Research Projects

Reinforcement Learning for Robotic Control and Information Sampling

I am currently working on the research titled, "Analysis of Deep Reinforcement Learning Algorithm applied on precision agriculture plume source tracking : A detailed experiment on DQN, DRQN and DTQN".

Advancement of robotic application have been applying to many fields including precision agriculture. In the real world scenario robots might distracted by environment parameters, for example heavy wind, fire or plume outbreak, unwanted obstacle etc. For precision agriculture application, robots need to collect data from environment with limited resource. By keeping all other constraints remain fixed, the motivation for this work to find a plume source on a regular crop field. To encounter this problem we are going to use reinforcement learning algorithm.

Manipulation of Multi-Robot Systems

My current research is going on the path-planning application of mobile and multi-robot systems. One of the application might be applied on precision agriculture in near future.

Path Planning of Mobile Robot

In my AI course project, I experimented with the basic to advanced path planning techniques including A*, RRT*, IRRT*. Our novel idea based on these algorithm is now being updated. We're looking forward to bring some flexible solutions for indoor robot activities with our idea.

Robotic Manipulation and HRI application

Here in picture showing a R&D is going on, to detect the defected bricks in a brick industry and to recycle those defected brick. In this project, I'm using ROS and RViz simulation and later on I will use machine learning algorithm also.

Aamra is one of the most leading and advanced technical solution company in Bangladesh. I'm developing pick & place, packaging application using UR5e 6DOF industrial cobot.

Circuitry to detect Touch Pad response

Processor unit of the system

Electronic Time-tracking device for Swimmers

The focus of this project was to make an alternate solution for time-tracking in the swimming pool, as the existed 'Omega' technology is too costly and difficult to keep maintenance by untrained workers. Inspired by our experience we introduced 'Teensy 3.6' as the main processor of this system. We also provided both wired/wireless feedback response incoming to the software.

Autonomous (Small one) and Semi-autonomous (large one) Robot

The task was to complete a wall-maze by the autonomous bot with help of the semi-autonomous robot. We used Gaussian path-planning algorithm & PID as conditioning learning. Later on, we developed another simulation based solution in ROS.

Autonomous wall-maze solver & Semi autonomous picker bot

The miniature is the autonomous one and the gripper enabled one is the semi- autonomous one. The task inspired to built these robot was to participate in 'International Robotics Challenge (IRC)', organized by Techfest, IIT, Mumbai.

Arena of the Final Event

Smart IoT power meter

IoT enabled smart power meter which be connected with the home device and controlled from the mobile device. We implemented BLE technology to reduce the internal power consumption and usable in home application

GSM & BLE enabled IoT Power meter

Sophomore Year Projects:
Line Following autonomous robot

Like many, my journey through robotics started with a baby step with Line following robot. In the sophomore years this simple robot could generate enough enthusiast which I'm still bearing. In the national level fest another eye-catchy event was battle of bots. The bottom-right corner is our own designed circuit for motor driver where we used L298N IC and relay to trig the command. These fun stuffs were really enjoyable.

The line following robot was the baby step during my journey to robotics. It was built for a competition, after numerous times failure, we as a team could managed top position in national level robotics competition

Customized motor driver based on L298N IC and relay switch which can take up to 7A current while operating

Page updated

Google Sites

Report abuse