SMART TRAFFIC MANAGEMENT SYSTEM FOR DETECTING TRAFFIC VIOLATIONS AND METHOD THEREOF

Abstract

Embodiments of the present disclosure relate to a smart traffic management system (100) for detecting traffic violations. A plurality of traffic signals is dynamically adjusted based on the received traffic data using a reinforcement learning (RL) engine. The RL engine is trained to optimize traffic flow by minimizing congestion, travel time, and traffic violations. An action space is defined by a set of traffic light signals such as Red, Yellow, Green for one or more intersections in a traffic network. An observation space comprising real-time data on vehicle types approaching each intersection, said vehicle types including cars, trucks, buses, motorcycles, and emergency vehicles. A penalty report is generated for the identified users using a report engine. Advantageously the present disclosure relates to a system that aims to achieve maximally efficient traffic flow and enhance road safety through intelligent traffic management and real-time violation detection.

Specification

Description:TECHNICAL FIELD
[0001] The present disclosure relates to the field of traffic management. More particularly, the present disclosure relates to a smart traffic management system and method for optimising traffic flow and detecting traffic violations.

BACKGROUND
[0002] Traffic problems are becoming more problematic. On one hand, during rush hour, there are usually more vehicles heading downtown than there are people leaving downtown when people go to work, and the situation is the opposite when people get off work. During this time, it is easy to be in a traffic jam. On the other hand, since the vehicles and traffic roads lack wireless communication, drivers cannot directly and conveniently get road condition information ahead, so that the drivers will find the traffic jam when they drive in a congested road. Then, the traffic jam may become more serious. Currently, traffic management is done very poorly in India, and an average Indian spends about 50 workdays stuck in traffic in a year.
[0003] Urban traffic congestion and violations of traffic regulations are growing concerns in modern cities, leading to increased travel times, higher pollution levels, and compromised road safety. Traditional traffic control systems rely on fixed-timed signals and static rules, which lack the adaptability required to respond to real-time traffic conditions. Furthermore, manual enforcement of traffic laws is resource-intensive and prone to human error, leading to inconsistent enforcement of traffic violations such as speeding, running red lights, or illegal turns.
[0004] To address these limitations, the present invention provides a smart traffic management system and method for detecting traffic violations that overcomes the shortcomings of the prior art.

OBJECTS OF THE PRESENT DISCLOSURE
[0005] It is a primary object of the present disclosure to provide a smart traffic management system and method for detecting traffic violations.
[0006] It is another object of the present disclosure to provide a smart traffic management system that controls traffic signals in order to minimize total travel time for all vehicles.
[0007] It is yet another object of the present disclosure to provide a dynamic tool that empowers urban planners and traffic engineers to experiment with and optimize traffic management strategies, visualizing the impact on congestion and traffic flow in specific areas, and hence, plan routes accordingly.
[0008] It is another object of the present disclosure to provide a system with Instant Environment Creation that eliminates the need to design maps and hence significantly improves computational efficiency.
[0009] It is another object of the present disclosure to provide a system that optimizes traffic management significantly compared to other traffic-based models.
[0010] It is another object of the present disclosure to provide a traffic management system with reinforcement learning model that prioritizes Special vehicles such as ambulances, police cars.
[0011] It is another object of the present disclosure to provide a system that incorporates reinforcement learning model in the traffic signals to optimize the time as that recalculates the traffic control at every instant and gives the best fit management.
[0012] It is another object of the present disclosure to provide a system that is integrated with Traffic Violation detection to make sure traffic discipline is followed.
[0013] It is another object of the present disclosure to provide a smart traffic management system and method for detecting traffic violations that automates traffic law enforcement.
[0014] The present disclosure allows provides a system which simulates the real maps using open-source tools allowing a complete traffic system.
 
