Mobility for All Can one size fit all

Mobility for All Can one size fit all?

Universal Access and the Universe of One • “There is no such thing as the average person. ” – Don Norman, The Design of everyday Things • Designers are often required to develop a single solution for everyone—but this is impossible • A little perspective – US Pop. = 281 million – People w/ severe disability = 12% or 33 million – Even if you could design for the 99 th percentile, 1% or 2. 8 million will be left out! • Universal access ≠ One size fits all

Designing with atoms vs. bits • The world of atoms is less malleable that the world of bits • Consider bus maps and schedules Reality of Printed Maps • Represent plans • Person must adapt to a one-size-fits-all representation • • Personalized maps are expensive—not scalable—not feasible? Cheap short-term cost—mass production

Designing with atoms vs. bits • The world of atoms is less malleable that the world of bits • Consider bus maps and schedules Reality of Printed Maps Potential of Digital Maps • • • Represent plans Person must adapt to a one-sizefits-all representation Personalized maps are expensive —not scalable—not feasible? Cheap short-term cost—mass production • • Represent reality Representation can dynamically adapt to the task and person Scalability is relatively cheap— distribution is simple Cheaper in the long-term?

User Models understanding the user • Different usage patterns between familiar and unfamiliar users Familiar Behavior Have learned (habituated) over time—they act Possible Errors • Capture errors—intended sequence is overridden by a familiar, unintended sequence Possible Fix • Evaluation—Plan vs. Actions • Timely feedback Unfamiliar Rely heavily on procedural knowledge—plan—wait— act • Errors in thought—error in the planning procedure • Eliminate need for planning • Guided planning • Claim: Usage patterns and needs are disability-dependant – Some people need constant reminders of the goal – Some people need to focus on the task steps and may be overloaded by excessive reminders

Using User and Task Models • Know the user – work with specialists to develop a taxonomy of cognitive (dis)ability. – Gardner’s Theory of Multiple Intelligences – develop an architecture to map representations onto a taxonomy of cognitive (dis)abilities • Know the user’s goals – context aware systems that will “know” what task they are trying to accomplish – what can be inferred from the environment?

Adaptable and Adaptive Systems • Systems need to be adaptive so they can keep pace as users improve and degrade in task performance. – routine tasks can become habituated—are some types of users more/less likely to habituate? – habituated tasks can be forgotten if not active or as disability degenerates—what types of disabilities are more/less likely to be affected? • Systems need to be adaptable so that caretakers can tailor the interaction experience for those in their care.

Envision model 1. Caretaker primes the system with formal and informal information about their charge Based on caretaker’s description, the user in mapped onto disability taxonomy and system acts based on that mapping Caretaker then acts as a moderator 2. 3. • • 4. continuously monitors the human-computer collaboration provides feedback to both human and computer. Caretaker’s feedback updates the user description, the person is remapped within the taxonomy, and the cycle continues

Envision model 1. Caretaker primes the system with formal and informal information about their charge Based on caretaker’s description, the user in mapped onto disability taxonomy and system acts based on that mapping Caretaker then acts as a moderator 2. 3. • • 4. Prescribe (2) Describe (1 & 4) continuously monitors the humancomputer collaboration provides feedback to both human and computer. Caretaker’s feedback updates the user description, the person is remapped within the taxonomy, and the cycle continues Evaluate (3)

Examples • Context Awareness – Knows where the user is located—GPS Coordinates -> Zip Code – Knows local weather – Knows that arriving bus looks similar to target bus -> avoid description errors – Knows that is Sunday—no work • User Model – Knows the user’s schedule – Knows what activities are available • At a given time • With given weather conditions – Knows that the users should not cross streets and generates appropriate route plan – Knows that the user tends to fall a sleep on buses—reminds bus driver to prompt user if necessary – Knows current activity is novel, but very similar to another highly familiar activity—get milk on the way home from work -> avoid capture errors

System Architecture (First Draft) Caretaker Describes Monitors Feeds back Activity d. B User Model Activity Planner User Selects/Performs Activities Zip code Data Activity Monitor Context Analyzer GPS Data Weather Data