Simulating Ionising Radiation in Gazebo for Robotic Nuclear Inspection Challenges

Round 1

Reviewer 1 Report

The paper presents the development of a tool that allows to incorporate, to the GAZEBO simulator, environments affected by sources of ionizing radiation.

Along with the ability to simulate the diffusion of radiation in the environment, considering the different penetration capabilities of radiation in different materials, it includes the possibility of simulating objects with geometric characteristics typical, of the environments affected by radiation. This allows to launch robotic applications that must operate in these scenarios.

Although the article does not represent a scientific advance, its contributions to researchers who develop applications of field robotics in inspection tasks, or even search and rescue of victims in contaminated environments (especially by ionizing radiation), are relevant and therefore are consider a work whose dissemination and publication is valuable.

After reading the work, the following modifications are considered necessary to improve its interest and, in this way, be accepted for publication.

-In section 2.1 World Building, reference is made to the V-REP simulator. Since November 2019 this simulator has been discontinued and replaced by CoppeliaSim, so it is more appropriate to cite this new simulator.

-Figure 1 represents the stages of generating an environment in Gazebo. The figure is valid and correct, but it would contribute more to the transmission of the idea if, instead of the boxes with the text, they incorporated an illustrative image of each element, in addition to the text.

-Something similar can be said of figure 6.

-In general, section 2.1 is excessively long for the contribution it represents, especially for a reader familiar with the use of Gazebo, it is advisable to reduce its length allowing the reader to arrive earlier, and less fatigued, to the reading of section 2.2, where the most interesting contribution of the work begins.

-Conversely, and since it is of great importance to have the best guarantees that the built-in plugin is valid for developing autonomous robot navigation algorithms under Gazebo, , more exhaustive comparisons of the simplified model with models documented in the literature should be incorporated. The benchmarking in section 3.1 is correct but  insufficient. More complex scenarios, that are documented in other publications, should be compared. This increase in the guarantees of the validity of the model is of great importance, in order that the proposal presented is really useful.

-In this sense, section 4.1 analyzes the limitations of the model and, among others, the lack of ability to simulate radiation scatters and reflections is recognized. It is necessary to renounce this ability in order to guarantee simulation in real time. A comparison should be made, or at least an argued reflection about how much this limitation can distort the results and consequently the usefulness of the plugin. Imagine, for example, that you are trying to develop an active scanning algorithm for an autonomous robot dedicated to locating the source of radiation in an environment with different radiation-reflective elements. Would the algorithm developed under the assumptions of the simulator be valid when used in the real environment?

-In section 4.2 it is convenient to add images of the environments that are presented. Scenario 4.2.1 is already described in section 3.3, it would be enough to cite the figure number (not just the section) in this section. The scenario in section 4.2.2 needs a figure to show it. It would even give more credibility to what has been explained, incorporating a real photo of the environment, if it is available.

-One last aspect that must be resolved so that it makes sense to publish the work and it really has some interest for researchers related to the subject: the links to github shown in the Supplementary Material section:

• github.com/EEEManchester/gazebosim_world_generator
• github.com/EEEManchester/gazebo_radiation_plugin

they do not work (consulted during the week of June 14 to 21, 2021)

Please, this must be correctly updated.  The world generator and radiation plugin must be available and operational

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

This paper presents a plugin tool for simulating ionising radiation in the Gazebo physics simulator. The proposed tool is very useful in the robotics community. However, the following two links provided on Page 23 are not available online:

github.com/EEEManchester/gazebosim_world_generator

github.com/EEEManchester/gazebo_radiation_plugin

Please check these links carefully to facilitate the contribution of this work to the community

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

The manuscript represent a very interesting simulation on the response of robots to radiation environments. The authors in particular studied three realistic scenarios of real-world nuclear sector environments in UK.

The simulation is compared with the de facto standard transport code MCNP6. A good agreement between the two is obtained, which demonstrates clearly the relibility of the simulator, but the present simulation can be 10^6 times faster than MCNP.

I believe such simulation can be very useful for increasing stakeholder confidence in robotics in the nuclear sector. I therefore recommend its prompt publication.

Two minor comments:

I was really confused by the title at first since I was not aware of the existence of such simulator called Gazebo. I would recommend the authors to be more specific.

I disagree with the statement "Radioactive decay is by its nature is a random process, described by a Poisson distribution [37]."

It may make sense to some extent by considering the fact that we cannot predict which specific  atom will decay at a given time. But the whole process and the fluctuations in decay process is well understood as a simple quantum phenomenon.

Author Response

Please see the attachment

Author Response File: Author Response.pdf