Modelling traffic flow through a road network

GB202414641D0Pending Publication Date: 2024-11-20CITISEN LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
GB2024014641
Authority / Receiving Office
GB · GB
Patent Type
Applications
Current Assignee / Owner
Filing Date
2024-10-04
Publication Date
2024-11-20

AI Technical Summary

Technical Problem

Existing traffic modelling systems fail to accurately predict journey times and adjust travel decisions in real-time due to insufficient consideration of real-world road conditions, leading to inefficiencies and environmental impacts from increased fuel consumption.

Method used

A computer-implemented method using ANPR cameras to identify vehicle routes, assess road popularity, and evaluate traffic flow changes, allowing for real-time adjustments to minimize delays and optimize traffic flow.

Benefits of technology

Enables real-time prediction of journey times and mitigation of traffic congestion, reducing fuel consumption and enhancing traffic efficiency across the road network.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

Computer-implemented modelling of traffic flow through a road network comprises receiving information from automatic number plate recognition (ANPR) cameras that identifies vehicles travelling between
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field This disclosure relates to the modelling of traffic flow through a road network. Background Traffic modelling allows predictions to be made of the expected duration of journeys. Simulating the flow of traffic allows the impact of a road closure to be calculated. This can be achieved taking into consideration the speed limit that vehicles are permitted to travel along each of the roads, as well as the typical level of congestion that is likely to be encountered for a particular time of day. There is a demand for systems and methods that process traffic data to accurately predict and journey times and help commuters take travel decisions in real time. Summary Aspects of the present invention are set out by the claims. Further aspects are also described. Advantageously, aspects of the invention process real-world data that takes account of the road conditions for the specific road network that is being evaluated. This information can be provided in real-time, allowing commuters to take decisions during their journeys, in order to prevent delays. According to a first aspect, disclosure is provided of a computer-implemented method (S10) of modelling traffic flow through a road network. The method comprises: receiving (S14) information from automatic number plate recognition, ANPR, cameras that identifies a plurality of vehicles travelling between nodes of the road network; determining (SI5) roads between the nodes that are travelled along for each vehicle, based on the information received from the ANPR cameras, to identify a route through the road network for each of the vehicles; assessing (SI 5) a popularity measure for each of the roads of the road network, by comparing the routes of each of the vehicles; and evaluating (SI6) how a change to a road of the road network would impact traffic flow through the road network, based on the popularity measure of each of the roads of the road network. A technical effect of modelling how a change of a road will impact traffic flow is that an assessment can be made in advance of how this change should be accommodated. This provides an opportunity for negative effects to be mitigated, and positive effects to be encouraged. The predictions allow an assessment to be achieved for the road network as a whole, so that decisions can be taken that accommodate all road users. Advantageously, this ensures that traffic can flow as efficiently as possible, which reduces journey times, as well as having a positive environmental impact due to vehicles emitting less fuel. The method may further comprise predicting a journey time, based on the popularity measure of each of the roads of the road network. Advantageously, congestion within the road network can be modelled, in order to establish the time taken for a vehicle to travel along its route. This can be used to plan journeys for individual vehicles. Furthermore, the effect on journey time for individual vehicles allows a prediction to be made of the level of congestion for the road network as a whole. The change to the road of the road network may includes any of: a closure of the road; a cycle lane being added to the road; a layout of the road being modified; and a restriction of the road due to a breakdown or a collision. Proposed changes to the road can be modelled to determine whether any plans should be changed. Unexpected changes to the road can be modelled in advance, so that contingency plans can be put in place. The impact of road works can be predicted in advance, which assists with planning. Optionally, the method may further comprise obtaining (SI 1-S13) a base model calculated using historic data, before the information from the ANPR cameras has been received. Advantageously, the initial base model can serve as a starting point for modelling the road network. Using a base model that has been calculated using historic data ensures that traffic flow can be evaluated