The Biorobotics Lab reduces complicated high-dimensional problems found in robotics to low-dimensional tractable ones for design, analysis, and planning. Often, we look to biology for inspiration and sometimes, we return the favor by providing analysis that models biology. Our research group runs the full pipeline from basic research, to applied research and all the way to system deployment, and in some cases commercialization. We find that basic theory informs system development and vice-versa. We strongly believe in working with stakeholders in our chosen application fields because they have a unique perspective that challenge our core assumptions, inform future research direction, and drive the development of practical solutions. Our ability to build a great variety of novel field-tested systems stems from advancing modularity, at all levels, both in software and hardware.
Displaying 39 Publications
The most important component in case of high-tempo, traumatic scenarios on the battlefield is inserting a needle in the vessel of the patient autonomously, as there is no professional medical person present.
Blaser is a compact ultra-short range 3D scanner for confined space robotic applications.
Robot solution for old pipes. Robots go inside the pipes to repair leakage by extruding epoxy pipe inside the original pipe without digging the old pipe otu of the ground.
CMU in-pipe mapping system is a self-sufficient payload consisting of sensors and computation devices, equipped with multi-modal Simultaneous Localization and Mapping (SLAM) algorithms.
Design robotic arm, manipulator, mobile base, and its control hardware/software for confined space applications.
Speeding up of the paint spray-gun path programming by automation, achieving uniform coverage.
Incremental construction of exact cellular decompositions that are formulated in terms of critical points of Morse functions.
An autonomy system that can handle tough terrain while moving fast.
The DARPA Subterranean or “SubT” Challenge seeks novel approaches to rapidly map, navigate, and search underground environments during time-sensitive combat operations or disaster response scenarios.
Predict deformation in ultrasound images and use it to create synthetic data to train other machine learning techniques.
Recycling Robotics
Investigating the different and rich ways in which agents interact in their environment, especially when that environment possesses complicated dynamics of its own.
We developed and tested a prototype based on an innovative approach of a highly articulated robotic probe.
This project seeks novel approaches to rapidly map, navigate, and search environments.
Two-meter tall hexapod for various terrain usage. It is also an exploration of building a set of self-sufficient modular joints unified by a novel control architecture.
This research seeks to develop the state of the art in fast moving autonomy while also exploring methods for human interaction with complex systems.
Recycling Robotics
This project aims to find novel ways of navigating and mapping the inside of the body
This project aims to allow ultrasound image segmentation to generalize across real-world patient diversity
This project is exploring methods of performing specialized bone surgery around the ear to make surgeries safer and more effective.
This project aims to create a device that reduces pain from needles.
An array of cells, each of which can induce a vectored force to an object resting on the cell. The cells collectively transport and manipulate objects resting on the array.
Self-contained sensing, computing, and a motor driver printed circuit board design provides onboard feedback control and motion planning.
Modular robotic systems have a core set of modules that can be combined and recombined to quickly form a customized robot.
We seek to obtain the benefits of both coupled and decoupled approaches by bypassing the curse of dimensionality, on average, for multi-agent planners.
Gradient-based inter-agent communication learning on bandwidth-limited communication networks.
We developed decentralized multi-agent RL to construct a structure out of blocks. Each agent vies for itself and the agents automatically build scaffolding structure as well.
A novel device that allows us to perform psychophysical experiments.
Motion planning algorithm for thr rod-shaped robots, based on distance measurements.
Giving the robots the sense of touch by developing compact, robust, and intelligent force/torque/tactile sensors and perception algorithms.
Research in ventilation control systems, tele-critical data collection, machine learning, and respiratory support hardware to enable tele-critical healthcare.
Giving Urban Search and Rescue workers more technological tools to help find and save victims of natural disasters.
Sensor based coverage of unknown environments using exact cellular decompositions.
Analyzing the factors that are of importance in designing a snake robot, and implementing new designs.
Dexterous On-Orbit Maintenance is used to perform repairs and replacements for satellites whose enabling functions - propellants, coolants, etc. - naturally deplete or whose subcomponents fail.
A tool for minimally invasive robotic surgery.
This work proposes a method for enhancing deep learning models’ capabilities by generating synthetic, or augmented, data which is transformed in a manner designed to account for various body types, injury scenarios, and imaging features.
Path-planning for a hyperredundant robot with low torque limits. Because the robot can't cantilever its full length, the planner finds bracing points in the environment to reduce the required torque.