System and method for monitoring road traffic congestion and traffic accident risk

Abstract

The application discloses a system and method for monitoring road traffic congestion and / or traffic accident risk, belonging to the field of intelligent transportation data system technology. In some embodiments, the system comprises: a plurality of vehicle-mounted sensors; a cloud server comprising a vehicle fleet management system, an information hub and a geographic information system platform; and at least one terminal. Other example embodiments are discussed herein. In some embodiments, the provided system and method support intelligent traffic / mobility and improve overall road traffic safety, further promoting the development of smart cities.

Description

Technical Field

[0001] This invention relates to the field of intelligent transportation data systems, and more particularly to systems and methods for monitoring road traffic congestion and / or the risk of traffic accidents. Background Technology

[0002] Many cities around the world have large populations but limited land resources. Limited road space coupled with a rapid increase in the number of vehicles has severely exacerbated urban traffic congestion and safety issues, a situation that warrants attention. Therefore, comprehensive dynamic urban traffic information is crucial, necessitating intelligent traffic information systems and methods that can significantly improve urban traffic efficiency, help prevent road accidents, and promote intelligent traffic management. Summary of the Invention

[0003] In view of this, in some embodiments, the technical problem to be solved by the present invention is to provide an intelligent traffic data system and method for monitoring road traffic congestion and / or traffic accident risks, thereby making traffic management more intelligent and ensuring road safety.

[0004] In some embodiments, one aspect of the present invention provides a system for monitoring road traffic congestion and / or traffic accident risks, comprising:

[0005] (1) Multiple vehicle-mounted sensors, each sensor is installed on a designated vehicle and configured to acquire real-time visual data obtained by the designated vehicle and process the real-time visual data to generate road and driving data and / or traffic accident risk data;

[0006] (2) A cloud server comprising a Fleet Management System (FMS), an Information Hub, and a Geographic Information System (GIS) platform, wherein the FMS, Information Hub, and GIS platform are configured to communicate with each other, wherein the FMS wirelessly communicates with multiple vehicle-mounted sensors and is configured to receive and process road and driving data and / or traffic accident risk data from the multiple vehicle-mounted sensors to generate processed driving data; wherein the Information Hub is configured to acquire and store at least one type of Information Hub data, including open data, processed driving data from the FMS, and / or stored data; wherein the GIS platform is configured to receive and process the at least one type of Information Hub data to generate at least one type of processed geospatial data; wherein the Information Hub is further configured to process the processed geospatial data to generate a Traffic Congestion Index (TCI) and a Traffic Accident Risk Index (TARI), and the GIS platform is configured to receive and process the TCI and TARI to generate visualized geospatial data; and

[0007] (3) At least one terminal, which communicates wirelessly with the cloud server and is configured to display at least the visualized geospatial data.

[0008] In some embodiments, another aspect of the present invention provides a method for monitoring road traffic congestion and traffic accident risks, comprising:

[0009] Real-time visual data obtained from a designated vehicle is acquired through multiple onboard sensors;

[0010] The real-time visual data is processed by the multiple vehicle-mounted sensors to generate road and driving data and / or traffic accident risk data;

[0011] The system receives and processes road and driving data and / or traffic accident risk data from the multiple on-board sensors via FMS to generate processed driving data.

[0012] At least one type of information hub data is acquired and stored through the information hub, including open data, processed driving data from the FMS, and / or stored data;

[0013] The system receives and processes at least one type of information hub data through a GIS platform to generate at least one type of processed geospatial data.

[0014] The processed geospatial data is processed through the information hub to generate the Traffic Congestion Index (TCI) and / or Traffic Accident Risk Index (TARI).

[0015] The TCI and / or TARI are received and processed through the GIS platform to generate visualized geospatial data; and

[0016] At least the aforementioned visualized geospatial data is displayed through at least one terminal.

[0017] This article discusses other example implementations.

[0018] This invention offers numerous advantages. In some embodiments, the systems and methods provided by this invention enable vehicle drivers, road users, public transport management and users, governments, and relevant stakeholders to understand and predict road congestion and traffic accident risks in real time, thereby improving overall road traffic safety and efficiency. In some embodiments, in the long term, a comprehensive and real-time updated database of road congestion and traffic accident risks will help improve intelligent transportation / intelligent transport systems, prepare for future autonomous driving, and provide a framework for data-based infrastructure, further promoting the development of smart cities. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of a system for monitoring road traffic congestion and / or traffic accident risks according to an embodiment of the present invention.

[0020] Figure 2 This is a schematic diagram of a method for monitoring road traffic congestion and / or traffic accident risks according to an embodiment of the present invention. Detailed Implementation

[0021] To make the technical problems, solutions, and beneficial effects of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of the invention.

[0022] As used herein and in the claims, “comprising” means including the following components, but does not exclude other components.

[0023] As used herein and in the claims, “comprising” means to include the following elements, but does not exclude other elements. The terms “comprising” (or any related form such as “comprise” and “comprises”), “including” (or any related form such as “include” or “includes”), “containing” (or any related form such as “contain” or “contains”), and “having” (or any related form such as “have” or “has”) mean to include the following elements, but do not exclude other elements. It should be understood that for each embodiment using the terms “comprising” (or any related form such as “comprise” and “comprises”), “including” (or any related form such as “include” or “includes”), “containing” (or any related form such as “contain” or “contains”), or “having” (or any related form such as “have” or “has”), this disclosure / application also includes alternative embodiments in which the terms “comprising”, “including”, “contains”, or “having” are replaced with “consisting essentially of” or “consisting of”. These alternative embodiments using “consisting of” or “consisting essentially of” are understood to be narrower embodiments of the “comprising”, “including”, “contains”, or “having” embodiments.