while the ANPR cameras are not functioning. This ensures that resources can be best deployed in order to enhance the base model. Optionally, the method may further comprise: partitioning the road network into a plurality of subcells; identifying boundaries of the subcells; and identifying roads within the road network that connect the subcells. Advantageously, processing resources can be invested in evaluating the impact of a chance for a road in a particular subcell. Once the impact of this change is understood for that subcell, the impact on the broader network can be investigated. Optionally, the method may further comprise selecting a desired level of granularity for the partitioning of the road network into the subcells. Advantageously, this allows a determination to be made of how to allocate processing resources, in order to evaluate the road network efficiently. Optionally, modelling traffic flow through the road network is implemented in real time (SI7). Advantageously, the effect of changes to roads can be mitigated in real time, by encouraging routes to be updated. This ensures that vehicles travelling through the network travel along efficient routes. Optionally, modelling traffic flow through the road network evaluates (SI6) how the impact of traffic flow will be affected by traffic volume. Advantageously, the effect of the change can be modelled for different times of the day, as well as for different road conditions. The impact at peak times can be compared with the impact at off-peak times, which assists with planning. According to a second aspect, disclosure is provided of a program which, when executed by a computer, causes the computer to perform a method according to the first aspect. According to a third aspect, disclosure is provided of a computer-readable storage medium storing a program according to the second aspect. According to a fourth aspect, disclosure is provided of a system for modelling traffic flow through a road network, the system comprising one or more processor configured to implement a method (S10) according to the first aspect. The modelling may be implemented my a number of processors, which may be distributed. Optionally, the system may further comprise the ANPR cameras. Brief Description of the Drawings Embodiments will now be described, by way of example only, with reference to the accompanying drawings, in which: • FIG. 1 is a flow chart illustrating a project flow for implementing traffic modelling, based on data received from automatic number plate recognition (ANPR) cameras; • FIGs. 2A-2D are annotated maps depicting a real-world environment that is to be simulated, wherein: o FIG. 2A identifies a network area; o FIG. 2B identifies subcells (labelled Al-El) of the network area; o FIG. 2C identifies major roads which serve as boundaries of the network area; and o FIG. 2D identifies some links and nodes of the network area; • FIGs. 3A-3C are annotated maps depicting the popularity of different routes within the real-world environment, wherein: o FIG. 3 A identifies desire lines for the network area; o FIG. 3B identifies typical traffic at a selected time; o FIG. 3C identifies a shortest path between two nodes of the network area; • FIGs. 4A-4E are annotated maps identifying major arterials for each of the subcells (labelled Al-El); • FIGs. 5A-5C are extracts of program code used to simulate the traffic flow, wherein: o FIG. 5A provides for extraction of the “nodes” data from the location data, as well as demonstrating how these nodes can be connected by links (i.e., roads); o FIG. 5B provides the location data of the nodes; and o FIG. 5C provides details of speeds that a vehicle is expected to travel along the roads (links); • FIGs. 6A-6C illustrate how raw data from the ANPR camera can be converted into demand values, wherein: o FIG. 6A illustrates a table identifying how raw data is converted into a formatted data frame; o FIG. 6B is a flow chart illustrating how ANPR camera data for an origin and ANPR camera data for a destination can be used to determine a travel time; and o FIG. 6C is an example of program code that can be used to determine how each vehicle registered through ANPR is modelled in the network; • FIGs. 7A-7B are examples of program code that can be implemented by the user interface to produce outputs, wherein: o FIG. 7A provides for visualisation of output events; and o FIG. 7B provides for visualisation of output plans; and • FIGs. 8A-8E are annotated maps illustrating execution of the simulation, wherein: o FIG. 8A illustrates the network being loaded; o FIG. 8B illustrates a base model; o FIG. 8C illustrates how to traffic flow is predicted to change if any link road is closed; and o FIG. 8D provides a heat map illustrating how a simulation is used to run the model on an updated network; and o FIG. 8E provides a heat map illustrating how once the simulation is run, the results can be visualised. Detailed Description Various exemplary embodiments, features, and aspects are described in detail below with reference to the drawings. Computer implemented methods have been developed in order to simulate the flow of traffic within a road network. Applications of the simulations include the development of a base model and a scenario