Ferdaus Ahmed Kawsar

e-ESAS: improving quality of life for breast cancer patients in developing countries

In this paper, we present e-ESAS, a mobile phone based remote monitoring tool for patients with palliative care need, carefully designed for developing countries. Most of the current remote monitoring systems are complex, obtrusive and expensive resulting in a system unsuitable to deploy in low-income countries. We here describe evolution and performance of e-ESAS within the contexts of breast cancer patients as these patients require management of pain and other symptoms. Edmonton Symptom Assessment Scale (ESAS) was developed to capture the important parameters where patients themselves report their level of ten symptoms. Our e-ESAS improves the current system by reducing visits by patients to clinics, providing more flexibility to both doctors and patients, improving the quality of data, accommodating doctors to fine tune interventions, and providing a convenient representation of data to doctors. User interface was designed according to feedbacks from users resulting in a UI with better performance. The system is intended to provide a platform for future research as large amount of real data is being accumulated from the deployment. The system demonstrates the feasibility of accessing quality health care through cell phones by rural, poor patients in developing countries. The system enables doctors to serve more patients as it saves time for doctors, requiring less time to view patient information.

e-ESAS: Evolution of a participatory design-based solution for breast cancer (BC) patients in rural Bangladesh

Abstract Healthcare facility is scarce for rural women in the developing world. The situation is worse for patients who are suffering from diseases that require long-term feedback-oriented monitoring such as breast cancer. Lack of motivation to go to the health centers on patients’ side due to sociocultural barriers, financial restrictions and transportation hazards results in inadequate data for proper assessment. Fortunately, mobile phones have penetrated the masses even in rural communities of the developing countries. In this scenario, a mobile phone-based remote symptom monitoring system (RSMS) with inspirational videos can serve the purpose of both patients and doctors. Here, we present the findings of our field study conducted on 39 breast cancer patients in rural Bangladesh. Based on the results of extensive field studies, we have categorized the challenges faced by patients in different phases of the treatment process. As a solution, we have designed, developed and deployed e-ESAS—the first mobile-based RSMS in rural context. Along with the detail need assessment of such a system, we describe the evolution of e-ESAS and the deployment results. We have included the unique and useful design lessons that we learned as e-ESAS evolved through participatory design process. The findings show how e-ESAS addresses several challenges faced by patients and doctors and positively impact their lives.

User Privacy Protection in Pervasive Social Networking Applications Using PCO

Privacy is the most often-cited criticism of context awareness in pervasive environment, and may be the utmost barrier to its enduring success. However, privacy implications associated with pervasive online community-based applications depend on the level of identifiability of the information provided, its possible recipients, and its possible uses. Unfortunately, conventional privacy preservation techniques are not suitable for these types of application. This paper describes our current work in developing a novel privacy sensitive architecture for context obfuscation (PCO) for privacy preservation in pervasive online community-based applications. More specifically, PCO preserves users’ privacy by generalising request parameters as well as the context data provided to the application. To support multiple levels of granularity for the released context data, the obfuscation procedure uses an ontological description that states the granularity of object type instances. We have developed and evaluated a contextual instant messaging application (PCO application) in Android platform that incorporates level-based privacy of the user’s contextual information. We also evaluate our prototype application through user evaluation survey. The PCO architecture can be extended to be used in diverse online community-based applications.

Smartphone based multimodal activity detection system using plantar pressure sensors

There have been numerous efforts to detect human physical activities automatically. Healthcare professionals who want to monitor patients remotely, people who want to know their measure of physical activity objectively, or people who develop context-sensitive systems are interested in such systems. A majority of such systems use accelerometers to collect data from different parts of the body. Recently, some systems have used the accelerometer and gyroscope sensors of smart phones to develop unobtrusive systems. Such systems require users to carry smart phones with them. Such requirement limits the practical usability of these systems because people often place their phones on the table while sitting and women usually carry phones in their purses. We have developed a multimodal system where we used pressure sensor data from shoes along with accelerometers and gyroscope data from smart phones to make a more robust system. In this paper, we present our novel activity detection system along with evaluation briefly.

Remote Monitoring Using Smartphone Based Plantar Pressure Sensors: Unimodal and Multimodal Activity Detection

Automatic activity detection is important for remote monitoring of elderly people or patients, for context-aware applications, or simply to measure one’s activity level. Recent studies have started to use accelerometers of smart phones. Such systems require users to carry smart phones with them which limit the practical usability of these systems as people place their phones in various locations depending on situation, activity, location, culture and gender. We developed a prototype for shoe based activity detection system that uses pressure data of shoe and showed how this can be used for remote monitoring. We also developed a multimodal system where we used pressure sensor data from shoes along with accelerometers and gyroscope data from smart phones to make a robust system. We present the details of our novel activity detection system, its architecture, algorithm and evaluation.

Activity Detection Using Time-Delay Embedding in Multi-modal Sensor System

About two billion people in this world are using smart devices where significant computational power, storage, connectivity, and built-in sensors are carried by them as part of their life style. In health telematics, smart phone based innovative solutions are motivated by rising health care cost in both the developed and developing countries. In this paper, systems and algorithms are developed for remote monitoring of human activities using smart phone devices. For this work, time-delay embedding with expectation-maximization for Gaussian Mixture Model is explored as a way of developing activity detection system. In this system, we have developed lower computational cost algorithm by reducing the number of sensors.

Identifying Phases of Gait and Development of Walking Model from Pressure and Accelerometer Data and It’s Ramifications in Elderly Walking

Locomotion is a feature of all animals. Whereas quadruped are fast and stable, human’s bipedal gait is less stable and less efficient. Human gait analysis is going on for a long time. Such analysis usually used force data applied on the ground during different phases of gait. In this paper, we have analyzed the pressure data collected from pressure sensors placed on shoes along with accelerometer data collected from cell phones during walking activity. We identified different phases of walking activity using the pressure data. We also have developed a biomechanical model of gait based on the pressure and acceleration data.

PryGuard: a secure distributed authentication protocol for pervasive computing environment

Handheld devices have become so commonplace nowadays that they are an integral part of our everyday life. Proliferation of these mobile handheld devices equipped with wide range of capabilities has bolstered widespread popularity of pervasive computing applications. In such applications many devices interact with each other by forming ad hoc wireless networks. The necessity of such unavoidable inter-device dependency along with volatile nature of connectivity and the lack of a fixed infrastructure for authentication and authorization, devices are susceptible and vulnerable to malicious active and passive snoopers. If a device registers a malicious device as its valid neighbor, the security and privacy of entire system might be jeopardized. Such sensitivity to malevolent activity necessitates the need for a robust mechanism to maintain a list of valid devices that will help to prevent malicious devices from authenticating successfully. In this paper, we present the feasibility of using a decentralized protocol in order to prevent malicious devices from participating illicitly into the ad hoc networks.