• Led perception and localization development of compute payload atop the Ghost Robotics Vision 60 robotic dog for autonomous capture and retrieval

• Building SLAM pipeline for a robotic ball

• Developing slope detection algorithm for rover

• Joint space control and cartesian control on manipulator

• Identifying contact points on manipulator

• Integrating identifying grasp candidates model into current manipulator architecture

• 6DOF pose estimation of satellite using stereo camera

• Designed and prototyped fiducial marker based localization system for indoor environments

• Conducted a comparative analysis of EKF and UKF sensor fusion algorithms for localization system, identifying key performance metrics and optimizing system efficiency

• Refactored localization codebase for outlier rejection in GPS data and produce better input for sensor fusion algorithm

• Developed a pose graph optimization-based localization approach utilizing GTSAM

• Implemented multi-camera tracking algorithms for safety and follow features

• Benchmarked neural net predictions to quantify training quality for 2D object detection

• Augmented and added 5 thousand images to existing dataset for object detection

• Designed and implemented calibration tests to visualize hardware longevity and bias across tractors

• Programmed 9 LED light-strip animations onboard tractor in Python and integrated in ROS to improve HRI (Human Robot Interaction)

• Configured and tested cheaper radios for base station and tractor to receive GPS RTK corrections

• Setup tracking and scheduling demo reservation system using Google Apps Script and JavaScript