SUMMARY
[0015] The present disclosure relates to the field of traffic management. More particularly, the present disclosure relates to a smart traffic management system and method for detecting traffic violations.
[0016] In an aspect of the present disclosure, a smart traffic management system for detecting traffic violations is disclosed. The system includes a reinforcement learning (RL) engine to determine position of a set of traffic cameras positioned at strategic intersections and along roads and receive traffic data from a set of traffic cameras to simulate a road network using a traffic simulation tool. The model controls traffic light phases at intersections to reduce congestion and improve traffic efficiency in real-time. The action space involves selecting a traffic light phase-red, yellow, or green-at each intersection, while the observation space tracks the number and types of vehicles approaching each intersection, including cars, trucks, buses, motorcycles, and emergency vehicles. The reward function is based on a modified Max Pressure approach, optimizing traffic flow by minimizing congestion and prioritizing emergency vehicles. The aggregate reward across intersections ensures that the entire traffic network is optimized, enhancing the flow of traffic and the rapid movement of high-priority vehicles.
[0017] In an aspect, the plurality of traffic violations is selected from vehicle overloading, helmetless riding, red-light running and any combination thereof.
[0018] In an aspect, the RL engine is configured to adjust the plurality of traffic signals based on real-time traffic density, pedestrian movements, vehicle types and any combination thereof.
[0019] In an aspect, the system includes a feedback loop unit with historical data used to increase future traffic signal decisions.
[0020] In an aspect, the report engine is configured to report and store the violations, including details of the offense, evidence, penalty information and any combination thereof.
[0021] In an aspect, the system comprising an emergency response unit configured to prioritize emergency vehicles including ambulances and fire trucks and to automatically adjust the plurality of traffic signals to create a set of green corridors for emergency vehicles.
[0022] In an aspect, the system is configured to dynamically allocate a timing to traffic data to optimize traffic flow, thereby ensuring efficient traffic management.
[0023] In another aspect of the present disclosure, a method of performing the smart traffic management system for detecting traffic violations is disclosed. The method begins with determining position of a set of traffic cameras at strategic intersections and along roads and receiving from the set of traffic cameras, traffic data to simulate a road network using a traffic simulation tool. The traffic data is interpreted from the set of traffic light cameras. A plurality of traffic violations is detected using the camera footage and the subsequent model. The method further includes identifying corresponding users associated with the detected violations and generating a penalty report for the identified users using a report engine.

BRIEF DESCRIPTION OF DRAWINGS
[0024] The accompanying drawings are included to provide a further understanding of the present disclosure and are incorporated in and constitute a part of this specification. The drawings illustrate exemplary embodiments of the present disclosure, and together with the description, serve to explain the principles of the present disclosure.
[0025] FIG. 1 illustrates an exemplary schematic diagram of the reinforcement learning model implementing the proposed smart traffic management system for detecting traffic violations, in accordance with an embodiment of the present disclosure.
[0026] FIG. 2 illustrates a schematic flow diagram of the Proximal Policy Optimization (PPO) model implementing the proposed smart traffic management system for detecting traffic violations, in accordance with an embodiment of the present disclosure.
[0027] FIG. 3 illustrates an exemplary view of a flow diagram of the proposed smart traffic management system and method for detecting traffic violations, in accordance with an embodiment of the present disclosure.
[0028] FIG. 4 illustrates an exemplary view of a flow diagram of the proposed method of performing the smart traffic management system for detecting traffic violations, in accordance with an embodiment of the present disclosure.

