Projects
Splat-Nav
Splat-Nav revolutionizes autonomous navigation by planning safe, non-conservative flight paths through complex, cluttered environments—at remarkable speed. By harnessing the power of 3D Gaussian Splatting primitives, we’ve drastically reduced the computational complexity of collision checking, enabling faster and more efficient path optimization.
This cutting-edge work is spearheaded by the Multi-Robot Systems Lab at Stanford University, pushes the boundaries of safe navigation for aerial robotics, paving the way for applications in dynamic and challenging settings.
Links: ArXiV
Generalized Traveling Salesperson Problem for Inspection Path Planning (GTSP-IPP)
GTSP-IPP is designed to enhance safety and efficiency in critical infrastructure inspections, such as bridges and skyscrapers, by eliminating the need for human involvement in dangerous tasks.
Our approach equips autonomous quadrotors with advanced safe coverage planners, enabling them to perform complex inspection missions with precision and reliability. By introducing an innovative alternative to conventional viewpoint inspection planning, we provide greater flexibility and customizability for solving inspection challenges in diverse scenarios.
This work represents a leap forward in autonomous inspection, making infrastructure evaluation safer and more efficient.
Links: Presentation
Robot Social Navigation Amongst Pedestrians (RoboSNAP)
This interdisciplinary project brings together experts in robotics and psychology to conduct groundbreaking human-robot interaction studies. By examining how pedestrians interact with robots in shared environments, we aim to develop robot planning strategies that prioritize safety while aligning with human expectations and social norms.
Links: Website
Embedded Analog & Digital Model Predictive Control
We successfully implemented real-time Model Predictive Control (MPC) on a coupled quadruple water-tank system, showcasing advanced control capabilities ondynamic, real-time processes. The control algorithm was tested on a Field Programmable Gate Array (FPGA) for high-speed digital processing and further augmented for analog control using a Field Programmable Analog Array (FPAA).
This project highlights the integration of cutting-edge hardware platforms to achieve precise and responsive control in complex systems, paving the way for applications in both academic research and industry.
Links: Analog Paper | Digital Paper
Ships at sea face constant deterioration from corrosion and biofouling, leading to significant wear and maintenance demands. To address this, we developed the Anti-Corrosion Inspection Drone (ACID)—a robust, specialized drone designed to perform surface detection and inspection in harsh maritime environments.
With its rugged design and advanced sensing capabilities, ACID empowers ship operators to identify and address maintenance needs efficiently, even under challenging conditions, enhancing safety and reducing downtime.