model. The base model is configured to perform agent based traffic modelling and simulation on a network (i.e., a study area) using a selected Traffic Modelling Engine &SmartLenz Traffic data on the SmartLenz portal. In these simulation models, the “agents” are vehicles. The scenario model is configured to study the traffic congestion and desired lines after rerouting the network due to road closures. FIG. 1 illustrates project flow, which serves as a method S10 for implementing traffic modelling through the road network, based on data received from automatic number plate recognition (ANPR) cameras. • In step SI 1, a network model is created. This network model serves as the base model, until the base model has been recalibrated based on real-world data. This data may be collected and analysed in real time, in order to provide up to date analysis of road congestion. • In step S12, a preliminary depiction is obtained of high priority and low priority routes by applying traffic engineering principles through traffic assignment algorithms. • In step SI3, model testing is performed using a sample of random data. Thus, a preliminary base model is obtained. This preliminary base model is calculated applying traffic engineering principles through traffic assignment algorithms using historic data, before the information from the ANPR cameras has been received. The base model may be subsequently updated using data received from the ANPR cameras. • In step S14, traffic data in the form of hashed unique identifiers is extracted from ANPR sensors. Information is received from ANPR cameras that identifies a plurality of vehicles travelling between nodes of the road network. • In step SI5, a traffic assignment is performed. This involves a determination being made of roads between the nodes that are travelled along for each vehicle. This determination is based on the information that has been received from the ANPR cameras. Thus, a route through the road network is identified for each of the vehicles. A popularity measure is assessed for each of the roads of the road network, by comparing the routes of each of the vehicles. As a result, traffic scenarios can be simulated by applying traffic engineering principles through traffic assignment algorithms based on real world data that has been extracted. • In step SI 6, the network model and the simulation of traffic scenarios are integrated by applying traffic engineering principles through traffic assignment algorithms. This is achieved using SmartLenz software. Accordingly, an evaluation is made of how a change to a road of the road network would impact traffic flow through the road network, based on the popularity measure of each of the roads of the road network. The change to the road of the road network may be expected to be temporary or permanent. Changes to the road of the road network that may be modelled include: a closure of the road, a cycle lane being added to the road, a layout of the road being modified, a restriction of the road due to a breakdown or a collision, etc. • In step SI7, the base model is periodically verified and recalibrated, based on current data. Advantageously, the modelling of traffic flow through the road network may be implemented in real time. An applications of this traffic modelling includes the prediction of journey time for a user travelling through the road network. A further application of this traffic modelling is the evaluating of how the closure of the road would impact traffic flow. The method S10 may be implemented by a system comprising one or more processor. FIGs. 2A-2D are annotated maps depicting a real-world environment that is to be simulated. FIG. 2A identifies a network area. The network area is addressed by its coordinates on the map, which in this example corresponds to (51.479, -0.209). FIG. 2B identifies the partitioning of the road network into a plurality of subcells. In this example, the subcells of the road network are shown labelled Al-El. The partitioning of the road network may be achieved by bifurcating the network area into regions (e.g., five regions in the described example). FIG. 2C identifies major roads which serve as boundaries of the network area. In this example, the network area is bounded by the following four major highways (or arterial roads): (A219 -Fulham Palace Road; A304 - Fulham Road; A3219 - Dawes Road; A3218 - Lillie Road). For each of the subcells, boundaries are identified for each of the subcells, which are used to identify roads within the road network that connect the subcells. FIG. 2D identifies some links and nodes of the network area. The map is shown overlaid with circles and lines, which represent the nodes and the links. The circles correspond to important nodes of the network area, and identify the locations of the ANPR cameras. The lines correspond to important links of the network area, and identify the locations of the roads of the network. It’s possible for subcell may be further subdivided, into further subcells. Similarly, each network may be merged with other networks. Thus, analysis of the road network can be achieved with a level of granularity that is desired for a particular model that is to be generated. The nodes represents typical intersections