All reviews of published articles are made public. This includes manuscript files, peer review comments, author rebuttals and revised materials. Note: This was optional for articles submitted before 13 February 2023.
Peer reviewers are encouraged (but not required) to provide their names to the authors when submitting their peer review. If they agree to provide their name, then their personal profile page will reflect a public acknowledgment that they performed a review (even if the article is rejected). If the article is accepted, then reviewers who provided their name will be associated with the article itself.
Both reviewers found the current manuscript acceptable.
[# PeerJ Staff Note - this decision was reviewed and approved by Jafri Abdullah, a PeerJ Section Editor covering this Section #]
No comment
No comment
No comment
I thank the authors for their thoughtful and thorough responses to my and the other reviewer's critiques, comments, and questions. I think the manuscript has been greatly improved.
no comment
no comment
no comment
Thank you for considering my comments and for making adjustments to the manuscript. I would like to offer one minor suggestion: for Figure 4, could you please consider adding an xlabel to the figure and using 'time' instead of 'frames' on the x-axis?
Both reviewers found merit in your manuscript. Please address both reviewers concerns, particularly related to the methods, citations, and address the concern related to your conclusion about effort.
[# PeerJ Staff Note: Please ensure that all review and editorial comments are addressed in a response letter and any edits or clarifications mentioned in the letter are also inserted into the revised manuscript where appropriate. #]
1. The article is well written, but there are some minor grammar issues throughout.
2. The article could better use existing knowledge about human-bicycle dynamics by making comparisons to other results and using prior knowledge to better motivate data analysis decisions. For example, recent work by G Dialynas, AL Schwab, and colleagues contains many measurements of human and bicycle dynamics; other work by Cain et al. (PLoS One 2016) contains significant data about steering, roll, and rider lean dynamics.
3. I was not able to find a few of the references and a few of them need additional details in the citations.
a. Additional details needed for:
i. Moore 2012
ii. Schwab et al. 2008
iii. Schwab et al. 2018
iv. Nieuwenhizen and Schwab 2017
b. I was unable to find the following references, either by general search or by searching the theses collection at TU Delft, which are both critical for understanding the steer assist design and algorithms:
i. Schwab et al. 2008
ii. Nieuwenhizen and Schwab 2017
4. The structure of the article is professional.
5. Figure 2 is not very useful and does not add much additional information to the article. In contrast, many figures are cited in Appendix 1. I would like to see more of the information from Appendix 1 in the main article. In particular, Figure S2 and Figure S6 panels A and B are very helpful for understanding the data collection and dataset.
6. Consider condensing Figure 3 and Figure 4, perhaps using grouped bar graphs to have Disturbed and Not Disturbed results on a single plot.
7. Figure S3 y-axis should be labeled.
8. Figure S5 panels A and B: provide units for ‘SteerMotor’ and please use the same y-axis limits between the left and right panels.
9. The submission is self-contained with results relevant to the hypotheses.
1. Research is original and falls withing the aims and scope of the journal.
2. Research questions are well defined.
3. Some concerns about how the research was conducted:
a. I was surprised that speed wasn’t more closely controlled (only specified cycling within a range of 2-5 m/s). Cain et al. 2016 shows that steering rate can change quite markedly for some riders in this speed range (Fig 8 in Cain et al. 2016). I am curious if the authors have any sense as to the influence of speed on the measured kinematics.
b. Was the order of the trials randomized? This is not clear from the description of the protocol.
c. Similarly, was the pedaling cadence controlled? Hopefully the authors can appreciate that pedaling cadence can have an enormous effect on roll rate and steering rate (and not necessarily roll angle and steer angle).
d. Depending on the posture of the riders, the z-axis angular rate captured by the IMU may not be an accurate estimate of torso lean rate.
