The author proposes a Piecewise Affine Reach-avoid Computation (PARC) method to efficiently compute trajectory plans for goal-reaching near danger scenarios, outperforming existing methods in safety and performance.
Learning-based Stackelberg game-theoretic approach using Koopman operator for effective multi-robot trajectory guidance.
Introducing PEP, a lightweight equivariant planning model that integrates prediction and planning in a joint approach.
본 논문에서는 두 지점 사이의 원하는 길이를 갖는 곡률 제한 궤적을 설계하기 위한 전략을 제시하며, 이는 세 개의 원형 호를 연결하여 구성되고, 내접하는 원의 개념을 사용하여 기존 {RLR, LRL} 경로에서 {LLL, LLR, LRR, RRL, RLL, RLR, RRR} 경로까지 확장합니다.
This paper introduces a novel end-to-end visual navigation system for quadrotors that combines the strengths of deep neural networks and traditional trajectory optimization to generate dynamically feasible, agile, and efficient flight paths directly from depth images.