or junctions of the roads within the network area. The nodes are the network objects having attribute features such as (Id, Positionx, Position_y). The links are important network objects having attribute features such as (Id, From_Node, To_Node, Length, Speed , Capacity, Number of lanes, Permissible lanes). For the example shown, the link network does not include blind alleys and dead-end streets. This simplifies the analysis by only taking into consideration the possibility of traffic movements from one node to another. FIGs. 3A-3C are annotated maps depicting the popularity of different routes within the real-world environment. Prioritized and non-prioritized links are identified, and displayed via the user interface. FIG. 3A identifies desire lines for the network area. The priority of the links can be demonstrated based on their colour. For example: • RED - High priority links • ORANGE - Low priority links From the example shown in FIG. 3A, the images are presented in greyscale. High &low priority links are shown demarcated, with the high priority links indicated by the lines that have a darker shade, and the low priority links indicated by the lines that have a lighter shade. Furthermore, the desire lines can be depicted with more colour gradation based on traffic data availability. If no traffic data is available, then the desire lines may be evaluated using Google’s traffic map timeline (see FIG. 3B) &QGIS’s shortest path algorithm (see FIG. 3C). (Google and QGIS are registered trademarks). Furthermore, the base model may be verified and / or calibrated SI7, by comparing the base model with data retrieved that corresponds to FIG. 3B or FIG. 3C. For this network area, which is sub-divided into subcells named Al, Bl, Cl, DI, El, each of the subcells were taken into analysis separately, to determine the traffic patterns and the heat map. Each sub cell is bounded by one or more major arterials or highway roads. For example, subcell Al is bounded by two major arterial roads (i.e., A219 &A3219). Three traffic / trip generation might be possible from A219 to other subcells, A3219 to other subcells &A219 to A3219. During this traffic / trip generation, there is a high possibility of commuters using internal residential links or network area. Out of this links, commuters most prioritized / used links are demarcated based on Google traffic map timeline (weekdays &weekends) (see FIG. 3B) &QGIS shortest path algorithm (see FIG. 3C). FIGs. 4A-4E are annotated maps identifying major arterials for each of the subcells (labelled Al-El). • FIG. 4A shows movements that are detected for vehicles passing through subcell Al. Subcell Al is bounded by two major arterials (i.e., A1_A3219 &A1_A219) to other subcells. • FIG. 4B shows movements that are detected for vehicles passing through subcell Bl. Subcell Bl is bounded by one major arterial (i.e., B1A219) to other subcells. • FIG. 4C shows movements that are detected for vehicles passing through subcell Cl. Subcell Cl is bounded by two major arterials (i.e., C1_A219 &Cl_A304) to other subcells. • FIG. 4D shows movements that are detected for vehicles passing through subcell DI. Subcell DI is bounded by one major arterial (i.e., D1A304) to other Sub cells. • FIG. 4E shows movements that are detected for vehicles passing through subcell El. Subcell El is bounded by two major arterials (i.e., E1A3219 &E1A304) to other Sub cells. As we already discussed, these is just a preliminary high priority routes in the internal network area. As we progress with the real traffic data these the actual high / low / most likely &less likely routes will be identified with traffic heat maps. The network is simulated using a traffic simulation engine. The traffic simulation engine may be an opensource traffic program / agent based simulation, such as MATSim (registered trademark). The traffic simulation engine provides a framework to implement large-scale agent-based transport simulations. The framework consists of several modules which can be combined or used stand-alone. Modules can be replaced by custom implementations to test single aspects of your own work. The traffic simulation engine offers a framework for demand-modelling, agent-based mobilitysimulation (traffic flow simulation), re-planning, a controller to iteratively run simulations as well as methods to analyse the output generated by the modules. FIGs. 5A-5C are extracts of program code used to simulate the traffic flow. FIG. 5A provides for extraction of the “nodes” data from the location data, as well as demonstrating how these nodes can be connected by links (i.e., roads). FIG. 5A illustrates that the traffic simulation engine network consists of nodes, links &its attributes in an xml formatted file. A “network” is a root with two sub elements “nodes” &“links” in xml. FIG. 5B provides the location data of the nodes. FIG. 5C provides details of speeds that a vehicle is expected to travel along the roads (i.e., links). FIGs. 6A-6C illustrate how raw data from the ANPR camera can be converted into demand values. FIG. 6A illustrates a table identifying how the raw data from the ANPR camera is converted into formatted data. The raw data includes the identify of the vehicle (“vehicleid”), which may correspond