A Neural Circuit for Acoustic Navigation combining Heterosynaptic and Non-synaptic Plasticity that learns Stable Trajectories

Publikation: Bidrag til bog/antologi/rapport/konference-proceedingKonferencebidrag i proceedingsForskningpeer review

Resumé

Reactive spatial robot navigation in goal-directed tasks such as phonotaxis requires generating consistent and stable trajectories towards an acoustic target while avoiding obstacles. High-level goal-directed steering behaviour can steer a robot towards the target by mapping sound direction information to appropriate wheel velocities. However, low-level obstacle avoidance behaviour based on distance sensors may significantly alter wheel velocities and temporarily direct the robot away from the sound source, creating conflict between the two behaviours. How can such a conflict in reactive controllers be resolved in a manner that generates consistent and stable robot trajectories? We propose a neural circuit that minimises this conflict by learning sensorimotor mappings as neuronal transfer functions between the perceived sound direction and wheel velocities of a simulated non-holonomic mobile robot. These mappings constitute the high-level goal-directed steering behaviour. Sound direction information is obtained from a model of the lizard peripheral auditory system. The parameters of the transfer functions are learned via an online unsupervised correlation learning algorithm through interaction with obstacles in the form of low-level obstacle avoidance behaviour in the environment. The simulated robot is able to navigate towards a virtual sound source placed 3m away that continuously emits a tone of frequency 2.2 kHz, while avoiding randomly placed obstacles in the environment. We demonstrate through two independent trials in simulation that in both cases the neural circuit learns consistent and stable trajectories as compared to navigation without learning.
OriginalsprogEngelsk
TitelEngineering Applications of Neural Networks : 18th International Conference Engineering Applications of Neural Networks
RedaktørerGiacomo Boracchi, Lazaros Iliadis, Chrisina Jayne, Aristidis Likas
Udgivelses stedCham
ForlagSpringer
Publikationsdato2017
Sider544-555
ISBN (Trykt)978-3-319-65171-2
ISBN (Elektronisk)978-3-319-65172-9
DOI
StatusUdgivet - 2017
Begivenhed18th International Conference on Engineering Applications of Neural Networks - Novus City Hotel, Athens, Grækenland
Varighed: 25. aug. 201727. aug. 2017
https://conferences.cwa.gr/eann2017/

Konference

Konference18th International Conference on Engineering Applications of Neural Networks
LokationNovus City Hotel
LandGrækenland
ByAthens
Periode25/08/201727/08/2017
Internetadresse
NavnCommunications in Computer and Information Science
Vol/bind744
ISSN1865-0929

Fingeraftryk

Plasticity
Navigation
Acoustics
Trajectories
Acoustic waves
Robots
Networks (circuits)
Wheels
Collision avoidance
Transfer functions
Mobile robots
Learning algorithms
Controllers
Sensors

Citer dette

Shaikh, D., & Manoonpong, P. (2017). A Neural Circuit for Acoustic Navigation combining Heterosynaptic and Non-synaptic Plasticity that learns Stable Trajectories. I G. Boracchi, L. Iliadis, C. Jayne, & A. Likas (red.), Engineering Applications of Neural Networks: 18th International Conference Engineering Applications of Neural Networks (s. 544-555). Cham: Springer. Communications in Computer and Information Science, Bind. 744 https://doi.org/10.1007/978-3-319-65172-9_46
Shaikh, Danish ; Manoonpong, Poramate. / A Neural Circuit for Acoustic Navigation combining Heterosynaptic and Non-synaptic Plasticity that learns Stable Trajectories. Engineering Applications of Neural Networks: 18th International Conference Engineering Applications of Neural Networks. red. / Giacomo Boracchi ; Lazaros Iliadis ; Chrisina Jayne ; Aristidis Likas. Cham : Springer, 2017. s. 544-555 (Communications in Computer and Information Science, Bind 744).
@inproceedings{f1b5805a203d4155aa19e02099edd6b4,
title = "A Neural Circuit for Acoustic Navigation combining Heterosynaptic and Non-synaptic Plasticity that learns Stable Trajectories",
abstract = "Reactive spatial robot navigation in goal-directed tasks such as phonotaxis requires generating consistent and stable trajectories towards an acoustic target while avoiding obstacles. High-level goal-directed steering behaviour can steer a robot towards the target by mapping sound direction information to appropriate wheel velocities. However, low-level obstacle avoidance behaviour based on distance sensors may significantly alter wheel velocities and temporarily direct the robot away from the sound source, creating conflict between the two behaviours. How can such a conflict in reactive controllers be resolved in a manner that generates consistent and stable robot trajectories? We propose a neural circuit that minimises this conflict by learning sensorimotor mappings as neuronal transfer functions between the perceived sound direction and wheel velocities of a simulated non-holonomic mobile robot. These mappings constitute the high-level goal-directed steering behaviour. Sound direction information is obtained from a model of the lizard peripheral auditory system. The parameters of the transfer functions are learned via an online unsupervised correlation learning algorithm through interaction with obstacles in the form of low-level obstacle avoidance behaviour in the environment. The simulated robot is able to navigate towards a virtual sound source placed 3m away that continuously emits a tone of frequency 2.2 kHz, while avoiding randomly placed obstacles in the environment. We demonstrate through two independent trials in simulation that in both cases the neural circuit learns consistent and stable trajectories as compared to navigation without learning.",
keywords = "behaviour-based robotics, reactive navigation, phonotaxis, lizard peripheral auditory system, synaptic plasticity, correlation-based learning",
author = "Danish Shaikh and Poramate Manoonpong",
year = "2017",
doi = "10.1007/978-3-319-65172-9_46",
language = "English",
isbn = "978-3-319-65171-2",
series = "Communications in Computer and Information Science",
publisher = "Springer",
pages = "544--555",
editor = "Giacomo Boracchi and Lazaros Iliadis and Chrisina Jayne and Aristidis Likas",
booktitle = "Engineering Applications of Neural Networks",
address = "Germany",