[0024] For clarity, “containing,” “contains,” “containing,” or “having,” and any related forms, are open-ended terms that allow for additional elements or features beyond the specified essential elements, while “consisting of” is a closed-ended term that is limited to the elements listed in the claims and excludes any elements, steps, or ingredients not specified in the claims.

[0025] As used herein and in the claims, unless the context clearly indicates otherwise, the singular forms “a,” “an,” and “the” also include their corresponding plural indicators. When a numerical range is mentioned in the specification, the numerical range is understood to include every discrete point within that range. For example, 1-7 represents 1, 2, 3, 4, 5, 6, and 7.

[0026] As used herein and in the claims, unless otherwise stated, "at least one" means one or more, and "more than one" means two or more. Furthermore, to facilitate a clear description of the technical solutions of the present invention, the terms "first," "second," etc., are used herein and in the claims to distinguish identical or similar items with substantially the same function and effect. Those skilled in the art will understand that the terms "first," "second," etc., do not limit the quantity or order of execution, and that "first," "second," etc., do not necessarily imply differences.

[0027] As used herein and in the claims, the terms “about,” “substantially,” and “approximately” are understood to be within the normal tolerance range in the art and not exceeding ±10% of the specified value. By way of example only, approximately 50 refers to all values ​​from 45 to 55, inclusive. As used herein, the phrase “about” also includes specific values, for example, approximately 50 includes 50.

[0028] As used herein and in the claims, the terms "generally" or "substantially" mean that it is not necessary to precisely achieve the listed features, angles, shapes, states, structures, or values, but deviations or variations, including, for example, tolerances, measurement errors, measurement accuracy limitations, and other factors known to those skilled in the art, may occur in an amount that does not preclude the effect that the feature is intended to provide. For example, an object having a "generally" cylindrical shape means that the object has a precisely cylindrical shape or a nearly precisely cylindrical shape. In another example, an object "substantively" perpendicular to a surface means that the object is either completely perpendicular to the surface or nearly completely perpendicular to the surface, for example, with a deviation of 5%.

[0029] For clarity, "characterized by, characterized in" (along with their related forms as described above) does not limit or change the nature of whether the following list of terms is open or closed. For example, in a claim for "an apparatus comprising A, B, C and characterized by D, E, and F," elements D, E, and F remain open-ended terms, and the claim is intended to include other elements due to the use of the word "comprising" preceding the claim.

[0030] As used herein and in the claims, an "intelligent transportation" system refers to a comprehensive management system that utilizes, for example, information technology, communication technology, electronic control technology and / or computer processing technology to improve the efficiency, safety and reliability of a transportation system.

[0031] As used herein and in the claims, “data” means digital information stored, processed, or transmitted in a computer system, for example, represented in binary (0 and 1) form. In some embodiments, data may include one or more types such as text, numbers, images, audio, video, etc., and may be processed by computer hardware and / or software, for example, for performing various tasks, making decisions, and communicating.

[0032] As used herein and in the claims, "visual data" refers to information captured through images or videos, which may originate from visual sensing devices (such as cameras, sensors) and may include details, colors, shapes, and / or motion of a scene. In some embodiments, visual data can be used to analyze, identify, and process an environment or objects, providing a foundation for computer vision applications. In some embodiments, visual data is real-time visual data obtained from a vehicle's field of view, including one or more types of visual data such as roads, traffic, environment, road users, and vehicles.

[0033] As used herein and in the claims, an "vehicle sensor" is a device that can be mounted on a vehicle for detecting / acquiring real-time visual data obtained from the vehicle's field of view. In some embodiments, the vehicle sensor further includes a processor to process the real-time visual data to generate road and driving data and / or traffic accident risk data.

[0034] As used herein and in the claims, "vehicle" means mechanical equipment used for transporting people, goods, or other articles, typically traveling on the ground. In some embodiments, a vehicle includes, but is not limited to, automobiles, trucks, motorcycles, buses, bicycles, etc. In some embodiments, a vehicle includes minibuses, taxis, logistics vans, trucks, etc.

[0035] As used herein and in the claims, "wireless communication" means the transmission of data and / or information without a physical connection (such as a cable or fiber optic cable). In some embodiments, wireless communication utilizes radio waves, microwaves, infrared light, or other wireless signals for data transmission.

[0036] As used herein and in the claims, “real-time” means the ability of a system or process to process and respond to input data or events immediately or nearly immediately, with virtually no delay. In some embodiments, in a real-time system, the time from data input to processing and feedback of results is very short to ensure timely response to dynamic environments or rapidly changing situations.

[0037] As used herein and in the claims, the “Global Positioning System (GPS)” is a system that provides location and time information over a global range via a network of satellites.

[0038] As used herein and in the claims, “Artificial Intelligence of Things (AIoT)” is a system that combines artificial intelligence (AI) technology with Internet of Things (IoT) devices to enhance the functionality and performance of IoT devices through intelligent analysis and decision-making. In some embodiments, AIoT enables devices not only to collect and transmit data, but also to analyze data in real time, identify patterns, make predictions, and automate control, thereby achieving more efficient management, operation, and user experience.

[0039] As used herein and in the claims, a “Geographic Information System” (GIS) is a system for collecting, storing, managing, analyzing, and / or displaying geospatial data. In some embodiments, GIS helps users understand and analyze geographic phenomena, spatial relationships, and trends through visualization tools such as maps and layers.

