Leader-Based Trajectory Following in Unstructured Environments—From Concept to Real-World Implementation

Round 1

Reviewer 1 Report

The authors need to take in consideration the following suggestions:

1. The introduction needs to be improved by discussing current articles because only a few new articles have been included in it.
2. The contribution is not clear.
3. A new section named “results discussion” needs to be added in order to condense all obtained results. I recommend presenting a table where the qualitative and quantitative features of your proposal and the other proposals reviewed in order to show the main advantages and disadvantages of your proposal.
4. The results are not explained in detail; hence, I cannot observe or read your contribution of your findings.
5. Fig 1 must be explained in detail.
6. The conclusion needs to be improved, adding quantitative results not only qualitative results. In addition, it is important to mention, what is the next with the investigation?

Author Response

1. The introduction was extended by referencing several other related works and discussing the similarities and differences with our work.
2. Also, the abstract and introduction were reworked to point out what the contribution is.
3. A comparison with results from other proposals would be an interesting result. Unfortunately, comparing results to literature is difficult as this would require a similar experimental setup, regarding the path of the leader vehicle but also the automated vehicle itself. We used a series production vehicle, while in some other publications off-road vehicles with a different sensor setup and vehicle dynamics were used. Additionally, we were limited by the size of the available proving ground and the field of view of the sensor for object detection. Due to these aspects, we would prefer a simulation-based approach to compare different vehicle following proposals.
4. We have extended the section on experimental results. Please consider the results as a proof of concept for the presented approach. Of course, there is still room for further improvements which will require additional time for implementation and testing.
5. We would have appreciated a detailed comment on what exactly is unclear. Nevertheless, this section was revised and extended including the figure explaining all used symbols.
6. It is mentioned that future work will include extensive testing of the presented vehicle following system which can then serve as a basis for the requested quantitative result evaluation. Unfortunately, the current amount of test results is not sufficient for a quantitative analysis.

Reviewer 2 Report

1. I do not get the motivation of the paper and I am confused with the sentence "Unfortunately, most of the cited work does not cover the path reconstruction process in detail, especially in the context of vehicle implementation, which requires consideration of real time performance requirements".  Path tracking is an old research topic, what are the disadvantages of the existing works??
2. what is the motivation for Smoothing the Reconstructed Path? what are the criteria?
3. What is the novelty when the author uses spine method?

Author Response

1. The motivation is to navigate a vehicle by means of a leader due to unavailability of typical guidance elements like lane markings. Some possible applications are listed in the introduction, like construction sites (lane markings invalid) or agriculture applications (no lane markings at all). To follow the path of the leader, the autonomous follower vehicle needs to estimate the leader path using on-board sensors. This reconstruction or estimation of the leader path is only covered briefly in other work. Just storing all leader positions to reconstruct the leader path would be an insufficient approach as every hardware (especially embedded real-time hardware) has limited resources (memory and computing capacity). Therefore, the focus of the paper is not on tracking the leader path, but obtaining an estimate of the leader path!
2. Typically, sensor measurements are noisy. The estimated leader path is constructed from such noisy sensor measurements. To improve the path tracking performance, the path is smoothed using a spline approximation algorithm.
3. The "novelty" lies in the fact of applying this technique to the presented problem. The main advantage of this approach is its computational efficiency.

Reviewer 3 Report

The manuscript proposed an interesting method for Leader vehicle-based trajectory following.

Notes for authors:

1- The introduction is very poor, please add more relevant references that use similar techniques, or different methods.

2- Illustrate the control of section 2.3.1 briefly.

3- The results need to compare with the literature.

 

Author Response

1. More relevant references were added and also the similarities/differences were discussed.
2. The control law was added and the meaning of all symbols explained.
3. I agree that this would be interesting results. Unfortunately, comparing results to literature is difficult as this would require a similar experimental setup, regarding the path of the leader vehicle but also the automated vehicle itself. We used a series production vehicle, while in a some other publications off-road vehicles with a different sensor setup and vehicle dynamics were used. Additionally, we were limited by the size of the available proving ground and the field of view of the sensor for object detection.

Reviewer 4 Report

This work aims to navigate a four-wheeled vehicle through unstructured environments characterized by the lack of availability of typical guidance infrastructure like lane markings or HD maps. The following approach is based on leader position and speed measurements combined with ego-motion estimates to reconstruct the leader’s path over time, which prevents the cutting corner phenomenon. A spline approximation technique is used to obtain a smooth reference path for the underlying lateral and longitudinal motion controllers. For longitudinal tracking, a constant time headway policy was implemented that allows following the leader with a constant time gap along the path. The algorithm was first developed and tested in a simulation framework and then deployed to a demonstrator vehicle for validation under real operating conditions. The presented experimental results were achieved using only on-board sensors of the demonstrator vehicle, while high-accuracy differential GPS based position measurements serve as the ground truth data for visualization

   In general terms, the paper under consideration contains sufficient novel information, which advance our existing knowledge on the field of Automation and Control of tracking vehicle’s trajectory. The results from both analytical estimates and numerical integration appear to be in sufficient agreement. On this basis, I would recommend the publication of the article. However, it is the reviewer's opinion that before publication the paper should be thoroughly scanned and its language should be improved as well as some minor remarks.

• The abstract must include only new obtained results
• The authors have to shed light on the similarities and differences among their work and the literatures of the problem. A clear explanation, what is the new result in their work, and how it is build up upon previous work in the field.
• The conclusion section should be extended to include more details and should avoid to repeat the abstract in the conclusion Section.

Author Response

1. Our honest opinion on this point is different as the abstract should summarize the content of the work, which usually does not contain newly obtained results only. But we agree that newly obtained results must be mentioned in the abstract, which was ensured during the review.
2. The introduction including the literature review was reworked and extended during the review. We put special focus on pointing out the differences and similarities between related literature and our work.
3. We totally agree with that point and made the according additions and changes.

Round 2

Reviewer 1 Report

The authors have addressed correctly the reviewer's concerns. Hence, I can recommend accepting the article

Reviewer 2 Report

Based on the authors' reply " The "novelty" lies in the fact of applying this technique to the presented problem. The main advantage of this approach is its computational efficiency", the paper can be considered as an experimental report rather than an academic paper.

 

1.  

Reviewer 3 Report

No more corrections are required.