ResearchGate Link: https://www.researchgate.net/profile/Swapnil_Sayan_Saha
Google Scholar Link: https://scholar.google.com/citations?user=hR5Yyx8AAAAJ
Google Scholar Link: https://scholar.google.com/citations?user=hR5Yyx8AAAAJ
Visit ResearchGate Link or Scholar Link for complete manuscripts (if available).
- S. S. Saha, S. S. Sandha and M. Srivastava, "Deep Convolutional Bidirectional LSTM for Complex Activity Recognition with Missing Data", to appear in Activity and Behavior Computing - Smart Innovations, Systems and Technologies, Springer (2020).
- S. S. Saha, S. Chakma and S. Vyawahare, "Inertial Reflex Model of Nightmare Actigraphy using Double Empirical Margins for the Blinded," 2019 IEEE International WIE Conference on Electrical and Computer Engineering (WIECON-ECE), Bangalore, India, 2019, pp. 1-3. DOI: 10.1109/WIECON-ECE48653.2019.9019977
- S. S. Saha, S. Rahman, Z. R. R. Haque, T. Hossain, S. Inoue and M. A. R. Ahad. "Position Independent Activity Recognition Using Shallow Neural Architecture and Empirical Modeling", In Adjunct Proceedings of the 2019 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2019 ACM International Symposium on Wearable Computers (UbiComp/ISWC '19 Adjunct). ACM, New York, NY, USA, 808-813. DOI: https://doi.org/10.1145/3341162.3345572
- S. S. Saha, S. Rahman, M. U. Ahmed and S. K. Aditya, "Ensuring Cybersecure Telemetry and Telecommand in Small Satellites: Recent Trends and Empirical Propositions," in IEEE Aerospace and Electronic Systems Magazine, vol. 34, no. 8, pp. 34-49, 1 Aug. 2019. DOI: 10.1109/MAES.2019.2927852
- S. S. Saha, S. Rahman, M. J. Rasna, T. Hossain, S. Inoue and M. A. R. Ahad. "Supervised and Neural Classifiers for Locomotion Analysis". In Proceedings of the 2018 ACM International Joint Conference and 2018 International Symposium on Pervasive and Ubiquitous Computing and Wearable Computers (UbiComp '18). ACM, New York, NY, USA, 1563-1570. DOI: 10.1145/3267305.3267524
- S. S. Saha, S. Rahman, M. J. Rasna, T. B. Zahid, A. K. M. M. Islam and M. A. R. Ahad, "Feature Extraction, Performance Analysis and System Design Using the DU Mobility Dataset," in IEEE Access, vol. 6, pp. 44776-44786, 2018. DOI: 10.1109/ACCESS.2018.2865093
- S. S. Saha, S. Rahman, M. J. Rasna, A. K. M. Mahfuzul Islam and M. A. Rahman Ahad, "DU-MD: An Open-Source Human Action Dataset for Ubiquitous Wearable Sensors," 2018 Joint 7th International Conference on Informatics, Electronics & Vision (ICIEV) and 2018 2nd International Conference on Imaging, Vision & Pattern Recognition (icIVPR), Kitakyushu, Japan, 2018, pp. 567-572. DOI: 10.1109/ICIEV.2018.8641051 Awarded the Best Paper Award Candidate
- Saha, Swapnil Sayan, and Ariful Islam. "Shape Memory Material- Concepts, Recent Trends and Future Directions." 7th Student Conference on Informatics, Electronics & Vision, Kitakyushu, Japan. 2018.
- S. S. Saha, M. S. Siraj and W. B. Habib, "FoodAlytics: A Formalin Detection System Incorporating a Supervised Learning Approach," 2017 IEEE Region 10 Humanitarian Technology Conference (R10-HTC), Dhaka, 2017, pp. 26-29. DOI: 10.1109/R10-HTC.2017.8288898
- Mamun, Abdullah Al and Saha, Swapnil Sayan. "Design and Implementation of a Sensorless Microcontroller Based Automated Residential Water Pump Operating System." Proceedings of 2nd Student Conference on Science and Engineering. IEEESBDU, DUSS, 2017. Awarded the Best Paper Award in Robotics and Electronics Category.
- Saha, Swapnil Sayan; Islam, Shekh Md. Mahmudul and Mashsharat, Anindita. "Design and Operation of an Economical and Infallible Domestic Gas Stove." Proceedings of 2nd Student Conference on Science and Engineering. IEEESBDU, DUSS, 2017.
- Zaman, Hasan U., et al. "Design, control & performance analysis of Baby Bot." 16th International Conference on Electronics, Information, and Communication (ICEIC). 2017.
- Habib, Walid Bin, et al. "Disaster Helpline - An Android Application for Tracking and Warning the User on Surrounding Natural Calamities and Man-Made Disasters." Students' Innovation Pavilion Special Track , Humanitarian Technology Conference (R10-HTC), 2016 IEEE Region 10. IEEE, 2016.