[0040] As used herein and in the claims, “geospatial data” refers to information relating to the location and characteristics of space on the Earth’s surface, including geographic location, shape, distribution, and spatial relationships. In some embodiments, geospatial data may exist in one or more forms such as maps, satellite imagery, geographic coordinates, and terrain models, and is used to describe and analyze geographic phenomena, environmental conditions, and spatial patterns.

[0041] As used herein and in the claims, "platform" refers to a computer platform, which may include a standalone or combined environment or system of hardware and software. In some embodiments, a platform includes one or more of an operating system, processor architecture, memory, storage devices, and system software, as well as possible development tools and / or runtime environments or systems. In some embodiments, a platform may be a desktop computer, server, mobile device, or cloud computing environment, providing a consistent environment for the deployment, management, and execution of applications.

[0042] As used herein and in the claims, “processing” refers to computer processing, which is the process by which a computer system manipulates, analyzes, and transforms input data to generate output results. In some embodiments, processing includes tasks such as reading, calculating, storing, modifying, and sorting data, performed by a central processing unit (CPU) or other processing unit. In some embodiments, the goal of computer processing is to transform raw or intermediate data into useful information or results.

[0043] As used herein and in the claims, "server" means a computer system or device that provides services, processes requests, and / or manages resources in a computer network.

[0044] As used herein and in the claims, a “cloud server” is a virtual server provided by cloud computing technology that runs on a physical server in a remote data center.

[0045] As used herein and in the claims, "terminal" means a device or interface through which a user interacts with a computer system, network, or service. In some embodiments, a terminal may be a computer, smartphone, tablet, monitor, printer, etc., used for inputting, displaying, processing, or receiving information. In some embodiments, a terminal may be a physical device or a virtual interface, allowing a user to access, control, and manage resources and services in an operating system or network. In some embodiments, a terminal may be a software application (APP) and / or a web browser, allowing a user to access, control, and operate an information control panel (dashboard) provided by a server and / or cloud server.

[0046] As used herein and in the claims, “Artificial Intelligence (AI)” means the technology that enables computers or systems to perform tasks that would normally require human intelligence by simulating human intelligence processes.

[0047] As used herein and in the claims, a “Graph Convolutional Network (GCN)” model is a neural network model for processing graph-structured data. In some embodiments, a GCN extracts local features of nodes in a graph by performing convolution operations on the nodes and edges of the graph, and utilizes the graph’s topology for information propagation and fusion. In some embodiments, a GCN aggregates features through adjacency relationships between nodes, thereby capturing complex structures and patterns in graph data.

[0048] As used herein and in the claims, the "Long Short-Term Memory (LSTM)" model is a recurrent neural network (RNN) designed to overcome the vanishing and exploding gradient problems of standard RNNs when dealing with long-term dependencies. In some embodiments, LSTM controls the flow and storage of information by introducing memory units and three gating mechanisms (input gate, forget gate, and output gate), thereby retaining and utilizing useful information over a longer time span.

[0049] As used herein and in the claims, a "Dynamic Bayesian Network (DBN)" is a probabilistic graphical model for modeling and analyzing time series data. In some embodiments, DBNs capture the changes and dependencies of variables over time by introducing a time dimension.

[0050] As used herein and in the claims, a “Recurrent Neural Network (RNN)” is a neural network model for processing sequential data whose structure allows information to be passed between different time steps of the network. In some embodiments, RNNs introduce recurrent connections, enabling the network to use the output of the previous time step as the input of the current time step, thereby capturing temporal dependencies and contextual information in the sequence.

[0051] In some embodiments, to address urban road traffic congestion and enhance road safety, certain embodiments of the present invention utilize advanced artificial intelligence of things (AIoT) sensors for real-time detection and high-precision analysis of road / traffic data. In some embodiments, cloud-based data analytics is employed, integrating multiple data sources such as weather, environmental, and historical datasets. In some embodiments, location-specific alerts, warnings, and road / traffic optimization strategies are provided using GIS technology and GeoAI algorithms. In some embodiments, the system operates near real-time, providing up-to-date information and leveraging big data analytics to improve traffic flow and overall transportation efficiency.

[0052] 1.1 Hardware and software components included in systems and methods for monitoring road traffic congestion and / or traffic accident risks. Item

[0053] In some embodiments, the system and method comprise two elements: (i) hardware: vehicle-mounted sensors equipped with real-time and accurate AIoT capabilities, suitable for installation on different types of vehicles, for road asset surveying / mapping and establishing a dynamic mobile database; and (ii) software: a cloud server-side software platform equipped with GIS and GeoAI capabilities, which combines AIoT mobile sensor data, various open data (such as real-time traffic / road / weather conditions for each road segment, surrounding environmental factors, etc.) and stored / historical data to provide functions for monitoring road traffic congestion and traffic accident risks, including prediction and visualization capabilities.

[0054] 1.1.1 Hardware

[0055] (1) AIoT Sensors

[0056] In some embodiments, the vehicle-mounted sensors include AIoT sensors compliant with the Mobileye ALL protocol for acquiring real-time visual data and generating road and driving data and / or traffic accident risk data.

[0057] In some embodiments, an AIoT sensor compliant with the Mobileye ALL protocol consists of three parts: 1) equipped with 1) AIoT camera device with on-chip system and speaker; 2) EyeWatch TM 3) Display screen; 4) GPS device.

[0058] In some embodiments, the AIoT sensor includes sensors based on cameras and radar-lidar.

