A CNN-Based System for Mobile Robot Navigation in Indoor Environments via Visual Localization with a Small Dataset

Abstract

Deep learning has made great advances in the field of image processing, which allows automotive devices to be more widely used in humans’ daily lives than ever before. Nowadays, the mobile robot navigation system is among the hottest topics that researchers are trying to develop by adopting deep learning methods. In this paper, we present a system that allows the mobile robot to localize and navigate autonomously in the accessible areas of an indoor environment. The proposed system exploits the Convolutional Neural Network (CNN) model’s advantage to extract data feature maps for image classification and visual localization, which attempts to precisely determine the location region of the mobile robot focusing on the topological maps of the real environment. The system attempts to precisely determine the location region of the mobile robot by integrating the CNN model and topological map of the robot workspace. A dataset with small numbers of images is acquired from the MYNT EYE camera. Furthermore, we introduce a new loss function to tackle the bounded generalization capability of the CNN model in small datasets. The proposed loss function not only considers the probability of the input data when it is allocated to its true class but also considers the probability of allocating the input data to other classes rather than its actual class. We investigate the capability of the proposed system by evaluating the empirical studies based on provided datasets. The results illustrate that the proposed system outperforms other state-of-the-art techniques in terms of accuracy and generalization capability.

Research on Intelligent Pick-Up Route Planning of a Logistics Cycle Automatic Robot

“…At first, the Voronoi diagram was used to solve the proximity problem of plane points in mathematics but some scholars soon introduced it into robot path planning. In this method, the Voronoi diagram can be used to describe the networked structure of feasible areas in the robot working environment [10]. Cai and Zheng proposed robot path planning based on the grid method, which was proposed relatively early and widely used.…”

“…In formula (10), cross represents the position of the intersection between the optimized two paths, beg is the starting If the d cross of the two robots are the same, it means that the two robots will collide when moving along the optimized path and the number of collisions between paths will be counted; otherwise, there will be no collision.…”

Base de datos abierta TRAM-FPV RACING. Secuencias completas de vuelo indoor para el estudio de drones de carreras.

“…Esto es para entrenar los movimientos de los drones de carreras autónomos o para validar sus comportamientos en espacios cerrados bajo ciertas condiciones de seguridad. [23,32,36].…”

Розроблення Автономної Системи Керування Мобільною Робототехнічною Системою З Використанням Моделей На Підставі Штучних Нейромереж

“…Results of starting the program on a real prototype Поширеним рішенням навігації мобільної роботизованої системи на підставі штучних нейромереж є вико-ристання набору фотографій потенційної місцевості, як вхідних даних для навчання [6]. Недоліком використання таких методів є необхідність створення та сортування великих масивів даних задля тренування моделей, що можуть приймати рішення в різних середовищах.…”