University of Pennsylvania GRASP Lab

University of Pennsylvania GRASP Lab
Name	GRASP Laboratory
Established	1979
Location	Philadelphia, Pennsylvania
Affiliation	University of Pennsylvania
Focus	Robotics, computer vision, autonomous systems, machine learning
Director	Raj Madhavan

University of Pennsylvania GRASP Lab The GRASP Laboratory at the University of Pennsylvania is a multidisciplinary research center specializing in robotics, computer vision, autonomous vehicles, machine learning, and robotic manipulation. Founded in 1979, the laboratory has contributed to advances in artificial intelligence, multi-agent systems, robot perception, and human-robot interaction, influencing research at institutions like Massachusetts Institute of Technology, Stanford University, Carnegie Mellon University, and companies such as Google, Amazon, Tesla, and Boston Dynamics.

History

The laboratory was founded in 1979 by a team including professors connected to University of Pennsylvania, interacting with researchers from MIT AI Lab, Stanford Artificial Intelligence Laboratory, Carnegie Mellon Robotics Institute, Bell Labs, and NASA Ames Research Center. Over decades the lab engaged with projects related to DARPA Grand Challenge, DARPA Robotics Challenge, Human Genome Project-era computing efforts, National Science Foundation initiatives, Defense Advanced Research Projects Agency programs, and collaborations with IBM Research, Microsoft Research, Intel Labs, and AT&T Bell Labs. Faculty and alumni moved between institutions such as Columbia University, California Institute of Technology, Princeton University, Johns Hopkins University, and Yale University while contributing to conferences like IEEE International Conference on Robotics and Automation, NeurIPS, ICCV, CVPR, and AAAI Conference on Artificial Intelligence.

Research Areas

Research areas include robotic manipulation connected to work at MIT CSAIL, perception research related to Stanford Vision Lab, multi-robot coordination rooted in studies from Caltech, navigation systems influenced by Carnegie Mellon University, and machine learning methodologies aligned with Google Brain, DeepMind, OpenAI, and Facebook AI Research. The lab addresses autonomous aerial systems inspired by Lockheed Martin and Northrop Grumman research, legged locomotion comparable to projects at Boston Dynamics and ETH Zurich, human-robot interaction building on MIT Media Lab and Honda Research Institute studies, and medical robotics paralleling efforts at Johns Hopkins Applied Physics Laboratory and Intuitive Surgical.

Facilities and Technology

Facilities include motion-capture arenas similar to those used by Walt Disney Animation Studios and Industrial Light & Magic, indoor flight rooms akin to MIT and ETH Zurich setups, testbeds for autonomous ground vehicles comparable to Stanford Racing Team and Carnegie Mellon Red Team environments, and machine learning clusters paralleling compute resources at Google Cloud, Amazon Web Services, NVIDIA, and Microsoft Azure. Instrumentation features high-speed cameras used by National Institutes of Health studies, force-torque sensors like those in ABB and KUKA labs, robotic arms comparable to products from Universal Robots and Fanuc, and GPU clusters reflecting infrastructure at NVIDIA Research, Intel Nervana, and Google TPU research efforts.

Notable Projects and Contributions

Notable projects include multi-robot coordination experiments that relate to the DARPA Network Challenge and DARPA SubT Challenge, quadrotor research comparable to work at University of Maryland and ETH Zurich, dexterous manipulation studies analogous to projects at OpenAI and Shadow Robot Company, and object recognition systems evaluated at ImageNet and presented at CVPR and ICCV. Contributions span algorithms that influenced ROS-based ecosystems adopted by Willow Garage, locomotion controllers resembling those from Boston Dynamics and MIT Biomimetics, and perception pipelines that informed systems at Tesla Autopilot, Waymo, and Cruise LLC.

Education and Outreach

The lab hosts graduate courses coordinated with departments linked to Wharton School, School of Engineering and Applied Science, and programs interfacing with Coursera, edX, IEEE Education Society, and AAAI workshops. Outreach includes K–12 initiatives modeled after programs at Carnegie Mellon, summer camps similar to MITx and Stanford Pre-Collegiate Studies, public demonstrations at venues like Smithsonian Institution and The Franklin Institute, and participation in robotics competitions such as FIRST Robotics Competition, RoboCup, and the International Aerial Robotics Competition.

Partnerships and Industry Collaborations

The lab maintains collaborations with industry partners including Google, Amazon, Microsoft, Intel, NVIDIA, Toyota Research Institute, Honda Research Institute, Boeing research groups, and defense partners such as DARPA and U.S. Department of Defense. Academic partnerships include exchanges with MIT, Stanford University, Carnegie Mellon University, ETH Zurich, Imperial College London, and University of Tokyo, plus consortiums involving National Science Foundation, Air Force Research Laboratory, and European Research Council projects.

Awards and Recognition

Faculty and alumni have received awards connected to IEEE, ACM, AAAI, NSF CAREER Awards, DARPA recognitions, and fellowships from MacArthur Foundation, Fulbright Program, Guggenheim Foundation, and prizes presented at NeurIPS and ICRA. The lab's publications have been honored at conferences including CVPR, ICCV, ICRA, RSS, and NeurIPS, and alumni have founded companies and taken leadership roles at organizations such as Google DeepMind, OpenAI, Boston Dynamics, Cruise LLC, Waymo, and NVIDIA.

Category:Robotics research institutes