Page 67 - Read Online

Boin et al. Intell Robot 2022;2(2):145­67 I http://dx.doi.org/10.20517/ir.2022.11 Page 165

that were controlled by DDPG agents trained without FRL. These results are backed by simulations performed
using models trained across four random seeds, and an additional simulation set with variable platoon sizes.
The focus of this paper was on decentralized platoon control, where each follower in the platoon trains locally
with respect to their individual reward.

In the future, improvements to the system could be made by implementing weighted averaging in the FRL
aggregation method. Moreover, in AV platooning, communication delays can be considered in the model to
give a more concrete real life example.

DECLARATIONS
Authors’ contributions
Made substantial contributions to the research, idea generation, testing, and software development. Solely pro-
grammed the Python AVDDPG application for conducting the DRL experiments, simulating and aggregating
experiment results. Wrote and edited the original draft: Boin C
Performed oversight and leadership responsibility for the research activity planning and execution, as well as
developed ideas and evolution of overarching research aims. Assisted editing the original draft: Lei L
Performed critical review, commentary and revision, as well as provided administrative, technical, and mate-
rial support: Yang S

Availability of data and materials
Not applicable.

Financial support and sponsorship
This work was supported by the Natural Sciences and Engineering Research Council (NSERC) of Canada and
the CARE-AI Seed Fund at the University of Guelph.

Conflicts of interest
The authors declared that there are no conflicts of interest.

Ethical approval and consent to participate
Not applicable.

Consent for publication
Not applicable.

Copyright
©The Author(s) 2022.