Scaling a Mixed-Feedback Approach to Controlling Neuromorphic Circuits

扩展混合反馈方法来控制神经形态电路

• 批准号: 2275448
• 金额: --
• 依托单位: University of Cambridge
• 依托单位国家: 英国
• 项目类别: Studentship
• 财政年份: 2019
• 资助国家: 英国
• 起止时间: 2019 至 无数据
• 项目状态: 未结题
• 来源: https://gtr.ukri.org/projects?ref=studentship-2275448
• 关键词: Scaling; Mixed; Feedback; Approach; Controlling

The research group has used mixed-feedback control theory to develop a neuromorphic circuit mimicking the behaviour of a single neuron. The novelty of this circuit is that its design and analysis can be done purely by shaping the current-voltage (I-V) curves. This input-output approach is simpler than tuning a state-space model, and allows robust and accurate control of non-linear behaviour.The focus of the PhD will be to design and characterise neuromorphic circuits comprising multiple such neuron elements connected together. These circuits potentially have engineering applications, including electronic sensing. They can also be used to model and investigate biological phenomena, such as the spontaneous emergence of rhythmic neural activity.

该研究小组利用混合反馈控制理论开发了一种模仿单个神经元行为的神经形态电路。该电路的新奇在于，其设计和分析可以完全通过对电流-电压（I-V）曲线进行整形来完成。这种输入-输出方法比调整状态空间模型更简单，并且允许对非线性行为进行鲁棒和精确的控制。博士学位的重点将是设计和识别包括多个连接在一起的神经元元件的神经形态电路。这些电路具有潜在的工程应用，包括电子传感。它们也可以用于模拟和研究生物现象，例如有节奏的神经活动的自发出现。