Vinayak Naik

SMSAssassin: crowdsourcing driven mobile-based system for SMS spam filtering

Due to increase in use of Short Message Service (SMS) over mobile phones in developing countries, there has been a burst of spam SMSes. Content-based machine learning approaches were effective in filtering email spams. Researchers have used topical and stylistic features of the SMS to classify spam and ham. SMS spam filtering can be largely influenced by the presence of regional words, abbreviations and idioms. We have tested the feasibility of applying Bayesian learning and Support Vector Machine(SVM) based machine learning techniques which were reported to be most effective in email spam filtering on a India centric dataset. In our ongoing research, as an exploratory step, we have developed a mobile-based system SMSAssassin that can filter SMS spam messages based on bayesian learning and sender blacklisting mechanism. Since the spam SMS keywords and patterns keep on changing, SMSAssassin uses crowd sourcing to keep itself updated. Using a dataset that we are collecting from users in the real-world, we evaluated our approaches and found some interesting results.

Low Energy and Sufficiently Accurate Localization for Non-smartphones

Location-aware mobile applications are steadily gaining popularity across the world. However, lack of Global Positioning System (GPS) and absence of Wi-Fi infrastructure prevent users with non-Smart phones (majority of population in developing countries) from using location-aware applications as their phones do not have access to their current location. Existing GSM based approaches such as Cell ID-based works on non-Smart phones but they require access to a comprehensive database of Cell IDs. Such a database either does not exist or is very limited in developing countries. In this paper, we propose a novel GSM-based approach of using Cell Broadcast Service (CBS) messages for getting current location on the phone. Proposed approach does not depend on a comprehensive database and can run on programmable low end phones. We demonstrate the effectiveness of our approach on data collected in New Delhi, India across two different operators and propose two space-time history based algorithms to improve upon the localization accuracy of our baseline CBS approach. The proposed algorithms provide up to 35% improvement in accuracy over the baseline method. Further, we compare accuracy of our CBS-based approach with that of Cell ID-based approach and also, present a multimodal approach that uses combination of both CBS and Cell ID (wherever available) to improve the localization accuracy.

Challenges and novelties while using mobile phones as ICT devices for Indian masses: short paper

Mobile phones have emerged as truly pervasive and affordable Information and Communication Technology (ICT) platform in the last decade. Large penetration of cellular networks and availability of advanced hardware platforms have inspired multiple innovative research opportunities in mobile computing domain. However, most of the research challenges have focused on typical scenarios existing in the developed economies. In this paper, we present research challenges and novelties in mobile computing domain that take account for differences between developing in particular India and developed economies. Our research is based on commonly available mobile platforms, communication cost, differences in user behavior and acceptable societal norms, among others.

Human sensors: Case-study of open-ended community sensing in developing regions

With the growing number of cities and population, continuous monitoring of citys infrastructure and automated collection of day-to-day events (such as traffic jam) is essential and can help in improving life style of citizens. It is extremely costly and ineffective to install hardware sensors to sense these events in developing regions. Due to advent of smartphones, citizens can play role of sensors and actively participate in collection of the events which can be shared with others for information or can be used in decisions which affects city development. In this paper, we describe an architecture of crowdsensing testbed for capturing and processing events affecting citizens in cities in India. One of the design principle of our testbed is that it encourages users to do an open-ended sensing under five broad categories: Civic complaints, traffic, neighbourhood issues, emergency and others. As part of testbed, we allow events submissions from different submission modes i.e. mobile application, SMSes and web. Our mobile application exploits different sensing interfaces provided by todays smartphones to add contextual data with event reports such as images, audio, fine-grained location etc. Proposed testbed is used by university students across India to report event happening around them. Finally, we describe the data collected and uncover some of challenges and opportunities which may help future designs of crowdsensing based systems.

Unity: Collaborative downloading content using co-located socially connected peers

Large proliferation of mobile phone applications result in extensive use of data intensive services such as multimedia download and social network communication. With limited penetration of 3G/4G networks in developing countries, it is common to use low bandwidth 2G services for data communication, resulting in larger download time and correspondingly high power consumption. In this paper, we present a system architecture, Unity, that enables collaborative downloading across co-located peers. Unity uses short range radio interfaces such as Bluetooth/WiFi for local coordination, while the actual content is downloaded using a cellular connection. Unity is designed to support mobile phones with diverse capabilities. End-to-end implementation and evaluation of Unity on Android based phones, with varying workload sizes and number of peers, show that Unity can result in multifold increase in download rate for the co-located peers. We also describe architecture of cloud-based Unity which uses principles of mobility prediction, social interactions, and opportunistic networking to make collaboration more pervasive and useful.

