office (+91) (0) 11 26907-458
fax (+91) (0) 11 26907-405

 

Current Research Projects

Saarthi

Organisation: IIIT-Delhi and Inria-Rocquencourt

Thanks to the increased abundance of mobile phones, the recent field of mobile participatory sensing could be leveraged towards providing a more fine-grain and up-to-date view of a city’s transportation  system.

Thus, in order to address problems like dynamicity (unexpected faults, stoppages, etc.) and unexpected load (number of people using the transportation), etc. in different societal contexts of France and India, Inria Paris-Rocquencourt and IIIT Delhi propose to work together for a project to produce an open source middleware platform called Sarathi, enriched with personalized mobility services for urban travelers and evaluated via real-life demonstrators.

 

Mobile Collaborative Applications

Organisation: IIIT-Delhi

Mobile phones are now nearly ubiquitous all over the world and come with a variety of applications which increase their utility. However, most of these applications are either standalone in nature or at most have cloud-based back-end to support extra functionalities. These applications inherently lack the ability of using the resources of other smartphones present in the vicinity, for shared processing data collection etc., mainly, because of the complexities involved in establishing interactions among multiple devices.

We believe that there is a needs of a  middleware framework which can facilitate the development of applications that take advantage of resources present on mobile devices in the surroundings. We call such applications, mobile collaborative applications. We are currently workign on to define requirements and challenges for building such a  framework that supports development of such applications and propose a flexible architecture supporting real-time communications.

 

Adaptable Interactive Voice Response (IVR) System:

Organisation: IIIT-Delhi

In recent years, the Internet has emerged as a primary source of diffusing information across the world. However, in developing countries like India, the Internet is limited to only urban and rich segment of the society (around 10% according to latest TRAI report). Moreover, illiteracy (especially with respect to English which is the primary language of Internet) coupled with fear of technology and weak economic status compared to their western counterparts will continue to limit use of internet and text-based information dissemination source for some time to come.  Use of Interactive Voice Response (IVR) system is emerging as an effective information dissemination platform for the masses. Unlike Internet, IVR system provides interaction in a natural language interface - often local language - over telephone to make information accessible. Researchers have used menu based IVR in developing regions with increased success. However, due to their rigid design and user interfaces IVR systems are often considered as frustrating to use. Difficulty in accessing information over IVR, has limited its usage for small tasks where user interaction and the amount of information requested by a user are limited.


We are working on so far untapped knowledge of context to make IVR systems self-adaptive. This will allow IVR systems to capture dynamic nature of usage, thus delivering better user experience to the callers. By taking into account both single user intricacies and global pattern of usage and combining knowledge of adaptive systems and context-aware systems, my research will focus on building IVR systems which have ease of access like the internet while adaptable enough to match human-like ability in dealing with caller intricacies. We are also developing tools to make better use of IVR systems.

Mareech: A Tool for Autonomous Testing of IVR systems: Mareech (previously known as MockTel) allows to model different users and test an IVR system from a user's point of interaction, specifically testing IVRs for optimal design and failure conditions. We are now enhancing Mareech to do advanced analysis on the logs of IVR systems and provide help in finding relevant/imporatnt information and their access pattern so that IVR systems can be enhanced

Deployments: Our IVR systems are deployed at several places. Most notably at 181-Women-in-distress Healpline of the Govt. of NCT. We have been managing that helpline since August 2013 and our systems reamins up for 24/7. We handle a large volume of calls and provide analytics to enhance the working of 181. We have also deployed our system at JJ Hospital, Mumbai to allow efficient follow-up study of CVD patients.

 

Utility Monitoring:

Organisation: IIIT-Delhi, India

 Buildings are the largest energy consumers of electricity, accounting for 47% of total national energy use in India. Within buildings lighting and Heating, Ventilation and Air Conditioning (HVAC) systems are the top most contributors to overall electricity consumption. With large number of buildings, even modest improvements in lighting and HVAC control within a building will have a significant aggregate impact at the national scale. Electric meters in most of the buildings do not provide segregated, appliance level and detailed consumption information. Such detailed information, if available, can be useful for the consumers to get detailed picture of their consumption patterns and thereafter optimize on the overall electricity usage. Additionally, such information can be used for smarter actuation of electrical loads (currently lacking in existing buildings).
This research work involves developing a scalable, end-to-end, fine-grained and low-cost utility (electricity and water) management system for buildings. We are working towards a scalable middleware architecture that follows a layered Service Oriented Architecture (SOA) with Representational state transfer (REST) and JavaScript Object Notation (JSON) based generic request and response method across the layers. Sensing layer at the bottom is responsible for interconnecting heterogeneous sensor nodes, modalities and network. The middleware will provide a set of generic core services and few adjunct services such as access control, data access methods, and profile management. Our aim is to develop a novel, low cost systems that can contribute to reduction in energy consumption in the buildings.

