Sanjay Viswanathan

Built a system integrating YOLO for real-time object detection, MiDaS for monocular depth estimation, and a ToF sensor for absolute distance measurement.
Designed an affordable alternative to LiDAR for industrial and field applications by combining AI-driven depth maps with physical sensor data, enabling robust performance without reliance on expensive LiDAR systems.
The fusion improves detection accuracy in challenging conditions where vision-only or sensor-only methods fall short.

Generated a synthetic dataset of 10,000+ samples in MATLAB/Simulink, modeling diverse leakage fault conditions to create a robust foundation for a machine learning solution.
Developed and trained an Artificial Neural Network (ANN) using Python and Scikit-learn to classify fault types and precisely predict their location from simulated electrical signal data.
Achieved an 82% prediction accuracy on the test dataset, proving the model's effectiveness for creating a rapid, data-driven fault detection system to enhance power grid reliability.

Developing a vision-guided wireless charging system for AGVs using OpenCV and ArUco marker detection, enabling precise autonomous docking and alignment of receiver/transmitter coils.
Implementing real-time feedback control with PID-based correction, improving charging efficiency and reducing coil misalignment errors by ensuring accurate AGV positioning.
Edge AI Deployment for edge processing, leveraging computer vision and lightweight algorithms for reliable, low-latency execution in industrial environments.

Developed Panda File Transfer Protocol using Python socket programming, implemented multi-threading and AES-256 encryption for secure, real-time file transfer across platforms.
Developed a GUI with automated key exchange and network stability features, ensuring smooth, secure file transfers across different operating systems.

Developed a real-time environmental monitoring system using Node MCU (Wifi SoC) with integrated temperature, humidity, and moisture sensors to gather and transmit farming data.
Created software integrated with the IoT hardware, enabling users to remotely monitor crop conditions and facilitating direct sales from farmers to streamline the supply chain.

Developed a sewage cleaning solution that detects blockages in drainage pipes, triggers a servo motor to clear the obstruction, and adjusts its movement based on real-time feedback.
Implemented IR sensors for obstacle detection and ultrasonic feedback to enable adaptive navigation, ensuring continuous operation with precise control using the Node MCU.

Currently Working in: