Wearable Sensor Data-Driven Walkability Assessment for Elderly People

Round 1

Reviewer 1 Report

The author presented a method to estimate terrain walkability via a wearable IMU sensor on a smartphone, and the maximum Lyapunov exponent (MaxLE) is used as a measurement. The goal of the paper is to assess age-friendly good walkable environment for elderly people. 30 elderly participants were recruited to walk through 4 sites, either with or without illegal parking or on sidewalk. The research aims at answering whether different walking environments can be distinguished by the elderly's walking patters and reactions. A twenties group is also used as a comparison. The results from MaxLE showed that the site with with more difficult walking environment reveals higher MaxLE. The paper concludes that the proposed measurement is able to distinguish different environments of age-friendly walking environments.

1. The author could take a look at the Project Sidewalk (https://sidewalk-sea.cs.washington.edu/), which took advantage of Google Street View and crowdsourcing to collect walkability. Although the Project Sidewalk is for wheelchair users, it is highly relevant to this work.
2. The author should better argue the novelty and advantage of the proposed method over other vision-based methods detecting pitfall, cracked bricks, etc.
3. Many elderly people may also suffer from visually impaired or need mobility aids. How is the proposed method going to address those? It is unclear if the recruited 30 elderly participants have normal vision or mobility.
4. There is no statistical analysis for the MaxLE values among different sites and two groups. The description of the MaxLE measurement by how many times between sites is not appropriate. 1.83 times in MaxLE does not mean the site C is 1.83 times easier to walk. A two-way ANOVA might be a good one to use here.
5. The author may need to describe how such walkability can be helpful for elderly users. Is the smartphone-based system allows to scale up the assessments to a city scale like the Project Sidewalk? 

Author Response

The author wishes to thank the reviewer for his/her time and effort in reviewing this manuscript. I hope the changes listed have made the manuscript suitable for publication and this author looks forward to your response. Please see the attachment file

Author Response File: Author Response.docx

Reviewer 2 Report

The authors have presented an interesting and novel use of body-worn sensors to enable automatic (as opposed to subjective) assessment of the walkability of the area.  The article could be improved by additional explanation of terminology, and discussion of the place of the tool within the wider context of walkability.

 

The article is not formatted with the suggested headings and contents of section from the Sustainability guidelines.  I found the current structure to be a little odd at times, in particular the separation of the introduction and literature review (which I feel could be combined), and the inclusion of results within the discussion section.  I recommend some (minor) restructuring to align with the journal guidelines.

 

Abstract:

Line 12:  “Existing studies have applied various methods …” this sentence should include something to answer “to do what?”.  The information is provided in the next sentence, but should be in the first one.

 

Line 19:  There is a lot in the abstract about the background, and very little on the methods.  In the sentence starting “walkability was measured …” please provide some additional information on what the authors mean by gait stability, what tool was used for measurement (which could just be tri-axial acceleration worn on the lower back).  Additionally, in my opinion there needs to be some brief explanation of what the maximum Lyapunov constant is, as it is not sufficiently well known to stand on its own.

 

Lines 22-23:  The final concluding sentence of the abstract should be made more tentatively.  The system shows promise for being able to assess the walking environment, but to do so in practice would require some scaling up –older adults would have to use it and it isn’t clear that the smartphone application described here would be immediately suitable and the results would need to be integrated into government systems.  And then demonstrating that continuous diagnosis would result in improvement of the active life of older adults is a logical link that is plausible, but not proven in this article.

 

Introduction:

Lines 33-34:  the phrase “major developed countries are looking for new opportunities through the change of perceptions” was not completely clear.  What sort of opportunities, for whom or to do what? And why does changing perceptions help with this?  Additionally, the examples placed in brackets did not clarify this for me –a little extra explanation is required.

 

Line 40:  I did not understand what was meant by “has a standard that is centred on the elderly”.  I think it is the use of the word standard in this context that is not clear to me.

 

Lines 57-59:  The terms “walking stability and uniformity” require additional explanation when they are introduced here.  They are key concepts for the study, and yet are never explicitly defined.  There are a range of aspects of gait that can have variability within them – for example, stride length, stride width and cadence - and it is not clear which aspect(s) the article is referring to in terms of stability.

 

Lines 59-60:  It would be good to have more detail about the types of wearable equipment that the author is referring to, as this the term is vague and could cover a lot of different options.

 

Literature review:

Lines 93-99:  The authors report on three separate studies here, but I find it very unclear what these studies actually found.  The authors focus on how much data was collected, but not the results.  For the Pikora et al article, especially, the phrase “and showed statistical significance of the results” needs more explanation.  What was being compared with what? And preferably some idea of the scale of the differences/similarities would be useful to interpret the study.

