Review Reports - A Viewpoint on the Challenges and Solutions for Driverless Last-Mile Delivery

Round 1

Reviewer 1 Report

Autonomous vehicles in city logistics is a relevant topic. The paper offers (not more than) a review of the literature. I miss the method of how this literature review was done. That needs to be explained. I see a lot of relevant and recent literature. But is this all there is?

I also miss a research framework that the researchers (might) have used for their review. How do the different topics relate?

Paragraph 4 is not liked to autonomous vehicles but to warehousing and RFID. Is this really 'on topic' with regard to driverless last mile?

The paper needs some work on these three issues.

Author Response

REVIEWER 1

Comment1:

Response:

We thank the reviewer for the constructive criticism and for giving us the opportunity to clarify better the contribution of our paper. To this end, we would like to point out that our paper summarises some of the most representative and recent research works in the domain of autonomous driving in supply chain management tasks. However, the main contributions of our work apart from studying the literature, are to highlight and stress the opportunities of introducing autonomous driving in last-mile delivery, to reveal the technological challenges and barriers of introducing such solutions in this domain, to suggest, where possible, some conceptual alterations that should take place in order to realise true autonomous driving in urban environments and to link the automated logistics section with the last-mile delivery needs.

In order to better clarify this statement, we have added the respective clarification in the introduction section.

Comment2:

I also miss a research framework that the researchers (might) have used for their review. How do the different topics relate?

Response:

Reviewer rightly points out that the framework of the proposed method should be better clarified. To this direction, we point out that the last mile challenge will start from the automation process in the warehouse το the complete automation and evolution of the robotic system, which was adopted to solve the last mile challenge. The fundamental framework of this work consists of:

the last mile definition
challenges of last-mile
automation of the last mile and technical challenges to achieve it
robotic process in logistics and the relationship between logistics, last-mile and autonomous driving

Firstly, we define that the transit of goods from a warehouse hub to its ultimate destination is denoted as the last-mile. In urban areas, the starting point is usually a city-hub and the final points are many urban delivery points following the door-to-door paradigm. City logistics encounter many difficulties, as the last-mile tends to be the stage exhibiting the least efficiency for different reasons (for example, the growingly nagging typical final consumer and the increased cost of delivery, traffic conditions etc.) Hence, the standard way to overcome issues imposed by human labor is automation. Such novelties can improve efficacy in terms of delivery times, cost, and environmental footprint (reduced CO2 footprint and efficient energy profiles). Simulations considering platoons of robots have indicated that mixed fleets are suitable for energy-efficient last-mile logistics. Furthermore, robots constitute the ideal agents for last-mile delivery, as they can conduct contactless deliveries, especially useful in pandemic situations. Specifically, further to the typical autonomous driving solutions, which so far mainly deal with navigation in well-structured streets, with clearly visible lanes and traffic signs, last-mile delivery includes also the need for the vehicle to move in further spaces, from which it will take the goods and at which it will deliver them. Hence, the navigation scenarios are rather complex, and in need of highly robust and efficient methods for scene segmentation, understanding and prediction, as well as for robot behavior and path planning, as further explained below. The challenges about perception and cognition of vehicles, navigation and localization with or without georeferenced information, behavior and high-level mission planning are analyzed in our work, reviewing recent methodology. Finally, our framework analyzes robotic process automation in logistics, pointing out many technologies that contribute to the automation of processes within warehouses are common and adopted by autonomous vehicles for last-mile achievement.

In order to better clarify this statement, we added the above mentioned clarifications to Section 1.

Comment3:

Paragraph 4 is not liked to autonomous vehicles but to warehousing and RFID. Is this really 'on topic' with regard to driverless last mile?

The paper needs some work on these three issues.

Response:

We understand the reviewer's concern about the pertinence between robotic process automation in logistics and the last mile challenge. According to related work, the automation of the processes from the stage of storing them within the warehouse or the city hub contributes to the achievement of the last mile automation. Hence, several processes in intelligent warehouses should be automated to address autonomous goods transportation in the entire supply chain. Also, many technologies that contribute to the automation of processes within warehouses are common and adopted by autonomous vehicles that achieve delivery at the final point, for example, the autonomous navigation of AGVs or AS/RS.

In order to better clarify this statement, we have added the respective clarification in Section 4.

Reviewer 2 Report

The paper complies with the topic of the Machines Journal but it isn’t orginal and doesn’t contain new and significant information adequate to justify publication.

The driverless cars in last mile delivery is nothing new. There are already such cars used in practice (e.g. JD.com), so there is no point in theorising about the sense of using them.

The paper's argument are not built well on an appropriate base of theory and the research framework has to be improved.

Results don’t contain sufficient contributions to the theory. This is mainly due to the weakness of the proposed approach.

The paper doesn’t identify implications for practice and science.

