Methodology

Authors have implemented Human-robot interaction torques methods and evaluated their performance and accuracy. The methods selected for analysis like motor currents or torques, with or without gravitational effects and/or friction, robot dynamics, inverse dynamics (ID), nonlinear disturbance observer (NDO) are efficient for the analysis. Rehabilitation exoskeleton, as well as dynamic and friction models, are used in the test bed is provided. The identification protocol that is used to improve the accuracy of the models, and torque/force estimation algorithms implemented. Good amount of previous research had been presented in the paper and with their results and there was enough background information available on the topic. The paper is nicely organized in sections. Authors clearly mention the results achieved and the contributions made of this paper. Results involve model identification, torque estimation, and sensitivity analyzes and simulations presented in the paper. I recommend the paper for the publication.

Strengths

1. Authors have implemented Human-robot interaction torques methods and evaluated their performance and accuracy.
2. Results involve model identification, torque estimation, and sensitivity analyzes and simulations presented in the paper.
3. There are a lot of references and background research.
4. The paper is nicely organized into coherent sections
5. Authors clearly mention the results achieved and the contributions made of this paper.
6. Authors estimated inertial parameters from SolidWorks and friction parameters estimated through experiments.

Weaknesses

1. The paper starts without giving adequate information about the context and background that makes paper difficult to understand.

Major points

1. The methods selected for analysis like motor currents or torques, with or without gravitational effects and/or friction, robot dynamics, inverse dynamics (ID), nonlinear disturbance observer (NDO) are efficient for the analysis.
2. Rehabilitation exoskeleton, as well as dynamic and friction models, are used in the test bed is provided.
3. The identification protocol that is used to improve the accuracy of the models, and torque/force estimation algorithms implemented.