Developed LR-Bot, an autonomous indoor litter-rescuing robot using Jetson Nano 2GB and a 4 DoF robotic arm. Employed YOLOv5s for high-accuracy object detection, achieving a significant success rate in targeted litter rescue tasks.

Developed "ChildTag," an IoT device for real-time child safety monitoring, featuring GPS location tracking and ESP32-powered activity alerts, supported by a camera system for visual monitoring.

We developed SACFomer, an online Reinforcement Learning framework that integrates a Decision Transformer into an actor-critic architecture, enabling effective online training with sequence modeling. This innovative approach demonstrated superior performance in various OpenAI Gym environments.