A Connected Autonomous Vehicle Testbed: Capabilities, Experimental Processes and Lessons Learned
Round 1
Reviewer 1 Report
The paper “A Connected Autonomous Vehicle Testbed: Capabilities, Experimental Processes and Lessons Learned” presents an overview of the technology development during the VENTURER project regarding Connected Autonomous Vehicle implementation. It describes the briefly the main hardware involved (the BAE Wildcat vehicle), the Decision-Making System (DMS), the objectives of conducted trials and some lessons learned from developing the project. The complexity of the project is impressive and the value for development of CAVs is incontestable.
The presented system uses a complete set of sensors for environment perception as radars, cameras, GPS, and LIDARs. This large amount of collected data is then used by the DMS to decide on decides on driving strategy, high-level navigation, and desired trajectories on the road. The DMS is implemented in Phyton as a behaviour tree to work under the Robot Operating System. The trials scenarios have included control handover between manual and autonomous operation, operation and response at junctions with other vehicles, autonomous vehicle interactions with cyclists and pedestrians, and a technology demonstration use case involving Wireless Communication between another vehicle and the CAV.
The paper is well written and well structured. It includes many details about project organization and management. The developed technologies are clearly summarized. The authors also succeed to describe the problems related interconnections and operational challenges. However, there are some points that can be improved. The abstract does not clearly point the subject and the purpose of the paper. The technical details are missing in many parts of the presentation. Apart of some important lesson learned presented at the end of the paper, no results are used to support the conclusions.
Author Response
We greatly appreciate the effort that has gone into reviewing and giving feedback on our manuscript. We have tried to address the points raised and have summarised them below. Changes specific to this reviewer are annoyed with "R1" in blue on the changes document.
1) The abstract does not clearly point the subject and the purpose of the paper.
> We have added some extra clarification to outline the aim of the paper.
2) The technical details are missing in many parts of the presentation. Apart of some important lesson learned presented at the end of the paper, no results are used to support the conclusions.
> The nature of this paper is somewhat subjective to our own experiences with the project, the focus has not been to detail results of the individual building blocks, but instead to create a high-level view of the project.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The paper deals with the outcome of the Venturer project which is a cooperation between multiple academic and industrial partners that aims to demonstrate the in both simulation and real experimental contexts the results of the deployement of the system.
The paper is well written, and even if there was not much details due to IP constraints, it explains the different aspect of the system going from a clear presentation of the experimental setup, research design, and results and some challenges that faced the project.
The conclusions gave previous insights not only for those who are interested in conducting similar research projects, but for mobile autonomous robots in general, and this was appreciated.
It could have been more benefical to the reader if the authors shared (if any) relevated publications that deals with the technical details of the different parts of the project: sensor fusion, vehicle state estimation, path planning and execution, obstacle detection, etc...
Author Response
Many thanks for taking the time to review out submission and give feedback. The only comment that we think needed feedback is as follows:
1) It could have been more beneficial to the reader if the authors shared (if any) related publications that deals with the technical details of the different parts of the project: sensor fusion, vehicle state estimation, path planning and execution, obstacle detection, etc...
> Currently there are no other publications of this type from the project, however they are ongoing. There is another linked paper from this project (Morgan et. al 2018), however this focuses more on the handover procedures between participants and the autonomous car (part of Trial 1).
Author Response File: Author Response.pdf
Reviewer 3 Report
At the manuscript entitled “A Connected Autonomous Vehicle Testbed: Capabilities, Experimental Processes and Lessons Learned”, the authors present the results of the funded research project VENTURER on the Connected Autonomous Vehicles. The project ended at September 2018. There are a lot of relevant resources at the project’s site https://www.venturer-cars.com/venturer-project/ . The manuscript is well written, and it explains the project’s objectives, the experimental processes and the lessons learned. Some minor comments about the text:
- Page 2, Line 61 BAE Systems is the name of partner, but it is not clear here (it is explained at page 3)
- At figure 2 please explain the Comms tag. It is not clear
- At line 239 I would prefer the Unmanned Aerial Vehicle definition for the UAV, but maybe the authors would like to give emphasis to another context
- At figure 3 and I would suggest the classic workflow diagram symbols (for example the diamond/rhombus shape for the decisions)
- Figure 8 is not of good quality and thus it is not clear
- Please explain every acronym before use for example Lines 388 and 389 please explain the acronyms V2I and V2V and the same for line 401 CCAV.
Author Response
The Authors really appreciate the reviews and feedback provided. We have outlined what we think were your key points below and attempted to answer them in turn.
1) Page 2, Line 61 BAE Systems is the name of partner, but it is not clear here (it is explained at page 3)
> We have added some extra context when we first use BAE systems.
2) At figure 2 please explain the Comms tag. It is not clear
> A more detailed description has been added to the text to more clearly describe each of blocks.
3) At line 239 I would prefer the Unmanned Aerial Vehicle definition for the UAV, but maybe the authors would like to give emphasis to another context
> This has been changed
4) At figure 3 and I would suggest the classic workflow diagram symbols (for example the diamond/rhombus shape for the decisions)
> The symbols used are what is typically used in behaviour tree literature. As its important to differentiate between Conditional and Selector blocks
5) Figure 8 is not of good quality and thus it is not clear
> The Sourc feed itself is only of that quality, however we have updated the image to make the text clearer and updated the gui to only show the relevant information (estimated distance and speed of detection).
6) Please explain every acronym before use for example Lines 388 and 389 please explain the acronyms V2I and V2V and the same for line 401 CCAV.
> These have now all been updated to explain each acronym
Author Response File: Author Response.pdf
