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

Modern autonomous vehicle systems use complex perception and controlcomponents. These components can rapidly change during development of suchsystems, requiring constant re-testing. Unfortunately, high-fidelitysimulations of these complex systems for evaluating vehicle safety are costly.The complexity also hinders the creation of less computationally intensivesurrogate models.  We present GAS, the first approach for creating surrogate models of complete(perception, control, and dynamics) autonomous vehicle systems containingcomplex perception and/or control components. GAS's two-stage approach firstreplaces complex perception components with a perception model. Then, GASconstructs a polynomial surrogate model of the complete vehicle system usingGeneralized Polynomial Chaos (GPC). We demonstrate the use of these surrogatemodels in two applications. First, we estimate the probability that the vehiclewill enter an unsafe state over time. Second, we perform global sensitivityanalysis of the vehicle system with respect to its state in a previous timestep. GAS's approach also allows for reuse of the perception model when vehiclecontrol and dynamics characteristics are altered during vehicle development,saving significant time.  We consider five scenarios concerning crop management vehicles that must notcrash into adjacent crops, self driving cars that must stay within their lane,and unmanned aircraft that must avoid collision. Each of the systems in thesescenarios contain a complex perception or control component. Using GAS, wegenerate surrogate models for these systems, and evaluate the generated modelsin the applications described above. GAS's surrogate models provide an averagespeedup of $3.7\times$ for safe state probability estimation (minimum$2.1\times$) and $1.4\times$ for sensitivity analysis (minimum $1.3\times$),while still maintaining high accuracy.

GAS: Generating Fast and Accurate Surrogate Models for Autonomous Vehicle Systems