Parth Doshi

Designed and implemented real-time autonomy software for AGVs using C++ and ROS on Linux, ensuring deterministic behavior, low-latency execution, and robustness under real-world operational conditions. Developed internal diagnostic and visualization tools in C#/.NET (Avalonia) and C++/Qt to support real-time debugging, data inspection, and operator interactions. Performed testing, commissioning, and iterative tuning on deployed fleets on customer sites.

"Localization of Automated Guided Vehicles in Indoor Environment" Reflector based localisation method was implemented to improve the navigational accuracy in indoor environment, achieving +/- 3 cm repeat accuracy using ROS (C++). Analysis and comparison of the state-of-the-art indoor localization methods by creating test cases. Successful implementation of reflector localization on the AGV hardware and in simulation (using Gazebo and RVIZ).

Developed realistic simulation environments for Autonomous Mobile Robots (AMRs) using Gazebo, RViz, and ROS2, enabling algorithm testing for navigation, localization, and path planning. Implemented and configured the Nav2 stack (SLAM, AMCL, Behavior Trees) to simulate multi-robot operations in indoor warehouse scenarios. Built automated test pipelines using GitHub Actions and Docker, enabling unit and integration testing for reproducible, containerized ROS2 builds and improving CI/CD reliability.