Smart control of traffic lights based on traffic density in the multi-intersection network by using Q learning
Abstract In urban areas, utilizing traffic lights to prioritize vehicles at the intersection is a solution to control traffic. Among the smart traffic light methods, the methods based on machine learning are particularly important due to their simplicity and performance. In this paper, Q-learning wi...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Springer
2023-11-01
|
| Series: | Discover Artificial Intelligence |
| Subjects: | |
| Online Access: | https://doi.org/10.1007/s44163-023-00087-z |
| _version_ | 1797414481198841856 |
|---|---|
| author | Seyedeh M. Mortazavi Azad A. Ramazani |
| author_facet | Seyedeh M. Mortazavi Azad A. Ramazani |
| author_sort | Seyedeh M. Mortazavi Azad |
| collection | DOAJ |
| description | Abstract In urban areas, utilizing traffic lights to prioritize vehicles at the intersection is a solution to control traffic. Among the smart traffic light methods, the methods based on machine learning are particularly important due to their simplicity and performance. In this paper, Q-learning with deep neural network are combined and used in two different intersection models. The first one is an individual intersection, and the second one is two intersections that are connected and shared their actions. By using this method, the traffic light can make an intelligent decision at the intersection to reduce vehicle consumption time by managing the allocation of phases. The proposed smart traffic light is studied and simulated via SUMO. The results illustrated that compared to fixed-time traffic lights, the average queue time of each vehicle in different traffic scenarios has been reduced by 34% in the individual intersection. In the case of two intersections, awareness and communication between agents led to a 24% reduction in the queue time of all cars in the heavy traffic scenario. |
| first_indexed | 2024-03-09T05:33:44Z |
| format | Article |
| id | doaj.art-d3de19947dd14c799f93a5ed55e97881 |
| institution | Directory Open Access Journal |
| issn | 2731-0809 |
| language | English |
| last_indexed | 2024-03-09T05:33:44Z |
| publishDate | 2023-11-01 |
| publisher | Springer |
| record_format | Article |
| series | Discover Artificial Intelligence |
| spelling | doaj.art-d3de19947dd14c799f93a5ed55e978812023-12-03T12:30:23ZengSpringerDiscover Artificial Intelligence2731-08092023-11-013111110.1007/s44163-023-00087-zSmart control of traffic lights based on traffic density in the multi-intersection network by using Q learningSeyedeh M. Mortazavi Azad0A. Ramazani1Electrical Engineering Department, Bu-Ali Sina UniversityElectrical Engineering Department, Bu-Ali Sina UniversityAbstract In urban areas, utilizing traffic lights to prioritize vehicles at the intersection is a solution to control traffic. Among the smart traffic light methods, the methods based on machine learning are particularly important due to their simplicity and performance. In this paper, Q-learning with deep neural network are combined and used in two different intersection models. The first one is an individual intersection, and the second one is two intersections that are connected and shared their actions. By using this method, the traffic light can make an intelligent decision at the intersection to reduce vehicle consumption time by managing the allocation of phases. The proposed smart traffic light is studied and simulated via SUMO. The results illustrated that compared to fixed-time traffic lights, the average queue time of each vehicle in different traffic scenarios has been reduced by 34% in the individual intersection. In the case of two intersections, awareness and communication between agents led to a 24% reduction in the queue time of all cars in the heavy traffic scenario.https://doi.org/10.1007/s44163-023-00087-zSmart trafficMachine learningQ-learningMulti-agentShared action |
| spellingShingle | Seyedeh M. Mortazavi Azad A. Ramazani Smart control of traffic lights based on traffic density in the multi-intersection network by using Q learning Discover Artificial Intelligence Smart traffic Machine learning Q-learning Multi-agent Shared action |
| title | Smart control of traffic lights based on traffic density in the multi-intersection network by using Q learning |
| title_full | Smart control of traffic lights based on traffic density in the multi-intersection network by using Q learning |
| title_fullStr | Smart control of traffic lights based on traffic density in the multi-intersection network by using Q learning |
| title_full_unstemmed | Smart control of traffic lights based on traffic density in the multi-intersection network by using Q learning |
| title_short | Smart control of traffic lights based on traffic density in the multi-intersection network by using Q learning |
| title_sort | smart control of traffic lights based on traffic density in the multi intersection network by using q learning |
| topic | Smart traffic Machine learning Q-learning Multi-agent Shared action |
| url | https://doi.org/10.1007/s44163-023-00087-z |
| work_keys_str_mv | AT seyedehmmortazaviazad smartcontroloftrafficlightsbasedontrafficdensityinthemultiintersectionnetworkbyusingqlearning AT aramazani smartcontroloftrafficlightsbasedontrafficdensityinthemultiintersectionnetworkbyusingqlearning |