A mobile robot platform for supervised machine learning applications

Abstract

In supervised machine learning applications, a data set of training and validation features and labels is required to train a neural network. In this paper, we present a remote-controlled, mobile robot and describe software used to generate a data set for vision-based, supervised machine learning applications. We present results from an experiment, which validates the developed platform, and also demonstrate the use of a generated data set for training of a convolutional neural network to detect objects. We have uploaded the platform's hardware and software source files to GitHub; where, this work is freely available.