English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Reinforcement learning techniques can provide substantial insights into thedesired behaviors of future autonomous driving systems. By optimizing forsocietal metrics of traffic such as increased throughput and reduced energyconsumption, such methods can derive maneuvers that, if adopted by even a smallportion of vehicles, may significantly improve the state of traffic for allvehicles involved. These methods, however, are hindered in practice by thedifficulty of designing efficient and accurate models of traffic, as well asthe challenges associated with optimizing for the behaviors of dozens ofinteracting agents. In response to these challenges, this paper tackles theproblem of learning generalizable traffic control strategies in simplerepresentations of vehicle driving dynamics. In particular, we look tomixed-autonomy ring roads as depictions of instabilities that result in theformation of congestion. Within this problem, we design a curriculum learningparadigm that exploits the natural extendability of the network to effectivelylearn behaviors that reduce congestion over long horizons. Next, we study theimplications of modeling lane changing on the transferability of policies. Ourfindings suggest that introducing lane change behaviors that even approximatelymatch trends in more complex systems can significantly improve thegeneralizability of subsequent learned models to more accurate multi-lanemodels of traffic.

Learning Generalizable Multi-Lane Mixed-Autonomy Behaviors in Single Lane Representations of Traffic