[0059] In some embodiments, the AIoT sensor includes an AIoT camera device based on a single camera that scans the road ahead to obtain real-time visual data.

[0060] In some embodiments, the AIoT camera is mounted on the inside of the windshield behind the rearview mirror of a designated vehicle. The AIoT camera continuously scans the road ahead of the designated vehicle, collecting real-time visual data.

[0061] In some embodiments, AIoT sensors can mimic how humans understand road landscapes, through... The on-chip system and its trained computer vision algorithm process the real-time visual data obtained from the specified vehicle in real time to generate road and driving data and / or traffic accident risk data.

[0062] Table 1 lists the names and types of exemplary variables included in road and driving data and / or traffic accident risk data generated by AIoT sensors.

[0063]

[0064]

[0065] Table 1 provides an exemplary list of variables included in road and driving data and / or traffic accident risk data.

[0066] In some embodiments, the AIoT sensor transmits the generated road and driving data and / or traffic accident risk data every minute via 4G / 5G to an Amazon Web Services (AWS) cloud server (or other cloud server).

[0067] In some embodiments, AIoT sensors only transmit road and driving data and / or traffic accident risk data to a cloud server for further analysis, without recording any raw real-time visual data. This ensures that vehicle journeys are not tracked, eliminating privacy concerns for drivers and pedestrians.

[0068] In some embodiments, the AIoT sensor is equipped with temporary data storage capabilities for road and driving data and / or traffic accident risk data, preventing the loss of any signal / connection due to tunnel crossings and / or other communication problems. Once communication / connection is restored, the stored data will be transmitted back to the cloud server.

[0069] (2) Engine Control Unit (ECU)

[0070] In some embodiments, the onboard sensors also include an engine control unit (ECU). The ECU is designed as an onboard network system that connects all in-vehicle hardware (AIoT sensors) and microcontrollers (including webcams). The purpose of the ECU is to monitor, capture, and collect real-time static and dynamic road data, as well as communications predicted by artificial intelligence models, and then push suggestions and alerts to the driver via a mobile application.

[0071] (3) Webcam

[0072] In some embodiments, the vehicle-mounted sensor also includes a webcam.

[0073] In some embodiments, real-time video data acquired by the webcam is clipped based on warnings from the ECU and transmitted to a local server for cross-validation of onboard sensors. In some embodiments, the local server is configured to store the real-time visual data and transmit the real-time visual data to at least one terminal.

[0074] In some embodiments, a total of 20 vehicles will be selected and equipped with network cameras in parallel for verification. Various video object segmentation tools will also be applied to the video and recordings to systematically track and identify key objects in the video for better evaluation. Detection and accuracy rates are expected to reach 90%.

[0075] 1.1.2 Software

[0076] (1) Cloud server software

[0077] In some embodiments, the system and method further include software running on a cloud server.

[0078] In some embodiments, the cloud server is an Amazon Web Services (AWS) cloud server or other cloud server.

[0079] In some embodiments, the software includes a Fleet Management System (FMS), an Information Hub, and a Geographic Information System (GIS) platform running on a cloud server, wherein the FMS, Information Hub, and GIS platform are configured to communicate with each other.

[0080] In some embodiments, the FMS communicates wirelessly with the plurality of vehicle-mounted sensors and is configured to receive and process road and driving data and / or traffic accident risk data from the plurality of vehicle-mounted sensors to generate processed driving data, providing vehicle tracking, route optimization, maintenance scheduling and driver performance monitoring functions.

[0081] In some embodiments, the information hub is configured to acquire and store at least one type of information hub data, including open data, processed driving data from the FMS, and / or stored data; in some embodiments, the information hub generates a Traffic Congestion Index (TCI) and / or a Traffic Accident Risk Index (TARI) based on the collected data.

[0082] In some embodiments, the FMS is further configured to store identity data of multiple vehicles, the vehicle identity data corresponding to the designated vehicle, and the FMS is configured to bind each vehicle identity data to the corresponding road and driving data and / or the corresponding traffic accident risk data.

[0083] In some embodiments, the GIS platform is configured to receive and process at least one of the information hub data to generate at least one processed geospatial data.

[0084] In some embodiments, the GIS platform is further configured to receive and process the TCI and / or TARI to generate visualized geospatial data for further visualization and decision-making.

[0085] In some embodiments, the cloud server further includes a first AI module, which includes a unified model algorithm comprising a graph convolutional network (GCN) and a long short-term memory (LSTM) for processing the at least one information hub data.

[0086] In some embodiments, the cloud server further includes a second AI module that combines the probabilistic reasoning capabilities of a Dynamic Bayesian Network (DBN) with the sequence modeling capabilities of a Recursive Neural Network (RNN) for processing the at least one information hub data.

[0087] In some embodiments, the software includes a website, which is a web application running on a cloud server and serves as the user interface of the system.

[0088] In some embodiments, the website is a map-based geographic information system (GIS) dashboard that displays in real time road and driving data and / or traffic accident risk data collected and aggregated from onboard sensors, and publishes road hazard and warning information to users and the public.

[0089] In some embodiments, the website is equipped with an interactive and user-friendly interface, allowing stakeholders to intuitively understand location-based road hazards and traffic conditions through maps.

[0090] In some embodiments, the purpose of website development is to establish front-end and back-end operations. The front-end includes a graphical user interface (GUI) that will function as a map-based website platform.

[0091] In some embodiments, the website front-end includes a user-friendly map-based dashboard that displays dynamic traffic patterns and road risk information, including TCI and TARI, by road segment. This allows users to easily identify congested areas, accident-prone locations, and receive real-time updates on road hazards, traffic flow conditions, and accident hotspots. In some embodiments, the website front-end also provides drivers with notifications of potential road risks or traffic flow conditions, including accident hotspots caused by weather phenomena within the next one to two hours.

