自己評価する知覚:視覚システムの失敗を予測する学習
Introspective Perception: Learning to Predict Failures in Vision Systems
視覚ベースのドローン飛行を対象に、センサ入力から直接システムの失敗を予測する汎用的な内省フレームワークを提案し、不確実な状況への有効性を示した。
著者: Shreyansh Daftry, Sam Zeng, J. Andrew Bagnell, Martial Hebert
分類: cs.RO, cs.AI, cs.CV
原文アブストラクト
As robots aspire for long-term autonomous operations in complex dynamic environments, the ability to reliably take mission-critical decisions in ambiguous situations becomes critical. This motivates the need to build systems that have situational awareness to assess how qualified they are at that moment to make a decision. We call this self-evaluating capability as introspection. In this paper, we take a small step in this direction and propose a generic framework for introspective behavior in perception systems. Our goal is to learn a model to reliably predict failures in a given system, with respect to a task, directly from input sensor data. We present this in the context of vision-based autonomous MAV flight in outdoor natural environments, and show that it effectively handles uncertain situations.
