HoLLiE – The Assistance Robot

A Flexible and Mobile Service Robot as a Support for Everyday Tasks

The robot HoLLIE

The combination of robust state-of-the-art industrial components with a flexible software framework for environment perception and robot control enables HoLLiE to manage complex tasks easily and support people in everyday situations.

General Information

HoLLiE is a mobile, bi-manual service robot that was developed at the FZI Karlsruhe as part of the “House of Living Labs” (HoLL). In the future the robot will carry out different tasks within the HoLL to accompany visitors, provide assistance and be of service. It comes as no surprise, then, how we arrived at the robot’s name: HoLLiE is an Acronym for “House of Living Labs intelligent Escort“. HoLLiE’s mobile platform has a footprint of 80 x 80 cm so she fits through almost all doors. Standing upright, the robot is 142 cm tall. Thanks to its active torso, the shoulder height can vary between 96 and 124 cm. The complete weight including all batteries is approximately 160 kg. Nevertheless, the slim design gives HoLLiE a graceful appearance. With, arms, hands and shoulders stretched out, HoLLiE can span up to a width of 250 cm. The 49 active degrees of freedom offer an impressively huge workspace (yellow area in the image on the right).

Innovations and targeted objectives

One of HoLLiE's outstanding features is her ability to reach the floor by bending the body forwards, allowing her to pick up objects from the ground. This requires a motion planning software that not only considers potential collisions but also the tip-over stability of the robot.

The extensive functionality of the hardware components gives us a basis for conducting a wide range of research topics. Amongs other things, HoLLiE is a test platform for cutting-edge GPU-based real-time motion planning algorithms. These allow the robot to quickly adapt her motions to a dynamic environment. Quick reactions allow us to employ HoLLiE next to humans.

With the help of speech synthesis software, a microphone array and an innovative LED-based information system within the body, HoLLiE offers possibilities for human-machine interaction that goes well beyond simple information services. New tasks can be commanded to the robot via intuitive 3D gesture recognition. This also enables users to teach HoLLiE new trajectories, as the robot can mimic and store human motion sequences.

By constantly enhancing the abilities that make HoLLiE an universal service robot, we can realize plenty of new interesting research opportunities.

Specifications and facts

HoLLIE
  • Omnidirectional, mobile platform with Mecanum wheels
  • 2 Schunk Powerball lightweight arms with 6 degrees of freedom each
  • 2 DLR/HIT II Anthropomorphic hands with 15 degrees of freedom each
  • 3 additional degrees of freedom in the torso, two are actively actuated
  • 2 joints in the neck to freely pan and tilt the head with the cameras
  • 2 powerful embedded onboard computers
  • 2 SICK LMS 100 laser range finders just above ground
  • Microsoft Kinect 3D Sensor and stereo vision system for user interaction and environment perception
  • 360° LED ring within the lower torso, displaying 2 million colors to visualize information intuitively
  • Pleasing Design with 3D printed covers combined with soft fabric surfaces
  • Wireless operation time of several hours through large internal batteries, or continous operation via external power supply. On-the-fly switching is possible.
  • Virtual testing of all functionalities within a software simulation environment for safe development of new features

Application Scenarios

HoLLIE

HoLLie showed off her capabilities of gesture based control for the first time at the international CeBIT fair 2012 in Hanover, Germany. At ICRA 2013 in Karlsruhe, we demonstrated the vast improvement of HoLLiE’s workspace using her new actuated torso to grasp laundry from the floor.

Further application scenarios include:

  • Pick-up and delivery services in domestic environments, e.g. in Smart Homes of the future
  • Support for nursing and care services by taking over repetitive standard tasks, transportation jobs or acting as a telemedicine interface
  • Information and attendant services in larger public facilities such as museums, fairs, etc.

In the video HoLLiE demonstrates one of its outstanding features: The articulated body that enables the robot to grasp rigid objects and clothes from the floor.

Contact

Dipl.-Ing. Arne Rönnau

Department Manager

more

Career

Arne Roennau studied Electrical Engineering at the University of Karlsruhe (TH), today's Karlsruhe Institute of Technology (KIT), from 2002 to 2008. The focus of his studies was on control and robotics. His thesis was focussed on sensor-based 3D environmental modelling and foot point planning for the six-legged walking robot LAURON.

From 2008 to 2011, Arne Roennau worked as a robotics expert in the department Interactive Diagnosis and Service Systems (IDS) at FZI. During this time, he worked on the localization of mobile robots, real-time control of AGVs as well as the optimization of multi-legged walking robots.

His current work focuses on Open Source Software (OSS), in particular ROS (Robot Operating System), human-robot collaboration (HRC), design of innovative service robotics applications such as the BratWurst Bot as well as learning deep neural networks/AI. His most important concern is the evaluation and transfer of these technologies into industrial, real world applications. Arne Rönnau is project manager in numerous public and industrial projects and coordinator of several national research projects.

Since 2011, Arne Roennau heads the department Interactive Diagnostic and Service Systems (IDS) within the research division Intelligent Systems and Production Engineering (ISPE). Besides, Arne Roennau is also the head of the FZI Living Lab Service Robotics since 2012.

Publications

Meine Veröffentlichungen sind in meinem Google Scholar Profil zu finden.

zu den Publikationen

Contact

Phone: +49 721 9654-228
Fax: +49 721 9654-229
E-Mail: roennau@dont-want-spam.fzi.de

hide details