[0092] In some embodiments, the dashboard also includes a dynamic map and optionally includes traffic snapshot images, accident types, and / or black spot locations to provide a comprehensive understanding of traffic conditions and accident patterns.

[0093] In some embodiments, the dashboard may optionally include an instrument panel for detecting vehicles in a video sequence, which may be approximately 8 seconds long.

[0094] In some embodiments, information hub data will be manipulated and displayed on the website's graphical user interface in the form of maps, tables, charts, icons, and alerts, depending on application requirements and system design.

[0095] In some embodiments, the website will be designed for a variety of devices equipped with the internet and a browser (i.e., computers, tablets, or any type of mobile device), and will be suitable for Android or iOS-based mobile devices.

[0096] In some embodiments, the website will utilize JavaScript, CSS (Cascading Style Sheets), HTML (HyperText Markup Language), or CRUD (Create, Read, Update, and Delete) to construct its information architecture. This process will properly define effective, goal-oriented interactions between the user and the system, during which the user can access, view, receive alerts, and interact with road and driving data and / or traffic accident risk data from onboard sensors and other information hub data.

[0097] In some embodiments, the website’s backend will be developed using server-side languages ​​such as Ruby or Python, enabling users to perform operations programmatically and make decisions based on the selected analytics and reporting types.

[0098] In some embodiments, the website employs various network security measures, including: identification or authentication, multi-access control, etc.

[0099] In some embodiments, the software portion of the system further includes a dynamic flow database for storing road and driving data and / or traffic accident risk data from multiple onboard sensors, generated processed driving data, information hub data (including open data, processed driving data from the FMS, and / or stored data), TCI and / or TARI, and visualized geospatial data.

[0100] In some embodiments, the open data includes geospatial data, i.e., data extracted from a general spatial data infrastructure portal and further processed / converted into dynamic traffic database data for further spatiotemporal and GeoAI modeling. In some embodiments, the open data also includes non-geospatial data, which will be extracted primarily from historical data, annual traffic reports, online media resources (such as crowdsourced high-risk accident hotspots), and reference travel characteristic surveys. Table 2 provides a list of exemplary types and primary sources of open data in some embodiments.

[0101] In some embodiments, spatial and non-spatial, real-time and historical data from multiple sources are collected and processed at the same frequency and resolution to build a comprehensive and accurate dynamic traffic database.

[0102] In some embodiments, all data in the dynamic traffic database is transformed into geospatial data (e.g., linked to each road segment or block level, depending on data quality) as attribute or feature data to enrich the dynamic traffic database.

[0103]

[0104]

[0105] Table 2 lists exemplary types of open data and their sources.

[0106] (2) Terminal software

[0107] In some embodiments, the software portion of the system further includes a terminal mobile application that provides a user-friendly interface for users such as drivers and stakeholders to access dynamic flow databases, TCI and TARI, and to receive visualized geospatial data generated by a GIS platform.

[0108] In some embodiments, the mobile application sends early safety alerts to users, such as drivers, when they are near potentially hazardous areas to improve driving safety. Table 3 provides a list of exemplary warnings and prompts issued by the mobile application to drivers and / or other users in some embodiments.

[0109] In some embodiments, the terminal mobile application also has basic route planning functions. In some embodiments, users such as drivers can plan routes in advance, and fleet managers can also pre-allocate routes according to different criteria and receive analysis alerts / road risk warnings based on GPS location during the journey.

[0110] In some embodiments, the terminal mobile application employs various network security measures, including: identification or authentication, multi-access control, etc.

[0111]

[0112]

[0113] Table 3 lists examples of alarms issued by exemplary mobile applications.

[0114] 1.2 Traffic Congestion Index (TCI) and Traffic Accident Risk Index (TARI)

[0115] In some embodiments, the Traffic Congestion Index (TCI) and the Traffic Accident Risk Index (TARI) are primary or one of the outputs of the system and method. TCI and TARI provide useful information to users such as drivers and road users, enabling them to make informed decisions about routes and driving behavior, thereby improving traffic safety and reducing road congestion.

[0116] In some embodiments, TCI represents the level of congestion on a particular road segment or street, measuring the severity of traffic congestion. It takes into account factors such as traffic flow, speed, and movement patterns, as well as historical data on congestion patterns. By assessing TCI, drivers can identify congested areas and make informed decisions about alternative routes, thereby helping to alleviate congestion and improve travel efficiency.

[0117] In some embodiments, TARI assesses the risk of traffic accidents occurring on specific road sections or blocks. TARI integrates various factors, such as historical accident data, weather conditions, driving behavior patterns, and road characteristics, to estimate the likelihood of an accident occurring in a specific area. TARI can help drivers identify high-risk areas and adjust their driving behavior accordingly for safer outcomes, reducing the likelihood of traffic accidents.

[0118] In some embodiments, TCI and TARI can also be used to alert other users, such as fleet managers, government agencies and other stakeholders, to take appropriate actions and measures to reduce traffic risks and improve road safety.

[0119] In some embodiments, TCI and TARI are obtained through a mobile application and a web browser accessing a website. This method allows different users, such as the general public, drivers, government officials, and stakeholders, to easily access TCI and TARI.

[0120] 1.3 Algorithm

[0121] In some embodiments, TCI and TARI are generated by processing and analyzing collected real-time and historical data using advanced artificial intelligence algorithms.