DETAILED DESCRIPTION
[0029] The following is a detailed description of embodiments of the disclosure depicted in the accompanying drawings. The embodiments are in such detail as to clearly communicate the disclosure. However, the amount of detail offered is not intended to limit the anticipated variations of embodiments. On the contrary, the intention is to cover all modifications, equivalents, and alternatives falling within the spirit, and scope of the present disclosure as defined by the appended claims.
[0030] Specific details are given in the following description to provide a thorough understanding of the embodiments. However, it will be understood by one of ordinary skill in the art that the embodiments may be practiced without these specific details. For example, circuits, systems, networks, processes, and other components may be shown as components in block diagram form in order not to obscure the embodiments in unnecessary detail. In other instances, well-known circuits, processes, algorithms, structures, and techniques may be shown without unnecessary detail to avoid obscuring the embodiments.
[0031] Also, it is noted that individual embodiments may be described as a process that is depicted as a flowchart, a flow diagram, a data flow diagram, a structure diagram, or a block diagram. Although a flowchart may describe the operations as a sequential process, many of the operations can be performed in parallel or concurrently. In addition, the order of the operations may be re-arranged. A process is terminated when its operations are completed but could have additional steps not included in a figure. A process may correspond to a method, a function, a procedure, a subroutine, a subprogram, etc. When a process corresponds to a function, its termination can correspond to a return of the function to the calling function or the main function.
[0032] The present disclosure relates to the field of traffic management. More particularly, the present disclosure relates to a smart traffic management system and method for detecting traffic violations.
[0033] The smart traffic management system for automatically detecting traffic violations is disclosed. The system includes a memory with processor-executable instructions, which on execution, causes a processor to determine position of a set of traffic cameras positioned at strategic intersections and along roads and receive traffic data from a set of traffic cameras to simulate a road network using a traffic simulation tool. A plurality of traffic signals is dynamically adjusted based on the received traffic data using a reinforcement learning (RL) engine. The RL engine is operatively coupled to the processor. The RL engine is trained to optimize traffic flow by minimizing congestion and travel time. The traffic data is interpreted from the set of traffic cameras. A plurality of traffic violations is detected by computer vision models utilizing the traffic camera footage. Corresponding users associated with the detected violations are identified via number plates. A penalty report is generated for the identified users using a report engine.
[0034] The present disclosure focuses on revolutionizing traffic management through an interactive web platform. Users can effortlessly pinpoint regions of interest for traffic interventions on the user-friendly website, initiating a sophisticated simulation process driven by a Reinforcement Learning (RL) algorithm developed with stable baselines, SUMO, and TRACI. This dynamic tool empowers urban planners and traffic engineers to experiment with and optimize traffic management strategies, visualizing the impact on congestion and traffic flow in specific areas, and hence, plan routes accordingly. To enhance accuracy, the proposed system incorporates state-of-the-art models like YOLOv8 for real-time vehicle tracking and detection from camera feeds within the simulations which is also used for real time traffic violation detection. This comprehensive solution offers a practical and intuitive approach to address urban traffic challenges, providing a valuable resource for decision-makers seeking efficient and data-driven traffic management solutions.
[0035] The proposed system with Instant Environment Creation eliminates the need to design maps and hence significantly improves efficiency. The system optimizes traffic management significantly compared to other traffic-based models. The system prioritizes Special vehicles such as ambulances, police cars through the Reinforcement model.
[0036] FIG. 1 illustrates an exemplary schematic diagram of the reinforcement learning model implementing the proposed smart traffic management system (100) for detecting traffic violations, in accordance with an embodiment of the present disclosure.
[0037] FIG. 2 illustrates a schematic flow diagram of the Proximal Policy Optimization (PPO) model (200) implementing the proposed smart traffic management system for detecting traffic violations, in accordance with an embodiment of the present disclosure
[0038] Referring to FIGs. 1 and 2, the model used for traffic light optimization is based on Proximal Policy Optimization (PPO), a reinforcement learning (RL) algorithm. PPO's on-policy mechanism ensures stability during policy updates by limiting large changes to the policy, making it particularly effective for traffic control in complex urban environments. The PPO model operates by optimizing the traffic light phases at intersections to reduce congestion and improve traffic efficiency in real-time.
[0039] In an embodiment, the action space consists of choosing the traffic light phase at each intersection, with three possible states: red, yellow, or green. Each action corresponds to selecting a phase for the entire intersection at a given time step. The action space for intersection i is defined as:
Ai= {T1T2T3T4 … Tn}
where i refers to the index of the intersection within the network and Tj refers to the current state of the Traffic light j at intersection i. Tj can take on the values:
{Red, Yellow, Green}.
The model determines the optimal action for each intersection based on real-time traffic observations to minimize delays and prioritize certain vehicle classes such as emergency vehicles.
[0040] In an embodiment, the observation space consists of the number of vehicles of each type approaching each intersection. These vehicle types include cars, trucks, buses, motorcycles, and emergency vehicles, ensuring that the model has a comprehensive understanding of the traffic load at every intersection. The observation vector o_t at time t for intersection i is represented as:

