Browse Technologies

Displaying 1 - 10 of 26

Robust Learning Algorithms in Adversarial Environments

Vanderbilt engineers have developed an algorithmic framework for machine learning under the threat of adversarial evasion. The framework leverages a game theoretic model of interaction between the learner and an evading attacker, and makes use of modern optimization tools to increase robustness of learning algorithms as they are used in operational settings.

Data-driven Agent-based Modeling Architecture

A team of engineers at Vanderbilt University has developed a data-driven agent-based modeling framework that can be used in forecasting consumer behavior for product adoption purposes. The architecture turns agent-based modeling into a reliable forecasting tool at both individual and population level resolutions.

Algorithms for Compliant Insertion and Motion Control of Continuum Robots

This technology enables continuum robots (aka snake robots) to precisely navigate the intricate structures of deep anatomical passages during minimally invasive or natural orifice surgery. Collateral surgical damage is minimized by the force sensing capabilities of the algorithms used.

Motion Generator to Transform Linear into Nutation Motion

This novel device converts linear motion into nutating motion and can create large angles from small linear displacements. The invention uniquely provides control and precision in the use of nutation motion making it particularly adaptable to micro-applications.

Highly Accurate Radio Chip Localization Technology

Summary: This technology, developed at Vanderbilt University's Institute for Software Integrated Systems, uses radio interferometry to locate tangible objects and attains, simultaneously, a higher degree of accuracy (within 3 centimeters), considerably longer range (up to 160 meters) and lower cost than other technologies.

TagDock: An Efficient Rigid Body Molecular Docking Algorithm For Three Dimensional Models of Oligomeric Biomolecular Complexes With Limited Experimental Restraint Data

TagDock is an efficient rigid body molecular docking algorithm that generates three-dimensional models of oligomeric biomolecular complexes in instances where there is limited experimental restraint data to guide the docking calculations. Through "distance difference analysis" TagDock additionally recommends followup experiments to further discriminate divergent (score-degenerate) clusters of TagDock's initial solution models

Ultrasonic Sensor for Non-intrusive Local Temperature, Transient Temperature and Heat Flux Measurements

An apparatus for measuring the temperature and heat flux of materials through the use of an ultrasonic sensor has been developed at Vanderbilt University. The sensor uses acoustic measurement techniques to determine the heat flux and temperature of material surfaces otherwise inaccessible in particular during system operation in order to enhance monitoring capabilities and reduce unsafe or impaired function due to extreme temperatures.

Brain Shift Compensation Using Computer Models

This technology eliminates the need to place cortical fiducial markers during image guided neurosurgery. As an additional and important feature, the technology is able to compensate for brain shift due to deformation of the brain during surgery.

Algorithms for Contact Detection and Contact Localization in Continuum Robots

This technology enhances the capabilities of continuum robots by not only detecting contact during movement but also estimating the position of the contact during the movements executed by the robot. An algorithmic feedback loop can then constrain the movement of the robot to avoid damage to its robot arm, damage to another robot arm or damage to surrounding structure. Applications for this technology include enhanced safe telemanipulation for multi-arm continuum robots in surgery, micro-assembly in confined spaces, and exploration in unknown environments.

Model-based Compression Correction Framework for Ultrasound

Vanderbilt researchers have developed a system that uses tracking and digitization information to detect the pose of an ultrasound probe during the imaging of soft tissue.  This information is used with a biomechanical model of the tissue to correct compressional effects during intraoperative imaging.