Research

 Intelligent Data Analytics Platform for a Metro Rail Transport System

Collaborating Institutions

3in1

Project Title: Intelligent CPS Data Analytics Platform for a Metro Rail Transport System

Priority area and sub-area : Computer Science and Engineering, CPS, Big Data Analytics, Transport

Duration: 36 months (Starting from December 2018)

Funding Agency: Department of Science and Technology, India

Total Fund Sanctioned : 24,96,000 Rupees

Principal Investigator : Fr Dr. Jaison Paul Mulerikkal CMI

Co-PIs: Dr Vinith Rajatlal (JEC), Dr Sminu Izudheen (RSET), Mr Binu A (RSET)

Research Assistants: Mr Joish George (RSET), Ms Sona CP (RSET)

Junior Research Fellows: Mr Sajan Raj & Ms Deepa Merlin Dixon

Project Summary

Through this project, we intend to develop an intelligent data analytical platform for a metro rail transport system integrating its various CPS and other data sources. We keep Kochi Metro Rail Limited (KMRL) as a specific use case for this project. This platform will be a middleware with input (data) APIs and output (information) APIs which is built on the top of an HDFS based software stack. The middleware will also implement a set of analytical solutions which are again exposed to the outside world through relevant APIs. Developers can leverage on these input and output APIs to develop internal or external applications. The project will also implement few Proof of Concept (POC) applications to prove the viability of the above middleware with its APIs. This middleware will be generic in nature so that it could be replicated in other Metro Rail systems with minimum modifications and customization.

Data is the new fuel. Kochi Metro was inaugurated on 17th June 2017 and there are four assured data streams from its IP enabled CPS modules. They are (1) Automated Fare Collection (AFC) system, (2) Trip Planner mobile application (Kochi 1 App), (3) CCTV cameras and (4) Automated Vehicle Location data. Moreover, there are two other external data streams which will complement the above data streams. They are (5) vehicle parking data associated with metro stations and (6) GPS data from the feeder city bus network, both are available to KMRL and thus made available to this project. This data, some of which are structured and the other unstructured, will be put together to create real-time (and near-real-time) information and valuable legacy insights to improve customer experience, customer trip planning, trip scheduling, ridership forecasting, commuter behaviour analysis, security and administration of a metro rail system. Details of these use cases are explained in the main body of the proposal.

KMRL Data Flow

The project proposes to develop a CPS data middleware built on the top of customized HDFS based software stack. The very basis of this middleware will be Hadoop Distributed File System (HDFS), which is an industry standard for big data platforms. Data will be stored in HDFS and will be analyzed using MapReduce programs. In order to process streaming and real-time data, the project proposes a combination of Spark, Flume and Kafka big data tools. Machine learning and pattern matching algorithms will be run on the top of this middleware using Google TensorFlow, OpenCV and Spark MLlib.

Picture2

The project will design a rich set of input APIs to PUT various stream of data into the middleware and output APIs to GET outputs from the middleware. Actionable insights and reports will be generated to meet project objectives, viz, improving customer experience, customer trip planning, trip scheduling, ridership forecasting, commuter behaviour analysis, security and administration of a metro rail transportation system.

 

Jyothi Engineering College (JEC) (www.jecc.ac.in) is a NAAC accredited institution with 4 NBA accredited engineering program, including B Tech Computer Science program. Rajagiri School of Engineering and Technology (RSET) (www.rajagiritech.ac.in) – a NAAC A grade accredited institution with 5 NBA accredited B Tech programs including CS and IT has signed a MoU with KMRL to do this research. RSET has a High Performance Computing (HPC) cluster with 7 TeraFLOPS capability at Sunya Labs and has also secured technical assistance from a Barcelona based consulting firm – HPCNow – to run this facility successfully.

Acknowledgement

We gratefully acknowledge that this project is financially supported by the Interdisciplinary Cyber Physical Systems Division of Department of Science and Technology for a total project amount of Rs. 24,96,000 under the order number: DST/ICPS/CPS-Individual/2018/1091(G).

Contact Form

Please feel free to contact, if you would like to contribute to this project or to know more about it using the following contact form:

48362482_10158690535509852_4690628857357664256_o

Figure: Joint KMRL, RSET and JEC meeting at Rajagiri School of Engineering & Technology, Kochi.

 

 

Comments are closed.