Work

Patents, publications, and projects — from research to production.

Patents

3

US Patent

Tool Calibration for Manufacturing Robots

Patented method for automated tool calibration in industrial robotic welding systems, enabling higher accuracy and reduced setup time in production environments.

RoboticsManufacturingCalibration

US Patent

Techniques for Path Clearance Planning

Novel path clearance planning techniques for robotic manipulators operating in constrained industrial workspaces, ensuring collision-free seam approach and retract motions.

Motion PlanningManufacturingCollision Avoidance

US Patent

Generating Simulated Weld Paths for a Welding Robot

System for generating and validating simulated weld paths, bridging the gap between offline programming and real-world robotic welding execution.

SimulationRoboticsWelding

Publications

5

Clearance Planning for Seam Approach and Retract in Autonomous Robotic Welding

Developed clearance planning algorithms deployed in production autonomous welding systems, solving the critical problem of collision-free torch approach and retract motions in cluttered weld environments.

Motion PlanningManufacturingDeployed

Motion Planning for Fully Autonomous Welding

End-to-end motion planning framework for fully autonomous robotic welding — from weld path generation to execution — deployed at scale in industrial manufacturing.

Motion PlanningAutonomyDeployed

IROS 2021

Search-Based Path Planning for a High Dimensional Manipulator in Cluttered Environments

Tackled path planning for a 21-DOF snake robot navigating cluttered gas turbines. Developed dynamic optimization-based actions, lazy generation strategies, and heuristics reasoning about homotopy classes with Dynamic Thompson Sampling.

Motion PlanningOptimization

Topology Guided Path Planning for a Snake Robot in Cluttered Environments

Developed topology-guided planning methods that leverage environmental structure to efficiently navigate high-dimensional snake robots through tight, cluttered spaces.

Motion PlanningTopology
Learning to Use Adaptive Motion Primitives in Search-Based Planning for Navigation

IROS 2020

Learning to Use Adaptive Motion Primitives in Search-Based Planning for Navigation

Proposed a learning-based approach for intelligent activation of adaptive motion primitives using deep learning, achieving over 2x speedup in planning times for 3-DOF navigation with Reeds-Shepp paths.

Motion PlanningDeep Learning

Projects

3

Motion Planning for Localization in Non-Gaussian Belief Spaces

Developed a method for motion planning under uncertainty to disambiguate multi-modal hypotheses on robot state, planning actions that achieve unimodal belief in finite time.

SLAMMotion Planning
Self-Driving Vehicle Controller

Self-Driving Vehicle Controller

Implemented a Stanley Controller for autonomous vehicles and tested it on a racetrack in the CARLA Simulator environment.

ControlSimulation

Motion Planning for an Autonomous Airship

Developed motion planning algorithms for autonomous airship navigation, combining search-based planning with vehicle dynamics constraints.

Motion PlanningControl