events such as accidents, weather statistics, road conditions, etc.

FIG 2A is a process flowchart representing a method 200 for processing data for adaptive traffic control, according to an embodiment of present disclosure. The method 200 shown in FIG 2A employs the data processing system 102 of the adaptive traffic control system 100 shown in FIGS 1A-1 B. At step 201 , the data processing system 102 receives data from the data collection system 101, the traffic management database 103 as disclosed in the detailed description of FIG 1 B, and/or one or more external sources. The data received mainly includes the aforementioned static and dynamic data. At step 202, the dynamic mode selection module 106 of the data processing system 102 dynamically selects one of three modes of input to be provided to the ATC 110 based on predefined parameters. The modes include: a mode ‘A’ being the video data input mode where the data received from the video data collection module 104 is used, a mode ‘B’ being the crowdsource data input mode where the data received from the crowdsource data collection module 105 is used, and a hybrid mode ‘C’ where the data received from both the video data collection module 104 and the crowdsource data collection module 105 is used.

At step 203, when mode ‘A’ is selected, the time division multiplexing module 107 is invoked by the data processing system 102. The time division multiplexing module 107, at step 206, processes the data received from the video data collection module 104. At step 209, the data processing system 102 transfers the output of the time division multiplexing module 107 to the ATC 110 for adaptive traffic signal plan generation.

At step 204, when mode ‘B’ is selected, the traffic density estimation module 108 is invoked by the data processing system 102. The traffic density estimation module 108, at step 207, processes the data received from the crowdsource data collection module 105 to calculate macroscopic traffic density of a given geographical area. At step 208, the traffic density estimation module 108 invokes the traffic queue length estimation module for effective estimation of macroscopic traffic density. At step 209, the data processing system 102 transmits the output of the traffic density estimation module 108 to the ATC 110 for adaptive traffic signal plan generation.

At step 205, when hybrid mode ‘C’ is selected, the time division multiplexing module 107 and the traffic density estimation module 108 are invoked by the data processing system 102. This invoking may happen in a sequential manner or simultaneously based on processing bandwidth available. In the hybrid mode ‘C’ of operation, the data processing system 102 transfers the output of the time division multiplexing module 107 to the ATC 110 and also transmits the output of the traffic density estimation module 108 to the ATC 110.

At step 210, the ATC 110 generates a traffic signal plan for traffic management and routing at a junction where the traffic signal(s) 110A are physically installed and to which the ATC 110 is physically connected.

FIG 2B is a process flowchart representing a method 202 for dynamically controlling a mode of input for adaptive traffic control, according to an embodiment of present disclosure. The method 202 shown in FIG 2B employs the dynamic mode selection module 106 of the data processing system 102 shown in FIGS 1A-1B. At step 202A, the dynamic mode selection module 106 receives data from the traffic management database 103 including the static data 103A and the dynamic data 103B. At step 202B, the dynamic mode selection module 106 receives data from the video data collection module 104. At step 202C, the dynamic mode selection module 106 receives data from the crowdsource data collection module 105. At step 202D, the dynamic mode selection module 106 based on the static data 103A and the dynamic data 103B determines whether the conditions associated with the traffic and the sensors, that is video cameras, are conducive. The conduciveness of conditions is decided based on clarity of weather, sufficient range of visibility, desired sensors operation, no physical damage to sensors, no occlusion of sensors, traffic pile up is within field of view of the sensors, etc. If all the conditions are found to be conducive, then the dynamic mode selection module 106 activates the mode ‘A’ of operation where the ATC 110 relies on the video data received from the video data collection module 104. If one or more of the conditions are not found to be conducive, then at step 202E, the dynamic mode selection module 106 determines whether there exist any performance problems with the sensors. The performance problems include, any physical damage to the camera(s) rendering it to be non-functional, heavy occlusions in field of view of the camera(s) caused, for example, due to heavy duty automobiles in its field of view, animals, birds, fallen trees, etc., and/or bad quality of images due to poor visibility either as a result of occlusion or due to worsening of weather such as rainy, foggy, stormy, or snowy weather. Basically, if there is a condition existing that renders the camera(s) to be completely non-functional for a certain period of time, then the dynamic mode selection module 106 activates the mode ‘B’ of operation where the ATC 110 relies completely on the crowdsourced data received from the crowdsource data collection module 105. If none of the above problems exist with respect to the cameras being rendered completely nonfunctional for a certain period of time, yet the conditions are not found to be conducive despite of camera performing as desired and clear weather, for example limitation in field of view of the camera(s) because of pile up of traffic extending beyond the field of view of the camera(s) as a result of road geometry or insufficient number of cameras installed, then the dynamic mode selection module 106 activates the mode ‘C’ of operation, also known as the hybrid mode, where the ATC 110 relies on both the video data received from the video data collection module 104 and the crowdsourced data received from the crowdsource data collection module 105.

