دورية أكاديمية

There and Back Again: Gene-Processing Hardware for the Evolution and Robotic Deployment of Robust Navigation Strategies

التفاصيل البيبلوغرافية
العنوان: There and Back Again: Gene-Processing Hardware for the Evolution and Robotic Deployment of Robust Navigation Strategies
المؤلفون: David M. Bryson, Aaron P. Wagner, Charles Ofria
المساهمون: The Pennsylvania State University CiteSeerX Archives
المصدر: http://devolab.msu.edu/files/2014/11/p689-bryson.pdfTest.
المجموعة: CiteSeerX
مصطلحات موضوعية: Algorithms, Design, Reliability. *These authors contributed equally to this work. Keywords Genetic architecture. Robustn
الوصف: Navigation strategies represent some of the most intriguing examples of complex and intelligent behaviors in nature. Accordingly, they have been the focus of extensive research in animal behavior and in evolutionary robotics. However, engineering successes in harnessing the evolutionary dynamics that shape sophisticated navigation strategies remain limited. Here we describe a novel gene-processing architecture for digital organisms that enables the evolution of central-place-foraging strategies, such as those seen in honeybees and striped hyena. While previous studies have evolved navigation de novo, the resulting algorithms have been relatively fragile and difficult to translate into physical systems. In contrast, the strategies evolved in this study are highly congruous with those seen in nature: a single evolved foraging strategy incorporates periods of directed travel, fixed pattern search, cue response, and reorientation when outcomes do not match expected results. Additionally, the genetic architecture enabled rapid extraction of the underlying behavioral algorithm and transference to a robotic system, proving to be robust to issues of noise and scale that commonly plague such attempts. Accordingly, we demonstrate that the flexibility and interpretability of the new gene-processing hardware readily facilitate the creation, study, and utilization of naturalistic and deployable algorithms for functionally complex behaviors.
نوع الوثيقة: text
وصف الملف: application/pdf
اللغة: English
العلاقة: http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.673.2638Test; http://devolab.msu.edu/files/2014/11/p689-bryson.pdfTest
الإتاحة: http://devolab.msu.edu/files/2014/11/p689-bryson.pdfTest
حقوق: Metadata may be used without restrictions as long as the oai identifier remains attached to it.
رقم الانضمام: edsbas.A16AC289
قاعدة البيانات: BASE