Roundabout Trajectory Planning: Integrating Human Driving Models for Autonomous Vehicles

Round 1

Reviewer 1 Report

The paper provides a detailed description of two key aspects of autonomous driving: the path model and the speed model for crossing a single-lane roundabout. The path model is designed to determine the optimal trajectory for the autonomous vehicle. and utilizes the Levenberg-Marquardt algorithm to optimize the trajectory and minimize the difference between the desired and actual trajectories.

The paper provides a description of the model calibration stage, which involves adjusting parameters for the speed model to best fit human driving data. This process is important to ensure that the autonomous vehicle exhibits behavior similar to human behavior while crossing the roundabout.

The model has a limitation, it relies solely on the geometric characteristics of the roundabout and does not consider external factors such as weather conditions or road surface conditions.

I would have liked to see a more detailed presentation of the experimental part. 

Overall, the paper offers a detailed and logically structured presentation of the process of modeling the trajectory and speed profile for autonomous vehicles while crossing single-lane roundabouts. It is a technical and scientific paper that could be used in the development of autonomous driving systems and in research in the field.

Author Response

See file in attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

I consider that the following key aspects of the paper are correctly developed:

1) The main goal is to develop a planning model for autonomous vehicles to cross single-lane roundabouts that mimics natural human driving behavior in the absence of traffic. 

2) A path model is proposed to represent the fastest vehicle trajectory through the roundabout based on geometry. It uses waypoints and least squares optimization.

3) A two-stage speed model is designed to capture human speed patterns - first generating a basic profile from path geometry, then iteratively smoothing to constrain jerk.

4) The speed model has 9 parameters that are calibrated using data from an experimental study with 15 drivers crossing 5 roundabouts. Linear regressions find correlations between parameters and path geometry.

5) The calibrated model is evaluated by comparing generated speed plans to additional human driving data. Results show good alignment, capturing essential features of human behavior statistically. 

6) Benefits include predictable/safe behavior, optimizing traffic flow, and adapting to different roundabout designs. Limitations include reliance on geometry and inability to account for external factors.

7) Future work aims to transition from theory to real-world implementation and incorporate interactions with other vehicles into the planning model.

In summary, the paper presents a promising human-inspired approach to planning autonomous vehicle trajectories through roundabouts, with applications for enhancing safety and user experience.

The figures, tables and equations are legible and of good quality.

Its publication is recommended.

Author Response

See file in attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

The article investigates trajectory planning， but has some major problems as follows，

1. The methodology proposed in the article is not compared with other methods and the exact contribution is unclear

2. Authors should highlight the significance of autonomous vehicles in the introduction section of the article, e.g., some accident data introduction, etc., e.g.,https://doi.org/10.1186/s10033-021-00675-z

3. The framework diagram of the author's article should be supplemented with a description to make it easier for the reader to understand it

4. Some of the writing and layout of the article needs improvement, such as this word in the abstract " MRoundabout“

need to be polished

Author Response

See file in attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

no more questions

well