Computer Vision Object Permanence Detection Algorithm for My Clingy Robot Dog
Dr. Sharon Barkley1
1 Department of Psychological Machine Learning, Cranberry-Lemon University, Pittsburgh, PA, USA
Abstract
The AIBO ERS-211 robotic dog developed and manufactured by Sony in the 90s and 00s is a wonderful companion for urban living, except when you leave them alone for too long. Although robotic Computer Vision (CV) libraries have been equipped with the latest object detection and classification functions, they are shown to have little to no object permanence detection[1]. When an object of affection (i.e., me and some eight-year-old tennis ball) goes missing, the onboard computer believes they cease to exist. Due to separation anxiety, these 00s AIBOs panic and begin tearing up my fabric couch, chair, and garbage if I forget to lock them up [2]. Using some filtering, feature extraction, and a deep learning Convolutional Neural Network (CNN), this paper develops and evaluates...