- S. S. Saha, S. M. M. Islam and A. Mashsharat, "Microcontroller Based Earthquake Detection System for Spontaneous Cut-off of Domestic Utility Lines for Safety Measures," 2016 9th International Conference on Electrical and Computer Engineering (ICECE), Dhaka, 2016, pp. 139-142. DOI: 10.1109/ICECE.2016.7853875
- H. U. Zaman et al., "Design, Control & Performance Analysis of Muktibot," 2016 IEEE 7th Annual Information Technology, Electronics and Mobile Communication Conference (IEMCON), Vancouver, BC, 2016, pp. 1-4. DOI: 10.1109/IEMCON.2016.7746321
- Saha, Swapnil Sayan, and Shekh Md Mahmudul Islam. "Microcontroller Based Automated Domestic Security System for Bangladesh Perspective." 5th Student Conference on Informatics, Electronics & Vision, University of Dhaka, Dhaka. 2016.
- Small Satellites, Big Future: Assertive Outlook
- Higher Studies Abroad (MS/PhD) - From Zero to Infinity
- The Art of Pitching
- Embedded System Design and Analysis for Small Satellite Applications
- Progress in Bangladesh on SSS Project (2018)
- Cracking the GRE
- Arduino in Space: Breaking the Frontiers
- Workshop on Basic Robotics and Automation
- Progress in Bangladesh on SSS Project (2017)
- Mission Analysis and System Design of SSS-1
- Effacing the Line between Bangladeshi Public Universities and the British Curricula
- SENIOR YEAR DISSERTATION PROJECT - Addressing Security Vulnerabilities of Cyberphysical Remote-Sensing Small Satellite Systems (2018): With the advent of a large number of small satellites in orbit performing critical missions, securing ground/space infrastructure against cyberattacks has become a blazing issue amongst operators and policy makers. In this dissertation, we explore the performance of selected encryption schemes in the cyberphysical space of small satellites, using throughput, memory footprint, security and energy consumption as key performance metrics. We propose an open-source cybersecurity analysis platform in context of small satellites and have found out that the Advanced Encryption Standard (128 bits) is most suitable for use in small satellites, based on experiments on microcomputers and our proposed platform. We have also investigated possibilities of telemetry compression, proposing a new telemetry/telecommand packet format to incorporate compression and encryption, and have found simple Huffman Coding to be suitable. In order to improve the resilience of ground infrastructure against cyberattacks, we have proposed a redundant array of ground stations to be architected.
- A Ground Station for Everyone: Design and Development of an Open-Source Portable Ground Station for Space Education in Developing Countries (2018): With a growing number of small satellites being launched every year, the humongous amount of data received from these satellites have become critical in catastrophe management as well as discovering new horizons in the big data and aerospace community. Unfortunately, the financial and infrastructural investments, as well as convoluted training exercises associated with traditional ground station architectures makes it impossible for enthusiastic students, professionals and engineers in underdeveloped countries to own or operate one, hence depriving them of a rudimentary academic research and educational tool. Thus, we are proposing a modular, portable, inexpensive, robust and customizable ground station architecture.
- SENIOR YEAR DISSERTATION PROJECT - Feature Extraction and Semi-Supervised Learning in a Wrist Mounted Motion Detection System for Elderly Citizens (2018): Human Action Recognition (HAR) in healthcare amongst senior citizens focuses on remote surveillance, healthcare monitoring and fall detection. As the population of senior citizens continue to increase, a rising cause of concern is in ensuring proper remote healthcare monitoring services that encompasses all of the coveted features. Impelled by global statistics regarding senior citizens, as well case studies from Japan and PepsiCo, this dissertation proposes a wrist-mounted HAR/fall detection system geared towards senior citizens that parallels the commercial landscape and user-acceptability. By performing comparative studies on 46 existing manuscripts/articles on HAR, as well as existing commercial ventures, we have established that the wearable approach, in particular, wrist-mounted sensors embracing machine learning models and accelerometry for HAR is most favorable when qualitative characteristics, parameter complexities and market projections are considered. An open-source dataset, named The University of Dhaka (DU) Mobility Dataset (MD) was built using 50 subjects with 10 ADL (7 basic ADL and 3 falls), amounting to 5000 training sets using a single wrist-mounted wearable sensor designed using the newly developed University of Tokyo (UTokyo) Internet of Things (IoT) platform. Some existing public databases have been compared extensively. 39 features of the DU-MD have been extracted and selected for training machine learning models. Existing machine learning models used in HAR, in particular, Support Vector Machines (SVM), Ensem- ble of Classifiers (EoC) and Subspace K-Nearest Neighbours (S-KNN) have been applied to our dataset for activity and fall classification, with outcomes being compared to benchmark and similar datasets. With a HAR classification accuracy of 93%, fall detection accuracy of 97% and fall classification of 68.3%, quantitative performance metrics have either approached or out- performed other datasets, making this dataset suitable for application in hardware-independent healthcare monitoring systems. We have also proposed our own version of HAR algorithm based on performance metrics, namely One-vs-one cubic SVM (1-1 CSVM) and bootstrap aggregated decision tree (Eoc 45). We have benchmarked these algorithms on the Sussex-Huawei Loco- motion (SHL) dataset. 91.7% accuracy was obtained using EoC 45 whilst 91% accuracy was obtained using 1-1 QSVM with 150 features. The algorithms were benchmarked against other established algorithms and our algorithms either exceeded or closely approached the accuracy and F score for other algorithms. With 226 features, accuracies of 91% and 92% were obtained using EoC 45 and 1-1 QSVM respectively on the SHL dataset. Best supervised classification accuracy was obtained using 1-1 CSVM, with an accuracy of 92.8%. Prospects of semi-supervised learning were tested by using a semi-supervised artificial neural network was applied on the SHL dataset. A regression coecient of 0.936, 0.8818 and 0.9005 was obtained for training, testing and validation phases, proving promising prospects of semi-supervised learning. Finally, we provided technical implementation details and business models of our proposed system in both external and internal environments considering both standalone client and multi-user environment. A real-time IoT cloud mesh architecture embracing the Microsoft Azure platform was proposed for multi-user environments.
