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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Terrain classification is a critical component of any autonomous mobile robotsystem operating in unknown real-world environments. Over the years, severalproprioceptive terrain classification techniques have been introduced toincrease robustness or act as a fallback for traditional vision basedapproaches. However, they lack widespread adaptation due to various factorsthat include inadequate accuracy, robustness and slow run-times. In this paper,we use vehicle-terrain interaction sounds as a proprioceptive modality andpropose a deep Long-Short Term Memory (LSTM) based recurrent model thatcaptures both the spatial and temporal dynamics of such a problem, therebyovercoming these past limitations. Our model consists of a new ConvolutionNeural Network (CNN) architecture that learns deep spatial features,complemented with LSTM units that learn complex temporal dynamics. Experimentson two extensive datasets collected with different microphones on variousindoor and outdoor terrains demonstrate state-of-the-art performance comparedto existing techniques. We additionally evaluate the performance in adverseacoustic conditions with high ambient noise and propose a noise-aware trainingscheme that enables learning of more generalizable models that are essentialfor robust real-world deployments.

Deep Spatiotemporal Models for Robust Proprioceptive Terrain Classification