College of Engineering, Pune

About: College of Engineering, Pune is a based out in . It is known for research contribution in the topics: Computer science & Sliding mode control. The organization has 4264 authors who have published 3492 publications receiving 19371 citations. The organization is also known as: COEP.

Scalable tracking system for public buses using IoT technologies

Abstract: Reliability in public transport is of great importance today. Millions of people travelling by public buses waste a lot of time waiting at bus stops. This paper focuses on presenting a solution to tackle the said problem by harnessing IoT technology stack. If the people travelling get accurate real time location of the buses along with estimate time for arrival at bus stop based on the real time traffic conditions, it will facilitate an overall increase in reliability on the public buses. The solution proposed in this paper involves using the existing internet enabled devices on the bus (like the e-ticketing system) or a simple android tablet to capture the real time location and send to the servers. Accessing this location data from servers will be facilitated by Representational State Transfer (REST) APIs which users can access through android application, SMS or web-portals. The system proposed will have distributed architecture in order to tackle high number of requests from users. Although there are existing solutions which harness the use of Global Positioning System (GPS) for bus tracking, they aren't ready to handle high demand on the backend which will exist in the near future. We have addressed this problem. The primary contribution of this paper is that it shows that a backend based on Message Queue Telemetry Transport (MQTT) instead of the traditionally used Hypertext transfer protocol (HTTP) based REST will be light weight, data efficient and scalable. We have proposed and implemented the backend as well as the front end required for the tracking system and presented the improvements.

Factors influencing decisions on delay claims in construction contracts for Indian scenario

Abstract: Construction industry in India is second largest next to agriculture. In current era of infrastructure development construction projects occupy a key position. In any construction project contract time and cost overrun is a common feature, which gives rise to claims leading to disputes. These disputes if not handled properly tend to consume time and money of all parties to the contract. To resolve the dispute in optimum time, it is essential to understand the root cause of disputes as early as possible. Hence there is a need of analyzing the disputes scientifically. The present study reveals from the study of arbitration awards that the causes for delay claims can be grouped in domains and the probable decisions to the disputes can be traced through the probing questions considered by decision makers. This paper attempts to identify questions related to disputes for Indian scenario through literature, arbitration awards, court cases and discussions with professionals.

Parametric studies of CI engine at various injection strategies using biodiesel blended nanoparticles as fuel

Abstract: This work shows the consequences of nickel oxide nanoparticle doping in NBD25 at various Injection opening pressures. Biodiesel blend NBD25 is a mixture of neem oil methyl ester (25% vol.) and mine...

Effective intrusion detection system using semi-supervised learning

Abstract: Network security is a very important aspect of internet enabled systems in the present world scenario. As the internet keeps developing the number of security attacks as well as their severity has shown a significant increase. Due to intricate chain of computers the opportunities for intrusions and attacks have increased. Therefore it is need of the hour to find the best ways possible to protect our systems. Every day new kind of attacks are being faced by industries. Hence intrusion detection system are playing vital role for computer security. The most effective method used to solve problem of IDS is machine learning. Getting labeled data does not only require more time but it is also expensive. Labeled data along with unlabeled data is used in semi-supervised methods. The rising field of semi-supervised learning offers a assured way for complementary research. In this paper, an effective semi-supervised method to reduce false alarm rate and to improve detection rate for IDS is proposed.