challenges when doing usability tests on physical devices af lars bo larsen, aalborg universitet

Temadag om evaluering af fysiske prototyper og produkter

Infinit Interessegruppen for usability og Interak8onsdesign

Delta, onsdag 16/1 2014

Infinit Workshop - Lars Bo Larsen

Program

9.30: Welcome / Infint and Delta 9.40: Brugerinvolvering og elektronisk skitsering af nye produkter

Morten Wagner, Delta 10.10: Challenges when doing usability tests on physical devices

Lars Bo Larsen, AAU 10.35 Pause 10.50 Usability of Medical Devices – Possibili8es and Challenges when

designing and tes8ng physical products Morten Purup, Radiometer Medical Aps

11.20 Exploratory user research with people that use drug delivery devices Chris Monnier og Peter Urban Novo Nordisk

12.00 Plenary Discussion 12.30 – 13. Lunch

16/01/14 p. 2 Infinit Workshop - Lars Bo Larsen

Challenges when doing usability tests on physical devices Outline:

•  Are there specific methods for usability tes8ng for physical devices?

•  What are the challenges when tes8ng on physical products?

•  Examples from mobile devices


Methods?

Looking at user tes8ng methods in textbooks, ar8cles, etc. reveals that:

While physical devices/products are oXen men8oned: •  No usability theories or methods are specifically targeted physical

devices •  Ie. “the usual” methods applies to physical products (which also

makes sense)

The conclusion must therefore be that the challenges are

opera8onal, rather than methodical: •  Logis8cs, costs, lab vs. field, mockups, test plan and tasks, etc.


Usability -‐ ISO 9241 defini8on

The effec,veness, efficiency and sa,sfac,on with which specified users achieve specified goals in par8cular environments.

•  effec,veness: the accuracy and completeness with which specified users can achieve specified goals in par8cular environments

•  efficiency: the resources expended in rela8on to the accuracy and completeness of goals achieved

•  sa,sfac,on: the comfort and acceptability of the work system to its users and other people affected by its use

-‐ universal


JUS Survey

A survey of UXPA’s Journal of Usability Studies for the past 8 years (130 ar8cles) years showed:

While about 10-‐15% of the papers present methodical approaches: •  Not one single paper addressed tes8ng of physical products as the

main methodical focus •  Very few papers (<5) present and discuss tests of physical products, (e.g.

“vo8ng machines”) -‐ however not from a methodological point of view •  A higher propor8on discussed mobile or medical devices, but mostly as

concrete studies, not methods

uxpa.org


Challenges

Physical products – contrary to soXware products, web, etc – typically require much more handling:

•  Difficult and expensive to distribute for field tes8ng and may en8rely prohibit remote tes8ng

•  Can pose problems for recording usability performance parameters

•  Prototypes can be quite expensive and 8me consuming to produce and might only exist in very few (one) instances – contrary to e.g. wireframing

•  Test users oXen need the whole device -‐ oXen not possible to separate the physical appearance from e.g. soXware driven interac8on (displays, etc)

•  SW and HW might not become available for tes8ng at the same 8me


Challenges -‐ prototyping

Building Mockups for tes8ng: •  Lo-‐Fi -‐ Very oXen done, not really a big problem:

16/01/14 p.8 Infinit Workshop - Lars Bo Larsen

Mockups

Solu8ons? •  Combining real device with simulated interac8on

Good solution, but only goes so far – and for some types of products


Challenges: Test design

Experimental designs become more complicated when both the hw and sw designs are subjected to usability tests simultaneously:

•  More variables to keep track of and decorrelate •  Solu8on would typically be to test separately as far as possible, but at some point everything must come together

16/01/14 Infinit Workshop - Lars Bo Larsen

Challenges – data recording

User’s Proximity Effects in Mobile Phones

Mauro Pelosi*, OndĜej Franek*, Mikael B. Knudsen**, Gert F. Pedersen* *Department of Electronic Systems, Aalborg University, Niels Jernes Vej 12, 9220 Aalborg Ø, Denmark,

[email protected], [email protected], [email protected] **Infineon Technologies, Denmark A/S, Alfred Nobels Vej 25, 9220 Aalborg Ø, Denmark,

[email protected]

Abstract— Thanks to a recent grip study, CAD models of the human hand have been generated, investigating user’s proximity effects in mobile phones. The simulation results show that the hand exhibits a major contribution in determining the total loss when compared to the phantom torso alone. The palm-handset gap is found to influence both absorption and mismatch loss.

Index Terms— Antenna proximity factors, body loss, efficiency, FDTD, hand phantom.

I. INTRODUCTION When mobile phones are used in close proximity with the human body, this results in a detrimental effect in its communication performances [1]. While it was shown that a SAM (Specific Anthropomorphic Mannequin) phantom can well represent the user’s torso in average sense [2], the hand modelization still encounters some practical difficulties [3], [4]. Though some standardization bodies have already proposed some preliminary hand phantoms, they utilize a hand grip that is not supported by grip studies [3]. Thanks to a recent contribution [5] it was possible to generate more detailed hand models. The objective of this work is to investigate through FDTD (Finite Difference Time Domain) simulations the user’s proximity effects for talk mode in mobile phones, focusing on both absorption and mismatch loss and isolating the contribution of both user’s torso and hand to the total loss. Moreover the influence of the palm-handset gap will be investigated too. This paper is structured as follows: Section II describes the grip study and its main results. Section III illustrates the used procedure to generate proper hand phantoms. In Section IV all the FDTD simulations are presented. Section V finally summarizes our conclusions.

