, s.mollahoseini02@umz.ac.ir
Abstract: (2650 Views)
Introduction and objectives: In the past decades, the control of hand tremors in neurological disorders such as Parkinson's has attracted a lot of attention. The theories of closed-loop deep brain stimulation method are increasing significantly. The purpose of this article is to provide an automatic closed-loop method for the rehabilitation of Parkinson's patients with hand tremor symptoms using machine learning.
Materials and methods: In this article, a mathematical model including muscle model, basal ganglia, cerebral cortex and supplementary motor area is used to simulate tremor. Also, to control hand tremors, a non-integer proportional-derivative-integral controller (non-integer PID) has been used, as well as using the smart Proximal Policy Optimization (PPO) algorithm as a subset of reinforcement learning to adjust the coefficients.
Findings: In addition to reducing hand tremors and automatic learning for use in different levels of the disease, which has given acceptable results compared to other control methods, among the advantages of the Prihadi method is the practical implementation of this method in the real world due to the simplicity of the controller. And also the easy implementation of the intelligent algorithm is due to the automatic adjustment of the coefficients of artificial intelligence networks.
Conclusion: In addition to optimizing output symptoms such as hand tremors compared to other controllers, the proposed intelligent system can also be used for all levels of the disease due to its adaptability, causing a significant reduction in the side effects caused by continuous brain stimulation in the brain stimulation method. It opens in the form of a ring.
Materials and methods: In this article, a mathematical model including muscle model, basal ganglia, cerebral cortex and supplementary motor area is used to simulate tremor. Also, to control hand tremors, a non-integer proportional-derivative-integral controller (non-integer PID) has been used, as well as using the smart Proximal Policy Optimization (PPO) algorithm as a subset of reinforcement learning to adjust the coefficients.
Findings: In addition to reducing hand tremors and automatic learning for use in different levels of the disease, which has given acceptable results compared to other control methods, among the advantages of the Prihadi method is the practical implementation of this method in the real world due to the simplicity of the controller. And also the easy implementation of the intelligent algorithm is due to the automatic adjustment of the coefficients of artificial intelligence networks.
Conclusion: In addition to optimizing output symptoms such as hand tremors compared to other controllers, the proposed intelligent system can also be used for all levels of the disease due to its adaptability, causing a significant reduction in the side effects caused by continuous brain stimulation in the brain stimulation method. It opens in the form of a ring.
Keywords: Parkinson's disorder, deep brain stimulation, machine learning, reinforcement learning, PPO
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