where Vcart,i,j represents the number of cars on road j connecting to intersection i, and similarly for other vehicle types. Here, i is the index of the intersection, and j is the index of the road connected to that intersection.
The reward function is based on a modified Max Pressure approach, where the objective is to optimize traffic flow by minimizing congestion. The reward for each intersection i at a given time step is computed based on the difference between the weighted sum of incoming and outgoing vehicles on all roads connected to that intersection. The reward for road j at intersection i at time t is defined as:

where Wvehiclein and Wvehicleout are the weighted sums of incoming and outgoing vehicles, respectively, and Lj is the number of lanes on road j. Higher weights are assigned to emergency vehicles to prioritize their movement.
The total reward for all intersections at time t is the sum of the rewards across all intersections:

where i is the index of the intersection and j is the index of the roads connected to intersection i. This aggregate reward ensures that traffic flow is optimized across the entire system, with priority given to emergency vehicles.
[0041] The reinforcement learning model is incorporated in the traffic signals to optimize the time as that recalculates the traffic control at every instant and gives the best fit management. It also prioritizes special vehicles like ambulances and police cars to help them travel faster. Additionally, it is integrated with Traffic Violation detection to make sure traffic discipline is followed. This could potentially be a game-changer for Traffic Management due to the incorporation of so many technologies together, which hasn't been seen before.
[0042] In an embodiment, the system 100 may be configured to dynamically adjust a plurality of traffic signals based on the received traffic data using the reinforcement learning (RL) engine. The RL engine is configured to adjust the plurality of traffic signals based on real-time traffic density, pedestrian movements, vehicle types and any combination thereof.
[0043] In an embodiment, the system 100 may be configured to train the RL engine to optimize traffic flow by minimizing congestion, travel time, and traffic violations by the training engine.
[0044] In an embodiment, the system 100 may be configured to interpret the traffic data from the set of traffic cameras by the traffic interpretation engine.
[0045] In an embodiment, the system 100 may be configured to detect a plurality of traffic violations by a plurality of users by the violation detection engine. The plurality of traffic violations is selected from vehicle overloading, helmetless riding, red-light running, high speed, illegal turns and any combination thereof.
[0046] In an embodiment, the system 100 may be configured to identify corresponding users associated with the detected violations.
[0047] FIG. 3 illustrates an exemplary view of a flow diagram 300 of the proposed smart traffic management system and method for detecting traffic violations, in accordance with an embodiment of the present disclosure.
[0048] With reference to FIG. 3, the process begins with initial training, where a specific region is selected, and relevant data from Open Street Maps is collected. This data is then used to simulate the road network using SUMO, a traffic simulation tool. In this simulation, vehicle routes are generated and applied, while past real-time traffic data is interpreted using CCTV footage. These two data sources-simulated and real-are then used to prepare training data and simulated Via SUMO simulation software. This data is utilized to train a Deep Reinforcement Learning Model, which undergoes continuous training using live data to enhance its accuracy and efficiency.
[0049] In an embodiment, once the model is trained, it is deployed for real-time use. YOLOv8m, a state-of-the-art object detection model, is employed to interpret live traffic data from CCTV footage. This interpretation helps in detecting violations such as vehicle overloading and helmetless riding. The number plates of violators are identified and stored for further action. The model also dynamically allocates traffic light timings to optimize traffic flow, leading to maximally efficient traffic management. Special priorities are given to emergency vehicles, helping to save lives by reducing their travel time. In real-time usage, live signal CCTV footage plays a crucial role in feeding data to the system, allowing it to make timely and accurate decisions. The overall system aims to achieve maximally efficient traffic flow and enhance road safety through intelligent traffic management and real-time violation detection.
[0050] In an implementation of an embodiment, a traffic violation detection module is integrated within the system. The cameras are integrated with the smart traffic signals to detect traffic violations such as red-light running, speeding, or illegal turns. A violation detection algorithm uses pattern recognition or machine learning to identify potential violations. Automatic generation of reports or alerts to the relevant authorities, including timestamped images or videos as evidence is performed. The ability to adapt traffic management decisions based on traffic conditions in nearby areas is achieved.
