Design of a UAV for Autonomous RFID-Based Dynamic Inventories Using Stigmergy for Mapless Indoor Environments

Round 1

Reviewer 1 Report

In the logistics industry, storage men often need to check the stock of the warehouse, which consumes lots of time and human labor. Therefore, this paper proposes a solution to automatically perform warehouse inventory missions using multi-rotor UAVs in a mapless environment, where drones need to read as much information about goods as possible by RFID tags. In the process of searching for RFID tags in the warehouse, the UAV adopts an RFID tag searching strategy similar to the greedy algorithm. That is, the UAV will move towards the direction where more RFID tags have been detected, to read more RFID tags during the inventory mission. The author also adopts an obstacle avoidance strategy based on 3D point cloud data to ensure safety. Finally, through UAV and simulation experiments, this paper compares the performance of using the proposed stigmergic and obstacle avoidance navigation system, static reading, and dead reckoning method, and verifies that the proposed method can read inventory information in the warehouse more comprehensively and efficiently.

 

In general, the RFID-based stigmergic obstacle avoidance navigation system(RFID-SOAN) proposed in this paper can effectively read inventory information in an unfamiliar warehouse environment and has a certain practical application value. However, this paper is still insufficient in some places.

 

The inadequacies of this paper are as follows:

 

1) The method proposed in this paper is an RFID-based stigmergic and obstacle avoidance navigation system (RFID-SOAN), but there lacks a detailed description of how to avoid obstacles. On page 5, line 162, if the distance between obstacles is less than the safe distance, the approach to avoiding obstacles needs to be further explained.

 

2) More detailed experimental condition needs to be given. The reading distance of different RFID tags varies greatly, ranging from a few centimeters to tens of meters. This variable has a great impact on the result of the inventory mission in the experiment. Therefore, the RFID reading distance used in the experiment needs to be supplemented.

Author Response

Please kindly see the attached document that contains the responses for the reviewer.

 

Kind Regards

Author Response File: Author Response.docx

Reviewer 2 Report

The authors chose a fairly relevant and essential elegance and avoidance of UAV obstacles in the premises. The authors consider this problem from the point of view of the use of RFID tags. This problem can be solved in various ways, including a vision system and the use of obstacle avoidance sensors. The authors do not make comparisons with such methods.At the same time, when describing the drone, the authors mention that the camera is used "Visual Simultaneous Localization and Mapping (VSLAM) based camera (Realsense intel t265) to supply the autopilot with self localization messages and act as a source of the odometry".The authors present an extensive experimental study. At the same time, there is not enough resulting information that would generalize all experiments. Which experiment was the most successful, which had problems and how to improve them. It is better to express this in the form of a table, where the test procedure programs are described and show pass/fail and notes. Otherwise, such a large number of experiments is confusing. Information needs to be structured more clearly.

Line 56 the authors use an interrogative sentence, which is not typical for the scientific and technical style.

The list of literature is incorrectly formatted, in particular, information about journals and conferences.

Author Response

Please find the attached document which contains the responses of the reviewer.

 

Kind Regards

Author Response File: Author Response.docx

Reviewer 3 Report

The authors present a UAV-based inventory scheme. The idea is based on taking the RFID tags present on the shelves as reference points for localization and mapping.

The paper describes both the theoretical part as well as the conducted experiments to illustrate the proposed scheme.

The reviewer has the following comments:

Line 32: How can drones help in automotive industry?

Line 107: calculation capacity is no longer a barrier when it comes to aerial vehicles. The deployed UAV can carry heavy loads, an embedded i5 intel based computer, or a jetson board can easily perform in real time when it comes to path planning and SLAM techniques.

Line 155: The Obstacle Avoidance system is based on a 3D PCD, however, the field of view of the used camera is crucial to define where the obstacles are. When flying up and down to scan the shelves, equations 2 and 3 are valid, however moving from one row to another, the drone may have obstacles present in its field of view (left and right shelves), but still have enough space to fly. Please include how you define the field of view, and how you take it in consideration.

Line 176: the new goal pose message is represented in the 2D plan, however the UAV has an extra degree of freedom, mainly on the Z axis. Since this is an inventory operation, flying up and down is also necessary, please clarify how you deal with that.

Please mention how the RFID tags are placed during the experiments.

How do you deal with interference (two tags detected at the same time).

In T shaped experiments, how does the UAV decides which direction to take (left of right), and how to fly towards the uncovered area after the decision is made?

As a general comment, the authors mentioned that UGVs can easily navigate inside the warehouses, which is valid. Thus it is worth mentioning that there are some works that perform the UAV inventory using a UGV as a reference for indoor navigation. This should be mentioned in the related works, while shedding the light on why the RFID method is better that taking a moving target as a reference.

Author Response

Please kindly see the attached document that contains the responses for the reviewer.

 

Kind Regards.

Author Response File: Author Response.docx

Round 2

Reviewer 3 Report

The authors answer the reviewer concerns in a satisfactory manner.