LATIDIA · Robótica

Espectrales y conformes: métodos de red neuronal para la predicción de trayectoria calibrada e intervalos de punto de control

arXiv:2611.00231v1 Anuncio Tipo: cross Resumen: Los sistemas autónomos que navegan por su entorno necesitan métodos para evaluar rápidamente la viabilidad y la seguridad de las trayectorias deseadas. Las soluciones tradicionales que se basan en n

La noticia

arXiv:2611.00231v1 Announce Type: cross Abstract: Autonomous systems that navigate their environment need methods to quickly assess the feasibility and safety of desired trajectories. Traditional solutions that rely on numerical integration can be computationally heavy, while fast data-driven surrogates lack the formal guarantees that numerical approaches provide. This work demonstrates a framework that uses (1) a neural network with a spectral backbone to learn trajectories and quantile bands with built-in constraints (initial conditions and stability); (2) a modified conformal quantile regression to provide calibrated, joint trajectory-level coverage; and (3) interval arithmetic to extend noise-free coverage guarantees to control points. After refactoring the network for inference, the trajectory inference time for a