Skip to main content
Fachgebiete
Automobil + Motoren Bauwesen + Immobilien Business IT + Informatik Elektrotechnik + Elektronik Energie + Nachhaltigkeit Finance + Banking Management + Führung Marketing + Vertrieb Maschinenbau + Werkstoffe Versicherung + Risiko
DE
EN
Bücher
Zeitschriften
Themenseiten
Organisationspsychologie Projektmanagement Marketing Smart Manufacturing
Weitere Formate
Podcasts Webinare Technik Webinare Wirtschaft Kongresse Veranstaltungskalender Awards
Jetzt Einzelzugang starten
Zugang für Unternehmen
Referenzkunden
SLX-Digitalkonferenz: Zukunftswerkstatt 2024

Mehr Umsatz mit Top-Kontakten

Am 7. Mai erfahren Sie, wie Vertriebsteams Leadgenerierung, Leadnurturing und Leadstrategien effizient steuern können.
Erweiterte Suche
Anmelden
nach oben

2014 | Buch

Introduction Kapitel lesen Erstes Kapitel lesen

Autonomous Robotics and Deep Learning

verfasst von: Vishnu Nath, Stephen E. Levinson

Verlag: Springer International Publishing

Buchreihe : SpringerBriefs in Computer Science

Enthalten in: Springer Professional "Wirtschaft+Technik" , Springer Professional "Technik" , Springer Professional "Wirtschaft"

Einloggen, um Zugang zu erhalten
insite
SUCHEN

Über dieses Buch

This Springer Brief examines the combination of computer vision techniques and machine learning algorithms necessary for humanoid robots to develop “true consciousness.” It illustrates the critical first step towards reaching “deep learning,” long considered the holy grail for machine learning scientists worldwide. Using the example of the iCub, a humanoid robot which learns to solve 3D mazes, the book explores the challenges to create a robot that can perceive its own surroundings. Rather than relying solely on human programming, the robot uses physical touch to develop a neural map of its environment and learns to change the environment for its own benefit. These techniques allow the iCub to accurately solve any maze, if a solution exists, within a few iterations. With clear analysis of the iCub experiments and its results, this Springer Brief is ideal for advanced level students, researchers and professionals focused on computer vision, AI and machine learning.

Inhaltsverzeichnis

Frontmatter
Chapter 1. Introduction
Abstract
Almost all of us have watched plenty of sci-fi movies that revolve around the same theme: scientists develop a robot that has true consciousness, observes the environment around it and realizes that human beings are an unnecessary part of the equation. Half an hour into the movie and we have an army of intelligent robots that are hell-bent on destroying the human civilization and the protagonist manages to explode a central server or insert a virus into it, deactivating all the robots and saving our civilization. In spite of all this, most of us would like robots to perform various day-to-day activities that we hate doing, freeing up our time for activities that we enjoy a lot more. The commercial success of Roomba is proof that consumers are willing to pay for personal assistants and more importantly, that they do not have a deep mistrust for robots. Perhaps the form factor of a Roomba might be crucial for people to think that in case of a robot uprising, there isn’t much to worry about from a Roomba!
Vishnu Nath, Stephen E. Levinson
Chapter 2. Overview of Probability and Statistics
Abstract
This chapter talks about the elementary concepts of probability and statistics that are needed to better comprehend this book. This appendix covers topics like basic probability, conditional probability, Bayes’ Theorem and various distributions like normal distribution (also called Gaussian distribution), Bernoulli distribution, Poisson distribution and binomial distribution.
Vishnu Nath, Stephen E. Levinson
Chapter 3. Primer on Matrices and Determinants
Abstract
In this chapter, we will be covering the basic concepts of matrices, determinants and, eigenvalues and eigenvectors in this chapter. If the reader is familiar with these concepts, then the reader can skip ahead to the next chapter without any loss of continuity.
Vishnu Nath, Stephen E. Levinson
Chapter 4. Robot Kinematics
Abstract
The robotic platform is the physical hardware on which the experiments have been conducted. All algorithms, by definition, should be replicable on any physical machine, irrespective of the individual hardware components. However, all other things being constant, there is no denying that algorithms perform better on more capable hardware. In this chapter, we provide an introduction to the physical characteristics of the iCub robot platform that was used to perform the experiments and benchmark it using parameters that are relevant to the domain of robotics.
Vishnu Nath, Stephen E. Levinson
Chapter 5. Computer Vision
Abstract
In this chapter, we present the various components of the computer vision algorithms that were used for the various aspects of the project. Initially, the chapter discusses the underlying algorithms of computer vision from a mathematical standpoint. Once this aspect has been completed, the next step would be to demonstrate to the reader how we incorporated the algorithms to fir the specific problem that the research project intended to solve.
Vishnu Nath, Stephen E. Levinson
Chapter 6. Machine Learning
Abstract
Whenever a problem seems extremely open ended with a large variety of random variables that have an effect on the process, it is impossible for a human programmer to be able to account for every single case. The number of cases increases dramatically with an additional parameter. In such scenarios, probabilistic algorithms have the greatest applicability. The algorithms need to be given a couple of examples of scenarios it might come across and the algorithm would be able to handle a new scenario with reasonable accuracy. The key word in the previous statement is “reasonable”. There is no probabilistic algorithm that will always return the optimum result with a probability of 1. That would make it a deterministic algorithm which, as has just been discussed, cannot handle every potential case. In this chapter, we discuss the algorithms that were employed to successfully complete the experiment.
Vishnu Nath, Stephen E. Levinson
Chapter 7. Experimental Results
Abstract
In this chapter, we discuss the results that were obtained while trying to solve an unknown maze by the iCub using the algorithm discussed in Chap. 6.
Vishnu Nath, Stephen E. Levinson
Chapter 8. Future Direction
Abstract
A simple literature review will reveal a plethora of maze solving algorithms. However, merely executing an algorithm in a sequence of steps is not worthy of present day research, simply because there is nothing to differentiate it from the millions of sequential executions taking place in an ordinary CPU. Furthermore, if the system finds itself in an unknown state, most algorithms do not have a fallback mechanism built into the algorithm. Engineers will have to add some sort of safety check just to get the system back to a known condition. There are times when even this is not sufficient since the exact sequence of states and timing is crucial for the successful termination of the algorithm.
Vishnu Nath, Stephen E. Levinson
Metadaten
Titel
Autonomous Robotics and Deep Learning
verfasst von
Vishnu Nath
Stephen E. Levinson
Copyright-Jahr
2014
Electronic ISBN
978-3-319-05603-6
Print ISBN
978-3-319-05602-9
DOI
https://doi.org/10.1007/978-3-319-05603-6