[0122] In some embodiments, the method processes the at least one information hub data through a unified model algorithm module that combines Graph Convolutional Networks (GCN) and Long Short-Term Memory (LSTM) to improve the accuracy of TCI and TARI.

[0123] In some embodiments, GCN is used to capture spatial dependencies in information hub data, thereby understanding and interpreting how traffic conditions in one area affect nearby areas, thus providing a holistic view of traffic patterns across the entire network.

[0124] In some embodiments, LSTM is used to process sequence data in information hub data to capture the temporal patterns of the information hub data.

[0125] In some embodiments, the method further processes the at least one information hub data through an AI algorithm module that combines the probabilistic reasoning capabilities of a dynamic Bayesian network (DBN) with the sequence modeling capabilities of a recursive neural network (RNN).

[0126] In some embodiments, the method learns at least one information hub data through DBN, thereby modeling the hidden interrelationships among meteorological / environmental determinants related to traffic / road conditions as probabilistic conditional dependencies.

[0127] In some embodiments, the DBN model generates a predictive model based on the temporal dependencies of spatial and non-spatial variables in at least one information hub dataset. After DBN learning, the network structure is studied to ensure no residual confusion occurs and to reveal reasonable paths between important variables in the final regression model.

[0128] In some embodiments, the at least one information hub data is processed by DBN to calculate the probability of different traffic conditions (such as traffic congestion and traffic accidents).

[0129] In some embodiments, by learning regression models, such as recurrent neural networks (RNNs), the hidden interactions between meteorological / environmental factors in different information hub data are discovered, and then future road conditions are predicted based on current road conditions.

[0130] In some embodiments, the method can also automatically generate and optimize rules using machine learning algorithms such as genetic programming to interpret road conditions and issue alerts to road users, thereby optimizing road efficiency and alleviating traffic congestion.

[0131] 1.4 Pilot study method for implementing the system and method described

[0132] In some embodiments, the system and method are implemented through a phased pilot study approach.

[0133] In some embodiments, pilot studies will be conducted in areas with heavy urban traffic congestion and a high number of serious accidents, with approximately 150 commercial vehicles of different types (e.g., buses, taxis, logistics vans / trucks) participating in the program.

[0134] In some embodiments, before the onboard sensors process the real-time visual data obtained from the specified vehicle to generate road and driving data and / or traffic accident risk data, it is recommended to initialize the onboard sensors based on the road and traffic characteristics of the city where the specified vehicle is located in order to improve the accuracy of the computer vision algorithm.

[0135] In some embodiments, during the vehicle sensor initialization phase (Phase 1), the initialization process is performed by 5-10 vehicles, and data should be collected for at least 20 trips per road. The vehicle sensors can be deployed on any type of vehicle traveling on the road at any ground clearance.

[0136] In some embodiments, during the pilot phase (Phase 2) of the vehicle-mounted sensor, small-scale pilot tests are conducted on 10-20 vehicles of different types to ensure that the vehicle-mounted sensor configuration, ground clearance customization, and network camera function properly before the vehicle-mounted sensor is deployed to a fleet of 150 vehicles.

[0137] In some embodiments, during the vehicle sensor baseline data collection phase (Phase 3), the road and driving data and / or traffic accident risk data generated by the vehicle sensors installed in 150 vehicles are evaluated, cleaned, and filtered for use in the artificial intelligence model training process.

[0138] In some embodiments, during the comprehensive data collection phase of the vehicle sensors (Phase 4), after reviewing and improving the data of the vehicle sensors installed in 150 vehicles, data collection is performed again on the vehicle sensors installed in 150 vehicles.

[0139] In some embodiments, a total of 150 valid satisfaction surveys will be conducted during the vehicle sensor user survey phase (Phase 5).

[0140] In some embodiments, the scalability of the system and method allows for the expansion of onboard sensor installation on different types of vehicles in a city. As more vehicles are equipped with onboard sensors, the system's coverage and the accuracy of generating TCI and TARI will improve, thus providing more reliable information for drivers and road users.

[0141] The various aspects of the present invention are further described below with reference to the accompanying drawings:

[0142] like Figure 1The exemplary system 100 for monitoring road traffic congestion and traffic accident risk shown basically includes multiple vehicle-mounted sensors 110, a cloud server 120, and at least one terminal 130. The cloud server 120 is communicatively connected to the multiple vehicle-mounted sensors 110 and the at least one terminal 130. In this embodiment, the system 100 also includes a local server 111, an FMS database 124, and an information hub database 125. In this embodiment, each vehicle-mounted sensor 110 is installed on a designated vehicle and configured to acquire real-time visual data obtained by the designated vehicle, and process the real-time visual data to generate road and driving data and / or traffic accident risk data.

[0143] In this embodiment, the cloud server 120 is an AWS cloud server, including the following modules: a fleet management system (FMS) 121, an information hub 122, and a geographic information system (GIS) platform 123, which are communicatively connected to each other for data exchange. In this embodiment, multiple vehicle-mounted sensors 110 communicate with the FMS 121 of the cloud server 120 using the Message Queuing Telemetry Transport (MQTT) protocol. The FMS 121 communicates with the information hub 122 using the Hypertext Transfer Protocol Secure (HTTPS) protocol. The information hub 122 communicates with the GIS platform 123 using the Hypertext Transfer Protocol Secure (HTTPS) protocol.