FIG 2C is a process flowchart representing a method 206 for time division multiplexing of video data for adaptive traffic control, according to an embodiment of present disclosure. The method 206 shown in FIG 2C employs the time division multiplexing module 107 of the data processing system 102 shown in FIGS 1A-1 B. The time division multiplexing is employed when either the mode ‘A’ for vide data or the hybrid mode ‘C’ combining both the video data and the crowdsource data is selected. The time division multiplexing ensures synching of the video data processing timings with the available signal plan timings thereby, saving substantial computational effort and allowing the same adaptive signal plan generation to be done with much lesser hardware requirements.

At step 206A, the time division multiplexing module 107 receives data from the video data collection module 104. At step 206B, the time division multiplexing module computes a number of directions in which traffic is inflowing at a junction for which the traffic signal plan is to be generated by the ATC 110, based on the video data received. For example, normally there are three or more directions of traffic flow at any given junction. At step 206C, the time division multiplexing module 107 checks whether the traffic signal for a particular direction is green. If yes, then at step 206H it considers the next direction thereby, continuing the check until the signal is found to be red for a particular direction. At step 206D, the time division multiplexing module 107 checks how much time ‘Te’ has elapsed since the signal turned red and whether this time ‘Te’ is equal to a predefined threshold. If not, then at step 206G the next direction for which the signal is red is considered until at step 206D the direction having a red signal for a time ‘Te’ equal to the threshold is found. The threshold is a time at which the signal is about to turn green from red. This comparison is performed on the premise that a peak traffic queue pile up would occur at a time just before the signal turns green at any road section and the processing of video data corresponding to this road section should be performed just before the peak traffic queue pile up takes place. At step 206E, the video data from the video data collection module 104 pertaining to the direction of traffic for which the signal is about to turn green is acquired by activating the data transfer from the camera and transferred to the ATC 110 at step 209 for adaptive traffic signal plan generation.

Thus, the image acquisition can be synced with the signal plan timings so that the image is acquired just a few seconds before the signal for that direction is supposed to turn green. This reduction in acquisitional and computational redundancy results in effectively a large number of applications being deployable in limited computational resources on field.

FIG 2D is a process flowchart representing a method 207 for determining traffic density based on crowdsourced data for adaptive traffic control, according to an embodiment of present disclosure. The method 207 shown in FIG 2D employs the traffic density estimation module 108 and traffic queue length estimation module 109 of the data processing system 102 shown in FIGS 1A-1B. At step 207A, the traffic density estimation module 108 receives the crowdsourced data from the crowdsource data collection module 105. At step 208A, the traffic queue length estimation module 109 determines lane geometries from the crowdsource data received. At step 208B, the traffic queue length estimation module 109 checks whether the lane is relatively straight or has curves therein. If no, that is if not straight, at step 208C the traffic queue length estimation module 109 virtually segments the lane by map-matching mechanism illustrated and described in FIG 3B. At step 208D, the traffic queue length estimation module 109 derives discrete co-ordinates Xo to Xn of the lane having a running length n based on the virtual segmentation. If the lane geometry is relatively straight, then at step 208E, a plain discretization of the lane having a length n is performed by dividing it into co-ordinates Xo to Xn of fixed lengths.

A counter i is set to be equal 0 and at step 207B, the traffic density estimation module 108 computes a travel time TT taken by an automobile plying on the lane to travel from a coordinate Xi to a co-ordinate Xi+1 within the respective lane, based on the crowdsourced data received. At step 207C, the traffic density estimation module 108 determines based on historical data stored in the traffic management database 103, a minimum travel time TTmin required to travel from Xi to Xi+1. At step 207D, the traffic density estimation module 108 computes a congestion time Tc as a difference between current travel time TT and minimum travel time TTmin.

