|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Interaction Lab |
Click here for information on undergraduate research funding
The Interaction Lab also supports undergraduate student researchers through NSF Research Experience for Undergraduates (REU) awards, for those students with relevant expertise and working on projects associated with our NSF-supported grants.
To get involved, you need to provide us with the following information:
Email the above requirements to the project contact for the project you are interested in:
- Amin Atrash (atrash@usc.edu)
- Eric Wade (ericwade@usc.edu)
- Aaron St. Clair (astclair@usc.edu)
- Juan Fasola (fasola@usc.edu)
- Ross Mead (rossmead@usc.edu)
- Elaine Short (elaine.g.short@usc.edu)
- Jillian Greczek (greczek@usc.edu)
- Katelyn Swift-Spong (swiftspo@usc.edu)
Open Projects
Contact: Ross Mead (rossmead@usc.edu)
Set up system infrastructure for a moderately-sized space (e.g., the pen in the Interaction Lab) with multiple markerless motion capture devices (i.e., Microsoft Kinects) for continuous multi-user tracking.
- Recommended Skills: moderate understanding of C++
Contact: Ross Mead (rossmead@usc.edu)
- Develop models of eye gaze for human-robot interactions.
- Implement gesture animations/behaviors for human-robot interactions.
- Parameterize eye gaze and gesture behaviors based on social (e.g., people, speech, gestures, etc.) and environmental (e.g., visually, aurally, or contextually salient objects and/or interference) factors.
- Recommended Skills: moderate understanding of C++
Contact: Ross Mead (rossmead@usc.edu)
- Set up hardware and software infrastructure for "supermarionette" robot (pulley-driven humanoid robot on loan from Disney Imagineering) mounted atop a Pioneer mobile robot base.
- Map (in 3D) and label a tour environment (e.g., the Interaction Lab).
- Develop socially situated navigation and interaction strategies for robot.
- Recommended Skills: moderate understanding of C++
Contact: Ross Mead (rossmead@usc.edu)
- Investigate audio (speech) and visual (gesture) factors that contribute to social spacing (interpersonal distance and orientation) in human-robot interactions.
- Investigate physiological factors (e.g., heart rate, respiration rate, skin conductance level, skin temperature, etc.) that contribute to social spacing (interpersonal distance and orientation) in human-robot interactions.
- Investigate other factors (e.g., amount of eye contact, amount of smiling, arm configurations, relative height, etc.) that contribute to social spacing (interpersonal distance and orientation) in human-robot interactions.
- Recommended Skills: moderate understanding of C++
Contact: Ross Mead (rossmead@usc.edu)
- Implement a graphical user interface for generating motions and behaviors for socially interactive robots.
- Implement a graphical user interface for annotating and analyzing data from audio, visual, and other sensor sources.
- Recommended Skills: moderate understanding of C++ and/or Python
Contact: Ross Mead (rossmead@usc.edu)
- Design skins and skeletons for virtual persons for simulating human-robot interactions.
- Implement simulated human behaviors (both hand-animated and data-driven).
- Recommended Skills: computer animation, graphics, or gaming background; moderate understanding of C++
Contact: Ross Mead (rossmead@usc.edu)
- Develop intuitive interfaces to enable K-12 students to use complex robotics software tools and services, such as mapping, localization, path-planning, navigation, object recognition, and obstacle avoidance.
- Implement wrappers around software tools and services available in the ROS (Robot Operating System) framework (www.ros.org).
- Implement wrappers around the Microsoft Kinect and OpenNI (Open Natural Interaction) framework (openni.org/).
- Recommended Skills: moderate understanding of C, C++, Python, and/or Java
Contact: Elaine Short (elaine.g.short@usc.edu), Ross Mead (rossmead@usc.edu), Amin Atrash (atrash@usc.edu)
We are currently running experiments to determine the efficacy of a socially assistive robot for stroke rehabilitation and autism therapy. We would like someone to annotate the video recordings of the sessions to encode and classify the interactions that take place between the participants and the robot.
Contact: Elaine Short (elaine.g.short@usc.edu)
We are interested in enabling our PR2 robot to manipulate objects in realisitic interaction scenarios. A student with a moderate to strong programming background (C++ and/or Python) will work on the following:
- Improve existing manipulation pipeline for manipulating objects in human-robot interaction scenarios.
- Document and extend open-source code for manipulation from various institutions.
- Develop and implement models for human-like manipulation in realistic in-home human-robot interaction scenarios.
Contact: Eric Wade (ericwade@usc.edu)
Using our in-house motion capture suit, develop code to do some of the following:
Software
- come up with a gesture library for use in robot control and interaction
- create a recognizer using the motion capture suit to recognize gestures from that library
- create a manual for the motion capture suit to allow new users to be able to install and use software and hardware properly
- implement software "reset" for sensors in receiver
Contact: Jillian Greczek (greczek@usc.edu)
We will be running an experiment this semester where an imitative game is played between a robot (the Aldebaran Nao) and a child with autism spectrum disorder (ASD) to improve social and motor skills. This work ties into the overall goal of personalized care over long-term interactions. We are interested in assistants for running these experiments and collecting data.
