BAC: Neural Optimization Algorithms in Respiratory Control

本研究的目的是 来揭示 被哺乳动物大脑用来 动态优化与控制 特定电机的电机输出 任务 研究人员将研究 神经计算算法 基于自适应控制 哺乳动物呼吸运动输出 并开发基于神经网络的 Hebbian学习算法 动态最优控制 问题. 工作重点将放在 神经模型的发展 和算法来适应 生理和 神经生理学数据， 描述呼吸道 系统 具体方面要 解决的是：1）神经 架构和自适应 算法用于 无限时间与有限时间 动态控制; 2）影响 系统不确定性及其 自适应补偿; 3）原因 系统不稳定， 稳定化方法;以及4） 神经策略， 鲁棒性和容错性， 噪声抑制和精度 网络的回应。 这项工作很重要，因为它可能 导致了一个新的范式， 智能控制设计 适用于各种 工程和机器人 应用.

The objective of this research is to uncover the neural strategies utilized by the mammalian brain for dynamic optimization and control of motor output in a specific motor task. The investigator will study the neural computation algorithms underlying the adaptive control of respiratory motor output in mammals and develop neural-network based Hebbian learning algorithms for this dynamic optimal control problem. The work will focus on the development of neural models and algorithms to fit the extensive physiological and neurophysiological data that characterize the respiratory system. The specific aspects to be addressed are: 1) the neural architectures and adaptive algorithms to use for both infinite-time and finite-time dynamic control; 2) the effects of system uncertainties and their adaptive compensation; 3) causes for system instabilities and methods for stabilization; and 4) neural strategies for increasing robustness and fault tolerance, noise rejection, and precision of the network response. This work is important since it may lead to a novel paradigm for the design of intelligent control systems for a wide range of engineering and robotic applications.