Tc = TT current — TT min

At step 207E, the traffic density estimation module 108 increases the counter i by 1 and at step 207F it checks whether the counter i is greater than the total number of co-ordinates ‘n’. If No, then the steps 207B to 207E are repeated for the next section of the lane that is Xi+1 to Xi+2. Upon completing computation of congestion times Tc for entire lane, at step 208F, the traffic queue length estimation module 109 prunes thorough all the congestion times Tc computed to determine the largest congestion time Tcmax and its corresponding co-ordinate Xm having traffic piled up therefrom.

Tcmax = max {Tel, Tc2 ... . Ten}

It is to be understood that m is a co-ordinate on the lane from where the traffic congestion starts such that Xo to Xm is congested and Xm to Xn is not congested. At step 208G, the traffic queue length estimation module 109 stores the travel times TT with their respective co-ordinates Xi to Xi+1 and the congestions times Tc in the traffic management database 103 for future computations.

At step 207G, the traffic density estimation module 108 computes a crawl velocity Vc by dividing length of the lane, that is ‘n’, by maximum travel time TTmax required for travelling in the lane. The maximum travel time TTmax is found from the current travel times TT computed for the lane between various co-ordinates Xo to Xn of the lane.

Vc = n/TTmax At step 207H, a length of the congestion Lc in a lane is computed by multiplying the crawl velocity Vc by the maximum congestion time Tcmax. The maximum congestion time Tcmax is considered here because the congestion time Tc for a subsection longer than the congestion traffic queue, that is from Xo to Xm, will be same and the congestion time Tc for a subsection shorter than the congestion traffic queue will vary because the automobiles would have already spent some time in the congestion traffic queue.

Lc = Vc * Tcmax

At step 209, the ATC 110 receives the congestion length Lc from the traffic density estimation module 108 representing the traffic density in that lane. The ATC 110 is primarily a mapping function that maps traffic densities in each lane at a junction to the traffic signal plans. This mapping function is generated, for example, by formulating the problem of traffic signal control as that of a partially observable Markov Decision Process and trained by exposing to various scenarios with intermittent adversarial inputs. The mapping function thus, learns to adapt to varying traffic conditions and becomes robust to adversarial conditions such as extreme traffic scenarios.

FIGS 3A-3B illustrate schematic representations showing traffic queue lengths required for determining the traffic density for adaptive traffic control, according to an embodiment of present disclosure. FIG 3A shows a relatively straight lane geometry having a length ‘n’ and it is discretized into n number sections by co-ordinates Xo to Xn such that Xo is closest to the traffic signal or to a junction. A formula used by the traffic queue length estimation module 109 in discretizing the lane having a relatively straight geometry is given below:

X_{i =} X₀ + (X_n - X₀)/n

As shown in FIG 3A and as disclosed in the detailed description of FIG 2D, Xm is the coordinate from where the congestion begins and Lc represent a length of congestion in the lane which is a metric of macroscopic traffic density in the lane.

FIG 3B shows a curved lane having a running length ‘n’. Discretizing a curved lane is not as simple as shown above for a relatively straight lane. Hence, as shown in FIG 3B the traffic queue length estimation algorithm 109 employs a map-matching technique by projecting a virtual plane of ‘n’ co-ordinates Pviroto Pvim onto the curved lane to derive the map-matched co-ordinates P mmo tO P mmn, P mmo being closest to the junction. FIG 4 illustrates a schematic representation of adaptive traffic control based on time division multiplexed video data, according to an embodiment of present disclosure. As shown in FIG 4, there are four traffic signals S1, S2, S3 and S4 routing traffic in four lanes at a particular junction. There are four cameras 401 , 402, 403, and 404 mounted to capture data respective to each of the four lanes. Phases P1 , P2, P3 and P4 denote the sequence in which the traffic signals S1 , S2, S3 and S4 turn green, for example, the Signal S1 turns green in the phase P1. By applying the method 206 shown in FIG 2C, the time division multiplexing module 107 synchronizes acquisition of data from the cameras 401-404 with the phases P1-P4 of operation of the traffic signals S1-S4 installed at a junction. This synchronization helps in optimizing the real-time processing of data received by the ATC 110 via the video data collection module, that is, the cameras 401-404. Thus, the data processing system 102 of the adaptive traffic control system 100 shown in FIG 1A, optimizes data processing required for adaptive traffic signal plan generation while ensuring minimal computational resources are used. The computational resources mainly are field computers. With this approach, multiple functionalities can be deployed on a single field computer, an example of which is provided in Table 1 below:

Table 1

It may be appreciated that aforementioned communication exchange happening between the components of the adaptive traffic control system 100 may employ available means allowing for a speedy yet secure communication to take place. Such means, may involve usage of protocols supported by V2X communication including but not limited to Transmission Control Protocol (TCP), Internet Protocol (IP), User Datagram Protocol (UDP), OPC Unified Architecture (OPC-UA) Protocol, etc., and usage of networks involving wireless networks such as 4G, LTE or 5G that meet desired requirements and are compliant with the standards laid down for traffic management such as IEEE 802.11.

Where databases are described such as the traffic management database 103, it will be understood by one of ordinary skill in the art that (i) alternative database structures to those described may be readily employed, and (ii) other memory structures besides databases may be readily employed. Any illustrations or descriptions of any sample databases disclosed herein are illustrative arrangements for stored representations of information. Any number of other arrangements may be employed besides those suggested by tables illustrated in the drawings or elsewhere. Similarly, any illustrated entries of the databases represent exemplary information only; one of ordinary skill in the art will understand that the number and content of the entries can be different from those disclosed herein. Further, despite any depiction of the databases as tables, other formats including relational databases, object-based models, and/or distributed databases may be used to store and manipulate the data types disclosed herein. Likewise, object methods or behaviors of a database can be used to implement various processes such as those disclosed herein. In addition, the databases may, in a known manner, be stored locally or remotely from a device that accesses data in such a database. In embodiments where there are multiple databases in the system, the databases may be integrated to communicate with each other for enabling simultaneous updates of data linked across the databases, when there are any updates to the data in one of the databases.

The present disclosure can be configured to work in a network environment comprising one or more computers that are in communication with one or more devices via a network. The computers may communicate with the devices directly or indirectly, via a wired medium or a wireless medium such as the Internet, a local area network (LAN), a wide area network (WAN) or the Ethernet, a token ring, or via any appropriate communications mediums or combination of communications mediums. Each of the devices comprises processors, some examples of which are disclosed above, that are adapted to communicate with the computers. In an embodiment, each of the computers is equipped with a network communication device, for example, a network interface card, a modem, or other network connection device suitable for connecting to a network. Each of the computers and the devices executes an operating system, some examples of which are disclosed above. While the operating system may differ depending on the type of computer, the operating system will continue to provide the appropriate communications protocols to establish communication links with the network. Any number and type of machines may be in communication with the computers.

The present disclosure is not limited to a particular computer system platform, processor, operating system, or network. One or more aspects of the present disclosure may be distributed among one or more computer systems, for example, servers configured to provide one or more services to one or more client computers, or to perform a complete task in a distributed system. For example, one or more aspects of the present disclosure may be performed on a client-server system that comprises components distributed among one or more server systems that perform multiple functions according to various embodiments. These components comprise, for example, executable, intermediate, or interpreted code, which communicate over a network using a communication protocol. The present disclosure is not limited to be executable on any particular system or group of systems, and is not limited to any particular distributed architecture, network, or communication protocol.

The foregoing examples have been provided merely for the purpose of explanation and are in no way to be construed as limiting of the present disclosure disclosed herein. While the disclosure has been described with reference to various embodiments, it is understood that the words, which have been used herein, are words of description and illustration, rather than words of limitation. Further, although the disclosure has been described herein with reference to particular means, materials, and embodiments, the disclosure is not intended to be limited to the particulars disclosed herein; rather, the disclosure ex- tends to all functionally equivalent structures, methods and uses, such as are within the scope of the appended claims. Those skilled in the art, having the benefit of the teachings of this specification, may affect numerous modifications thereto and changes may be made without departing from the scope of the disclosure in its aspects.

Claims

Patent Claims

1 . An adaptive traffic control system (100), comprising:

- a data collection system (101) configured to obtain traffic management data and comprising a video data collection module (104) and a crowdsource data collection module (105);

- an adaptive traffic signal plan controller (110) operably connected to one or more traffic lights (110A), configured to generate a traffic signal plan for the one or more traffic lights (110A) based on the traffic management data received from the data collection system (101); characterized by:

- a data processing system (102) comprising: o a dynamic mode selection module (106) configured to dynamically select one or more of the video data collection module (104) and the crowdsource data collection module (105) of the data collection system (101), for obtaining the traffic management data, based on one or more predefined parameters.

