AN AFFORDABLE MOBILE ROBOTIC MANIPULATOR.

Dr Michael Hillman CEng. Bath Institute of Medical Engineering, Wolfson Centre, Royal United Hospital, Bath BA1 3NG. UK

ABSTRACT

A robotic manipulator has been mounted on a wheeled trolley. This novel approach aims to overcome some of the limitations of a fixed site workstation system, while keeping the cost within reasonable limits. Preliminary user experience with the system is encouraging.

BACKGROUND

There is obvious potential for using robotic technology to assist those who because of accident, illness or congenital defects do not have full use of their hands or arms. Although the application of robotics to help people with disabilities has been investigated for many year, very few devices have left the research environment and even fewer can claim to have been commercially successful. Several reasons cam be suggested for this failure. One reason may be that few of the devices have been designed as commercial products.

RATIONALE

Our own experience has come from the development of a workstation system [1]. While being able to successfully move and manipulate objects on and around the workstation there were obviously limitations. The main barrier was the physical size of the system. While the size of the workstation was based on a standard office desk, this was still too large for a domestic environment. Additionally it was found that within a home it was limiting for a robotic system to be used in a single room. Rather it should be available for use for different tasks in different rooms.

The idea of a workstation system may be extended in several ways. The aim of mobility has been often achieved either by the use of a mobile robot or by mounting a robotic manipulator on a wheelchair. The approach which we have been following, in the hope of achieving a substantially cheaper product, has been to mount a robotic manipulator on a simple trolley. This may be moved around the home to various local "work sites" by a carer. Alternatively it has been suggested that it might be possible for the trolley to be clipped to a wheelchair so that it might be moved around by a wheelchair user.

DESIGN IMPLEMENTATION

The aim of the design is to produce a stand alone assistive robot which is mobile, yet low cost.

The overall geometrical arrangement was retained from the earlier workstation system. This uses a SCARA geometry for the upper arm with a linear vertical actuator. The manipulator is mounted on a wheeled trolley. 5cm diameter braked castors are adequate for use over carpet, slightly offset from the cabinet at one end for stability when the manipulator is extended.

Vertical actuator

The use of a linear bearing, incorporating a carriage running on a rectangular profiled rail gives low resistance, yet can cope with high off-centre loads. The vertical travel is 0.4m. A constant tension spring balances the weight of the whole manipulator.

Upper arm

The upper arm structure consists of three vertical axes, all to a common design. Since none of the actuators act against gravity low powered motors can be used (4.2W). Thus the manipulator is not powerful enough to harm anyone, yet is able to carry a reasonable load of 2kg. The internal aluminium structure has a cosmetic cover which allows easy access to the mechanical and electrical components.

Wrist pitch/roll assembly

Pitch and roll are implemented by a reverse differential gear system. Currently the end effector is a two motor prosthetic hand, loaned to us by Hugh Steeper Ltd (London, UK).

Electronic control

The motor control boards are mounted within the upper arm structure. The power supply and I2C serial link from processor to motor control boards run within the manipulator structure. The motor control boards are 5cm x 5cm and use the HP HCTL1100 motor controller. The Pulse Width Modulated output from this is taken to a current limited amplifier to drive the motor.

Aesthetics

Since the manipulator uses an internal structure covered by a cosmetic vacuum moulded cover, it is relatively easy to change the appearance of the upper arm. The trolley is designed so that different colours and finishes can be investigated simply by sliding in new side panels. The manipulator currently has a white and grey trolley with white upper arm and grey vertical post.

User Interface

The user interface uses a scanning menu system on a LCD display. Various parameters of the scanning system (eg number of switches used for input, speed of scanning) may be varied to suit the control abilities of individual users. The LCD unit (19cm x 13cm) is attached to the trolley by a 2m cable, and may be positioned as near to the user as necessary. The manipulator may be controlled directly in real time, though this is a slow and cumbersome method. When the manipulator is "docked" at a local work site, pre-set routines may be used to decrease the control burden on the user. These pre-set routines may be created by the user himself. When replaying a routine the movement may be interrupted to allow for adjustment of the position of the manipulator.

EVALUATION

One of the priorities of the design of the system was that it should be a stand alone product which could be easily transported to, and left at, the home of a user. The system should be usable without an engineer or other professional present.

Initial evaluation results have confirmed that the system meets these requirements. Several improvements have been recommended. In particular it is necessary that the user interface parameters should be variable by the user himself. Continuing trials should show whether such a system can cost effectively perform a useful function for disabled users.

DISCUSSION

The aim of this project has been to build a relatively low cost, mobile manipulator which can perform a useful function in the home of a disabled user.

Mobility

Although systems have been built using advanced autonomous mobile robots, such sophistication is difficult to justify for an assistive robot. A more useful comparison of the current system is with a wheelchair mounted manipulator. The advantages of working in a structured work site must be compared with the advantages of having the manipulator always available to the user.

Ease of use

The system does not require complicated setting up or subsequent intervention by engineers or other professionals.

User interface

While not being sophisticated, the user inteface can be readily used by people with a wide range of abilities. Pre-set routines are widely used, but with the facility for the disabled user to create and modify such routines. Although not implemented yet, it is intended that the user should be able to set up and adjust the parameters of the user interface system.

Low cost

The system has been designed so that it can be produced in small batch quanitities, at relatively low cost. Although production development still needs to be finalised we are hopeful of reaching our target price of under oe10000.

Usefulness

Ultimately any assistive device must be judged on whether it can cost effectively perform a useful function for disabled users. Continuing evaluations should provide a realistic answer to this question.

REFERENCES

[1] Further development of a robot workstation for the disabled. G Pullin, M Hillman, A Gammie, R Orpwood, C Stammers. Proc RESNA 14th Annual Conference, Kansas City, 1991, pp21,22

ACKNOWLEDGEMENTS

The author wishes to acknowledge the contribution made to the project by Graham Pullin in originating the concept of a trolley mounted assistive robot. Acknowledgements are also due to others at the Institute who have been part of the project - Martin Rouse (Mechanical Workshop), Peter Laidler (Electronics Workshop), Andrew Gammie and Sean Hagan (Electronics Design), Jill Jepson (Occupational Therapist) and Roger Orpwood (Head of Engineering).

Dr Michael Hillman CEng. Bath Institute of Medical Engineering, The Wolfson Centre, Royal United Hospital, Bath BA1 3NG. UK. Tel. 01225-824103 Fax. 01225-824111 email. bime@bath.ac.uk