 

Previous Projects:

 

Alternative Localization-based applications:

Organisation: IIIT-Delhi, India

It is very clear that location is the most important contextual factor for a mobile application. GPS has become the primary mean of accessing location on smart-phones in developed countries, however, the cost of GPS has prohibited its use in developing countries. In India, most of the phones do not have GPS. Besides being costly, GPS also consumes high energy. Constant use of GPS results in depletion of phone battery in couple of hours.
We proposed a novel Cell Broadcast Service (CBS) message based system to get location. while keeping energy consumption minimum. In future work, I would like to work on improving the accuracy provided by our method as well as development of smart mobile applications to allow use of our approach on widespread basis.

CVDMagic:

Organisation: IIIT-Delhi, India

Cardio-Vascular Diseases (CVD) are one of the major healthcare problems across the world, causing deaths for nearly 17 million people every year. With more than 80% of all CVD cases occurring in developing countries, it is a big challenge that needs immediate attention. Specifically for India, World Health Organization (WHO) and many other organizations have predicted rapid growth of CVD patients in near future. It is known that CVD can be prevented or deferred, if detected in its earlier stages and by subsequently adapting to appropriate preventive methods. Cost and availability of lab equipments - for early diagnosis of CVD - act as deterrents in controlling the spread of CVD cases in India, particularly in the rural parts. Non-laboratory based methods overcome the factor of cost while mobile technology provides the availability to allow for approaches that can detect CVD risk early even in the remotest part of the country.
In this work, We worked on to develop CVDMagic - a mobile phone based study for CVD risk detection. Our study, a mixed-method approach, uses two non-laboratory based approaches (including one proposed by WHO) together with inputs from local doctors corresponding to Indian context. We also present analysis from initial survey (of 169 people) from a pilot deployment of CVDMagic . The preliminary analysis suggests that mobile-based approaches can be used for efficiently collecting required data leading to accurate, low-cost, non-laboratory based early detection of CVD risk in Indian context.

 iBICOOP:

Organization: Inria-Rocquencourt, France

Description: Current day users own multiple devices, e.g. PC, Phone, PDA etc. and frequently collaborate with colleagues or friends which includes exchanging, accessing, and sharing of data in multi-user and multi-device scenarios. Sometimes collaboration can even occur with people who are not trusted or known a priori to user thus demanding enhanced privacy and security requirements. In iBICOOP, we created a middleware to minimize user effort in collaborating with other users and also providing seamless access to user data across different networks and devices. The middleware helps users overcome the device or network specific limitations thus facilitating user collaboration and making seamless access to data possible. My work was concentrated towards developing privacy and security, replication, data sharing, user and data management modules.

iBICOOP now forms core of a new startup - Ambientic.

Environment: Java, Symbian, Windows Mobile, Web Services, Android.

 Immersive Video

Organization: Newcastle University, U.K.

Description: Developers of ubiquitous computing applications have to rely on low-fidelity prototyping measures to develop the applications because often the existing infrastructure is not available at the time of development. Similarly evaluation of such applications is difficult and time-consuming because their testing is to be done in real world. We designed and developed a high-fidelity prototyping tool to assist development and evaluation of ubiquitous computing applications. We use videos of real environment, taken with a wide view angle (typically between 140-360 degrees) using multiple cameras, and augment them with real sensor values, such as GPS, Bluetooth, or IR to create a realistic simulation of environment. Sensor abstraction layer hide the application from simulation environment and application behaves the same way it would in real world while the videos provide a good visual and aural impression. We used immersive video to prototype two real-world applications - developed for Ask-IT project and Future Traveler Information System project.

Environment: C++, C#, Windows Mobile 5.0, RFID reader, IR beacons.

