Neural Network-Based Self-Tuning PID Control for Underwater Vehicles

논문 리뷰

이 포스트의 모든 저작권은 원작자에게 있습니다.

 

Sensors 2016

by Rodrigo Hernández-Alvarado 1,*,†,‡,Luis Govinda García-Valdovinos 1,‡,Tomás Salgado-Jiménez 1,‡,Alfonso Gómez-Espinosa 2,‡ORCID and Fernando Fonseca-Navarro 1,‡

---

subject:

Comparison between Conventional PID-like controller and Auto-tune PID-like controller based on Neural Networks (NN)

 

keyword:

neural networks
auto-tuning PID
ROV(Remotely Operated Vehicles) control
disturbances

 

result:

Auto-tune PID-like controller based on Neural Networks attained the best performance with less energy.

 

background knowledge:

Neural Network

PID

 

Model

Block diagram of an auto-tuned PID with artificial NN control

Block diagram of the implemented backpropagation NN

parameter

• u(n), u(n-1)
  • reference inputs
  • desired trajectory
• y(n), y(n-1)
  • reference outputs
  • real trajectory
• C(n), C(n-1)
  • control signals
• wji
  • weights of the hidden layer
• vji
  • weights of the output layer.