Research and Projects
Robotic Learning My current research focuses on learning based planning and controls of robots. You can more about my research on our lab website.
Multi Robot Mapping Another project that I have worked towards is the multi-robot mapping. Mapping large environment can be painstaking and time consuming process. In our project we wroked towards developing a multirobot setup for the purpose of mapping and planning tasks with a team of robots.
Motion Planning Stack Here some videos of my motion planning stack implmentation on a Crzyflie. This employes a minimum acceleartion trajectory generator with A* path planner and SE(3) Attitutde controller. Video in Sim Video on Physical Setup
- Path Planning Planning and controls are at the heart of autonomous systems. The problems in this field can vary a lot and the biggest task being able to handle the steep climb in the complexity, both in terms of implementation and conceptualization. The outcome is equally satistfying. I have had the opportunity to work with planning algorithms at both the levels. In brief I have worked with 3 major types of planning algorithms:
- Vector Field (Attraction based)
- Graph Search (A-star, Dijsktra, RRT, RRT-star)
- Optimal Path (Planning in continous state space)
NN Based System Identification System Identification has been studied in great depth in the field of controls theory. In this project my aim to discover the ability of a neural netwrok to learn the dynamics of a Unamanned Underwater Vehicle. The dynamics of a UUV can very complicated. For the purpose of the study, the NN was trained with simple predictable trajectories which was tested against the PRBS excited model.
- Mechatronics Here a few videos showing the mechatronic projects I have worked on in the past.
- Under Grad Projects
- NSIT Solar Car Link to video
- Face Recognition and Tracking Link to Video