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

We investigate what grade of sensor data is required for training animitation-learning-based AV planner on human expert demonstration.Machine-learned planners are very hungry for training data, which is usuallycollected using vehicles equipped with the same sensors used for autonomousoperation. This is costly and non-scalable. If cheaper sensors could be usedfor collection instead, data availability would go up, which is crucial in afield where data volume requirements are large and availability is small. Wepresent experiments using up to 1000 hours worth of expert demonstration andfind that training with 10x lower-quality data outperforms 1x AV-grade data interms of planner performance. The important implication of this is that cheapersensors can indeed be used. This serves to improve data access and democratizethe field of imitation-based motion planning. Alongside this, we perform asensitivity analysis of planner performance as a function of perception range,field-of-view, accuracy, and data volume, and the reason why lower-quality datastill provide good planning results.

What data do we need for training an AV motion planner?