to the number plate of the vehicle that has been recognised. Activity of the vehicle is recorded in the data, including its origin and destination. For both the origin and the destination, the formatted data includes the name of the location, together with its coordinates, and the time when the vehicle was detected. Depending upon the mode of travel (e.g., car), the timing of the leg of the journey is calculated, including departure time, travel time, and arrival time. Thus, the raw ANPR data is converted into a traffic simulation engine supportable population demand xml version. To do so, first the raw data is converted into data frames, and then this raw data is inputted to Python script. FIG. 6B is a flow chart illustrating a method for calculation travel time for a vehicle (e.g., a car) travelling through the network. The ANPR camera data includes details of the origin activity, such as position and duration, in order to determine the departure time. The ANPR camera data includes details of the destination activity, such as position and duration, in order to determine the arrival time. The travel time is calculated by comparing the departure time and the arrival time. Each activity has a start time, a duration, and an end time. For the present example, the end time of the activity is nothing but the departure time of the leg mode using car. Likewise, the arrival time of the leg mode using car is nothing but the start time of another activity. As a buffer we can give activity duration from 10 minutes to 30 minutes. This formatted data frame is used to generate the traffic simulation engine population demand in xml format. The disclosed model is a particle-oriented approach, which is nothing but importance given for every each and every vehicle and its route path. Each vehicle registered through the ANPR sensor will be modelled in the network, rather than some volume of vehicles going from one origin to another destination in some period of time. FIG. 6C is an extract of program code used to determine how each vehicle registered through ANPR is modelled in the network. Each person with unique id (e.g., a vehicle number plate) will have a plan with two activities and one leg between them with the mode of car. FIGs. 7A-7B are extracts of program code that can be implemented by the user interface to produce outputs. FIG. 7A provides for visualisation of output events. FIG. 7B provides for visualisation of output plans. Once the network and population demand xml filei s ready, this a configuration file is prepared. INPUT (network,Population) -—> TRAFFIC SIMULATION ENGINE CONFIG —> TRAFFIC SIMULATION ENGINE ------>OUTPUT(Output Event xml) The outputs include: 1. Output Event file 2. Output Plans 3. Output network etc... FIGs. 8A-8E are annotated maps illustrating execution of the simulation. The user interface (UI) is setup with the traffic simulation engine running in the background. • FIG. 8 A illustrates the network being loaded into the UI, which corresponds to step S11 of FIG. 1. • FIG. 8B illustrates a base model being loaded into the UI, which corresponds to steps S12-S15 ofFIG. 1. • FIG. 8C illustrates how to traffic flow is predicted to change if any link road is closed, which corresponds to steps S16-S17 ofFIG. 1. The user can draw and edit the links in the traffic network. The user can visualize and predict how any link road being closed results in traffic flow being affected for the network. • FIGs. 8D &8E provide heat maps. FIG. 8D illustrates how a simulation is used to run the model on an updated network. FIG. 8E illustrates how once the simulation is run, the results can be visualised. Thus, once the user submits a simulation, the traffic simulation engine runs the model on the updated network computing the trafficflows. Once the simulation is run, the results can be visualised thru various representation and compared to the base model. The disclosed examples can be realised by a computer of a system or apparatus. These examples can be implemented by any device configured to execute instructions, or any dedicated hardware that is capable of carrying out all or a portion of the functionality. Disclosure is provided of hardware (e.g., a processor such as a central processing unit (CPU) or a microprocessor unit (MPU)) configured to read out and executes a program recorded on a memory device to perform the functions of the disclosed examples. For this purpose, the program is provided to the computer for example via a network or from a recording medium of various types serving as the memory device (e.g., a computer-readable medium such as a non-transitory computer-readable medium). The steps of the disclosed methods may be performed in any suitable order, or simultaneously where possible. Although the disclosed subject matter has been described using specific terminology relating to apparatus features and / or method features, it is to be understood that the claimed subject matter is not necessarily limited to the examples disclosed. The scope of the following claims is to be accorded the broadest interpretation so as to encompass all such modifications and equivalent structures and functions. The advantages disclosed may relate to several of the examples that are disclosed.