Author Response

REVIEWER 2

Comment1:

The paper complies with the topic of the Machines Journal but it isn’t original and doesn’t contain new and significant information adequate to justify publication.

Response:

We thank the reviewer for his/her candid criticism and for giving us the opportunity to revise and improve our work. Firstly, we would like to express our dissent regarding the received comment for the novelty and the impact of our work. Our work constitutes a position and conceptual paper and we apologise if this was not clear in the submitted version of the manuscript however, in order to better clarify this, we have added the respective statements in the very beginning of the manuscript and stressed this also in the introduction section. Regarding the originality of our work we wish to state that the aggregated information is adequately cited in the paper and that we have double checked that there is no kind of plagiarism both in the initially submitted version as well as in the revised one. In addition, the main purpose and objectives of our work is to highlight the significance of the automation for the last mile delivery application through autonomous vehicles and the consistent presentation of the existing challenges in robotic perception, cognition, navigation and the warehouse automation that hinder the wide adoption of autonomous vehicles in the respective domain. To the best of our knowledge there is no other recent publication that concentrates and discusses these open issues for the respective domain. We wish to realise our work in the scientific community as a guidance for the scientific subtopics on which the attention should be concentrated in order to expedite the introduction of this technology in the examined logistics domain. Yet, in order to better clarify this we have revised the abstract and the introduction section so as to avoid any further misunderstandings.

Comment2:

The driverless cars in last mile delivery is nothing new. There are already such cars used in practice (e.g. JD.com), so there is no point in theorising about the sense of using them.

Response:

We thank the reviewer for giving us the opportunity to respond to his/her concerns. Firstly, we need to clarify that our aim is not to introduce the already known technology of the autonomous vehicles in logistics tasks. We certainly agree that autonomous vehicles is a known technology. Our aim is to aggregate and reveal those challenges that do not allow Level 5 of driving automation. Considering the very well paradigm of Jd.com which is utilised in China from the homonymous corporation, it should be mentioned that the announced driverless bots are of Level 4 of autonomous driving, meaning that theses vehicles were designed to operate under specific circumstances with very well controlled environment (e.g. the 618 Grand Promotion festival which will take place in a particular space where Geofencing is applied) and stills requires the human supervision. Our work goes one step beyond this and highlights all these challenges introduced in the uncontrolled environment that the existing infrastructures impose and need to be overcomed so as to enable Level 5 of autonomy.

For the sake of completeness of our work we also added the very good example of JD.com in our manuscript.

Comment3:

The paper's argument are not built well on an appropriate base of theory and the research framework has to be improved.

Response:

We thank the reviewer for his/her constructive criticism on which we certainly agree. In the revised version of our work (following also the Comment 2 of Reviewer 1) we have clarified the research framework we utilised and we also numbered the contributions of our work in the introduction section.

Comment4:

Results don’t contain sufficient contributions to the theory. This is mainly due to the weakness of the proposed approach.

Response:

We wish to point our reviewer at our response in his/her first comment, where we stated that we aim at a position and conceptual paper. Considering the fact that we have added the research backbone and rationale followed to present our work, we believe that the revised version of the manuscript has been significantly improved and sufficiently contributes to the theory.

Comment5:

The paper doesn’t identify implications for practice and science.

Response:

We believe that the major impact of our work is to concentrate the attention and, hence, the research efforts of the scholars towards the weaknesses of robotic perception, cognition, behavioural navigation and supply chain organisation where major flaws and limitations exist so as to expedite the introduction of Level 5 driverless vehicles in unconstrained urban and rural environments. The statement has been also added in the conclusions section.

Reviewer 3 Report

This paper attempts to present a review of the literature supporting automated last-mile delivery of goods. It introduces the idea of warehouse 4.0 and reviews some of the literature on warehouse automation, then turns to autonomous vehicles and assesses their fitness for completely autonomous delivery of goods.

The English language grammar and spelling will need significant attention (there are about a dozen major points of confusion and another dozen more minor spelling/grammatical errors or word choices that, in the opinion of this reviewer, hinder understanding.

Vol. 12, No. 1 (2019), pp. 45-72 https://www.jstor.org/stable/26911258]. Indeed, this task seems better suited to this than many others such as the scaled experiments in robo-taxis where passengers won’t wait the minutes that this escalation would take, but goods would be fine with it.

Comments for author File: Comments.pdf

Author Response

REVIEWER 3

Comment1

Response:

We have reviewed the entire manuscript for typographical and grammatical errors. As the reviewer noted, we found and corrected them in the whole manuscript. Also, we added the explanation of Figure 1 in Section 1.

