System level hybrid functional and power modeling methodology to extend the longevity of implantable seizure controllers

Abstract

An implantable seizure controller needs a system level model to predict power generation and consumption as well as function because it has a limited power source. Methods/Statistical analysis: The targeted system is divided into several sub-blocks according to functions. Each sub-block is redefined in respect of energy. They have information about both their functional operation and role of energy. This model is developed in Simulink/Matlab environment that supports multi-domain components and customizable libraries efficiently. It was examined for an implantable seizure controller using data with seizure and non-seizure. Findings: Most system models for power estimation have usually considered physical-level and instruction-level models based on processors. They are used to measure power consumption of system while programs are running. Few system level architectures have been presented that consider both power consumption and power generation. Because a closed loop seizure controller may be implanted with limited power sources, it needs to be designed with consideration for power consumption and generation together. The proposed system level modeling methodology supports power generation and consumption description, and a dynamic power management scheme in a single domain as well as functional behavior check. It enables system designers to consider system specifications within a given power budget at the early stage of design. Improvements/Applications: This hybrid function and power modeling methodology can be applied to a system with a limited power source, such as implantable medical devices and sensor network modules.