MERLIN a Cognitive Architecture for Service Robots
Round 1
Reviewer 1 Report
The paper is written in suitable way. The main idea as well as its presentation are interesting. There are my suggestions fo the authors:
- There is any description of abbrevitation PDDL in the 76th row.
- Please consider, if all readers of journal Applied Sciences have deep knowledge about ROS (Robot Operating System). From this point of view, you should maybe describe ROS and its working in a few sentences / images, before you describe ROSPlan and SMACH.
- Figure 1: Please, describe all points from Figure 1. There is necessary for not deep interested reader to understand what is "TTS", "STT", etc. Next, the relations between individial blocks of Figure 1 should be explainder in more details. From the visual as well as text view.
- Row 180: Explain abbrevitation FSM. It should be done for all abbrevitations in the paper.
- Some figures such as figure 4: consider if it is not better to express some ideas by flowchart.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 2 Report
The paper presents a software engineering approach for normalizing the process of integrating hybrid architectures in a fully cognitive architecture. In particular, the proposed approach seeks to overcome the issues that arise when deliberative and behavioral-based approaches are combined for robot behavior generation.
The paper is interesting and promising, it provides enough details and diagrams of the architecture, the experiment aims to validate the approach by comparing the proposed method with the Naive approach. Nevertheless, the results section somehow appears fragmented.
The results which should validate the usability of the approach are not properly argued. Figures and Tables are cited hastily. Extend this section by giving more details with a wider description of the experimental scenarios by pointing out the benefits of the approach with respect to the issues that it aspires to resolve.
For instance, it might be useful to discuss the metrics chosen for evaluating the performance of the architecture:
(1) Why are these indicators relevant to measure performance?
(2) What metric might be used to reveal the benefit of the proposed approach in terms of robot behavior generation?
(3) It might be useful to plot data in comparison charts to compare experimental results. In general, data visualization is the most clear way to compare the results and can provide insight that representation in tables cannot.
For what concerns the background information, the paper deals with a more detailed description of the related solutions used as starting points for the proposed work. Indeed, in the last years, many European projects aim to face similar purposes, especially for elderly and vulnerable people where a generic assistive robot has to dynamically infer knowledge about the status of users and the environment, and provide personalised supporting actions accordingly.
For sake of completeness, it would be great if the authors could evaluate the possibility to add comparisons with the most pertinent cognitive architectures developed within such projects (e.g. Giraff+ Project, MARIO and ACCRA Projects, ENRICHME).
Here follow some references that may be relevant as an overview:
- Toward the integration of perception and knowledge reasoning: An adaptive rehabilitation scenario.
- Assistive Robots for the Elderly: Innovative Tools to Gather Health Relevant Data.
In conclusion, the paper is promising but requires significant revision and some additions. In its current version, the major flaws spotted are related to the readability of data.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
I appreciate the effort the authors have invested in addressing all the comments. The final result is a good and inspiring work.
The authors claim that they have decided to avoid the comparison given the absence of information about the implementation mechanisms. Notwithstanding, in preparing for the future works, I suggest to enhance the study of different approaches presented in these projects also in the terms of use of ontological and task manager approaches.
In conclusion, the manuscript has been significantly improved and now warrants publication but a minor revision of few typos and correction of a duplicate reference (i.e., [3] and [31]) is needed.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf