Path planning and navigation of agriculture robots.

Development of a complete pipeline for an autonomous Agri robot from low levelcontrol to high level behavioral planning.Implementation of perception pipeline and navigational pipeline in ROS2. Designbehavior trees for high level behavioral planning.Improvement in the existing navigation controller and swerve controller (low level) and writing new controller plugins. Created different models of the rover and the environment for simulation in Gazebo/Ignition

Development of several robotics system like drones and rovers right from the scratch to the final product along with the implementation of autonomous control software. The project starts with the selection, testing and validation of hardware (motors, sensors like IMUs, Lase scanner, depth sensors , cameras, controllers, RTK GPS e.t.c.). After integration of the entire system, components of autonomous control algorithm are being tested like EKF fusion and SLAM. The present work is mainly focused on the evaluation of different localization and mapping algorithms i.e. Lidar based SLAM, VisualSLAM or fusion of both. Though the EKF fusion based estimation has been tested in the real field, most of the evaluation of localization algorithms are done in a simulated framework of GAZEBO and ROS which is then extended to the real field soon.

An online education platform for robotics education.

Safe and friendly human robotics interaction.Thesis Title: "Design, modelling and control of inherently compliant actuators in agonistantagonistanthropomorphic configuration". The research project aims at design andmodel inherently compliant actuators for anthropomorphic systems. The first part ofthe work focuses on the choice of design and construction of the actuator and theintegration into the robotic system. In parallel, a model of the actuator dynamicsand the model-based controller (MPC and optimal control) have been synthesizedand integrated within the software environment developed by the team. Using themechanical system and the numerical simulator, tests are done in order to evaluatethe potential of this actuator and compare with the human body capabilities.

Investigating the Intelligent control methods for Welding control system. The major area of research is to apply the fuzzy and neural network based learning techniques to make the control system adaptive to the variation in welding conditions. Presently working on the design of Neuro-Fuzyy classifier and the fuzyy-neuro control system in MATLAB/Simulink and implement the model in Freescal 32 bit Power PC using Real time workshop and Simulink coder. The learning, training and validation are based on the real field data from the manufacturing plants.Implementing the algorithms developed in MATLAB/SIMULINK into FPGA module with NIOS-II Cyclone-II embedded system.

Real time control model development of actuators for prototype lunar rover of Thales Alenia Space. Integration on Linux embedded platform and hardware in Loop Testing (HIL) using RT-LAB and CANalyser. Extensive use of real time workshop in MATLAB for automatic code generation in C/C++.Design, analysis and simulation of dynamical model and implementing control algorithm in MATLAB/Simulink.

Characterization of Surface relative attitude sensors systems using Laser altimeters and IMU for attitude control scheme of TALARIS Hopping Lunar rover of MIT. The main responsibility is to interface the laser sensors and IMU and design an algorithm to predict the attitude of the rover. Some time had also been dedicated in developing user interface for Avionics system testing using LABVIEW