[0051] In an implementation of an embodiment, a method for training the reinforcement learning model is disclosed. Various traffic scenarios are simulated and learnt from the outcomes of previous signal adjustments. The model is rewarded for minimizing congestion, reducing violation occurrences, and optimizing pedestrian and vehicle flow. Detected traffic violations trigger an automated penalty system. The system is configured to send violation notifications to vehicle owners, including details of the offense, evidence, and penalty information.
[0052] In an implementation of an embodiment, if a particular traffic signal adjustment leads to lower congestion, the agent or user is rewarded, and this decision is reinforced. Conversely, if congestion increases, the agent penalizes the action and adjusts future decisions accordingly. Using computer vision and machine learning techniques, the violation detection module analyses the footage from cameras to identify violations. When a vehicle runs a red light or speeds, the system captures images or videos, records the vehicle's license plate, and timestamps the violation. Upon detecting a violation, the system generates a report containing evidence such as the time, location, vehicle identification, and nature of the violation. This information is sent to traffic authorities or an automated penalty system for further action.
[0053] FIG. 4 illustrates an exemplary view of a flow diagram of the proposed method of performing the smart traffic management system for detecting traffic violations, in accordance with an embodiment of the present disclosure.
[0054] In an embodiment, the proposed method 400 of performing the smart traffic management system for detecting traffic violations is disclosed. At step 402, training, by the system 100, a reinforcement learning agent based on Proximal Policy Optimization (PPO) to control traffic light phases at one or more intersections in a traffic network. Cameras are used to visually monitor vehicles and record potential traffic violations.
[0055] At step 404, interpreting, by the system 100, the traffic data from the set of traffic cameras.
[0056] At step 406, detecting, by the system 100, a plurality of traffic violations by a plurality of users. This module utilizes computer vision techniques, including object detection and pattern recognition, to analyze the data collected from cameras. Deep learning models are employed to distinguish normal traffic behaviour from violations, enabling automatic detection and reporting.
[0057] At step 408, identifying, by the system 100, corresponding users associated with the detected violations.
[0058] At step 410, generating, by the system 100, a penalty report for the identified users using the report engine.
[0059] In an embodiment, a central server or control unit integrates data from all cameras, and RL agents. This unit processes the incoming data, runs the RL algorithms, and controls traffic signals accordingly. The violation detection module also reports detected violations to this unit, which logs violations and communicates with enforcement authorities or systems. A communication framework connects traffic signals, cameras, and the central control unit. This network ensures the real-time transfer of data, allowing immediate response to traffic conditions and violations. The system can be integrated with city-wide IoT infrastructure or existing smart city systems.
[0060] In an embodiment, a Reinforcement Learning based Traffic Manager using Computer Vision is disclosed that uses OpenStreetMap to make instant environments and simulations for the Reinforcement Learning model to train. The present disclosure uses computer vision to obtain traffic data for the Reinforcement Model as well as for Traffic Violation Detection, which detects traffic violations and stores the respective License Plate information.
[0061] The present disclosure provides a Reinforcement Learning Powered Smart Traffic Manager with Traffic Violation Detector, designed to improve the efficiency of urban traffic management while ensuring compliance with traffic laws. The system leverages real-time traffic data and reinforcement learning (RL) algorithms to dynamically control traffic signals and manage traffic flow in an optimized manner. Additionally, the system integrates a traffic violation detection module, capable of identifying and reporting violations such as red-light running.
[0062] The system continuously learns and adapts to traffic conditions, using a reward-based RL model to make decisions that minimize congestion, reduce travel times, and improve overall road safety. The violation detection system utilizes cameras to monitor traffic behaviour, with machine learning algorithms analyzing the data to accurately detect and classify violations. Upon detection, the system automatically logs and reports violations, including relevant evidence for enforcement purposes.
[0063] While the foregoing describes various embodiments of the invention, other and further embodiments of the invention may be devised without departing from the basic scope thereof. The scope of the invention is determined by the claims that follow. The invention is not limited to the described embodiments, versions or examples, which are comprised to enable a person having ordinary skill in the art to make and use the invention when combined with information and knowledge available to the person having ordinary skill in the art.

