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

Autonomous vehicles should be able to generate accurate probabilisticpredictions for uncertain behavior of other road users. Moreover, reactivepredictions are necessary in highly interactive driving scenarios to answer"what if I take this action in the future" for autonomous vehicles. There is noexisting unified framework to homogenize the problem formulation,representation simplification, and evaluation metric for various predictionmethods, such as probabilistic graphical models (PGM), neural networks (NN) andinverse reinforcement learning (IRL). In this paper, we formulate aprobabilistic reaction prediction problem, and reveal the relationship betweenreaction and situation prediction problems. We employ prototype trajectorieswith designated motion patterns other than "intention" to homogenize therepresentation so that probabilities corresponding to each trajectory generatedby different methods can be evaluated. We also discuss the reasons why"intention" is not suitable to serve as a motion indicator in highlyinteractive scenarios. We propose to use Brier score as the baseline metric forevaluation. In order to reveal the fatality of the consequences when thepredictions are adopted by decision-making and planning, we propose afatality-aware metric, which is a weighted Brier score based on the criticalityof the trajectory pairs of the interacting entities. Conservatism andnon-defensiveness are defined from the weighted Brier score to indicate theconsequences caused by inaccurate predictions. Modified methods based on PGM,NN and IRL are provided to generate probabilistic reaction predictions in anexemplar scenario of nudging from a highway ramp. The results are evaluated bythe baseline and proposed metrics to construct a mini benchmark. Analysis onthe properties of each method is also provided by comparing the baseline andproposed metric scores.

Towards a Fatality-Aware Benchmark of Probabilistic Reaction Prediction in Highly Interactive Driving Scenarios