Passivity-based task space (robot tip motion) tracking controllers for soft robotic manipulators - using adaptive control and robust control to mitigate uncertainty

Task space bilateral tele-operation frameworks for soft robotic manipulators - using a rigid leader and a soft follower robot. 

Soft inverted pendulum - novel system with a a revolute base and a soft body. Upright stabilization achieved using energy-based control

Rigid-soft hybrid (HyRiSo) robot control - novel manipulator with rigid links and soft links serially attached together. Task space tracking achieved using adaptive and robust controllers.

Extensive simulations (MATLAB/Simulink)

Experimental validation - on a pneumatically actuated soft robot (fabricated in-house from silicone rubber using 3D printed molds) using MATLAB/Simulink.

Soft sensing skin - retrofitting a spray-coated piezoresistive sensing layer on a latex membrane

ML for bending estimation - trained and used a recurrent neural network (LSTM)

Cartographer_glass - a SLAM algorithm to detect and map transparent objects such as glass walls utilizing 2D LiDAR data leveraging Google Cartographer

CARDYNET - Trained and utilized a deep neural network for safe autonomous robot navigation among dynamic obstacles 

Experimentation - on a RACECAR platform using ROS with LiDAR sensors

Periodic stable walking gaits - for three-link and five-link biped robots using geometric control

Stable uphill walking - for three link bipedal robot

Simulations (MATLAB)

Position estimation and a mapping system - for a mobile robot using a single monocular vision camera through image processing techniques and perspective geometry