An Active Learning Framework for Constructing High-Fidelity Mobility Maps

Gary R. Marple

doi:10.1109/TVT.2021.3107338

An Active Learning Framework for Constructing High-Fidelity Mobility Maps

2021-10-01

5

OA

AI

Shravan Veerapaneni *

DOI：10.1109/TVT.2021.3107338

Original

Original request for help

Save

Abstract

En 中文

Recent workat the U.S. Army CCDC Ground Vehicle Systems Center has shown that machine learning classifiers can quickly construct high-fidelity mobility maps. Training these classifiers, on the other hand, is still a challenge, since each data instance is labeled by performing a computationally intensive, physics-based simulation. In this paper we introduce an active learning framework, based on the query-by-bagging algorithm, that substantially reduces the number of simulations needed to train a classifier. Experimental results suggest that our sampling algorithm can train a neural network, with higher accuracy, using less than half the number of simulations when compared to random sampling.

Keywords:

Training

Computational modeling

Predictive models

Uncertainty

Prediction algorithms

Data models

Soil

Autonomous vehicles

machine learning

path planning

An Active Learning Framework for Constructing High-Fidelity Mobility Maps

Abstract

Journal

Organization