Comment2:

I’m not convinced of one major premise of the paper. That is, that automated vehicles must be level 5 autonomous and capable of solving all driving/parking problems before deployment for this task - I think that some level of supervised autonomy would be acceptable for this task, with one human intervening in cases when the AV is not sure what to do [see e.g. “Hail, robo-taxi, IEEE Spectrum Jan 2017 and/or “Understanding autonomous vehicles: A systematic literature review on capability, impact, planning and policy” by Asif Faisal, Md Kamruzzaman, Tan Yigitcanlar and Graham Currie in Journal of Transport and Land Use Vol. 12, No. 1 (2019), pp. 45-72 https://www.jstor.org/stable/26911258]. Indeed, this task seems better suited to this than many others such as the scaled experiments in robo-taxis where passengers won’t wait the minutes that this escalation would take, but goods would be fine with it.

Response:

We thank the reviewer for raising our attention to the suggested alternatives regarding the existence of remote human operators for handling ambiguous cases in the navigation of automated delivery vehicles. As we already mentioned in our text, the lack of human supervision in complex environments poses important liability issues in the operation of autonomous vehicles in urban areas. The current solutions on such technologies, as it is also mentioned on the references that the reviewer suggested, require the existence of remote operators that can intervene and handle dangerous situations. However, the existence of automated vehicles along with pedestrians and other vehicles (either autonomous or not) can cause severe conflicts in the normal flow of traffic in urban scenarios. Even the escalation process may lead to a delay that will create an uncomfortable situation for the rest of the people that are involved in the case. For example, AGVs crossing an intersection interfere with multiple traffic actors and any malfunction that is not directly solved is a problem affecting everyone and not only for the goods’ transportation. In addition, companies that deploy such automated vehicles in order to optimise their benefits aim to reduce as much as possible the remote operator-to-fleet size ratio, ideally making this ratio equal to 0. These create the need for Level-5 autonomous systems, that even though they cannot be deployed at this period of time, will be a reality for the last-mile delivery services, once the issues that are mentioned in our article are solved.

Moreover, following the reviewer’s suggestion we have added the respective concern as part of the the section 3.2 related to the cognition challenges and we also included the suggested citations in our work.

Round 2

Reviewer 1 Report

I feel chapter 4 still does not belong in this paper. It is off topic.

Author Response

REVIEWER 1

Comment1:

I feel chapter 4 still does not belong in this paper. It is off topic.

Response:

The above comment was raised again by the reviewer. However, we believe that our answer was clear and we try to emphasize that in this response as well. Hence, the automation of the processes from the stage of storing them within the warehouse or the city hub contributes to the achievement of the last mile automation. Several processes in intelligent warehouses should be automated to address autonomous goods transportation in the entire supply chain. Also, many technologies that contribute to the automation of processes within warehouses are common and adopted by autonomous vehicles that achieve delivery at the final point, for example, the autonomous navigation of AGVs or AS/RS. Thus, we create Section 4 with the aim to describe the relationship between the automation of processes within the warehouse and last mile.

In order to better clarify this statement, we have added the respective clarification in Section 4.

Reviewer 2 Report

The paper complies with the topic of the Machines Journal but it isn’t original and doesn’t contain new and significant information adequate to justify publication.

Author Response

REVIEWER 2

Comment1:

The paper complies with the topic of the Machines Journal but it isn’t original and doesn’t contain new and significant information adequate to justify publication.

Response:

We wish to state the above comment has already been answered in the first round of review. Since there is not a clear direction on what exactly should be added/ modified in our paper round and the reviewer’s comment is the same, we are stressing the same arguments. We should mention that we have improved the quality of our work based on the initial posed comments and we have clarified the contributions and novel aspects of our work. In order to address the concern regarding the originality we emphasize in the discussion of the limitations and barriers of driverless last-mile delivery due to the challenges in robotic perception, cognition and navigation. To the best of our knowledge these issues have been barely discussed in other works in the relevant topic.

To this end, we summarize once again that our work constitutes a position and conceptual paper and we apologize if this was not clear in the initial submitted version of the manuscript however, in order to better clarify this, we have added the respective statements in the very beginning of the manuscript and stressed this also in the introduction section. Regarding the originality of our work we wish to state that the aggregated information is adequately cited in the paper and that we have double checked that there is no kind of plagiarism both in the initially submitted version as well as in the revised one. In addition, the main purpose and objectives of our work is to highlight the significance of the automation for the last mile delivery application through autonomous vehicles and the consistent presentation of the existing challenges in robotic perception, cognition, navigation and the warehouse automation that hinder the wide adoption of autonomous vehicles in the respective domain. To the best of our knowledge there is no other recent publication that concentrates and discusses these open issues for the respective domain. We wish to realize our work in the scientific community as a guidance for the scientific subtopics on which the attention should be concentrated in order to expedite the introduction of this technology in the examined logistics domain. Yet, in order to better clarify this we have revised the abstract and the introduction section so as to avoid any further misunderstandings.