Designing Adaptive Controller based on Spiking Neural Network
کد مقاله : 1136-CFIS (R2)
نویسندگان:
مینا زینالی *1، محمد منثوری2، محمد تشنه لب1
1دانشگاه خواجه نصیرالدین طوسی
2گروه برق-کنترل، دانشکده فنی و مهندسی، دانشگاه شاهد
چکیده مقاله:
This paper presents the design of a Spiking Neural Network (SNN) structure for control applications. A SNN Controller is designed for the wheel slip regulation problem of an Antilock Braking System (ABS). A gradient descent based learning algorithm is used for training of the network. The Spike Response Model (SRM) is employed to denote the effect of the incoming spikes on the postsynaptic membrane potential. Population coding is utilized to convert analog inputs into neural network inputs and for converting the output of the network (control signal) to a real number, Delay coding is employed. To verify the performance of the designed controller, it is applied to a quarter vehicle model. Simulations indicate fast convergence and good performance of the proposed control algorithm.
کلیدواژه ها:
Spiking Neural Network, Antilock Braking System, Spike Response Model, Population Coding, Delay Coding
وضعیت : مقاله برای ارائه شفاهی پذیرفته شده است