My Portfolio
MY LATEST WORK
AIP
2025 -
One of the major challenges in applying artificial intelligence (AI) to power systems is the lack of reliable, high-quality data. This limitation has slowed the integration of AI into power-related applications. To overcome this issue, there is a need for cost-effective sensing technologies that can capture accurate data over long periods. This project focuses on designing and implementing a real-time sensing system for power applications, capable of monitoring current and voltage parameters and transmitting the collected information to a central server for AI-based analysis. Working on a team brilliant fellow students, developing the hardware , the enclosure, the AI API model , the user interface and the hardware(the sensor subsystem and processing )
LASER MICROPHONE
2025-
The laser microphone project is designed to capture audio signals by modulating a low-cost red laser and detecting its reflections with a fast photodiode, specifically the SFH203, which offers good sensitivity and response in the audio-kHz range. A Raspberry Pi Pico serves as the core controller, generating a high-frequency PWM carrier to amplitude-modulate the laser while simultaneously digitizing the photodiode output at high speed using its built-in ADC with DMA support. On the digital signal processing (DSP) side, the Pico implements subtraction techniques to remove DC offsets and common-mode interference, while gating and band-pass filtering are applied to isolate the modulated audio band and reject ambient noise and light flicker.
Dettera - User Interface for a
predator deterring system
This project is a smart wildlife monitoring(specifically penguin predators) and deterrence system built with Next.js. It helps track, classify, and manage wildlife interactions by combining automated detection with configurable deterrence mechanisms. The system is designed for flexible deployment in conservation areas, farms, and residential zones where human-wildlife conflicts occur.
Features
-
🦊 Wildlife Detection – Automatically detects and classifies different animals (Penguin, Deer, Fox).
-
📸 Image Management – Stores detection images with timestamps for review and analysis.
-
🚨 Deterrence System – Configurable methods include:
-
LED controls (adjustable frequency and duty cycle)
-
Speaker controls (adjustable frequency and duty cycle)
-
Motor controls (ON/OFF switching)
-
-
📊 Detection Records – Logs all detection events and deterrence actions with metadata.
-
🌐 REST API – Provides endpoints for retrieving detection data and adding new records, enabling easy integration with other systems.
About
The micromouse project also incorporated MATLAB Stateflow for modeling the control logic and simulating system behavior before hardware implementation. An STM32 microcontroller served as the core processing unit, executing real-time decisions and interfacing with the sensors and motors. To achieve precise motion and stability, PID control algorithms were applied to regulate the motor speeds, ensuring smooth line-following and accurate maze navigation. This integration of simulation, embedded programming, and control theory provided practical experience in developing an optimized autonomous system capable of solving and traversing mazes efficiently.
LINE FOLLOWING MICROMOUSE
2024
SENSOR BOARD FOR THE MICROMOUSE
2024
The Micromouse sensor board was designed as a custom PCB inspired by UCLA’s design but tailored for the project’s requirements to ensure reliable wall detection and efficient performance. IR333C emitters and TEFT4300 receivers were selected for their matched wavelengths and fast response, with the emitters powered from the 3.7 V battery through 47 Ω resistors to achieve strong radiant intensity. MMBF170 MOSFETs, driven by STM32L476 GPIO PWM outputs, handle switching efficiently, aided by 100 Ω gate resistors and 47 kΩ pulldowns for stability. On the receiver side, an LM324 op-amp provides buffering to protect the ADC from noise spikes, while voltage dividers scale the signals safely within the 3.3 V input range. With bypass capacitors ensuring clean LED drive signals, the final PCB integrates compact, low-noise, and robust circuitry to deliver accurate and responsive sensor readings for the Micromouse.