out when I felt like it, sort the laundry and start the washing machine, tidy up, get a book ... [Online]. URL: http://icorr99.stanford.edu/attendees/papers/eftring.html.
The Swedish experience of Rehabilitation Robotics Håkan Neveryd 1, Håkan Eftring 1 and Gunnar Bolmsjö 2 1
Certec, Division of Rehabilitation Engineering, Department of Design Sciences, Lund University 2 Division of Robotics, Department of Mechanical Engineering, Lund University
Over the last ten years, there has been a great deal of research and development in Sweden in the field of rehabilitation robotics. The research effort has often focused on workplace adaptations. This paper deals with four robotic adaptations, all of which are in use. The current state of the research and development is also described.
Mr Christer Evaldsson’s robotic workstation As a result of a fall at the construction site where he was working as a carpenter, Christer Evaldsson suffered a spinal cord injury that left him quadriplegic. After his injury, Christer continued to work for Skanska in Helsingborg, doing office work, such as invoicing. This work involved producing an invoice based on reference documents. In the office where Christer was working, there was a sheet conveyor which, although designed and redesigned several times, would sometimes send two or more sheets (reference documents) at the same time. The only solution was for Christer’s personal assistant to put out two or three documents in front of him. In two to four minutes, Christer would produce the invoices and the assistant had to remove the documents and put out new ones. After the robotization of the workstation the robot handled the coming and going previously done by the assistant, see figure 1. The input device to the company’s computer system was a scanning system controlled by means of a sip and puff device. The robot’s computer was controlled through a microphone with the aid of a voice recognition system and an early version of CURL for DOS. To make it possible to change end-effectors, a tool changing system was modified. The end-effector used was a specially designed “suction device” whose task was to hold the reference document. The “suction device” consisted of three suction cups. The robot was installed in October 1990 and is used for about four hours each day.
Figure 1. Digitized video images from VHS showing Christer Evaldsson at his robotic workstation.
Partly on the basis of experience gained from the RAID project, the workstation was modified to some extent in 1994, see figure 2. Voice control was removed. The input device to the computer systems was modified from a scanning system to a manually controlled system by means of a combined mouth joystick and sip and puff device. The tool changing system was removed and a new end-effector was manufactured. The new end-effector is based on the principle of first lifting a reference document using a suction cup and then clamping it. A readerboard was mounted on the workstation. A whole sheaf of reference documents can be placed on the readerboard. When Christer has finished producing an invoice from the first reference document, the robot begins to lift it off the readerboard. This means that Christer can start producing a new invoice immediately.
Figure 2. Christer Evaldsson’s modified workstation.
What we learned from the project: • It is important to establish at the outset who is responsible for service. Train service staff, for example one or several technicians working at the company in question. • Ensure that funding is available for updating the system if/when the user’s needs, disability, etc., change. Funding should also be available for updating the system in case of other external changes, for instance when the computerized robotic system must be made compatible with new versions of the company’s computer system. • No time limit should be imposed on the project. The user should be able to use the equipment during its entire useful life. • It is an advantage if the user can act as a teacher in training programs and as a test user in other robot projects. • The user has an important role as a sounding board for researchers and developers on the subject of the possibilities afforded by robotics for people with a motor disability.
Ms Ann-Christine Olsson’s robotic workstation Ann-Christine Olsson works part-time sorting and redirecting mail in the mailroom at Volvo Transport in Göteborg. Ann-Christine has limited mobility in her arms and hands and a speech impairment as a result of cerebral palsy. Because of restructuring, Ann-Christine’s job was at risk. The National Insurance Office and the
Labor Market Institute made an inventory of possible jobs for Ann-Christine and found sorting and redirecting mail to be a suitable task. Her task has been robotized so that a CRS robot picks up an envelope from a letter tray and shows it to her. Ann-Christine reads the incorrect address and fetches the correct address from the address register in the computer. A printer prints out a new address label. Ann-Christine can choose where she wants the robot to place the label. The robot then presses the envelope against the label and puts the letter in the out-tray. Mr Henrik Lundblad’s robotic workstation Ever since he was a child, Henrik Lundblad has been very interested in machines. Henrik suffers from severe cerebral palsy since birth, which means that he only has the use of his left hand and that he is unable to speak. When he was 6 years old his parents bought him a go-cart which he drove around in the garden. After a while he learned how to drive a snowplough in front of and a trailer behind the go-cart. When Henrik was in secondary school, the Labor Market Institute in Göteborg began looking for a suitable job for him and found one at Alingsås Environmental Station splitting wood and sorting garbage. Henrik controls a crane with a grapple, which puts logs in a hydraulic wood splitting machine.
A National Rehabilitation Robotics Center Certec at Lund University and the Department of Rehabilitation at Lund University Hospital have established a National Rehabilitation Robotics Center. At the Center, people with severe physical disabilities, such as high-level spinal cord injuries, have the opportunity to try out the RAID workstation and the wheelchair-mounted Manus manipulator during their rehabilitation. Between May 1997 and May 1998, eight individuals tested the Manus manipulator for carrying out various tasks [1]. Seven of eight users answered a questionnaire. Only one user wanted to have a Manus arm as it looks and works today. The following improvements were mentioned: It should be mounted on the back of the wheelchair. It should be possible to use the wheelchair joystick to control the Manus arm. It should be smaller, lighter, easier to use and have more reach. It should be possible to lift heavier things. It should be possible to detect the weight of a grasped object. It should be possible to do short movements with high acceleration.
