Current Research | Past Research

Robotic Telepresence

During Summer 2010, we conducted a series of user studies of two telepresence robots (Anybots' QB and VGo Communications' VGo) in an office environment at Google in Mountain View, CA. One study focused on virtual teams in which a remote teammate (n=6) used a telepresence robot to attend his/her regularly scheduled team meetings. We found that people who used to be in the same building as their teammates and then moved to a different location had the best experiences recreating this closeness with their teams using telepresence robots.

Autonomous Robots and Trust

Research in the field of human-automation interaction (HAI) has shown that trust is a key factor that influences operator's interaction with an autonomous system. Researchers also found that proper calibration of trust is critical to safe operation of an autonomous system. Too much trust on the system can lead to abuse of automation and conversely too little trust can lead to disuse of automation. In dynamic systems, operators need a control allocation strategy that optimizes performance. Hence mis-calibrated trust can lead to inefficient operation or even catastrophic failures.

Neural Networks and Mild Traumatic Brain Injury

We have just started an interdisciplinary project in cooperation with UML neurobiology researchers. In the course of this work, our lab will be working on artificial neural net (ANN) models of brain damage and drawing inspiration from biological mechanisms of brain damage and recovery to attempt to optimize neural network training and minimize the computational resources required to use ANNs.

SUBTLE Multi-University Research Initiative

SUBTLE stands for Situation Understanding Bot Through Language and Environment. The goal of the SUBTLE Multi-University Research Initiative (MURI) is to build a robot that can accept and understand spoken commands and report its situation and past activities to a human user in English. This project is being developed in coordination with several other research teams at other universities.

Our team is focused on providing a platform for the linguistic technologies that provides an embodiment of the system in the real world, and allows the conversation between the user and the robot to be grounded in the robot's perception of the world around it. This will allow for a richer development environment, with the complexity of the real world, rather than the somewhat impoverished world of simulations.

Visual Control Interface of a Wheelchair Mounted Robotic Arm for Cognitively Impaired Wheelchair Users

The Exact Dynamics' Manus ARM, a 6+2 degree of freedom wheelchair mounted robot arm, is able to function in unstructured environments. However, it is awkwardly controlled through a menu system using a keypad, a joystick or a single switch. These controls are not intuitive or natural because they require a high level of cognitive awareness. Also, the input devices may not correlate well to the user’s physical and cognitive abilities.

Our research investigates visual control of a robot arm. We leverage all of the Manus ARM’s benefits while eliminating its weaknesses. Our vision-based system draws inspiration from people’s innate abilities to see and touch. Because the wheelchair occupant is collocated with the ARM, the occupant's view is the same as a camera mounted over the ARM's shoulder. The occupant selects the desired object using from a flexible interface and touch screen or mouse-emulating joystick. Our goal is to allow the occupant to acquire the object by unfolding the ARM, then reaching and grasping the object in a manner emulating human kinematics. This human-in-the-loop control will provide simpler and effective interaction.

Improved Robot Interfaces for Urban Search and Rescue

Situation awareness (SA) is critical to successfully operate an unmanned vehicle. Since 2003, we have studied multiple systems designed for USAR. We have both designed evaluation methods and produced design guidelines for HRI.

Door Opening Robotic Arm (DORA)

A low cost robot arm system is developed that will increase a person’s accessibility to indoor spaces by unlatching door knobs and door handles. Implemented is a minimized arm configuration for use with a wheelchair or mobile platform and a gripper design that utilizes only a single motor to turn door knobs and door handles. This proof of concept prototype demonstrates how an arm with many degrees of freedom is not required if we target the expectations for its use.

Variable Geometry Tracked Vehicle (VGTV)

The VGTV is a tele-operated, tethered robot. It is capable of changing its shape, depth rated to 100 feet, and equipped with a color zoom camera, two-way audio, and lights. Sensors and small equipment can be carried by the VGTV into search areas. Our VGTV was used in Biloxi, Mississippi after Hurricane Katrina by Florida Task Force Three. Rescuers were able to search unsafe structures and experience a first person view from the robot.

Computer Vision: Phission

Phission is a concurrent cross-platform, multiple language vision system software development kit (SDK). The SDK constructs a processing sub-system for computer vision applications. Phission abstracts low-level image capture and display primitives, which allows researchers to focus on their primary work.


Consisting of an electric wheelchair outfitted with a computer and sensors and a Macintosh Powerbook, this robot can travel semi-autonomously in an indoor environment. This allows the user to issue general directional commands and to rely upon the robot to carry out the low level routines such as object avoidance and wall following.

Serial Sense

Serial Sense allows anyone to interface simple digital and analog sensors to a computer. It also gives you the ability to control custom hardware circuits with the digital outputs. This was designed to expand the sensor capabilities of ActivMedia's Pioneer robots. However, it will work with any PC that has a serial port.


Pyro stands for Python Robotics. The goal of the project is to provide a programming environment for easily exploring advanced topics in artificial intelligence and robotics without having to worry about the low-level details of the underlying hardware.

Street Crossing

Video - Car tracking from a stationary curbside camera (AVI)
Microsoft PowerPointThe Seventh Annual Student Research Symposium