II. GRIP STUDY DESCRIPTION A recent contribution within the COST Action 2100 [5] reports a first grip study for talk and data modes in mobile phones, where a rigorous investigation methodology was used over a sample population of 100 subjects; thanks to an unobtrusive data acquisition system (figures 1, 2) and a proper investigation protocol most of the experimental biases were minimized, allowing the collection of stable and comprehensive statistics. The index finger’s location was confirmed to be in the back region of the handset in most cases (figure 3). The palm-handset distance was indirectly

estimated based on grip style, fingers’ contact points and relative anthropometric properties. A proper categorization procedure led to the identification of two main ways of holding mobiles while talking, naming them “firm” and “soft” grip styles respectively. In the “firm” grip style the fingers are placed around the handset so that while the intermediate phalanges touch its side, the distal ones reach its front region, with a corresponding palm-handset gap that does not exceed the length of the longest proximal phalanx. In the “soft” grip style the hand holds the handset only with the distal phalanges, creating an air gap between the palm and the handset that does not exceed the length of the thumb.

Fig. 1. Test room [5].

Fig. 2: Example of videotape screenshot for talk mode (only 12/21 webcams are displayed here) [5].

1022

Grip study of mobile phones at AAU labs •  Data capture using 21 webcams •  100 test persons. Tasks were texting, making and answering calls From: “User's proximity effects in mobile phones” by: Mauro Pelosi, Gert F. Pedersen et al. 3rd European Conference on Antennas and Propagation, 2009

User’s Proximity Effects in Mobile Phones

Mauro Pelosi*, OndĜej Franek*, Mikael B. Knudsen**, Gert F. Pedersen* *Department of Electronic Systems, Aalborg University, Niels Jernes Vej 12, 9220 Aalborg Ø, Denmark,

[email protected], [email protected], [email protected] **Infineon Technologies, Denmark A/S, Alfred Nobels Vej 25, 9220 Aalborg Ø, Denmark,

[email protected]

Abstract— Thanks to a recent grip study, CAD models of the human hand have been generated, investigating user’s proximity effects in mobile phones. The simulation results show that the hand exhibits a major contribution in determining the total loss when compared to the phantom torso alone. The palm-handset gap is found to influence both absorption and mismatch loss.

Index Terms— Antenna proximity factors, body loss, efficiency, FDTD, hand phantom.

I. INTRODUCTION When mobile phones are used in close proximity with the human body, this results in a detrimental effect in its communication performances [1]. While it was shown that a SAM (Specific Anthropomorphic Mannequin) phantom can well represent the user’s torso in average sense [2], the hand modelization still encounters some practical difficulties [3], [4]. Though some standardization bodies have already proposed some preliminary hand phantoms, they utilize a hand grip that is not supported by grip studies [3]. Thanks to a recent contribution [5] it was possible to generate more detailed hand models. The objective of this work is to investigate through FDTD (Finite Difference Time Domain) simulations the user’s proximity effects for talk mode in mobile phones, focusing on both absorption and mismatch loss and isolating the contribution of both user’s torso and hand to the total loss. Moreover the influence of the palm-handset gap will be investigated too. This paper is structured as follows: Section II describes the grip study and its main results. Section III illustrates the used procedure to generate proper hand phantoms. In Section IV all the FDTD simulations are presented. Section V finally summarizes our conclusions.

II. GRIP STUDY DESCRIPTION A recent contribution within the COST Action 2100 [5] reports a first grip study for talk and data modes in mobile phones, where a rigorous investigation methodology was used over a sample population of 100 subjects; thanks to an unobtrusive data acquisition system (figures 1, 2) and a proper investigation protocol most of the experimental biases were minimized, allowing the collection of stable and comprehensive statistics. The index finger’s location was confirmed to be in the back region of the handset in most cases (figure 3). The palm-handset distance was indirectly

estimated based on grip style, fingers’ contact points and relative anthropometric properties. A proper categorization procedure led to the identification of two main ways of holding mobiles while talking, naming them “firm” and “soft” grip styles respectively. In the “firm” grip style the fingers are placed around the handset so that while the intermediate phalanges touch its side, the distal ones reach its front region, with a corresponding palm-handset gap that does not exceed the length of the longest proximal phalanx. In the “soft” grip style the hand holds the handset only with the distal phalanges, creating an air gap between the palm and the handset that does not exceed the length of the thumb.

Fig. 1. Test room [5].

Fig. 2: Example of videotape screenshot for talk mode (only 12/21 webcams are displayed here) [5].

1022


Summing up

Usability tes8ng of physical products does not really call for specific theore8c or methodical approaches and has not drawn much special alen8on in the community However, there are many prac8cal problems not present with pure sw products:

•  Logis8cs •  Mockups •  Data recording •  Costs •  Complexity •  …..


Plenary Discussion

•  Are you going to do a usability test on a physical product – what problems are you facing?

•  From your experiences, can you recommend solu8ons, tools etc for some of the challenges you’ve met?

16/01/14 13

Infinit Workshop - Lars Bo Larsen

challenges when doing usability tests on physical devices af lars bo larsen, aalborg universitet

Technology