Performance Evaluation of Autonomous Driving Control Algorithm for a Crawler-Type Agricultural Vehicle Based on Low-Cost Multi-Sensor Fusion Positioning

Round 1

Reviewer 1 Report

Introduction

• the lack of review of the world literature in the field of development of the science of construction and use of autonomous vehicles;
• the lack of description of the current state of knowledge in the use of autonomous vehicles in agriculture

Test:

• - lack of information on how the algorithm will behave when the autonomous vehicle will pull a farm machine (what will be the levels of accuracy of maintaining own position of set). In agriculture, the tractor does not move alone - only as a set with an agricultural machine. Therefore it is more interesting to use of autonomous vehicle + agriculture machine as the set in the function of the task to be performed (for example task: plowing fields or fertilising agricultural fields)

Conclusions

• the lack of a clear indication of the authors of their own contribution to the development of science in the field of construction and use of autonomous vehicles

References:

• lack of literature from other countries and continents (e.g. Europe, America), which describe achievements in the area described in the article.

Author Response

Please find the attached file.

Author Response File: Author Response.pdf

Reviewer 2 Report

Paper review
------------

This paper presents a localization and path tracking algorithms for agricultural vehicles using low cost sensors. A low cost GPS-RTK and MEMS-IMU are used for localization, where measurements from IMU are integrated in through INS mechanization to provide vehicle motion which are later combined with GNSS measurements in a Extended Kalman Filter.
Using the positioning given by the previous algorithm, a path-tracking control method based on waypoint navigation is proposed.
A set of rules are defined to command the right and left wheels of the vehicle based on distances and orientations to the next waypoint.
The system is implemented in a real vehicle where 3 different experiments have been performed showing that the vehicle is able to follow predefined trajectories with high fildelity.

The paper is well written and easy to follow, the approach and experiments are well described.
Regarding novelty, I did not found the proposed system theoretically novel, since INS mechanization in combination with GPS is a well founded approach for localization, and path planning uses commonsense heuristics to drive the vehicle wheels.

The proposed system, demonstrated in the experiments, has high applicability for outdoor navigation.
I think the paper could be interesting for other readers, however more details about the EKF algorithm and the path tracking control should be provided to facilitate the implementation and replication of the system.

I recommend the authors not to write in the abstract the limits of your system and future work to be done to address these limitations (i.e., last sentence of the abstract).

Typos:
-----
Abstract: is currently faced -> is currently facing
Line 139: the design matrix should be HMAG not HGNSS

Author Response

Please find the attached file.

Author Response File: Author Response.pdf

Reviewer 3 Report

The paper describes a driving algorithm using low-cost GNSS-RTK and low-cost motion sernsor module for crawler-type atricultural automous vehicles.
The paper is well written and well organized. The experimental results are good and the algorithm implementation is clear.

The reviewer needs only some clarification:
- How do the authors calculate the errors? How do they measure the real position of the vehicle?
- I suggest to expand the conclusions adding more informations about the limitation of the work and the improvement needed.

Minor:
- check the trajectory length of the experiment #2.
- check the figure caption layout (Fig. 7)

Author Response

Please find the attached file.

Author Response File: Author Response.pdf

Reviewer 4 Report

The paper is well organied and the topic is interesting but there are some issues that should be accounted before publication.

A careful revision of the text is necessary, as some sentences are not formulated correctly (e.g. check lines 265, 284) and there are other errors such as wrong equations numbers (e.g. 7, 9 repeated).

Please add the following additional contents:
- In section 2, a detailed description of the differences and similarities in the algorithms compared to the other papers mentioned.
- In section 4, a table summarizing the differences in performance, costs and algorithms of this and other literature solutions.

I would like some clarification on the following points and, if necessary, appropriate additions to the text:
- Is there any particular criteria in the choice of test paths?
- The tests were carried out in a car park, but the terrain of a typical agricultural environment is different. Is the algorithm able to compensate for any movement errors of the vehicle due to the terrain?

Author Response

Please find the attached file.

Author Response File: Author Response.pdf