Modeling people with motor disabilities to empower the automatic accessibility and ergonomic assessment of new products Nikolaos Kaklanis, Georgios Stavropoulos, Dimitrios Tzovaras * Information Technologies Institute, Centre for Research and Technology Hellas, 6th Klm. Charilaou e Thermi Road, P.O. BOX 60361, GR-570 01, Thessaloniki, Greece article info Article history: Received 11 September 2013 Accepted 27 April 2015 Available online 20 May 2015 Keywords: User modeling Virtual User Model Regression analysis abstract Virtual User Models (VUMs) can be a valuable tool for accessibility and ergonomic evaluation of designs in simulation environments. As increasing the accessibility of a design is usually translated into addi- tional costs and increased development time, the need for specifying the percentage of population for which the design will be accessible is crucial. This paper addresses the development of VUMs repre- senting specific groups of people with disabilities. In order to create such VUMs, we need to know the functional limitations, i.e. disability parameters, caused by each disability and their variability over the population. Measurements were obtained from 90 subjects with motor disabilities and were analyzed using both parametric and nonparametric regression methods as well as a proposed hybrid regression method able to handle small sample sizes. Validation results showed that in most cases the proposed regression analysis can produce valid estimations on the variability of each disability parameter. © 2015 Elsevier Ltd and The Ergonomics Society. All rights reserved. 1. Introduction It is important to realize that people with disabilities are not a tiny minority of the population. Around 640 million people worldwide live with a disability (Enable, 2008). Many strategies have been undertaken worldwide to enforce accessibility for peo- ple with disabilities. For instance, the European Commission has launched a number of initiatives (Klironomos et al., 2006; Bühler and Stephanidis, 2004), such as the eEurope 2002, the eEurope 2005, the EDeAN network and the DfA@eInclusion Coordination Action (  Sim sík and Galajdov a, 2008) to promote the “Design For All” (DfA) principles that will lead to more accessible services and systems. The Commission has also advocated the use of standard- ization to improve the effectiveness and uptake of DfA. It mandated CEN, CENELEC and ETSI to include the “Design for All” concept in relevant standardization initiatives. Digital Human Modeling (DHM) can significantly support designers and developers on staying compliant with the aforementioned strategies, as it reduces the need for physical prototypes, and also enables the incorpora- tion of ergonomics science and human factors engineering princi- ples earlier in the design process. DHM and simulation has gained importance in the past few years (Phillips and Badler, 1988; van der Meulen and Seidl, 2007; Porter et al., 2004; Lind et al., 2008; Feyen et al., 2000) and al- lows designers to easily observe and evaluate the interaction of the designed product with a virtual user having specific needs and/or preferences. Simulation offers designers the opportunity to explore how a new system would behave before the real prototype is developed, or how an existing system would perform if altered, thus, reducing development time and costs. Moreover, there are cases where simulation is the only method of verifying that a design concept is acceptable to a prescribed population, since hardware prototypes are not available (Chaffin, 2001). The core concept of this paper is to empower the accessibility and ergonomics of new products by introducing a novel user modeling technique for people with disabilities, based on statistical analysis. The main goal of the paper is the development of accurate Virtual User Models (VUMs) representing specific disabled popu- lation groups. Measurements were initially obtained from a total of 90 subjects having the following disabilities: Parkinson's disease, stroke, multiple sclerosis and cerebral palsy. Statistical analysis was performed on these measurements using both parametric and nonparametric regression analysis, as well as a proposed hybrid regression method able to handle small sample sizes by using both real and virtual samples along with information found in the literature. The accuracy and validity of the final regression results * Corresponding author. Tel.: þ30 2311 257777; fax: þ30 2310 474128. E-mail addresses: nkak@iti.gr (N. Kaklanis), stavrop@iti.gr (G. Stavropoulos), Dimitrios.Tzovaras@iti.gr (D. Tzovaras). Contents lists available at ScienceDirect Applied Ergonomics journal homepage: www.elsevier.com/locate/apergo http://dx.doi.org/10.1016/j.apergo.2015.04.016 0003-6870/© 2015 Elsevier Ltd and The Ergonomics Society. All rights reserved. Applied Ergonomics 51 (2015) 120e136