Claims

1. A computer-implemented method (S10) of modelling traffic flow through a road network, the method comprising:receiving (SI4) information from automatic number plate recognition, ANPR, cameras that identifies a plurality of vehicles travelling between nodes of the road network;determining (SI5) roads between the nodes that are travelled along for each vehicle, based on the information received from the ANPR cameras, to identify a route through the road network for each of the vehicles;assessing (SI5) a popularity measure for each of the roads of the road network, by comparing the routes of each of the vehicles; andevaluating (SI 6) how a change to a road of the road network would impact traffic flow through the road network, based on the popularity measure of each of the roads of the road network.

2. The method according to claim 1, further comprising:predicting a journey time, based on the popularity measure of each of the roads of the road network.

3. The method according to claim 1 or claim 2, wherein:the change to the road of the road network includes any of:a closure of the road;a cycle lane being added to the road;a layout of the road being modified; anda restriction of the road due to a breakdown or a collision.

4. The method according to any preceding claim, further comprising:obtaining (SI 1-SI3) a base model calculated using historic data, before the information from the ANPR cameras has been received.

5. The method according to any preceding claim, further comprising:partitioning the road network into a plurality of subcells;identifying boundaries of the subcells; andidentifying roads within the road network that connect the subcells.

6. The method according to claim 5, further comprising:selecting a desired level of granularity for the partitioning of the road network into the subcells.

7. The method according to any preceding claim, wherein modelling traffic flow through the road network is implemented in real time (SI7).

8. The method according to any preceding claim, wherein modelling traffic flow through the road network evaluates (SI6) how the impact of traffic flow will be affected by traffic volume.

9. A program which, when executed by a computer, causes the computer to perform a method according to any preceding claim.

10. A computer-readable storage medium storing a program according to claim 9.

11. A system for modelling traffic flow through a road network, the system comprising one or more processor configured to implement a method (S10) according to any one of claims 1 to 8.

12. The system according to claim 11, further comprising the ANPR cameras.Claims17 09 251. A computer-implemented method (S10) of modelling traffic flow through a road network, the method comprising:5 receiving (SI4) information from automatic number plate recognition, ANPR, camerasthat identifies a plurality of vehicles travelling between nodes of the road network;determining (SI5) roads between the nodes that are travelled along for each vehicle, based on the information received from the ANPR cameras, to identify a route through the road network for each of the vehicles;10 assessing (SI5) a popularity measure for each of the roads of the road network, bycomparing the routes of each of the vehicles; andevaluating (SI 6) how a change to a road of the road network would impact traffic flow through the road network, based on the popularity measure of each of the roads of the road network.

152. The method according to claim 1, further comprising:predicting a journey time, based on the popularity measure of each of the roads of the road network.20 3. The method according to claim 1 or claim 2, wherein:the change to the road of the road network includes any of:a closure of the road;a cycle lane being added to the road;a layout of the road being modified; and25 a restriction of the road due to a breakdown or a collision.

4. The method according to any preceding claim, further comprising:obtaining (SI 1-SI3) a base model calculated using historic data, before the information from the ANPR cameras has been received.

305. The method according to any preceding claim, further comprising:partitioning the road network into a plurality of subcells;identifying boundaries of the subcells; andidentifying roads within the road network that connect the subcells.17 09 256. The method according to claim 5, further comprising:selecting a desired level of granularity for the partitioning of the road network into the 5 subcells.

7. The method according to any preceding claim, wherein a base model is periodicallyverified and recalibrated, based on current data, so that modelling traffic flow through the road network is implemented in real time (SI7).

108. The method according to any preceding claim, wherein modelling traffic flow through the road network evaluates (SI6) how the impact of traffic flow will be affected by traffic volume.15 9. A program which, when executed by a computer, causes the computer to perform amethod according to any preceding claim.

10. A computer-readable storage medium storing a program according to claim 9.20 11. A system for modelling traffic flow through a road network, the system comprising oneor more processor configured to implement a method (S10) according to any one of claims 1 to 8.

12. The system according to claim 11, further comprising the ANPR cameras.s