2. The adaptive traffic control system (100) according to claim 1 , wherein the video data collection module (104) employs one or more sensors (401-404) configured to record the traffic management data in real-time.

3. The adaptive traffic control system (100) according to claim 1 , wherein the crowdsource data collection module (105) is configured to obtain the traffic management data from plurality of infrastructure-devoid sources comprising automobiles and portable electronic devices travelling along with the automobiles.

4. The adaptive traffic control system (100) according to claim 1 , wherein the data collection system (101) is configured to store the traffic management data in a traffic management database (103) in form of static data (103A) and dynamic data (103B).

5. The adaptive traffic control system (100) according to claim 1 , wherein the predefined parameters comprise one or more characteristics associated with one or more sensors (401-404) employed by the video data collection module (104) and the real-time traffic management data recorded by the sensors (401-404). The adaptive traffic control system (100) according to claim 1 , wherein the data processing system (102) further comprises a traffic density estimation module (108) configured to determine a macroscopic traffic density in a geographical area, based on the traffic management data obtained by the crowdsource data collection module (105). The adaptive traffic control system (100) according to claim 6, wherein the traffic density estimation module (108) comprises a traffic queue length estimation module (109) configured to determine a peak traffic queue length in the geographical area for determining the macroscopic traffic density. The adaptive traffic control system (100) according to claim 1 , wherein the data processing system (102) further comprises a time division multiplexing module (107) configured to synchronize acquisition of the traffic management data via one or more sensors (401-404) of the video data collection module (104) with the traffic signal plan generated by the adaptive traffic signal plan controller (110). A data processing system (102) for adaptive traffic signal planning and control according to the claims 1-8 operably communicating with a data collection system (101) comprising a video data collection module (104) and a crowdsource data collection module (105), and an adaptive traffic signal plan controller (110), the data processing system (102) characterized by:

- a dynamic mode selection module (106) configured to dynamically select one or more of the video data collection module (104) and the crowdsource data collection module (105) of the data collection system (101), for obtaining the traffic management data, based on one or more predefined parameters. . A method (200) for adaptive traffic signal planning and control, said method (200) employing an adaptive traffic control system (100) and comprising:

- obtaining (201) traffic management data by one or more of a video data collection module (104) and a crowdsource data collection module (105) of a data collection system (101) of the adaptive traffic control system (100); characterized by:

- dynamically determining (202) an input mode for obtaining the traffic management data by a dynamic mode selection module (106) of a data processing system (102) of the adaptive traffic control system (100), wherein the input mode comprises one or more of the video data collection module (104) and the crowdsource data collection module (105);

- transmitting (209) the traffic management data from the dynamically selected input mode to an adaptive traffic signal plan controller (110) of the adaptive traffic control system (100), by the data processing system (102); and

- generating (210), by the adaptive traffic signal plan controller (110), a traffic signal plan for one or more traffic lights (110A) operably connected to the adaptive traffic signal plan controller (110) based on the traffic management data received from the data processing system (102). The method according to claim 1 , wherein dynamically determining the input mode for obtaining the traffic management data comprises performing one of:

- selecting the video data collection module (104) for obtaining the traffic management data based on predefined parameters;

- selecting the video data collection module (104) and the crowdsource data collection module (105) for obtaining the traffic management data based on based on predefined parameters; and

- the crowdsource data collection module (105) for obtaining the traffic management data based predefined parameters. The method according to claim 11 , wherein the predefined parameters comprise at least on one or more characteristics associated with sensors (401-404) employed by the video data collection module (104). The method according to claim 10, further comprising synchronizing acquisition of traffic management data from the video data collection module (104) with the traffic signal plan, when the video data collection module (104) is selected as an input mode, by a time division multiplexing module (107) of the data processing system (102). The method according to claim 10, further comprising determining traffic density in a geographical area proximal to the traffic lights (110A), when the crowdsource data collection module (105) is selected as an input mode, by a traffic density estimation module (109) of the data processing system (102). The method according to claim 14, wherein determining the traffic density in a geographical area comprises computing traffic queue length in the geographical area, by a traffic queue length estimation module (109) of the traffic density estimation module (108).