Smart Town Traffic Management System Using LoRa and Machine Learning Mechanism

By Seung Byum Seo and Dhananjay Singh, ReSENSE Lab Hankuk University of Foreign Studies, Global Campus – Yongin, South Korea

November 2018

In this article we propose a customized Low Power Wide Area Networks (LPWAN) architecture and machine learning mechanism adaptable to a smart town’s traffic management system. LoRa, also known as Long Range Wide Area Network, is a new technology used for LPWAN [1]. This proposal focuses on a smart town, a smaller concept than a smart city, in order to detour the challenges for simulation and provide high-quality service to a small community. By integrating concepts from wireless communication, traffic theory, and machine learning, the proposed cloud platform provides a powerful traffic management model for the smart town. The proposed customized LoRa architecture is not only suitable for manageability, but also for scalability. Our goal is to develop a real-time testbed solution in order to conduct performance analysis and verification of the competency of our model in a real-world scenario.

Figure 1: System Overview of LoRa based Intelligent Traffic Management System

Driving has become one of the critical factors of everyday stress. Even though driving should provide a convenient experience as private transportation, it turns out that everyday drivers experience not only negative moods but also frustration and irritation due to high traffic congestion [2]. In order to alleviate traffic congestion and provide a better driving experience to drivers, this research introduces a logistic regression mechanism for an Intelligent Traffic Management System (ITMS), which is designed explicitly for a LoRa wireless network. Figure 1 shows how the proposed system works in general. The data collected from smart sensors is transferred to the proposed LoRa cloud platform. Then the platform runs data analysis and machine learning algorithm, which are used as an input of ITMS. As long as solid ITMS is built, any applications including adaptive navigation system can be implemented based on ITMS [4]. Therefore, the ultimate goal of this research is to establish a basis of ITMS.

ITMS should be able to control the traffic flow in real-time demand [3]. That is because its function includes controlling the traffic flow through LoRa installed traffic signals, managing parking lots, providing public services such as fire and medical services, and protecting the town with public surveillance. Therefore, it is essential to collect the traffic data, to manage the collected data, and to predict the best solution that fits the needs of the town residents. That is why ITMS should be activated through the LoRa cloud platform, which has strengths on performance and availability. In addition, we give a brief explanation of the advantages of LoRa over other wireless technologies and explain why LoRa is suitable for smart town traffic management [5]. Thanks to LoRa, the two beacons installed at each corner of a road transfer the data to the cloud server. Then, the cloud server sends this data to the machine learning algorithm, which predicts whether the road is congested or not. Figure 2 illustrates how the proposed mechanism runs above the ITMS:

Figure 2: Pseudo-code of the Proposed Machine Learning Mechanism

This code makes our system simple but powerful thanks to the logistic regression algorithm. Logistic regression algorithm is a machine learning algorithm, which is one of the popular, supervised learning methods. Based on the training set, the algorithm learns how to predict the real-valued output. The main inputs we use for the training set are traveling time and density. They are supposed to be calculated from the cloud server. We predict traffic congestion by these two inputs. Regarding implementation and visualization, Scikit-learn library for python and Matplotlib library are used respectively. The response value obtained from this algorithm also has the potential to be used in an adaptive navigation system.

The benefits of smart traffic management are numerous. Economically, it reduces the everyday pollution emitted from vehicles by providing the shortest routes to drivers, leading to less fuel consumption. From the drivers’ perspective, they become free from everyday stress caused by traffic congestion. When a medical emergency occurs, this system can contribute to saving lives by giving priority to the ambulance and the trauma center by providing the fastest path to the victim [7]. Lastly, a smart traffic system can contribute to preventing and catching criminals. This contribution is possible by smart analytics keeping track of data from surveillance cameras [5]. Nonetheless, the main feature of using smart traffic management is to provide faster and better driving experiences to drivers. That is why understanding traffic flow theory is important to implement ITMS.

The proposed model aims to be used as an intelligent traffic management system in a smart town. Based on simple but powerful performance, this model will provide a high-quality service, which is suitable for smart town traffic management. By converging benefits of supervised learning and the LoRa network, this new mechanism will contribute to the driving experience by predicting the overall traffic flow of the driver’s route and providing adaptive routing services for each driver. By using the cloud platform, we also expect for traffic-related organizations, such as city traffic management, traffic analysts, and public service organizations, to utilize the collected data for signal management, routing, smart parking, tracking, and emergency service. As a result, drivers will benefit from saving fuel and being free from driving stress. High-quality public service will be delivered to town residents leading to the increase of satisfaction both in public and private sectors.

