Review Reports

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript described some good technical work toward a clever mechanical solution to reducing barriers to adoption for robotic gait assistance technology. The design, analysis, and characterization of the system were well executed, and the paper is well-written. Still, the potential impact of the design is overstated in the paper due to several significant practical limitations:

• The system requires custom fabrication (of each of the 8 bars) for each user, which is impractical for a device intended for use in gait therapy, which typically occurs in one-hour sessions. While the paper suggests this system represents an improvement over traditional exoskeletons in terms of setup time, I would argue that adjusting a few software settings on a device is much quicker than measuring and fabricating 8 new structural members, and then assembling a new mechanical system for each new patient.
• The system only supports level ground walking, which means it is not appropriate for home or community use, cannot help patients practice sit-to-stand or stand-to-sit transitions, and may require patients to be upright before the system can even be donned.
• The form factor of the system makes it difficult, and perhaps unsafe, for a clinician to walk closely behind the patient to provide spotting support, which they typically would in most gait therapy sessions.
• The use of a worm-gear transmission means the system will not be backdrivable, preventing true cooperative control between person and machine. The ability to modulate the level of assistance is crucial in gait rehabilitation therapy so that the patient is encouraged and able to contribute to the walking motion. Without a backdrivable drive system, this ability is severely restricted. Furthermore, a system intended to support neuromuscular reeducation should allow the users to “make mistakes” during walking, specifically allowing small but normal deviations from the nominal joint angle trajectories. The ability to make these small errors and correct them is an important part of reconnecting the neural pathways during recovery. A high-impedance transmission prevents this.

Some specific comments:

Line 13: The abstract claims the new system is simple and compact. In fact, I think many would argue that even though you’ve eliminated a degree of freedom, this system is (from a practical standpoint) more complex, and less compact that commercially available exoskeletons.

Line 30: “Nowadays” is colloquial language.

Line 39: “…difference with EksoNR” should be “difference from EksoNR”

Line 39: The current version of the Indego does have a passive spring-loaded ankle joint.

Line 44: “gauge” is probably not the right word. Perhaps “lock” would be better.

Lines 47-48: The phrase “of these exoskeletons” is redundant.

Line 52: There are many more disadvantages associated with the passive Zhou exoskeleton which you have not mentioned (for instance, the inability of the system to actively assist flexion, extension, etc).

Line 56-58: This seems to suggest that the Rex does not have actuated knee joints, which is incorrect.

Line 66: “with the three previous…” should be “from the three previous…”

Line 89: “forward” would be better worded “overground”

Line 90: describing the 8-bar mechanism as “compact” seems incorrect, particularly when comparing to commercial exoskeletons like Indego or Ekso

Line 91-93: Meaning is unclear

Line 124-125: Here you compare cost to selling price, which are two different things. Consider making this point more clearly.

Line 153: Modern exoskeletons support higher gait speeds. It’s unclear why 0.42m/s was selected as the target gait speed, or what the upper limit of comfortable walking support would be in this 8-bar system (See: S. A. Dalley, C. Hartigan, C. Kandilakis and R. J. Farris, "Increased Walking Speed and Speed Control in Exoskeleton Enabled Gait," 2018 7th IEEE International Conference on Biomedical Robotics and Biomechatronics (Biorob), Enschede, Netherlands, 2018, pp. 689-694, doi: 10.1109/BIOROB.2018.8488065.).

Line 210, 216, 221, 270, 271, 323, 324, 363, 381,409, 410: “On the other hand” and “on one hand” are used multiple times. This is colloquial language, and actually means that two pieces of information are being contrasted, so it does not seem like this expression is being used correctly in the paper. All occurrences of these expressions should be removed.

Line 293: “…box plot of data” should instead say “…box plot of _____ data”, denoting what data is being displayed.

Line 327: “stablish” should be “establish”

Line 353: It’s unclear if the assembly weight includes the orthosis or just the 8-bar mechanisms.

Line 356: “study” and “conclude” should be past tense.

Line 365: It’s unclear how SCI patients will receive a higher quality of life, since this device would not be appropriate for home or community use. Probably should adjust the claims being made in the paper, or else defend them more fully.

Line 377: motor power is provided, but not torque.

Line 394: “simplicity of manufacture” seems debatable, given the requirement of creating 8 new bars (per side) for each new patient, and then assembling it all.

Comments on the Quality of English Language

Generally, the English is good, with only a few exceptions which I noted specifically in the Comments to Authors section.

Author Response

Attached is a file with comments for the reviewers (the file includes responses to reviewers 1 and 2).

