Evaluation of Machine Learning Techniques for Road Traffic and Related Air Quality Index in Smart Cities using IoT data.
MSc Data Analytics (MSCDA_B)
Research in Computing (CA1)
Submitted: 17th June 2018
Student Name – Pramit Kumar
Student ID – X01419895
School of Computing
National College of Ireland
Lecturer: Dr. Catherine Mulwa
By 2030, as indicated by the insight from United nation which shows that about 60 percent of total population will move to urban cities. With more addition of vehicles to the streets, worldwide governments are facing challenges in managing the traffic movement to keep roadways streamline and steady. Traffic activity is the best contribution to air quality index, with increase of nitrous oxides levels and other pollutants proceeding to rise. some zones where traffic congestion is high, for them it’s very hard to meet the new EU emission bench mark. With introduction of new smart city, we need smarter and efficient ways to maintain road traffic and control air quality index. Busy hour traffic jam is huge loss to the country’s economy and health of people. An answer for the regularly developing issue may in truth not be too far away. Because of new sources of information like Internet of Things (IoT) and excellent machine learning techniques can help us to find an efficient solution to solve the problem.
1.2 Research Question / or Problem Statement
“Can optimized routing of vehicles improves road traffic and maintain air quality index?”
Traffic congestion has bad effects on country’s economy, it causes slow movement and delay in transport services, wastage of fuel, vehicle-maintained cost increases, effecting the air quality index and many more. According to Irishtimes.com, Dublin is losing around 350 million of euro per year due to bad traffic congestion and if same situation continues in future the loss will be about 2 billion euro by 2033 and new public health research centre tests propose that, traffic pollution effects mental and physical health, knowledge and phycology of person, the harmful pollutant effect the person’s lungs and according to the World Health Organisation (WHO) said due to traffic pollution which cause death for 3.2 million every year. We need to fix this issue before it goes beyond our control and city planner or government bodies should take the help of Data Analytic to understand the cause of traffic congestion. The collections of Internet of Things (IoT) sensor data like road traffic and traffic pollution can be used for constructing a machine leaning techniques for smart traffic management. As more and more smart cities evolve worldwide, IoT data like images, sensors, pollution, weather and many more will well get integrated and can be vital source for information for data analytic
To implement machine learning technology, we need to consider some important aspects before starting the implementation process, like statistical model requires good amount of data having various pattern and training use-case and there is sufficient availability of road traffic and pollution pre-validate data (supervised) which will fulfil the implementation need. The fundamental of the smart city traffic would be decision-making system which will help to route the traffic as per different situations and there is wide scope for exploration of various decision machine learning algorithm like decision tree, clustering, Bayesian etc. with the provided dataset. The good metrices in the provided dataset will help us in building machine learning model with more accuracy. After exploring many research paper and journals (please see reference section) and studying different implementation methodologies, different perspectives helps to understand the techniques in faster pace and gives clearer picture of implementation approaches. The dataset which has been considered for the research is taken from “CityPulse Dataset Collection” which General Data Protection Regulation (GRPR) compliant and don’t comes under any other types of compliance and ethics rules. It has been concluded that the proposed research area is feasible and achievable for a given timeline and readily available pre-validate dataset will give more confidence to test the accuracy of the model.
1. Evaluation of Air Pollution Measurements in Urban Environment Considering Traffic Intensity.
Atis Zari?š1, J?nis Smirnovs2, Raitis L?cis3, 1–2 Riga Technical University, 3 SIA Inženierb?ve, doi: 10.2478/cons-2014-0005, 2014/15
2. A Machine Learning Approach for Air Quality Prediction: Model Regularization and Optimization
Dixian Zhu 1,*, Changjie Cai 2, Tianbao Yang 1 and Xun Zhou, 28 December 2017; Accepted: 19 February 2018; Published: 24 February 2018.
3. Traffic management for smart cities
Andreas Allström, Jaume Barceló, Joakim Ekström, Ellen Grumert, David Gundlegård and Clas Rydergren, Part of: Designing, developing, and facilitating smart cities: urban design to IoT solutions. Part III. Vangelis Angelakis, Elias Tragos, Henrich C. Pöhls, Adam Kapovits and Alessandro Bassi (eds) , 2016, pp. 211-240
4. Traffic Jam Is Freezing Strong Economy and Healthy Environment: A Case Study of Dhaka City.
Malaya Tashbeen Barnamala Senior Lecturer Prime University Mirpur-1, Dhaka Bangladesh, IOSR Journal of Economics and Finance (IOSR-JEF) e-ISSN: 2321-5933, p-ISSN: 2321-5925.Volume 6, Issue 1. Ver. II (Jan.-Feb. 2015), PP 36-40