PR2 (robot)

PR2 (robot)
Name	PR2
Manufacturer	Willow Garage
Year	2010
Type	Mobile manipulation robot
Autonomy	Semi-autonomous

PR2 (robot) is a mobile manipulation research platform developed to accelerate robotics research and application in domestic and service environments. Designed and produced by Willow Garage, the platform integrated sensing, computation, manipulation, and mobile base capabilities to enable advanced experiments in perception, cognition, and human-robot interaction. The robot served as a common hardware and software baseline for universities, research institutes, and companies worldwide, influencing standards in open-source robotics and collaborative research initiatives.

Overview

PR2 was intended as a general-purpose research platform bridging efforts in robotics labs such as Stanford University, University of California, Berkeley, Carnegie Mellon University, Massachusetts Institute of Technology, and industrial partners like Google and NASA. The platform combined mobility features influenced by autonomous vehicle research at DARPA Grand Challenge teams and manipulation design practices from industrial manipulators like those by KUKA and ABB. PR2 deployments often participated in competitions and workshops associated with organizations such as IEEE and ACM, informing curricula at institutions including University of Washington and Caltech.

Design and Specifications

PR2 featured a differential-drive mobile base, dual 7-degree-of-freedom arms, a pan-tilt head, and a payload bay for computation and batteries. Its kinematics were compatible with planning frameworks used by researchers from University of Oxford and ETH Zurich, enabling integration with motion planners developed in labs such as MIT CSAIL and CMU Robotics Institute. The platform supported ROS middleware versions advanced by contributors from Willow Garage and the Open Source Robotics Foundation, facilitating interoperability with tools from Google DeepMind collaborators and vision stacks inspired by work at University of Toronto and University of Montreal.

Hardware Components

PR2's sensor suite included high-resolution cameras, depth sensors influenced by innovations from Microsoft Research and hardware like PrimeSense modules, laser rangefinders akin to those used by teams at Stanford Racing Team, and inertial measurement units similar to those in iRobot products. Manipulation hardware incorporated custom end-effectors comparable to designs by Barrett Technology and Schunk, while motor controllers and actuators reflected engineering practices from Yaskawa and Mitsubishi Electric. Onboard compute used processors and GPUs paralleling platforms from Intel and NVIDIA, with batteries and power systems designed following guidelines from IEEE Standards Association committees.

Software and Control Systems

PR2 was tightly integrated with the Robot Operating System developed at Willow Garage and supported community contributions from research groups at University of Tokyo, University of Freiburg, and University of Pennsylvania. Control stacks implemented inverse kinematics libraries comparable to those in work by Kavraki Lab and motion planning algorithms like those from the OMPL project and researchers at ETH Zurich. Perception pipelines utilized machine learning models influenced by breakthroughs at University of Oxford (DeepMind collaboration), University College London, and Carnegie Mellon University vision groups, incorporating object recognition, simultaneous localization and mapping techniques from SLAM research and datasets curated by teams at Stanford and MPI-SHH. Development practices tied into version-control workflows promoted by GitHub and continuous-integration methods used at Google Research.

Applications and Research Contributions

PR2 enabled experiments in household manipulation, assistive robotics, long-term autonomy, and human-robot collaboration that informed projects at Harvard University, Yale University, Brown University, and Johns Hopkins University. The platform contributed to robotic learning studies influenced by algorithms from DeepMind, reinforcement learning benchmarks from OpenAI-adjacent research, and transferable grasping strategies studied at Cornell University and UC Berkeley AI Research (BAIR). Use cases included service robotics demonstrations at conferences like IROS and ICRA, participation in challenges organized by DARPA and industry showcases at CES.

Development History and Variants

PR2 originated within Willow Garage following earlier robotic efforts such as the Stanford Personal Robotics Project and collaborations with teams at SRI International and HRI labs. Over time, variants and successor platforms drew inspiration from efforts at Toyota Research Institute, Facebook AI Research, and startups emerging from Silicon Valley incubators. Spin-off platforms adapted PR2 concepts for industrial settings at companies like Amazon Robotics and healthcare robotics at institutions such as Mayo Clinic and Johns Hopkins Hospital. The open dissemination of design files and software fostered forks and derivatives produced by research groups at Tokyo Institute of Technology, Seoul National University, and Tsinghua University.

Reception and Impact on Robotics

PR2's release catalyzed adoption of ROS across academia and industry, influencing standards and pedagogy in robotics programs at Imperial College London, Technische Universität München, and KTH Royal Institute of Technology. Reviews in outlets covering technology and engineering referenced its role in democratizing access to advanced manipulation hardware and accelerating reproducible research, echoing policy discussions at venues such as National Academy of Engineering forums. The platform's legacy persists in contemporary service robots developed by Boston Dynamics, research frameworks at the Open Source Robotics Foundation, and curricular materials used in courses at University of Michigan and Purdue University.

Category:Robots