This project was completed as a part of my undergraduate thesis. Check out the video for a detailed project demonstration.

The project performs a comparative analysis between four distinct RL algorithms (DDPG, TRPO, PPO, A2C) on the highway driving task. Other tenchniques incorporated into the project include -

• Incorporation of a safety-net around RL model to avoid rear-end collisions

• Using a set of continuous-time polynomial trajectories as the action space instead of discrete throttle and steering actions

• Use of Frenet-space co-ordinates for trajectory generation

Check out detailed project report here.

This project was completed for our "EEE 304: Digital Electronics Laboratory" course. Check out the video for a detailed project demonstration.

The project designs a dyanmic Traffic Light Controller (TLC) that is responsive to real-time vehicle density on a four-way junction. The TLC is designed based on a finite state machine that consists of 27 states. The controller is fed by vehicle detecting IR sensors & emergency vehicle detecting sound sensors. Based on the sensor input, our system provides an optimum sequence for the traffic lights to turn on or off. The system also enables emergency vehicles to pass through with highest priority and can detect and record traffic rule violations. The intelligent traffic light control system proposed in this project is a much more efficient alternative to currently used fixed time sequence traffic light controllers.

Check out detailed project report here.

This project was completed for our "EEE 416: Microprocessors and Embedded Systems Laboratory" course. Check out the video for a detailed project demonstration.

In this project we created a model for a voice-controlled wheelchair. The project was performed at two levels. A hardware prototype for the wheelchair was created using cardboard, wheels, motor, motor drivers, battery etc., and a simulation of a wheelchair was also done on Proteus. To perform voice to text conversion we used Google API. To perform the control tasks we used ARM Cortex Based STM32 microcontroller. Voice commands were transferred from mobile App to microcontroller via HC-05 Bluetooth module. The STM32 microcontroller was programmed in the Arduino IDE. 

Check out detailed project slides here.

This project was completed for "EEE 310: Communication System I Laboratory" course. We developed a trainer board that could be used to both perform and demonstrate amplitude modulation and demodulation. Here, at first, we generated a high-frequency carrier signal and modulated a random message signal with it. After the modulation, we transmitted & finally demodulated the overall signal with a demodulator circuit. We used a switching modulator as the modulator circuit and a diode detector as the demodulator circuit. We designed our trainer board on a PCB layout.

Check out the video for a detailed project demonstration.

The goal of this project is to take input data from a person through our website, use a machine-learning algorithm to analyze that data, and predict whether the person is at risk of Covid-19 or not. This is a form of pre-testing which will enable doctors to prioritize which patients to test first, and allow patients to know if it is necessary for them to go for a Covid-19 test. Our project can be regarded as a classification problem, where the machine learning algorithm takes a particular set of input data (which are known as input features) from a person and then categorizes them into positive or negative classes. The set of input features we supplied to our machine learning algorithm are:

•  Age

• Travel History

• The presence of symptoms such as fever, cough and sore throat, breathing problems, pneumonia, headache, Weakness and nausea

According to the World Health Organization(WHO), 285 women die from preventable diseases related to pregnancy and childbirth every day all over Bangladesh. Most of it is due to a lack of proper information, quick service, and nearby healthcare facilities. On the other hand, 1 in 5 new specialist female doctors is currently unemployed in Bangladesh. So, We built a mobile application that works as a link between doctors and female patients and solves both these problems at once. The aim of this project is to digitalize our existing analog health care system & create a revolution in the medical sector. Our team is still continuously working on this project & we also won some awards regarding this.

Check out the video to know the details.

This project was completed for our "EEE 415: Microprocessors and Embedded Systems" course. Check out the video for a detailed project demonstration.

In this project we modelled a 4-bit computer using VerilogHDL that could perform a pre-defined set of operations by taking in Assembly Language input. The computer was modelled on the SAP-1 architecture with a (16X4) RAM for stroing both data and instructions. It also contained an ALU capable of performing addition, subtraction, exchange and rotation operations. The system also contained two registers to hold temporary data. Data was transferred throughout the system using a 4-bit bus. Finally an assembler was built to convert Assembly Language into Machine Language. Check out the detailed project report here.

This project was completed for "EEE 306: Power System I Laboratory" course. Power systems in most countries around the world today are based on analog, back-dated methods with no measures for complex data measurements (such as power factor), data storage, and data interpretation. To solve all these problems, we have developed a smart energy meter that is basically a digital electric meter integrated into an IoT (Internet of Things) based system enabling it to transfer data onto a server over a local area (our current project) or wide area (future large scale application) network.

In our current project, we have focused more on the consumer side. We have built an electric meter that incorporates a voltage measurement and current measurement circuit. This circuit is operated using an Arduino Nano microcontroller unit, step-down voltage transformer, current sensor, voltage regulator, capacitor, and resistors. The readings on this meter (real power, RMS voltage, and power factor) are displayed on an LED display and simultaneously transmitted from the microcontroller unit to a local server website via a NodeMCU (ESP 8266) using a local WiFi network where the data for each load connected to the system can be continuously viewed by the customer in a smart device connected to the same local network. This data can be stored and examined for later purposes.

Check out detailed project report here.

This project was completed for our "EEE 206: Energy Conversion Laboratory" course. In this project, we performed rewinding & testing of a squirrel-cage induction motor. We performed the following procedures:

• Inserting new insulators

• Making a demo coil with new wires

• Making coils & inserting them into the motor

• Final insulation

• Providing connection between coils

• Applying varnish & reconstruction of the motor

• DC testing, No-load testing & Locked-rotor testing