Abstract
A self-repairing robot utilising a spiking astrocyteneuronnetwork is presented in this paper. It uses the outputspike frequency of neurons to control the motor speed and robotactivation. A software model of the astrocyte-neuron networkpreviously demonstrated self-detection of faults and its selfrepairingcapability. In this paper the application demonstratorof mobile robotics is employed to evaluate the fault-tolerant capabilitiesof the astrocyte-neuron network when implemented ina hardware-based robotic car system. Results demonstrated thatwhen 20% or less synapses associated with a neuron are faulty,the robot car can maintain system performance and complete thetask of forward motion correctly. If 80% synapses are faulty, thesystem performance shows a marginal degradation, however thisdegradation is much smaller than that of conventional faulttoleranttechniques under the same levels of faults. This is thefirst time that astrocyte cells merged within spiking neuronsdemonstrates a self-repairing capabilities in the hardware systemfor a real application
Original language | English |
---|---|
Title of host publication | Unknown Host Publication |
Publisher | IEEE |
Number of pages | 8 |
Publication status | Accepted/In press - 15 Mar 2016 |
Event | International Joint Conference on Neural Networks (IJCNN) - Canda Duration: 15 Mar 2016 → … |
Conference
Conference | International Joint Conference on Neural Networks (IJCNN) |
---|---|
Period | 15/03/16 → … |
Keywords
- self-repair
- adapt
- astrocyte
- glia
- robotic
- FPGAs
- hardware
- error
- fault
- detection
- adaptive