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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

In Autonomous Driving (AD), real-time perception is a critical componentresponsible for detecting surrounding objects to ensure safe driving. Whileresearchers have extensively explored the integrity of AD perception due to itssafety and security implications, the aspect of availability (real-timeperformance) or latency has received limited attention. Existing works onlatency-based attack have focused mainly on object detection, i.e., a componentin camera-based AD perception, overlooking the entire camera-based ADperception, which hinders them to achieve effective system-level effects, suchas vehicle crashes. In this paper, we propose SlowTrack, a novel framework forgenerating adversarial attacks to increase the execution time of camera-basedAD perception. We propose a novel two-stage attack strategy along with thethree new loss function designs. Our evaluation is conducted on four popularcamera-based AD perception pipelines, and the results demonstrate thatSlowTrack significantly outperforms existing latency-based attacks whilemaintaining comparable imperceptibility levels. Furthermore, we perform theevaluation on Baidu Apollo, an industry-grade full-stack AD system, and LGSVL,a production-grade AD simulator, with two scenarios to compare the system-leveleffects of SlowTrack and existing attacks. Our evaluation results show that thesystem-level effects can be significantly improved, i.e., the vehicle crashrate of SlowTrack is around 95% on average while existing works only havearound 30%.

SlowTrack: Increasing the Latency of Camera-based Perception in Autonomous Driving Using Adversarial Examples