Ms Eva Almberg, a Swedish Manus user After the trials one of the users, Eva Almberg, decided to buy a Manus arm, and she received her Manus arm in November 1998, see figure 3. Eva has a congenital spinal muscular atrophy and only limited mobility in her right arm and hand. She has a personal assistant 24 hours a day.
Figure 3. Ms Eva Almberg and her Manus.
Here is an excerpt from a two-hour in-depth interview with Eva Almberg [2]: Q: What makes the Manus manipulator worth using in your opinion? A: It is very important to me to be able to do something without having to tell someone what it is I want to do. Being able to think, without having to talk about it. I can do that with the Manus manipulator. Even if it takes me three hours, at least I can do it. That means I don’t have to talk to someone for three hours, if I don’t want to. It might have taken me ten minutes to do it with the help of my assistant, but it doesn’t matter, because then I would have had to give instructions, and that might be just as tiring as spending those three hours. Q: What where your expectations before you bought the Manus? A: I had enormous expectations. I thought I would be able to get something to drink from the fridge and drink it, make tea, cook simple meals, make a sandwich, go in and out of my apartment, shut the door if I wanted to be left in peace and get out when I felt like it, sort the laundry and start the washing machine, tidy up, get a book and read it, put on a record, put on a movie, have coffee at a café, feed my cats, do practical things for my family and go shopping. I really couldn’t see any limit to the things I’d be able to do. I thought the main limitation at first would be my own fear of trying new things. I thought I’d be able to do anything a person in wheelchair using one arm is capable of doing. Q: And how did it turn out? A: When I thought about having a robotic arm I imagined it would bring a great deal of independence. I thought I would be able to manage on my own to a much greater extent than I am. On the other hand, I wouldn’t want to be without the Manus, since it has given me a certain amount of independence that I didn’t have before. One thing I didn’t think I would use it for is for drinking coffee even though my assistant is sitting next to me. I can spend more time on my own with the aid of the Manus, but not as spontaneously or as long as I thought I would. One thing the Manus has done for me is to inspire me to look for new technology that would make me more physically independent of other people. A modular wheelchair-mounted robot There is a need of developing new robotics concepts where people are working together with robots in highly unstructured environments. A way to meet these new conditions is to apply a modularized approach that allows for customization on an
individual basis. However, this concept reveal a need for new developments related to innovative solutions in both mechanical and controller design. One primary requirement was the possibility to mount the robot on an electrical wheelchair. The basic framework was defined as follows: Payload 1 kg not including the gripper in all positions. A change of the value up to 2 kg will not affect the design principle significantly. Workspace reach 0.75 m in all positions (floor, table) measured from the user. Low weight of the robot arm. Battery operation using the power of the wheelchair and low consumption. Motion control via standard interface, such as joystick as well as pre-defined motions. The design should take safety issues into consideration.
Figure 4. One of the robot’s links as an independent module.
The features required for integrating the drives in the robot links are: Make each joint as an independent module, see figure 4. Solution: Integrate the drive systems in the joints. Use high performance brushless motors with integrated planetary reducers. Minimizing the weight of each part. Solution: Select carbon fiber reinforced polymer materials and aluminum alloys. Integrate the motor controllers for each drive unit in each link to minimize the cabling in the robot arm. Solution: develop a small motor control unit for servo control and commutation of each motor. To facilitate “intelligent” distributed control, every two motor controllers share an 8-bit micro controller with CAN-bus interface, with exception for the wrist where all three wrist axes share one micro controller. The CAN-bus makes it possible to reduce the cabling through the robot arm to one 24 V power line, one ground line and two digital signal lines for the distributed joint level motor control. Precise simulations were used to evaluate effects of different designs with respect to required torques for the drive units of a “standard” work cycle. Simulations are used throughout the project to show the use of the robot in realistic situations, see figure 5.
Figure 5. Simulation of the modular wheelchair-mounted robot.
The weight of the robot arm is less than 12 kg, which will be possible to decrease. Furthermore, the workspace is larger than originally specified. The experimental platform has been built, including a stand-alone robot, and is in a validation phase for later integration with a wheelchair. During the validation, the simulation tool is used as a high level controller which makes it possible to test the motions defined in the user scenarios as well as the integration with different human machine interfaces. References [1] Eftring, H., & Boschian, K. (1999). Technical Results from Manus User Trials. Proc. International Conference on Rehabilitation Robotics (ICORR), 136-141. [Online]. URL: http://icorr99.stanford.edu/attendees/papers/eftring.html [2]
Eftring, H. (1999). The Useworthiness of Robots for People with Physical Disabilities. PhD Thesis, Certec, Department of Design Sciences, Lund University. ISBN 91-628-3711-7. [Online]. URL: http://www.certec.lth.se/doc/useworthiness/