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This manuscript mainly addresses the issues of high cost and complex control in current rehabilitation exoskeletons by proposing an eight-bar exoskeleton mechanism based on a single degree of freedom (DOF). The device is characterized by low cost and simple control. The research process is relatively complete, covering theoretical modeling, mechanism optimization, and finally prototype experiments. However, the manuscript does not highlight its unique contributions to rehabilitation exoskeletons and presents the following problems:

1. Numerous studies worldwide have already reported on single-DOF rehabilitation exoskeleton mechanisms, including planar four-, five-, six-, and eight-bar linkages, as well as spatial single-DOF linkage mechanisms, all of which share the property of requiring few actuators. What distinguishes the proposed single-DOF mechanism from these existing designs? Where is the unique contribution? This requires clear explanation.
2. For rehabilitation exoskeletons, lighter weight is naturally preferable. Why did the authors choose an eight-bar mechanism rather than a four-bar, six-bar, or even a geared five-bar mechanism, which are also single-DOF linkages?
3. The manuscript’s stated objectives are vague, oscillating between customization and low cost. While low-cost customization is not impossible, the current design can only generate a fixed gait. Since different patients have different gait characteristics and hip-knee-ankle motion patterns, true customization would likely increase development costs.
4. With only one degree of freedom, the mechanism can adapt to only a single motion pattern. It cannot accommodate different walking speeds, slopes, or environments. For patients at different stages of recovery, would this mean new devices are required each time?
5. The authors assume that hip joint stabilization torque is provided by crutches, but they do not present quantitative torque values, stability margins, or related analyses. Such explanations should be added.
6. In optimization, the objective function only considers femur/tibia angle errors, while neglecting other critical indices such as ankle/knee trajectory, energy consumption, and gait symmetry. This results in an overly narrow evaluation framework. Additionally, while the authors collected gait data from over 50 individuals, they ultimately chose the 12 with the greatest differences, but did not justify the selection criteria or reasons for exclusion.
7. The dynamic modeling involves several simplifications, but the impact of these simplifications is not analyzed. For example, the hip is simplified to a pure translational connection with the world frame, without analyzing the influence of roll and pitch DOFs. The use of four contact spheres to model the foot sole, with parameters of 70 N/mm stiffness and 0.5/0.3 friction coefficients, is also not validated through sensitivity analysis or compared against real plantar pressure distributions. Such omissions could significantly affect model accuracy and experimental reliability.
8. The experimental data only provide qualitative gait demonstrations, lacking quantitative analysis. It is therefore difficult to compare experimental outcomes with simulation results—for instance, trajectory error statistics or energy consumption comparisons.
9. The authors fail to report the total weight of the device. From the experimental images, it is evident that the device is relatively heavy and bulky. Whether such weight benefits patient rehabilitation should be a central factor in optimization, but this is not addressed.
10. In the conclusion, the authors claim that the proposed method “provides an effective rehabilitation solution.” However, based on the presented validation, the system only demonstrates a pre-defined gait under ideal conditions. This makes the conclusion overly optimistic and insufficiently convincing.

Author Response

Attached is a file with comments for the reviewers (the file includes responses to reviewers 1 and 2).

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The paper is significantly improved by the changes implemented in this revised manuscript. Thank you for your careful attention to my original review comments.

One final point of clarification I want to make about one of my original comments:

Original comment:
"The system requires custom fabrication (of each of the 8 bars) for each user, which is impractical for a device intended for use in gait therapy, which typically occurs in one-hour sessions. While the paper suggests this system represents an improvement over traditional exoskeletons in terms of setup time, I would argue that adjusting a few software settings on a device is much quicker than measuring and fabricating 8 new structural members, and then assembling a new mechanical system for each new patient."

Authors' response:
"The exoskeleton is custom-made for each patient. Therefore, the exoskeleton is not adjustable; each patient has his/her own. To start the rehabilitation session, all the patient has to do is put on the device." 

Reviewer response:
My point was that it is problematic to have to make a custom exoskeleton for each patient. This would require a patient coming in for an initial screening visit, taking measurements of the patient, the sending the patient home for days or weeks while the exoskeleton is fabricated to their dimensions before they can receive gait therapy. This is a difficult process to follow for every patient, and is not a typical process used in rehabilitation clinics. Additional complications exist in the form of payment for the device. Does the patient pay for this device, which they will only use in clinical rehabilitation sessions? Does the clinic pay for the device, which can't be used with anyone else after the patient completes their gait training sessions? The model of creating a custom exoskeleton for each patient for use in clinical rehabilitation seems challenging, and the paper would be strengthened by proposing a future route to addressing this challenge.

Author Response

We have answered your questions in an attached file.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

All issues have been resolved and can be published

Author Response

Thank you very much for your comments, which have helped to improve the article.