On-Going:
Learning Based Control : An Application to Transportation Networks
We develop deep reinforcement learning based control strategies for collaborative multi-robot networks. We then apply them on local transportation networks of Dhaka.
Memes in Engineering Education
We explore the effective use of memes as a teaching and learning tool in tertiary level engineering education.
Nonlinear Control of Networked UAV Systems
We have developed robust control techniques for quadcopter systems and are now exploring the robustness and scalability issues in multi-UAV setups, with heterogeneous agents.
Past Projects:
Electrical Systems and Electronics Curriculum for Female Garments Workers of Bangladesh
We are developing visual literacy based technical training materials for female garments workers of Bangladesh as a was to re-skill and provide an alternate opportunity of employment in the near future for this hardworking backbone community of our economy.
Open-Source Course Management Systems
We work towards finding effective strategies to use free social media platforms as fully stand-alone course management systems for undergraduate engineering courses
Female Students in STEM in Bangladesh
We explored the situation of female undergraduate students studying in Computer and Electrical Engineering programs in urban universities of Bangladesh. We looked into why they seemed to have been doing great in theory classes, yet are less stellar in laboratory setting and national programming competitions and other hands-on extracurricular activities as well as why are they less visible right after graduation.
Adaptive Control of Underactuated Surface Vehicles
We designed control protocols for finite-time path-following task for an underactuated surface vehicle, with consideration to parametric uncertainties, unknown disturbances and involving input-control saturation. This project was led by Dr. Jawhar Ghommam.
Area-Coverage Algorithms for Multirobot Systems
We developed flocking-based algorithms for scalable network of mobile robots that can spread over and cover any convex area with unknown boundaries. The algorithms also allowed the robots to efficiently avoid obstacles of arbitrary shapes and minimize overlapping of sensed regions.
Driving Experience on the Roads of Dhaka
We explored the stressful situation faced by Dhaka’s drivers given the unique nature of the roads and the multimodal vehicle fabric. We used simple wearable electronics to measure various aspects of drivers as they travelled through the roads at different times of the day.