- Project Shadhinota (2017): S-Cube is the first nanosatellite architecture of the University of Dhaka, and the cheapest one to date built in Bangladesh, jointly developed by the University of Dhaka, Bangladesh University of Engineering and Technology and Khulna University of Engineering and Technology. The satellite is intended for disaster management and scientific experimentation. Goals: 1. Establish a technological way to partially fulfil Strategic Goal 6.0 of the Disaster Management Action Matrix (2010-2015) of the Government of the People’s Republic of Bangladesh, by strengthening the early hazard warning systems present in Bangladesh through the establishment of a Low-Earth Orbit 1U CubeSat, which will monitor the weather and climatic conditions of Bangladesh. 2. Share the data obtained from the CubeSat payload sensors and cameras through universal, reliable and secure information sharing services to necessary clients (universities, agricultural personnel, residents of disaster prone areas) to allow them perform scientific research, agricultural practices, disaster preparedness and so on, thus, contribute to the socioeconomic development of the country and assist fulfilling the mission of the ICT Division, Government of the People’s Republic of Bangladesh.
- FoodAlytics - An Android Application Based Formalin Detection System for Food Items (2017): FoodAlytics is the world’s first truly portable and replicable formalin testing station aimed at the general mass. The package consists of a light, stylish and eye-catching portable formalin sensing kit that can be clipped to any smartphone like those portable smartphone camera lenses. The sensor can detect up to 50 ppm of formalin in food items. The sensor sends real-time data to a microcontroller which then establishes connection with a smartphone via Bluetooth. Our very own android application then performs detailed mathematical analysis for each food item very differently, using the data to plot plots graphs, charts and show whether the concentration of formalin in the particular food item is within safe limits. The android application can also keep track of your daily formalin intake, analytics on your daily diet and food intake, your daily calorie intake and medical history. The application also features reporting to concerned authorities about harmful levels of formalin in food items. The package is the first of its kind, introducing an inexpensive, portable and rapid way to check one’s favorite food item for contaminants in seconds.
- Project MEDSys (2017): Project MEDSys is a completely automated system for use in earthquake-prone areas. The hardware prototype consists of a microcontroller based earthquake detection system that performs four actions:
1. Turn off the electricity supply during earthquakes.
2. Turn off the gas supply permanently during and after earthquakes.
3. Illuminate evacuation paths.
4. Send a warning message to the client’s cell phone.
The software counterpart is a crowd-funded android application that fully prepares the user on how to act before, during and after earthquakes. Both of these systems work in conjunction with one another in order to provide a seamless solution against both absolute and subsidiary earthquake hazards.
- Project P.E.G.A.S.U.S. (Pragmatic Earthquake and Gesture Automated System for Ubiquitous Security) (2016): Project P.E.G.A.S.U.S. provides the ultimate low-cost, feasible solution against intruders and natural disasters. It can detect specifically intruders breaking into properties, as well as detect earthquakes and take some necessary steps to minimize loss of life and property.
- Microcontrollers, A New Way to Sense the World (Bangladesh Perspective) (2015): As our society becomes more and more technologically advanced, there is a conspicuous necessity for everyday routines to become completely or at least, partially autonomous. Digitization has not just halted within offices and industries, but is rapidly curtaining our homes as well. Swift developments in domestic technologies have opened marvelous realms and have made our lives longer, simpler and safer.
One of the most prolific implementations of domestic digitization is the use of micro-controllers to sense the world. A microcontroller is a microcomputer mounted on an integrated circuit that contains a central processing unit, a memory chip and peripherals. Micro-controllers can be programmed to virtually automate or implement any domestic, industrial and office systems, limited only by our imagination. Virtually, all
domestic electronic devices are now equipped with micro-controllers; let it be kitchenware, televisions, stereo systems, remote controlled toys, vacuum cleaners, security systems and much more.
This project intends to showcase how several sophisticated and complicated but indispensable electronic devices can be made cheaply and simply using several different microcontroller platforms from a point of view of a developing country such as Bangladesh. The project includes:
1. A microcontroller based android controlled automated domestic security system
2. A microcontroller based radar.
3. A three in one robot (TV Remote/RC/Autonomous)
4. A two in one robot (Android Controlled/Autonomous)
5. A Line Follower
Cipherduino, an open-source cybersecurity analysis platform for small satellites.