References

[1] L. Alliance, LoRa alliance technology, https://lora-alliance.org/resource-hub/what-lorawantm, 2018.

[2] D. A. Hennessy, D. L. Wiesenthal, Traffic congestion, driver stress, and driver aggression, aggress.
Behav. 25 (1999) 409–423.

[3] Smarter Cambridge Transport, Smart traffic management, http://www.smartertransport.uk/ smart-t traffic-management/, 2017.

[4] D. Singh, G. Tripathi, S. C. Shah, R.R.Righi, “Cyber-Physical Surveillance System for Internet of Vehicles”, IEEE World Forum on Internet of Things WF-IoT 2018, 5-8 February 2018 — Singapore, pp. 551-556.

[5] J. de Carvalho Silva, J. Rodrigues, A. Alberti, P. Soli´c, A. L. L. Aquino, LoRaWAN – A Low Powerˇ WAN Protocol for Internet of Things: A Review and Opportunities, 2017.

[6] R. P. Singh, Smart traffic management with real time data analysis, https://www.cisco.com/c/ en_in/about/knowledge-network/smart-traffic.html2014.

[7] D. Singh, M. Singh, “Internet of Vehicles for Smart and Safe Driving”, International Conference on Connected Vehicles and Expo (ICCVE), Shenzhen, 19-23 Oct. 2015.

Seung Byum (Bruno) Seo, Research Intern, ReSENSE Lab Seung Byum (Bruno) Seo is a graduate student in the department of computer science at University of Illinois Urbana-Champaign, Illinois. Currently, he is a research intern at ReSENSE Lab, Hankuk University of Foreign Studies (HUFS), Global campus. He received his B. E. degree in Software Convergence from Hankuk University of Foreign Studies (HUFS), Seoul, South Korea. His research interests lie in the field of Machine Learning, IoT, and Natural Language Processing.

Dhananjay Singh, Director of ReSENSE Lab Dhananjay Singh received his B. Tech. degree in Computer Science and Engineering from United College of Engineering and Research (UCER), Allahabad. M. Tech. degree in IT with the spec. in Wireless Communication and Computing from Indian Institute of Information Technology (IIIT), Allahabad, India. Ph.D. degree in Ubiquitous IT from Dongseo University (DSU), Busan, South Korea in 2003, 2006 and 2010 respectively. In September 2012, Dr. Singh joined Hankuk University of Foreign Studies (HUFS), South Korea where he is an Associate Professor in the Department of Electronics Engineering. He was also chairperson/head in the division of Global IT at HUFS from Feb. 2013 to Feb. 2018. Before joining the HUFS, he worked in the division of Future Internet Architecture at Electronics and Telecommunication Research Institute (ETRI) and National Institute of Mathematical Sciences (NIMS), Daejeon, South Korea from 2010 to 2012. He has published more than 100 research papers and an author of 3 books, 7 chapters and an inventor of 7 patents. He is an ACM distinguished speaker as well as delivered more than 50 invited talks in the areas of Internet of Things, Future Internet, wireless sensor networks, Information communication and networks. He is a Senior Member of IEEE & ACM society. He is also an Editor of PlosOne and Heliyon Journals.

Editor:

Stephan S. Jones, Ph.D. Stephan S. Jones, Ph.D., Director, Center for Information and Communication Sciences, Ball State University joined the Center for Information and Communication Sciences faculty in August of 1998. He came to Ball State University (BSU) from completing his doctoral studies at Bowling Green State University where he served the Dean of Continuing Education developing a distance-learning program for the College of Technology’s undergraduate Technology Education program. Dr. Jones was instrumental in bringing the new program on board because of his technical background and extensive research in the distance-learning field. Prior to coming to higher education, Dr. Jones spent over sixteen and a half years in the communication technology industry. He owned his own teleconnect, providing high-end commercial voice and data networks to a broad range of end users. Dr. Jones provided all the engineering and technical support for his organization that grew to over twenty employees and two and a half million dollars per year revenue. Selling his portion of the organization in December of 1994, Dr. Jones worked briefly for Panasonic Communications and Systems Company as a district sales manager providing application engineering and product support to distributors in a five-state area prior to starting doctoral studies.