Development and Assist-As-Needed Control of an End-Effector Upper Limb Rehabilitation Robot
Round 1
Reviewer 1 Report
This paper describes an Assist-as-Needed control for a novel end-effector robot for assistance purposes. Honestly, I didn't find the paper super interesting because the subject (a virtual channel to guide the user arm) is something quite well known and studied in the field. However, the research is well carried out and the results are coherent with the hypothesis. A bit of work could be done to improve the state of the art, by citing other relevant papers about this topic, and by improving the quality of pictures.
Following, a list of major and minor points to review. Major points are mostly like suggestion to improve a next study, rather than to modify necessarily this one.
Major Points:
- How do you control or better guide the z axis with the "flat" virtual interface? It seems from figure 11/13 the subject was wandering a lot along that dimension without assistance. I think having a 3D figure instead of two flat images could have helped more the subject to achieve a better result.
- It would be interesting to study how the three conditions (no control, trajectory tracking and virtual channel) affect the movements on several repetitions of the tests, potentially on several days. It could be interesting to see if the subject learns to stay within the channel, reducing with time the assistance needed by the robot to track the reference.
Minor Points:
- 12: "Due to the plasticity.." I don't think this sentence makes sense in English.
- 39: What do you mean by domestic and foreign research institutions? Domestic and foreign to who?
- 92: I like the idea of spatial freedom, and I do believe it is something very central to control for upper-limb rehabilitation robots. I would suggest to cite other papers that attempted spatial freedom, maybe not in an AAN or Impedance control manner, but still worthy to mention: Brokaw et al. (Retraining of interjoint arm coordination after stroke using robot-assisted time-independent functional training, JRRD 2011), Proietti et al. (Modifying upper-limb inter-joint coordination in healthy subjects by training with a robotic exoskeleton, JNER 2017).
158: I don't understand how the screen can be used for showing training scenes if the user is not looking at the screen when within the robot.
238: Double "stiffer" word.
Figure 10: I don't think showing the 7 DOFs of the robot is that useful/meaningful or add anything to the comprehension of the results.
Figure 13: I guess you didn't update the caption, which is the same as figure 12.
Figure 13: I like when the different plots share the same axes, cause it's easier to spot the differences. I know that having 0-240 mm of deviation on all the three plots would make (b) and (c) smaller, but I think it is still worth trying.
- 436: any specific reason why that larger error happened from 50s on?
- 444: same question as before: what happened from 70s? Is it made on purpose or was a natural behavior by the participant? Did you ask him on purpose to fight against the robot?
Author Response
Dear reviewer,
Please see the attachment.
Best regard,
Author Response File: 
Author Response.docx
Reviewer 2 Report
The paper presents the development and assist-as needed control of an end-effector upper limb rehabilitation robot. The first part of the paper presents an extended introduction in the field of the upper limb rehabilitation robots denoting the need of performing spatial rehabilitation motion instead of simple planar motion rehabilitation.
The aspect of spatial rehabilitation motion though is a quite an interesting instrument in performing daily tasks, it cannot be used in the rehabilitation of acute patients that sometimes cannot move their limb at all so the targeted patients for this kind of rehabilitation should be better indicated in the introduction of the paper.
Second part of the paper presents a robotic system for upper limb rehabilitation consisting of two kuka arms joined through a pillar. The presented solution is quite simple and easy to implement however a few questions arise:
- Why are there two robotic arms? Usually, the rehabilitation of the upper limb is performed one limb at a time, and each affected joint is given special attention by the kinetotherapeut. Neurological diseases patients usually are paralyzed on one side of the body.
- Is the handle presented enough for the rehabilitation task? Usually the rehabilitation motion is performed by applying force in the direction of the motion of the impaired limb and in the same time some anchor points are used to isolate the rehabilitation motion.
In the last part of the paper some experiments are provided. Whatever in order to prove the effectiveness of the robotic system some extended experiments should carried on. Also some antropomorphic data regarding the subjects of the experimental tests should be provided.
Overall the paper presents interesting research however some aspects should be better explained in the text and the English language should be revised.
Author Response
Dear reviewer,
Please see the attachment.
Best regard,
Author Response File: Author Response.docx
This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.
Round 1
Reviewer 1 Report
The paper describes a bilateral upper-limb end-effector robot for rehabilitation of patients with peripheral nerve injury. An assist-as-needed controller has been developed, to assist the movement of the patient if his/her trajectory is out of the boundaries of a virtual cylinder channel around the desired trajectory. The proposed controller has been tested on a single healthy subject.
I would recommend an extensive editing of the English language, since some concepts are not conveyed adequately, resulting to be misleading or incorrect sometimes, e.g.: "Most of the traditional passive rehabilitation training methods always hope that the patient's movement can completely coincide with the predetermined trajectory to ensure the accuracy of the patient's movement". Passive training does not require any active contribution from the patient to accomplish the movement, but this sentence is not very clear about it.
In addition, the paper has some limitations, as explained in the following:
- The introduction is focused on the rehabilitation of patients with peripheral nerve injuries. However, it is never explained how the proposed robotic platform and controllers have been designed to specifically address this kind of patients (i.e. which are the functional and technical requirements that have been taken into account)
- The verification of the controller has been done with a single healthy subject, so there is poor evidence about the performance of the system when used by a person with reduced upper-limb mobility
- The platform is presented as a bilateral upper-limb exoskeleton, but the proposed controller does not take into account interaction effects between the two arms, or bilateral exercises. This is a strong limitation that needs to be discussed and justified in the text.
- The description of the robotic platform can be improved. It is not clear how the motion capture module has been used in the experiments.
Other minor comments:
- The acronyms ADL and DOF are not explicit
- What are "n","o","a" and "p" in eq. (1)?
- Which filters are used to calculate velocities and accelerations?
Reviewer 2 Report
In this paper, the authors proposed the design and control of a bilateral arm exercise robot using two 7-DOF KUKA robot arms. From a design and control point of view, there is a need for distinguishing problem formulation and some idea with regard to the practical control of a two-arm robot system.
Please highlight the contribution of this paper with respect to several previous teleoperation and/or rehabilitation robot control studies.
This paper is about the design and control of a bilateral robot, but does not seriously deal with AAN control in a bilateral (two-arm) robot.
Overall, it is ambiguous whether it is PNS or CNS rehabilitation.
