Research

I aspire to build a career in high impact research and am actively seeking PhD positions. My primary research interest is in Control Systems, Biomedical Systems & Control, Robotics, Cyber-Physical Systems, Rehabilitation Engineering and AI. I am profoundly intrigued by the multidisciplinary nature of these areas and their impact on the evolution of technology. Being involved in research for the last 5 years in different capacities has concretely galvanized my trajectory to be fixated in research.

Developing a Low Cost Wearable For Upper Limb Rehabilitation Of Stroke Patients

A GCRF funded project by North South University and Cardiff University. The project is supervised by Dr. Katarzyna Stawarz, Dr. Nova Ahmed and Dr. Valerie Sparkes.

Stroke patients or patients with acquired brain injury (ABI) often left with severe health complications. Sometimes, upper limbs are affected causing weakness and function impairment resulting in difficulty to perform everyday tasks. Rehabilitation regimes can help these patients regain upper limb function. However, in countries like Bangladesh, rehabilitation access can be limited and costly causing patients to turn away from rehabilitation. This causes patients to loose the opportunity to undergo optimal recovery. In this project, we aim to develop a low cost wearable to aid the upper limb rehabilitation process through technology. Ideally, the final prototype of the wearable would be able to recognize exercises for the rehabilitation and give a feedback to the user.

Primarily we built a prototype with the basic assumption that it will recognize hand movements. Afterwards, through interviews of Physiotherapy experts in UK and Bangladesh, we have understood the nature of upper limb rehabilitation, rehabilitation requirement and basic exercises. From the expert's insights, we are upgrading the design of the prototype to meet the requirements and needs. We will conduct interviews of patients and possibly their caregivers and try to understand their perspectives and requirements. From there, we will initiate another iterative design improvement.

Through this project we will be able to help patients and their caregivers and assist in healthcare using technology.

Alor Akash : Intersection of Gender and Technology in Bangladesh

A Bill & Melinda Gates Foundation funded project by North South University. The project is supervised by Dr. Nova Ahmed.

The research project aims to understand the intersection of technology usage and financial inclusion through a deep understanding of barriers and opportunities for advancing women's financial inclusion, including the experiences and perceptions of low-income women from across various regions and occupations.

Through this project we will be able to identify the gaps between technology and fintech usage among the women and suggest interventions that can ameliorate access and usage.

Learning-based control of Transportation Networks

A CTRG funded project supervised by Dr. Lamia Iftekhar.

This project aims to design control algorithms for multiagent systems by employing deep reinforcement learning techniques. We will consider multiple models for the system, apply deep reinforcement learning independently to these model and obtain optimized parameters for these models. The obtained results shall be numerically verified and the efficacy will be tested on a group of robot networks. The algorithms will be fine tuned and updated from the process.

The devised theory will be applied to local transportation problems focusing on paratransit, shared mobility and congestion modelling and control.

Through this project we will be able to precisely address the unique local transportation challenges.

Sensor Based Human Arm Motion and Gesture Replication

The project started out as part of an undergraduate design course and was funded by CTRG and extended. The project was supervised by Dr. K. M. A. Salam.

The aim of the project was to "capture" the motion and gesture of human arm and replicate the motion in a 3D arm model. The signals that drive the human arm model could be employed to control features of any arbitrary hardware or software applications, e.g. a robotic arm, a game etc.

Using flex sensors and accelerometer a sensor array was developed and mounted on a glove. The sensors were connected to an arduino, the arduino acted as a server and relayed the sensor data to MATLAB. The signals were calibrated and accelerometer data was stabilized using PID controllers. The processed signals were provided as an input to a 3D Arm Model developed in Solidworks. From MATLAB Control System Designer Toolbox the state-space matrices of the system were also obtained.

Under the CTRG fund, the project was extended and with the help of Dr. Mohammad Monir Uddin. We applied Lyapunov State Feedback Stabilization Technique and Riccati based Optimal Control algorithms the to the system state-space matrix and achieved optimal control. We also found that the hyperparameters of the Riccati based Optimal Control Algorithms could be altered to change the pole locations of the system and we were able to fully eliminate oscillations from the system response by shifting the complex conjugate poles to the real axis.

Through this project we were able to translate human arm motion to signals that could be manipulated and applied to any arbitrary application. Additionally, we learned the affect of hyperparameter values of Riccati based stabilization algorithm on the poles of the system.

PLC Based Traffic Control System With Emergency Vehicle Detection And Management

Dhaka city has a very unique traffic congestion problem and like me every inhabitant of this city is affected by it. This recurring transportation problem has been one of my primary provenience of motivation for being involved in research and search for solutions that disperse impact. As an instructor of Programmable Logic Controllers (PLC) and automation, I was inspired to work on a practically implementable solution to the transportation problem of Dhaka while preparing PLC examples for my lab class. I wanted to work on a solution that would reduce traffic congestion and reduce deaths on ambulances due to congestion. I proposed my idea to my enthusiastic peers and Dr. K. M. A. Salam. With their support and contribution the project came to fruition.

A traffic light system for a four-lane junction was considered and a PLC based time controlled traffic light switching system was developed. The system was then modified to clear a lane with emergency vehicle after detection of an emergency via a sensor, e,g, RFID sensors to detect emergency vehicle by RFID tags of the license plates. The state diagram of the system was developed and from there state equations were derived. Based on these state equations, the ladder logic code was developed for an Allen Bradley Micrologix 1200 model PLC in RsLogix 500 software. The outcome of the code were then verified by running simulations in RsLogix and generating Timing Diagrams of the states of the traffic lights and the emergency condition detection in each road.

Through this project we were able to propose a practically implementable solution to improve the transportation congestion situation in Dhaka city and provide an easy passage to the emergency vehicles. The solution is extendable and applicable to any city in the world.

Engineering Education

I regularly collaborate with Dr. Lamia Iftekhar on her Engineering Education research.

A survey on Tumor Growth Modelling And Development Of An Incremental Framework: This project focused on surveying existing literature and arrange a incremental framework for Tumor Growth Models. Th most popular tumor growth models were extracted, the notations were streamlined and a simple discourse of mathematical modelling of tumor growth was developed. Afterwards, a systematic framework was developed by mapping the learning outcomes of a control engineering curriculum to various aspects and parameters of tumor growth models.

Facebook Social Learning Group as a Course Management System (CMS) : This project focused on gauging the effectiveness of Facebook Social Learning Groups as a CMS. Facebbok is popular among the engineering students of Bangladesh based on this fact students can potentially be engaged more actively by employing Facebook Groups as a CMS. We investigate through student survey and literature review if Facebook Social Learning groups would be an effective CMS platform. We found that Facebook performed excellently in 2 of the 4 key features of a CMS, but was poor in the other 2.

A case study of Emergency Remote Teaching amidst Digital Divide due to the onset of COVID-19 global pandemic: With the onset of global COVD-19 pandemic, Bangladesh closed educational institutions from March 2020. Emergency Remote Teaching had to be adapted and we opted to use Facebook Social Learning Groups as a CMS during this period. We tried to evaluate the level of digital divide amidst the students and the effectiveness of Facebook Social Learning Groups as a CMS platform in this situation. We found that there was some level of digital divide amidst the students and Facebook group was effective in this emergency remote teaching situation.