Reinforcement Learning Traffic Signal Control for Sub-Area Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current traffic signal control systems operate statically based on predefined time-of-day schedules, lacking real-time optimization due to limitations in data collection and processing, especially in urban areas with high traffic volumes.
Innovation Solution
The implementation of a reinforcement learning model using a neural network to configure state information from downstream traffic data, allowing for dynamic adjustment of green times and offsets for traffic lights in a sub-area.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traffic signals are operated in a static manner based on predefined time of day, then the control system is simple and easy to implement, but the system cannot optimize traffic flow in real-time due to limitations in data collection and processing
Solution Approach 1:
The patent transforms the static traffic signal control system into a dynamic one by implementing real-time signal control based on current traffic conditions. The system uses vehicle detection data, queue length information, and waiting time metrics to dynamically adjust green and red signal durations, allowing the control system to adapt to changing traffic patterns rather than following fixed time-of-day schedules
Solution Approach 2:
The patent implements a feedback mechanism where the traffic signal control system continuously receives information about actual traffic conditions (vehicle presence, queue length, waiting time) and uses this feedback to adjust signal timings. The processor analyzes real-time data from detection devices and modifies green/red signal durations accordingly, creating a closed-loop control system that optimizes traffic flow based on actual conditions rather than predetermined schedules
2Measurement precision
If AI image analysis technology is introduced to obtain higher quality traffic data, then data quality improves, but real-time optimal signal control remains inadequate for practical implementation
Solution Approach 1:
The patent segments the complex AI image analysis system into simpler, more practical components. Instead of relying on full AI image analysis, the system uses dedicated detection devices to measure specific parameters (vehicle presence, queue length, waiting time) at key locations. This segmentation approach maintains data quality while improving real-time processing capability and practical implementability
Solution Approach 2:
The patent introduces an intermediary processing layer between data collection and signal control. The processor acts as an intermediary that receives various traffic parameters (vehicle detection, queue length, waiting time), processes this information according to predetermined algorithms, and generates optimized signal timings. This intermediary approach simplifies the overall system while maintaining real-time control capability
3Productivity
If multiple traffic signals are optimized simultaneously, then overall traffic flow improves, but the complexity of optimizing multiple signals increases significantly
Solution Approach 1:
The patent merges the control of multiple traffic signals into a unified control system. The processor coordinates green and red signal timings across multiple intersections by considering cumulative queue lengths and waiting times. This merging approach allows simultaneous optimization of multiple signals while managing complexity through integrated processing rather than separate independent controls
Solution Approach 2:
The patent implements a universal control algorithm that can be applied to multiple traffic signals with different configurations. The predetermined processing methods and optimization criteria are designed to work across various intersection types and traffic patterns, allowing the system to optimize multiple signals simultaneously using a single versatile framework rather than requiring separate specialized controls for each signal
Data Source
AI summary
Provided are methods and apparatuses for controlling traffic signals of traffic lights in a sub-area by using a neural network model. The method according to an embodiment of the present disclosure may configure state information of a sub-area by using downstream information obtained in a current cycle time for each of a plurality of intersections included in the sub-area. In addition, the method may input the state information to a trained reinforcement learning model, and obtain action information of the sub-area including green times and offsets, by using an output from the trained reinforcement learning model. Furthermore, the method may generate coordinated signal values for applying the action to traffic lights in the sub-area in a subsequent cycle time.