Autonomous Mobile Manipulation for Healthcare

Charlie Kemp, Georgia Tech

Robots have the potential to revolutionize healthcare by providing economical and reliable assistance 24 hours a day, 7 days a week. In this talk, I will give an overview of my lab's research, which seeks to develop autonomous mobile robots capable of providing physical assistance to patients and healthcare workers. Our efforts have resulted in three prototype robots that can perform a variety of assistive manipulation tasks, including retrieving objects, delivering objects, opening doors, opening drawers, and turning on lights.

Three key questions drive this research: what healthcare tasks would be valuable for robots to perform; how can users effectively direct robots to perform these tasks; and how can robots reliably perform these tasks in environments relevant to healthcare, such as the home? To help answer these questions, we have taken inspiration from helper monkeys, service dogs, and human assistants. We have also conducted user studies with ALS patients and Atlanta area nurses. In our experiments, ALS patients and nurses intuitively and effectively used our robots. Reliable operation in real healthcare settings, however, remains a challenge. To overcome this challenge, we are exploring a variety of approaches, including characterizing real-world variation, modifying the environment, and improving robotic perception.

Building Blocks for Mobile Manipulation

Brian Gerkey, Willow Garage

Mobile manipulation is perhaps the most integrated, multi-disciplinary area of robotics research to date. Experimental work in this field requires expertise in all aspects of robotic perception, planning, and control. Yet most researchers have very focused areas of experience and interest, outside of which their time is not well-spent. A key problem in accelerating the pace of research in mobile manipulation is the construction and maintenance of a commons of building blocks that allow researchers to remain focused on their specialties and yet synthesize and experiment with fully-integrated systems. I will describe the ROS commons, and some of the key building blocks that we have developed to date.

Justin - a Test Platform for Mobile Manipulation in Space and Service Applications

Alin Albu-Schaeffer, DLR

For interacting with unknown environments, the DLR robots have features which distinguish them substantially from other mobile manipulation systems. Such features are a high load/weight ratio of 1:1 for arms and hands, link side torque measurements, active vibration damping, sensitive collision detection and reaction strategies as well as joint and Cartesian impedance control in all joints of the upper body. Moreover, a novel kinematic solution for the mobile platform allows stability during manipulation of high loads as well as high manoeuvrability in narrow passages. The talk gives an overview on the hardware design, on the soft robotics control approaches used for interaction and manipulation as well as on some application examples.

Manipulation Planning

Siddartha Srinivasa, Intel Research

I will present my lab's work on planning and executing manipulation tasks in real-world environments with high clutter, on the HERB platform. The talk will bring together our work on adapting grasp to clutter, extending randomized planners to constraint manifolds, as well as some new work on functional gradient optimization in high-dimensional spaces.

Robots for the Human Environment

Oussama Khatib, Stanford University

Robotics is rapidly expanding into the human environment and vigorously engaged in its new emerging challenges. From a largely dominant industrial focus, robotics has undergone, by the turn of the new millennium, a major transformation in scope and dimensions. This expansion has been brought about by the maturity of the field and the advances in its related technologies to address the pressing needs for human-centered robotic applications. Interacting, exploring, and working with humans, the new generation of robots will increasingly touch people and their lives, in homes, workplaces, and communities, providing support in services, entertainment, education, health care, and assistance. The discussion focuses on new design concepts, novel sensing modalities, efficient planning and control strategies, modeling of human motion and skills, which are among the key requirements for safe, dependable, and competent robots.

Mobile Manipulation for Everyday Setting

Henrik Christensen, Georgia Tech

Gradually we are deploying mobile manipulation system for operation in everyday environments. This poses a number of interesting challenges in terms of navigation, integration with manipulation, detection and modelling of objects, grasp planning and execution of entire missions. In this presentation we will discuss an overall architecture for mobile manipulation for robust execution of mission. In addition we will cover methods for autonomous object modeling, model based visual servoing, grasp planning and integrated control of mobility and manipulation. The system has bene implemented using a Segway RMP200 with an on-board KUKA LWR manipulator. The system uses an on-board stereo system in combination with several laser scanners. Real-world examples will be demonstrated in realistic home environments.