2003 Conference Proceedings

Go to previous article 
Go to next article 
Return to 2003 Table of Contents 


Ifedayo A., Adelola *, #
Email: ifedayo.adelola@ul.ie

Abdur Rahman *
Email: abdur.rahman@ul.ie

Sara L., Cox #
Email: sara.cox@ul.ie

*Department of Electronics and Computer Engineering
#Department of Mechanical and Aeronautical Engineering
University of Limerick, Limerick, Ireland.


The use of a powered wheelchair has been shown to help children with severe disabilities to explore and interact with their environment. Their ability to interact with the environment will improve their cognitive and physical development. These children, however, need to be trained to use electric wheelchairs. Conventional wheelchair training and rehabilitation is expensive and sometimes unsafe. A virtual reality (VR) based training system will resolve the problem of cost and safety. The wide variety of disabilities and preferences requires that the specific needs of each child be recognised and acknowledged during VR based training. It is vital to evaluate and model the performance and abilities of users in order to fulfil this requirement. This paper therefore proposes the use of conjoint analysis technique, in the Virtual Environment Mobility Simulator (VEMS), to assess and evaluate the needs of each individual in order to tailor the virtual environments to provide the most appropriate training scheme.


Childhood development has been shown to depend on the ability to independently explore, manoeuvre and interact with the environment (Rosenbloom, 1975; Bertenthal et. al., 1984). Spatial perceptual and cognitive abilities are developed during self-locomotion (Downs and Stea, 1973). Children with severe disabilities often experience hindered motor development due to deficient self-locomotion. The use of a powered wheelchair has been shown to help reduce the loss of such developmental experiences (Furumasu et. al., 1996; Nisbet et. al., 1996). Children, however, need to be trained to use electric wheelchairs effectively.

Conventional wheelchair training and rehabilitation, for these children, is expensive and sometimes unsafe. A virtual reality based training system will reduce the cost and increase safety. Research has shown that there is a correlation between training in VR and the real world (Standen and Cromby, 1995; McComas et. al, 1998). The medical conditions of some children who have severe disabilities and the possible implications of VR technology compel the use of a non-immersion type VR based system for wheelchair training and evaluation. Generally, the varied nature of disability requires that the virtual environments be adapted to the specific training needs of each child. Evaluating and modelling the performance and abilities of users, as their wheelchair control needs are recognised during training, will help fulfil this requirement. Consequently, the research currently relates to the development of the VEMS incorporating an experimental and statistical technique known as Conjoint Analysis.


The VEMS system will deliver a simple and cost effective wheelchair training alternative, designed to help children learn while they play and vice versa. Accordingly, the goals set out for the VEMS development are to


Conjoint Analysis is a generic term coined and first introduced by Green and Srinivasan (1978) to explain the "models and techniques that emphasise the transformation of subjective responses into estimated parameters". These responses are due to variations in objective stimuli. Conjoint analysis is a natural development with a strong theoretical basis in mathematical psychology. It has wide applications in academics, marketing, economics and health care. The word conjoint indicates "jointly". Thus, Conjoint analysis is the study of the joint effects of varying attributes (or independent "factors") on an individual's ultimate decision or judgement. Each factor is classified into varying "levels". The subjects are required to make decision or judgements on several possible "profiles" that are obtained from the combination of the varying factors. The fundamental interest of using this methodology relates to the combination rule that guides an individual to judge between factor levels to yield the importance attached to a given profile. That combination rule produces the functional model, which facilitates the development of simulation models to optimally predict future judgement and decisions. Similar judgement abilities are required for powered wheelchair control.

Conjoint experiments traditionally use verbal presentations of profiles, though some have been pictorial. Notably, verbal presentations may inadequately portray realism, which is often required when making judgements. However, virtual reality is characterised by verbal and pictorial techniques that enhance presentations to create realism. VR technology provides immersion and non-immersion type three-dimensional environments that could help the subjects to experience the various features and challenges of the real world. Such features or challenges may be reproduced in the profile settings of the virtual environment.


Experimental procedures in conjoint analysis require factorial design approach to combine the given independent factors to yield the profiles of interest. Thus, fractional factorial design is applied. The subject's performance rating or perceived importance of a given multi-factor level profile combination can be measured using statistical techniques. Analytical procedures in conjoint analysis often involve multiple regression techniques. Hence, the multiple linear regression equation below suffices to evaluate the psychological judgement and performance of the subject.

Where,Y = the dependent variable (i.e. user rating)XN = the Nth independent variable (i.e. given factor level)d N = the regression coefficient for respective independent variableN  = the observed factor

The VR application of these principles involves the use of several basic tasks to test the user's ability to track the primary and secondary targets that are located in the three-dimensional environment. Each factor is characterised by a specific task requiring a particular wheelchair control skill. And every task contains three target levels (0, 1, 2). The number of secondary targets successfully tracked within a specific factor task determines the factor level. The number of primary targets successfully tracked within the profile determines the performance rating.


The preliminary results show significant correlation in the evaluation of performance during wheelchair training in the VEMS. Wheelchair control abilities and performance are measurable, and can be modelled using conjoint analysis technique. The predictability of the wheelchair driving performance of children with severe disabilities in VR systems will prove beneficial to future research in cognitive rehabilitation and assessment.


The proposed technique in the VEMS system will facilitate the


Research grant from the Sir Halley Stewart Trust, United Kingdom is greatly appreciated.


Bertenthal B.I., Campos J.J. & Barret K.C., (1984) Self-Produced Locomotion: An Organiser of Emotional, Cognitive and Social Development in Infancy, Continuities and Discontinuities in Development, pp. 175 - 209.

Cox S.L., Rahman A. & Desbonnet M., (1999) An Evaluation of VR as a Tool to Assist in Training Children with Disabilities to Use Electric Wheelchairs, Assistive Technology on the Threshold of the New Millenium. (C.Buhler and H Knops Eds.) IOS Press, pp. 233 - 238.

Downs R.M. & Stea D., (1973). Cognitive Maps and Spatial Behaviour: Process and Products, In Image and Environment. (Downs, R.M. and Stea, D. Eds.) Aldine Publishing Co. Chicago, pp. 8-26.

Furumasu J., Guerette P. & Tefft D., (1996). The Development of a Powered Wheelchair Mobility Program for Young Children, Technology and Disability 5, pp. 41 - 48.

Nisbet P., Craig J., Odor P. & Aitken S., (1996). Smart Wheelchairs for Mobility Training, Technology and Disability 5 (1996) pp. 49 - 62.

Rosenbloom L., (1975) The Consequences of Impaired Movement-A Hypothesis and Review, Movement and Child Development, Clinics in Developmental Medicine. (Holt, K.S. Ed) No 55. S.I.M.P. and Heinemann Medical, London.

Standen P.J. & Cromby J.J., (1995) Can Students with Developmental Disability use VR to Learn Skills Which Will Transfer to the Real World. Proceedings CSUN VR Conference. California State University, Northridge.

Go to previous article 
Go to next article 
Return to 2003 Table of Contents 
Return to Table of Proceedings

Reprinted with author(s) permission. Author(s) retain copyright.