Contact: Aaron St. Clair (astclair@usc.edu)
We are currently developing a system to support effective collaboration between a human and a robot during the course of collocated tasks. We would like someone to help with the following aspects of the project:
- Development of interactive task simulations/games for use with our existing projector-based augmented reality system
- Implementation of robot behaviors and task controllers enabling the robot to perform tasks with real or simulated objects (e.g. herding, simple object manipulation)
- Development and integration of on- and off-board person tracking and gesture recognition systems using cameras/Kinect
- Assistance in running a series of data collections and user studies aimed at evaluating the communication system in the above environments
Contact: Elaine Short (elaine.g.short@usc.edu)
Reciprocity is the compulsion people feel to return a favor or gift. As part of their work on computers as social actors, other researchers have demonstrated that reciprocity functions in Human-Computer Interaction (HCI). We hypothesize that this principle functions within HRI as well.
We are looking for a student to be be involved in the execution and data anal\ ysis of a study to validate a computational model of reciprocity in Human-Robot Interaction (HRI).
The student should be comfortable with technology, although programming skills are not necessary. Experience in experimental psychology is preferred but not required. The student should have completed classwork in statistics or have a strong background in mathematics.
Contact: Elaine Short (elaine.g.short@usc.edu)
A student with a moderate to strong programming background (C++ and/or Python) will do the following:
- Develop a plugin for existing annotation software to interface with collected human-robot interaction data, including automatic annotation of robot behaviors and collected sensor data.
- Collect feedback on annotation software user interface and functionality, as specifically related to the annotation of interaction data.
- Design and program any new functionality required for annotation of human-robot interactions.
Contact: Katelyn Swift-Spong (swiftspo@usc.edu)
One of our robots, the Dragonbot, uses an Android phone to display the eyes and mouth that make up the face of the robot. We are currently in the process of redesigning the the Android application used to display this face. This project involves redesigning and developing the backend software for this app. It also involves collaborating with an animator who will develop the animations for the face. The skills required for this project include knowledge of Android development, Python or C++, and Java. Familiarity with ROS (Robot Operating System) is preferred.
Contact: Katelyn Swift-Spong (swiftspo@usc.edu)
This project involves developing software for recognizing various exercises with data from a Kinect. Examples of possible exercises are squats, push ups, and bicep curls. The skills required for this project include knowledge of Python or C++. Familiarity with ROS (Robot Operating System) and the Kinect is preferred.
Contact: Edward Kaszubski (kaszubsk@usc.edu) and Ross Mead (rossmead@usc.edu)
Design and implement unit tests for other software components.
Recommended Skills:
- Experience with testing/debugging software
- Understanding of C++ and/or Python
- Understanding of ROS (Robot Operating System)
- Familiarity with the Linux operating system
Contact: Edward Kaszubski (kaszubsk@usc.edu) and Ross Mead (rossmead@usc.edu)
Ensure inline documentation is complete and descriptive, and write intuitive tutorials for completed software components.
Recommended Skills:
- Excellent communication skills
- Understanding of C++ and Doxygen
- Basic understanding of ROS (Robot Operating System)
- Familiarity with the Linux operating system
Contact: Edward Kaszubski (kaszubsk@usc.edu) and Ross Mead (rossmead@usc.edu)
Develop and execute testing procedures for a wide range of robot platforms, maintain a list of hardware states for these platforms, and assist in debugging and basic maintenance.
Recommended Skills:
- Hardware or electronics background
- Basic understanding of ROS (Robot Operating System)
- Familiarity with the Linux operating system
Contact: Edward Kaszubski (kaszubsk@usc.edu) and Ross Mead (rossmead@usc.edu)
Design and implement demonstrations within the lab's software framework that highlight the functionality of hardware and software platforms, and evaluate the quality of documentation and tutorials.
Recommended Skills:
- Experience with C++ and/or Python
- Understanding of ROS (Robot Operating System)
- Familiarity with the Linux operating system
Contact: Edward Kaszubski (kaszubsk@usc.edu) and Ross Mead (rossmead@usc.edu)
Add packaging/build system components to existing projects, and ensure the build system and package archive are functional and up-to-date.
Recommended Skills:
- Understanding of Debian packages and package management systems
- Basic understanding of and the Debian build toolchain and isolated build environments
- Basic understanding of shell scripts
- Basic understanding of ROS (Robot Operating System)
- Familiarity with the Linux operating system
Contact: Edward Kaszubski (kaszubsk@usc.edu) and Ross Mead (rossmead@usc.edu)
Design and implement core software infrastructure components.
Recommended Skills:
- Strong understanding of C++
- Strong understanding of software design principles
- Familiarity with version control systems
- Moderate understanding of ROS (Robot Operating System)
- Familiarity with the Linux operating system