CVDMagic: a mobile based study for CVD risk detection in rural 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 paper we present 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.

Characterizing mobility patterns of people in developing countries using their mobile phone data

Human mobility patterns give insights into how people travel in their day-to-day lives. With availability of cellular data, either at large-scale but with low location accuracy or at small-scale but with high location accuracy, studying mobility patterns is now possible. An example of former dataset is CDRs (Call Detail Records) and that of latter is GSM/WiFi/GPS traces collected from mobile phones. So far the studies have been focussed on data collected in developed countries. In this paper, we make an attempt in finding and analyzing mobility patterns of people in developing countries using both the categories of data. We use publicly available CDRs data and we collect our own data for capturing fine-grained location. Ours is the first dataset of its kind that is publicly available. We analyze this data to find movement as well as place visiting patterns, compare our findings with existing studies, and discuss their implications. For example, urban people in developing countries travel farther distances in their day to day life as compared to people living in non-urban areas. Also, distance travelled by urban people in developing countries is as much as six times lower compared to developed countries.

Improving the efficiency of healthcare delivery system in underdeveloped rural areas

A low doctor-to-patient ratio in rural areas of under-developed regions results in an inefficient and expensive delivery of healthcare. Information and Communication Technology (ICT) could play an important role in improving the efficiency and making healthcare more affordable. In this position paper, we present an architectural framework to use ICT (specifically mobile technology) for efficient delivery of healthcare to masses. The proposed framework is (1) comprehensive to cover majority of critical diseases (2) sound from medical science point of view, (3) interfacable to Electronic Medical Record (EMR) system, and (4) self-learning in order automatically diagnose and predict future outbreaks. We present preliminary data and share social experiences from a pilot conducted for preliminary detection of Cardio Vascular Disease (CVD) in rural areas of Punjab, India. We also propose an outline for future study for diagnosing Leptospirosis in rural areas of Gujarat, India.

Alternative localization approach for mobile phones without GPS

Location is a primary indicator of context and forms the core basis of several context-aware applications. Most common way of getting location information is to use specialized hardware like GPS. However, GPS is expensive and is available only on high-end phones restricting its use to a smaller population in developing countries. Further, GPS also consumes a lot of battery power during its operation, thereby making it infeasible to run for longer durations with limited mobile phone battery. An alternative to GPS-based localization is GSM-based localization that is more suitable for developing countries due to much lower power consumption and ability to run even on low-end phones. Currently available, network-operator independent, GSM-based solutions require building perceptual map of cell towers in a city using war-driving. In this paper, we present a novel low cost GSM-based solution based on Cell Broadcast (CBS) Messages. Location accuracy in our approach does not depend on building extensive cell ID database, typically built using war-driving. We present empirical studies (performed in the sub-city of Dwarka, New Delhi, India) comparing location accuracy of our approach with other GSM-based localization scheme that uses one of the most extensive open source database of cell IDs. We also compare power consumption of our proposed solution with GPS-based localization leading to energy-accuracy tradeoff that can be further exploited for a hybrid solution.

BluePark: tracking parking and un-parking events in indoor garages

Finding a parking spot in a busy indoor parking lot is a daunting task. Retracing a parked vehicle can be equally frustrating. We present BluePark, a collaborative sensing mechanism using smartphone sensors to solve these problems in real-time, without any input from user. We propose a novel technique of combining accelerometer and WiFi data to detect and localize parking and un-parking events in indoor parking lot. We validate our approach at the basement parking of a popular shopping mall. The proposed method outperforms Google Activity Recognition API by 20% in detecting drive state in indoor parking lot. Our experiments show 100% precision and recall for parking and un-parking detection events at low accelerometer sampling rate of 15Hz, irrespective of phone?s position. It has a low detection latency of 20s with probability of 0.9 and good location accuracy of 10m.