Other Projects

We developed a controller for a social tour-guide robot. Our controller drives the robot close to the line of sight of the tourist, making it easier for them to follow. We conducted a virtual human-participant study to analyze the effect of our controller on constructs such as follow-ability, perceived safety, and perceived intelligence.

The goal of this project was to estimate the orientation of vehicles from their respective images. Our proposed model consists of two steps: vehicle detection and regression. We evaluated our model on the widely used KITTI dataset and reported the performance of state-of-the-art models by comparing the mean-squared error and the throughput rate.

Our goal was to develop a controller for enabling a UUV to follow 3D waypoints underwater. We simplified the problem into two univariable control problems: depth control and steering control. We proposed a sliding mode control (SMC) strategy for both the tasks and performed simulations showing resilience to water currents.

In this project, we wanted to build a coherent representation of the environment with the help of multiple ground robots. Our approach uses algorithms like Hough peak matching and iterative closest point matching to identify similar regions in the local map obtained from different robots. This helps in understanding the relative coordinate transformation among the robots, eventually allowing us to fuse the maps together. We tested our strategy on a real-world dataset from two laser-equipped robots.  

We built a waterproof hull for the institute’s autonomous underwater vehicle (AUV) team. The hull was propelled by 4 self-made thrusters that provided it 4 degrees of freedom. Lastly, we assembled the AUV with a torpedo that utilized a pneumatic launching mechanism.