4. The methods are not described with sufficient detail and information to replicate or to allow the reader to understand how the research addresses existing knowledge gaps.
a. The description of the design and implementation of the controller is insufficient. I was unable to find either paper referenced regarding the previous design (Nieuwenhuizen and Schwab 2017) or the ‘steer-into-the-fall' (Schwab et al., 2008). While the basic concept of the controller is understood, the details about the exact design and tuning process are lacking, which prevents others from replicating the design and prevents others from identifying ways to improve bicycle balance assist systems. Details must be added to the manuscript regarding the controller.
b. The researchers chose mean absolute angular rates to quantify performance rather than the standard deviation. Can the authors provide their motivation for this choice?
1. The analysis of the data set was appropriate.
2. The processed data (outcome metrics for each trial) on which the conclusions are based are provided. All data files can be opened and appear complete.
3. Figures do not appear to be manipulated.
4. Most conclusions are supported by the data, but I take issue with making conclusions about rider effort when the authors did not directly measure effort – they have only measured kinematics (steer, roll, and lean rates), which may or may not have direct and clear relationships to effort. While no statements about effort appear in the conclusions section, they do appear in the Discussion and Abstract. Unless the authors have evidence that steer rate is strongly correlated to effort/steering torque (please provide a reference), statements about effort should be excluded from the manuscript.
1. Overall, I enjoyed reading this manuscript. However, I feel the impact and value of the results is limited by the limited description of the balance assist algorithm.
2. Design and interpretation of the controller: If I interpret this correctly, the controller is using a bicycle model that uses the benchmark bicycle parameters, correct? Is the controller adapted in any way to account for different bicycle geometries or riders?
3. Figure S4: Can you provide some additional details regarding the ‘Current command from Teensey to Steering motor’ (what is it and how does it relate to the experiment)? Are these data from a disturbed trial?
4. The main revisions I would like to see are as follows (as discussed above):
a. Provide additional details about the controller for the balance assist
b. Provide additional details for the citations that I was unable to locate
c. Do not make claims about effort unless you can provide convincing evidence that steer rate is highly correlated to steering torque
d. Provide additional details about experimental design (randomization of trials, cadence, rider posture)
no comments
no comments
no comments
This paper evaluates the effectiveness of a balance assist system in older and young adults while cycling a straight line (with or without a dual task and perturbations). This is a well written paper. I have several minor comments/suggestions that might improve the readability of the paper.
Minor comments:
1. Methods L 88 – 95: You mention that you manually tuned the gains, but it is unclear how this was done ? It understand that you select gains that result in a locally stable controller (negative real parts eigenvalues) but how can you minimize required steering torque ? Do you mean minimize the feedback gains (and hence minimize required steering torque). You can maybe make this more explicit for the reader
2. Methods L95: slightly different results are to be expected in real scenarios. I assume that you only expect slightly different results when the subjects are indeed moving on the bicycle but do not control the steering angle. I guess that you expect large difference when the subject is also controlling the steering angle. You might want to clarify this to the reader…
3. L77-95: maybe make it more explicit that you are talking about local stability of the system (around the state of upright position of the bike with zero velocities) and not stability in general.
4. Figure 1: it might be easier for the reader if you add legends for different colours (eigenvalue 1, 2, 3). The reader might also like a black line for the maximal value of the real parts of eigenvalues (you use this to determine if the system is locally stable) and maybe highlight locally unstable from 1 - +/- 4.5 m/s in the controller off condition.
5. L103-110: You did not explain how you measured the roll angle and rate at this point (I assume using an IMU on the frame, is this correct ?).
6. L111: simple equation for feedback laws + report feedback gains would be nice in the main manuscript.
7. Figure 2: you might want to illustrate roll axis in this figure ? (this might help some readers without a mechanics background)
8. Figure 3: I think that you can improve this figure. Given the relatively small number of subjects I think that you have to show the individual datapoints or visualise the distribution (using for example a violin plot). As the balance assist on-off is a within- subjects variable (repeated measures) you also might want to connect the datapoints per subjects with a line as you do the statics on the difference between assist on-off for each subject.
9. L364-366: You might want to rewrite this sentence. I understand that the effect of the balance assist system in general is independent on the test location as this is a within-subjects variable. However, this sentence now suggests that the conclusions about age are also independent of the test location (which is not true).
All text and materials provided via this peer-review history page are made available under a Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.