[0144] In this embodiment, the vehicle sensor 110 includes an AIoT sensor, such as a camera- and radar-LiDAR-based sensor equipped with the Mobileye ALL protocol. The vehicle sensor 110 includes an AIoT camera unit, an on-chip system processor, a speaker, a display, a GPS unit, and memory for temporary data storage. The vehicle sensor 110 captures and processes real-time visual data regarding speed, distance, lane departure, collision warning, and other relevant factors.

[0145] In this embodiment, the FMS 121 wirelessly communicates with multiple on-board sensors 110 and is configured to receive and process road and driving data and / or traffic accident risk data from the multiple on-board sensors to generate processed driving data. The FMS 121 collects real-time driving data related to a designated vehicle from the on-board sensors 110 installed on the vehicle, providing vehicle tracking, route optimization, maintenance scheduling, and driver performance monitoring. In this embodiment, the FMS is further configured to store identification data for multiple vehicles, each vehicle identification data corresponding to the designated vehicle, and the FMS is configured to bind each vehicle identification data to the corresponding road and driving data and / or the corresponding traffic accident risk data.

[0146] In this embodiment, the FMS database 124, which communicates with FMS 121, is specifically used to store data (e.g., real-time visual data) collected by multiple vehicle sensors 110. The information hub database 125, which communicates with information hub 122, has a broader purpose. It serves as a central data repository for storing various types of data, including spatial and non-spatial data, real-time and historical information, and geospatial and non-geospatial data. The data can be acquired, generated, and / or stored from different modules in system 100, or it can be acquired, generated, and / or stored from other sources, such as open data listed in Table 2.

[0147] In this embodiment, the information hub 122 is used for cloud storage and as a data hub. It is a repository for acquiring and storing at least one type of information hub data, including open data, processed driving data (such as road and driving data and / or traffic accident risk data) and / or stored data (such as road data) from the FMS. The collected data (including processed geospatial data from the Geographic Information System (GIS) platform 123) is processed to generate the Traffic Congestion Index (TCI) and the Traffic Accident Risk Index (TARI). The information hub 122 is further configured to store the processed geospatial data from the GIS platform as new stored data for future processing.

[0148] In this embodiment, the Geographic Information System (GIS) platform 123 integrates data from different modules and systems outside the system and processes the analysis results to further enable visualization and decision-making. For example, the GIS platform 123 processes and summarizes at least one type of information hub data, such as each dataset, real-time data from vehicle sensors, and road design and environmental settings data around each road segment, to generate at least one type of processed geospatial data with the same frequency and resolution, thereby deriving spatially continuous surfaces through various spatial interpolations and analyses. These surfaces will also be further analyzed based on factors such as proximity analysis, spatial filtering of potential hazards / traffic events, and microclimate conditions along the road segment. Before the analysis, a simple exploratory spatial data analysis will be performed to study the intuitive relationship between road / traffic conditions and various meteorological / environmental / historical data. The summarized data (including the processed post-geospatial data) will be transmitted to the information hub for further processing to generate the Traffic Congestion Index (TCI) and the Traffic Accident Risk Index (TARI).

[0149] In this embodiment, terminal 130 includes a mobile phone and a computer, which provide user-friendly interfaces for the driver and stakeholders, respectively. The mobile application accesses the system to obtain data such as TCI, TARI, and visualized geospatial data, and receives road hazard and safety alerts generated by the system (such as the safety alerts listed in Table 3), and displays relevant information or data.

[0150] In this embodiment, the local server 111 is communicatively connected to multiple vehicle sensors 110 and the terminal 130, and the local server 111 is configured to store real-time visual data from the vehicle sensors 110 and can transmit the real-time visual data to the terminal 130 so that users with access rights can access the raw real-time visual data.

[0151] Now for reference Figure 2 The exemplary method 200 for monitoring road traffic congestion and traffic accident risk, as shown, includes the following steps:

[0152] Step 210: Acquire real-time visual data from a designated vehicle using multiple onboard sensors. Each onboard sensor 110 is mounted on the designated vehicle. As an example, the onboard sensors include AIoT sensors, such as camera- and radar-LiDAR-based sensors equipped with the Mobileye ALL protocol. As an example, the onboard sensors include an AIoT camera unit, an on-chip system processor, a speaker, a display, a GPS unit, and memory for temporary data storage.

[0153] Step 220: Process the real-time visual data using the multiple onboard sensors to generate road and driving data and / or traffic accident risk data. As an example, the real-time visual data is processed by an on-chip system processor.

[0154] Step 230: Receive and process road and driving data and / or traffic accident risk data from multiple onboard sensors via FMS to generate processed driving data.

[0155] Step 240: Acquire and store at least one type of information hub data through the information hub, including open data, processed driving data from the FMS, and / or stored data.

[0156] Step 250: Receive and process the at least one information hub data through a GIS platform to generate at least one processed geospatial data.

[0157] Step 260: Process the processed geospatial data through the information hub to generate a Traffic Congestion Index (TCI) and / or a Traffic Accident Risk Index (TARI).

[0158] Step 270: Receive and process the TCI and / or TARI through the GIS platform to generate visualized geospatial data.

[0159] Step 280: Display at least the aforementioned visualized geospatial data via at least one terminal.

[0160] In some embodiments, method 200 further includes the step of: providing any system as described in this disclosure, such as Figure 1 The system 100 shown basically includes multiple vehicle-mounted sensors 110, a cloud server 120, and at least one terminal 130. The cloud server 120 is an AWS cloud server and includes the following modules: a fleet management system (FMS) 121, an information hub 122, and a geographic information system (GIS) platform 123, which communicate with each other to exchange data.