 NERVE (NorthEast Regional Visualisation Environment)

Organizations: Newcastle University, IMASS, 1Spatial, Ordnance Survey.

Description: Today, different regional organizations, e.g. public works, water, electricity board, have their own databases which they use for operation purposes. However, often cooperation is required to achieve their aims, for example a street work needs to have information about electricity cables that pass underneath. In NERVE, we aim to provide a platform to share data of participating organizations and ability to visualize it in real-time in 2D or 3D on PC as well as mobile devices. The data can come from water, gas, or electricity companies as well as services like fire, police and emergency medical services. I developed a 3D visualization client for mobile devices to help visualize the data in 3D. The client fetches data and other real-time feeds, such as GPS, using web service from server and presents it to user. At the same time, user could receive notifications and updates on his/her wearable display. In future, NERVE can be also be integrated with sensor network deployed at the site and can provide real-time information about different sensor feed such as hazardous gases, high temperature etc. More information can be found at: Part1, Part2.


Environment: C#, Java, DirectX, Visual Studio 2005, Windows Mobile 5.0, XML, .Net Web services.

 Future Traveler Information System

Organization: Newcastle University.

Description: Future Traveler Information System (FTIS) was in collaboration with Transport Operations Research Group of Newcastle University. The aim of FTIS is to provide users with real-time information about public transports on their mobile devices to encourage them to use public transportation rather than private vehicles to help reduce congestion/pollution on roads. I developed a mobile prototype application to display real-time information about public transports (timing and details of bus and metro services) with maps of current location of user and suggesting routes to destination. In future, FTIS can also integrate with transport sensor network and can provide real-time information for authorities about pollution etc. A striped-down version of FTIS was also developed for wearable displays and both of the approaches were compared. We also used FTIS as a test case for “Immersive Video” and conducted evaluation studies in both real and prototype environment.

Environment: C#, Visual Studio 2005, Windows Mobile 5.0.

 Ask-IT project

Organization: Newcastle University.

Description: I developed a prototype of context-aware museum mobile guide for Ask-IT project which has an aim of providing navigational help to mobility-impaired people. The museum mobile guide, running on PDA, used to detect current location of user using GPS and other proximity technologies such as Bluetooth or Infrared sensing and suggest easiest route to a museum and help user navigate in the museum and download information about the artifacts. Users also received context-aware notifications such as special offers from a near-by restaurant during lunch time. Apart from displaying map for navigation, we also experimented with providing user panoramic images (360 degree) of the surroundings from his current location.

Environment: C#, Visual Studio 2005, Windows Mobile 5.0.

Multimedia/video Software Integration and Testing

Organization: Nokia, Tokyo, Japan

Description: Objective methods are needed to evaluate quality of video films obtained by mobile phone cameras. Such methods should help in evaluating camera characteristics, such as sharpness, noise, colour saturation etc., without human intervention and should be combined to give a single metric of quality to facilitate easy and rapid evaluation of overall video quality. I used commercial software, Imatest and Genista, to evaluate various characteristics of video quality and combined them into a single metric which helped us comparing the quality of video from different mobile phones. I also developed a part of H.263 parser to compute frame drops.

Environment: C++, Visual Studio 2005, Windows XP.

 Mobile Adaptive Call (MAC): An adaptive software for computer assisting language learning

Organization: Portsmouth University, Portsmouth, UK

Description: Mobile learning has two particular advantages over class-room learning: availability of application (i.e. pervasiveness) and scope of adaptation according to the learner’s needs. MAC is aimed to help Japanese-English speakers in perceptually distinguishing the non-native /r/ vs. /l/ English phonemic contrast with a view to improving their own English pronunciation. MAC is implemented in Java (J2ME). This allows the application to be used anywhere and anytime, on devices that learners would commonly already own. The MAC software adapts according to the learner’s responses and presents to the learner a contrast of the type on which they will most need further practice on.

Environment: J2ME, Eclipse, Sun WTK, Netbeans, Nokia series 60 SDK, Php.

 Allocator

Organization: Veritas Software (now Symantec Corporation), Pune, India

Description: Veritas Volume Manager requires special mechanisms to store data in RAID environment to conserve memory space. My work was related to developing a persistent mechanism to keep track of ongoing requests to a persistent storage while maintaining atomicity of operations.

Environment: C++, Windows 2000.