}

Shaikh, D & Manoonpong, P 2017, A Neural Circuit for Acoustic Navigation combining Heterosynaptic and Non-synaptic Plasticity that learns Stable Trajectories. i G Boracchi, L Iliadis, C Jayne & A Likas (red), Engineering Applications of Neural Networks: 18th International Conference Engineering Applications of Neural Networks. Springer, Cham, Communications in Computer and Information Science, bind 744, s. 544-555, 18th International Conference on Engineering Applications of Neural Networks, Athens, Grækenland, 25/08/2017. https://doi.org/10.1007/978-3-319-65172-9_46

A Neural Circuit for Acoustic Navigation combining Heterosynaptic and Non-synaptic Plasticity that learns Stable Trajectories. / Shaikh, Danish; Manoonpong, Poramate.

Engineering Applications of Neural Networks: 18th International Conference Engineering Applications of Neural Networks. red. / Giacomo Boracchi; Lazaros Iliadis; Chrisina Jayne; Aristidis Likas. Cham : Springer, 2017. s. 544-555 (Communications in Computer and Information Science, Bind 744).

Publikation: Bidrag til bog/antologi/rapport/konference-proceedingKonferencebidrag i proceedingsForskningpeer review

TY - GEN

T1 - A Neural Circuit for Acoustic Navigation combining Heterosynaptic and Non-synaptic Plasticity that learns Stable Trajectories

AU - Shaikh, Danish

AU - Manoonpong, Poramate

PY - 2017

Y1 - 2017

N2 - Reactive spatial robot navigation in goal-directed tasks such as phonotaxis requires generating consistent and stable trajectories towards an acoustic target while avoiding obstacles. High-level goal-directed steering behaviour can steer a robot towards the target by mapping sound direction information to appropriate wheel velocities. However, low-level obstacle avoidance behaviour based on distance sensors may significantly alter wheel velocities and temporarily direct the robot away from the sound source, creating conflict between the two behaviours. How can such a conflict in reactive controllers be resolved in a manner that generates consistent and stable robot trajectories? We propose a neural circuit that minimises this conflict by learning sensorimotor mappings as neuronal transfer functions between the perceived sound direction and wheel velocities of a simulated non-holonomic mobile robot. These mappings constitute the high-level goal-directed steering behaviour. Sound direction information is obtained from a model of the lizard peripheral auditory system. The parameters of the transfer functions are learned via an online unsupervised correlation learning algorithm through interaction with obstacles in the form of low-level obstacle avoidance behaviour in the environment. The simulated robot is able to navigate towards a virtual sound source placed 3m away that continuously emits a tone of frequency 2.2 kHz, while avoiding randomly placed obstacles in the environment. We demonstrate through two independent trials in simulation that in both cases the neural circuit learns consistent and stable trajectories as compared to navigation without learning.

AB - Reactive spatial robot navigation in goal-directed tasks such as phonotaxis requires generating consistent and stable trajectories towards an acoustic target while avoiding obstacles. High-level goal-directed steering behaviour can steer a robot towards the target by mapping sound direction information to appropriate wheel velocities. However, low-level obstacle avoidance behaviour based on distance sensors may significantly alter wheel velocities and temporarily direct the robot away from the sound source, creating conflict between the two behaviours. How can such a conflict in reactive controllers be resolved in a manner that generates consistent and stable robot trajectories? We propose a neural circuit that minimises this conflict by learning sensorimotor mappings as neuronal transfer functions between the perceived sound direction and wheel velocities of a simulated non-holonomic mobile robot. These mappings constitute the high-level goal-directed steering behaviour. Sound direction information is obtained from a model of the lizard peripheral auditory system. The parameters of the transfer functions are learned via an online unsupervised correlation learning algorithm through interaction with obstacles in the form of low-level obstacle avoidance behaviour in the environment. The simulated robot is able to navigate towards a virtual sound source placed 3m away that continuously emits a tone of frequency 2.2 kHz, while avoiding randomly placed obstacles in the environment. We demonstrate through two independent trials in simulation that in both cases the neural circuit learns consistent and stable trajectories as compared to navigation without learning.

KW - behaviour-based robotics

KW - reactive navigation

KW - phonotaxis

KW - lizard peripheral auditory system

KW - synaptic plasticity

KW - correlation-based learning

U2 - 10.1007/978-3-319-65172-9_46

DO - 10.1007/978-3-319-65172-9_46

M3 - Article in proceedings

SN - 978-3-319-65171-2

T3 - Communications in Computer and Information Science

SP - 544

EP - 555

BT - Engineering Applications of Neural Networks

A2 - Boracchi, Giacomo

A2 - Iliadis, Lazaros

A2 - Jayne, Chrisina

A2 - Likas, Aristidis

PB - Springer

CY - Cham

ER -

Shaikh D, Manoonpong P. A Neural Circuit for Acoustic Navigation combining Heterosynaptic and Non-synaptic Plasticity that learns Stable Trajectories. I Boracchi G, Iliadis L, Jayne C, Likas A, red., Engineering Applications of Neural Networks: 18th International Conference Engineering Applications of Neural Networks. Cham: Springer. 2017. s. 544-555. (Communications in Computer and Information Science, Bind 744). https://doi.org/10.1007/978-3-319-65172-9_46