ADVANTAGES OF THE INVENTION
[0064] The present disclosure provides a smart traffic management system and method for detecting traffic violations.
[0065] The present disclosure provides a dynamic tool that empowers urban planners and traffic engineers to experiment with and optimize traffic management strategies, visualizing the impact on congestion and traffic flow in specific areas, and hence, plan routes accordingly.
[0066] The present disclosure provides a system that optimizes traffic management significantly compared to other traffic-based models.
[0067] The present disclosure provides a traffic management system with reinforcement learning model that prioritizes Special vehicles such as ambulances, police cars.
[0068] The present disclosure allows provides a system which simulates the real maps using open-source tools allowing a complete traffic system.
[0069] The present disclosure provides a system that incorporates reinforcement learning model in the traffic signals to optimize the time as that recalculates the traffic control at every instant and gives the best fit management.
[0070] The present disclosure provides a system that is integrated with Traffic Violation detection to make sure traffic discipline is followed.
[0071] The present disclosure provides a smart traffic management system and method for detecting traffic violations that automates traffic law enforcement
, Claims:1. A smart traffic management system (100) for detecting traffic violations, the system (100) comprising:
an action space defined by a set of traffic signals such as Red, Yellow and Green for one or more intersections in a traffic network;
a set of traffic cameras positioned at strategic one or more intersections and along roads and simulate a road network using a traffic simulation tool;
an observation space comprising real-time data on vehicle types approaching each intersection, wherein the vehicle comprising cars, trucks, buses, motorcycles, emergency vehicles and any combination thereof;
a reward function that calculates a reward for each intersection based on a modified Max Pressure approach, wherein the reward is based on the difference between the weighted sums of incoming and outgoing vehicles for each road connected to said intersection, with higher weights assigned to emergency vehicles; and
a reinforcement learning (RL) engine configured to execute a Proximal Policy Optimization (PPO) reinforcement learning (RL) algorithm that determines the optimal traffic light phases for said intersections in real-time to minimize congestion and improve traffic efficiency,
wherein the RL engine is trained to optimize traffic flow by minimizing congestion and travel time. The traffic data is interpreted from the set of traffic cameras. A plurality of traffic violations is detected by computer vision models utilizing the traffic camera footage, corresponding users associated with the detected violations are identified via number plates and a penalty report is generated for the identified users using a report engine.

2. The system (100) as claimed in claim 1, wherein the plurality of traffic violations is selected from vehicle overloading, helmetless riding, red-light running and any combination thereof.
3. The system (100) as claimed in claim 1, wherein the RL engine is configured to adjust the plurality of traffic signals based on real-time traffic density, pedestrian movements, vehicle types and any combination thereof.

4. The system (100) as claimed in claim 1, wherein the system (100) comprising a feedback loop unit with historical data used to increase future traffic signal decisions.

5. The system (100) as claimed in claim 1, wherein the report engine is configured to report and store the violations, including details of the offense, evidence, penalty information and any combination thereof.

6. The system (100) as claimed in claim 1, wherein the system (100) comprising an emergency response unit configured to prioritize emergency vehicles including ambulances and fire trucks and to automatically adjust the plurality of traffic signals to create a set of green corridors for emergency vehicles.

7. The system (100) as claimed in claim 1, wherein the system (100) is configured to dynamically allocate a timing to traffic data to optimize traffic flow, thereby ensuring efficient traffic management.

8. A method (400) of performing the smart traffic management system (100) for detecting traffic violations, the method (400) comprising steps of:
training (402) a reinforcement learning agent based on proximal policy optimization (PPO) to control traffic light signals at one or more intersections in a traffic network;
interpreting (404), by the system (100), the traffic data from the set of traffic cameras;
detecting (406), by the system (100), a plurality of traffic violations using the camera footage and the subsequent model;
identifying (408), by the system (100), corresponding users associated with the detected violations; and
generating (410), by the system (100), a penalty report for the identified users using a report engine.

Documents