IEEE Industrial Application Society Robotics Contest (2019)
This work focuses on service robots that can be used in household works or in commercial places like restaurants or banks. This has been accepted for the finale of IEEE IAS CMD Robotics Contest 2019 as one of the four contestent around the globe. In the finale this work was awarded as the “Champion” of the IAS CMD Robotics Contest 2019, Baltimore, Maryland, USA. Furthermore, this work is now being prepared for publication.
IEEE Industrial Application Society Robotics Contest (2018)
Our aim is to develop a system that will be efficient, fast but robust for autonomous material delivery system with the help of image processing and AI. Our existing manual forklift system is outdated in industries. Although there are some autonomous line following based vehicles for material delivery but it suffers from different constrains. Line following based system is not portable enough and a robot has no capability of going out of the line. Hence it is limited to several constrains and hence we are proposing an intelligent and efficient system for that purpose. In this system the robot will be given a set of task (material pickup and delivery places) in a central computer. This computer will have a top view of the whole area where it will perform autonomous delivery without the help of any physical line or human operation. Once the process is started it will pick and place all the materials to its desired location.
This project is a developed version of of the SHUFFLE project. It uses IOT based communication for swarm capability. Also it has faster computation. Currently it is being extended further to incorporate on-board image for precision positioning of the objects.
This project was selected as one of the Four Finale project for IEEE IAS Annual meetin 2018 at Portland, Oregon, USA. Where this work was awarded as 2nd best project for industrial applications.
GPS BASED AUTONOMOUS TRAVERSAL (2017)
Abstract—Now a days, with the constant evolution of Artificial Intelligence and Machine Learning, robots are getting more perceptive than ever. For this quality they are being used in varying circumstances which humans cannot control. Rovers are special robots, capable of traversing through areas that are too difficult for humans. Even though it is a robust bot, lack of proper intelligence and automation are its basic shortcomings. As the main purpose of a rover is to traverse through areas of extreme difficulties, therefore an intelligent path generation and following system is highly required. Our research work aimed at developing an algorithm for autonomous path generation using GPS (Global Positioning System) based coordinate system and implementation of this algorithm in real life terrain, which in our case is MDRS, Utah, USA. Our prime focus was the development of a robust but easy to implement system. After developing such system, we have been able to successfully traverse our rover through that difficult terrain.
It uses GPS coordinates of target points that will be fed into the rover from a control station. The rover capturing its own GPS signal generates a path between the current location and the destination location on its own. It then finds the deviation in its current course of direction and position. And eventually it uses Proportional Integral Derivative control loop feedback mechanism (PID control algorithm) for compensating the error or deviation and thus following that path and reach destination. A low cost on board computer (Raspberry Pi in our case) handles all the calculations during the process and drives the rover fulfilling its task using a microcontroller (Arduino).
Published in Asian Symposium 2017 (Malaysia).
MARS ROVER VER-3 (2018)
This one is the upgraded version of the former one. It has lots of improvements starting from the suspension system to the control system. Its robotic arm is much more precise. It has more AI based system than the previous one. It has GPS foot print following system. It can analyze soil and environment more efficiently and can bring up more samples with a very little contamination.
MARS ROVER VER-2 (2017)
This project was done under my New Lab “Robotics and Rover Development Lab” at Independent University Bangladesh. Me along with my team “IUB Attendant” brought the top 10 (10th) place for Bangladesh at “University Rover Challenge-2017”, Utah, USA. This rover was fully built from scratch.
This was one of the biggest project of IUB and RRD Lab. It was supported by three other labs there. 12 people attended the Finale at Utah, USA. This project got massive publicity nationwide.
This rover utilizes autonomous traversal method that was published in an international conference further. Also it had 6 DOF robotic arm. It uses RPI and arduino as control unit. It has a long range video and telemetry data communication method in two different frequency bands. It also has different AI for science task and robotic hand movement. Along with that I has the capability of taking geo-tagged auto panorama capture. For the control station site, it uses Python and CPP based scripts. Further details can be found in the video.
MARS ROVER VER-1 (2016)
This was a Mars Rover Project by my ex-team, BUET Interplanetar. It was a revolutionary team to introduce Mars Rover in BUET by participating both ERC-15 at Poland and URC-16 at USA. This team secured 19th overall and 5th in Phobos Finale.
European Rover Challenge, ROVER VER-0 (2015)
This was the first ever Mars Rover by BUET team for European Rover Chellenge 2015, Poland. The rover was named after Wall-E robot. This was called Mongol-E. Mongol means Mars in Bengali. We worked hard to get a breakthrough towards rover development in Bangladesh. Following the footstep of this rover, successive rovers were developed. Development of Mars Rover Version-1 was the first outcome of this project.
International Robotics Challenge (IRC) –(2015)
International Robotics Challenge is a flagship competition of Techfest, IIT, Bombay, India. This challenge is based on both autonomous and manual robot. The problem statement is provided in the first video. My team was the champion of the regional round (second video) and secured second position globally in the finale at Techfest-2015, IIT, Bombay, India.
Shuffle-Bot is an image processing based robot. The task is to place all the red objects in yellow places, blue objects into pink and black object into black place in a sequential order. The robot works on a cloud based computing system which has been developed from scratches. Open-CV library along with C++ language aid the main algorithm. A single board computer captures the frame of the arena. It the performs image processing. Afterwards it computes and generates its path from its current position to grab the object. Its complex algorithm allows the robot the avoid any object and generate clear path optimizing the turns and distance. A wireless communication guides the robot to reach destination, grab the object and place it on the required place.
After starting the program the system itself calibrates and does not require any intervention. It performs all the task without any manual operation. Moreover the robot itself doesn’t contain any algorithmic part and hence it paves the way to perform swarm robotics and this work is under progress.
INTERNATIONAL AUTONOMOUS ROBOT CHALLENGE- IIT, KANPUR, INDIA (2014)
This was an international challenge at IIT, Kanpur, India. The task of the robot is to find the destination, revert back to the original position along the shortest path, pass through static and moving obstacles and then come out of the arena and stop.
There was only two people in our team. And we have been able to secure the championship both in Zonal (Bangladesh round) and International (Finale) round. This was my first international challenge and my maiden championship. There were so many hindrances ranging from administrative to technical. I took that as a challenge and successfully overcame it.