[0161] In some embodiments, method 200 further includes the following steps: storing the real-time visual data via a local server; and transmitting the real-time visual data to the at least one terminal via the local server.

[0162] In some embodiments, method 200 further includes the step of storing processed geospatial data from the GIS platform as new stored data for future processing.

[0163] In some embodiments, method 200 further includes the following steps: storing multiple vehicle identity data through the FMS, the vehicle identity data corresponding to multiple designated vehicles; and binding each vehicle identity data with the corresponding road and driving data and / or the corresponding traffic accident risk data through the FMS.

[0164] In some embodiments, method 200 further includes the step of processing the at least one information hub data through a first AI module comprising a unified model algorithm of a graph convolutional network (GCN) and long short-term memory (LSTM).

[0165] In some embodiments, method 200 further includes the step of processing the at least one information hub data by a second AI module that combines the probabilistic reasoning capabilities of a dynamic Bayesian network (DBN) with the sequence modeling capabilities of a recursive neural network (RNN).

[0166] The preferred embodiments of the present invention have been described above with reference to the accompanying drawings, but this does not limit the scope of the invention. Those skilled in the art can implement the present invention in various modifications without departing from its scope and spirit; for example, a feature of one embodiment can be used in another embodiment to obtain yet another embodiment. Any modifications, equivalent substitutions, and improvements made within the scope of the present invention should be within the scope of the present invention.

Claims

1. A system for monitoring road traffic congestion and / or traffic accident risks, comprising: (1) Multiple vehicle-mounted sensors, each sensor is installed on a designated vehicle and configured to acquire real-time visual data obtained by the designated vehicle and process the real-time visual data to generate road and driving data and / or traffic accident risk data; (2) A cloud server, including a fleet management system (FMS), an information hub, and a geographic information system (GIS) platform, wherein the FMS, information hub, and GIS platform communicate with each other. The FMS communicates wirelessly with the plurality of vehicle-mounted sensors and is configured to receive and process road and driving data and / or traffic accident risk data from the plurality of vehicle-mounted sensors to generate processed driving data. The information hub is configured to acquire and store at least one type of information hub data, including open data, processed driving data from the FMS, and / or stored data. The GIS platform is configured to receive and process at least one of the information hub data to generate at least one processed geospatial data. The information hub is further configured to process the processed geospatial data. To generate the Traffic Congestion Index (TCI) and / or the Traffic Accident Risk Index (TARI); and The GIS platform is further configured to receive and process the TCI and / or TARI to generate visualized geospatial data; and (3) At least one terminal, which communicates wirelessly with the cloud server and is configured to display at least the visualized geospatial data.

2. The system of claim 1, wherein the plurality of vehicle sensors include an AIoT camera unit, a system-on-a-chip processor, a speaker, a display, a GPS unit, and a memory for temporary data storage.

3. The system according to claim 1, further comprising a local server, the local server communicating with the plurality of vehicle-mounted sensors and the at least one terminal, and the local server configured to store the real-time visual data and transmit the real-time visual data to the at least one terminal.

4. The system of claim 1, wherein the information hub is further configured to store processed geospatial data from the GIS platform as new stored data for future processing.

5. The system according to claim 1, wherein the FMS is further configured to store identity data of multiple vehicles, the vehicle identity data corresponding to the designated vehicle, and the FMS is configured to bind each vehicle identity data with the corresponding road and driving data and / or the corresponding traffic accident risk data.

6. The system according to claim 1, wherein the cloud server further includes a first AI module, the first AI module including a unified model algorithm comprising a graph convolutional network (GCN) and a long short-term memory (LSTM) for processing the at least one information hub data.

7. The system according to claim 1, wherein the cloud server further includes a second AI module, the second AI module including a dynamic Bayesian network (DBN) and a recursive neural network (RNN) for processing the at least one information hub data.

8. A method for monitoring road traffic congestion and / or traffic accident risks, comprising: Real-time visual data obtained from a designated vehicle is acquired through multiple onboard sensors; The real-time visual data is processed by the multiple vehicle-mounted sensors to generate road and driving data and / or traffic accident risk data; The system receives and processes road and driving data and / or traffic accident risk data from the multiple on-board sensors via FMS to generate processed driving data. At least one type of information hub data is acquired and stored through the information hub, including open data, processed driving data from the FMS, and / or stored data; The system receives and processes at least one type of information hub data through a GIS platform to generate at least one type of processed geospatial data. The processed geospatial data is processed through the information hub to generate the Traffic Congestion Index (TCI) and / or Traffic Accident Risk Index (TARI). The TCI and / or TARI are received and processed through the GIS platform to generate visualized geospatial data; as well as At least the aforementioned visualized geospatial data is displayed through at least one terminal.

9. The method of claim 8, further comprising the following steps: The real-time visual data is stored on a local server; and The real-time visual data is transmitted to the at least one terminal via a local server.

10. The method of claim 8, further comprising the following steps: The processed geospatial data from the GIS platform is stored as new stored data for future processing.

11. The method of claim 8, further comprising the following steps: The FMS stores multiple vehicle identity data, which correspond to multiple specified vehicles. as well as The FMS binds each vehicle's identity data with the corresponding road and driving data and / or the corresponding traffic accident risk data.

12. The method of claim 8, further comprising the following steps: The at least one information hub data is processed by the first AI module, which includes a unified model algorithm of a graph convolutional network (GCN) and long short-term memory (LSTM).

13. The method of claim 8, further comprising the following steps: The at least one information hub data is processed by a second AI module that includes a Dynamic Bayesian Network (DBN) and a Recursive Neural Network (RNN).

Images