 

Lines 115-118:  More detail on this section would be useful.  What sort of wearable/smart devices does the People Centric Sensing technology use?  What “different elements of information” can it/has it been used to infer.  Additionally, in the examples in the second sentence, the term “events” needs additional explanation here, as it is a term that is used differently in different fields.

 

Experimental Design:

Lines 143-144:  The authors hypothesise that “different walking environments” can be distinguished.  But they have not yet described the types of differences in walking environment that they are aiming for, and that they hypothesise will be distinguished.  Walking environment is a multi-faceted concept, and the authors need to explicitly state which aspects they are interested in.

 

Lines 172-174:  I suggest “The four test sites are located a short distance from each other, …”

 

Lines 178-181:  It is unclear why both older and younger individuals were recruited.  The sentence structure of the two items listed does not work, and I cannot follow the rationale.

 

Lines 197-199:  By whom was the walking convenience measured?  It is clear from later it is the participant, but this should be mentioned here.  Was the convenience rated for each walk at the end of that walk, or were all four walks rated at the end of whole experiment?  Finally, what specific question was asked?  Walking convenience is mentioned here, but is different terminology from walking environment.  Why was a nine-point scale selected?  It is not a “standard” scale size (Likert scales are often five or seven, but rarely nine, and ten-point scales are also often used), is there a particular reason for its selection.

 

Lines 200-207:  The authors explain here why they used a smartphone as opposed to (for example) an accelerometer.  However, it is not clear to me why having the GPS and ability to transmit and receive data is useful for this experiment. If this section is related to potential future use of the device, then that should be clearly stated.  Also, if this does relate to future use, how does the fact the smartphone need to be attached to the back of the user (as opposed to held or carried in a pocket/bag) then potentially limit its wider use?

 

Lines 27-209:  Here it is stated that a Go Pro camera was used, but there is no mention of why, or how the data was used in the results or the discussion.

 

Line 212:  The term walking stability needs to be defined, either here or earlier in the article (see previous comment).

 

Line 217-219:  In the first sentence here, you are conducting a temporal analysis on “the multi-dimensional variability” of what?  Acceleration? Position? Or …?  At the moment I do not follow what is meant by the “attenuation of the multi-dimensional variability”, but that might be cleared up if the previous sentence is clarified.

 

Line 235-236:   The text states that figure 2 is from an IMU attached to a limb.  But in this experiment the smart phone was attached to the lower back.  So – what is being shown in the figure?  Ideally data from this experiment should be used to demonstrate the technique, and if it is not, this should be clearly explained.

 

Line 279:  I suggest you add here the fact that that the lower the value of MaxLE the higher the walking stability, to help with interpretation in the results.

 

Results:

Line 308:  Figure 5 should be figure 4.

 

Line 328:  The table title should make clear this data is just for the older adults.  In table 2 what are the asterisks for?

 

Line 360:  At the end of the results I felt there were parts of the study which had not been reported.  This included the data from the Go Pro, and the data from the younger participants.  The results for the younger participants are reported as part of the discussion.  To me this is an odd format, and it does not match the Sustainability author instructions.  The authors should rearrange so the results for the younger adults are reported in the results, and then relevant parts of the results can be discussed in the discussion (without introducing new information).

 

Discussion:

Lines 272-274.  I do not follow this sentence.  Two groups are mentioned, but then single values for differences between sites.  Does this mean that both groups had similar differences between sites (in terms of relative difference)?  If so, this should be made more clearly.  It might be better to start with the fact that the younger group have lower values (i.e. more stable walking) compared to the older group.  And then a comment that the relative differences were similar would really stand out.

 

Lines 386-395:  The authors make some really important points about how walkability is currently assessed.  However, it would be good to also include some information on how the authors envision that their tool could be rolled out to older adults.  Do you envisage one or two users for an area, or many people using the device.  Would people be expected to wear it all the time?  And if so, how would having to strap your smart phone to your back impact this? (or do the authors have suggestions for future improvements that might mean it could be worn elsewhere?).

 

Line 395:  I felt there could have been wider discussion about the multi-faceted nature of “walkability”.  Subjective tools for walkability cover a much wider set of concepts, including (for example) street connectivity, dwelling density and land-use mix.  It would be nice to have discussion of what aspects of walkability this objective measure covers, and how it fits into the wider context of the measurement of walkability.

 

Line 395:  The authors should include a section on the limitations of the research within the discussion section (as per author guidelines).

 

Line 395: It might be useful to also include some discussion on what next steps would be needed to move this research from an experimental demonstration to use in the real world. 

 

Author Response

The author wishes to thank the reviewer for his/her time and effort in reviewing this manuscript. I hope the changes listed have made the manuscript suitable for publication and this author looks forward to your response. Please see the attachment. 

Author Response File: Author Response.docx