In powered ankle–foot prostheses, multilevel hierarchical control systems are usually used with predetermined parameters (tuned during the prosthesis trial). Therefore, control systems cannot adaptively interact with terrains variation where control systems is most effective for ground level walking and less effective for ascending or descending stair/slop. In order to address the control system performance in ever-changing terrains, an adaptive mechanism should be included the control system structure. Here, we present a pilot study to illustrate the applicability of a genetic algorithm-based adaptive fuzzy logic control system. The design method could be divided into two stages: initial knowledge base and membership functions for the genetic pool on the basis of the analysis of biological ankle–foot behaviour. Additionally, the construction of genetic optimization mechanism rules and constraints (fitness function, mutation rats, replacement rate, etc.). Takagi–Sugeno–Kang fuzzy (TSK-fuzzy) inference system is selected because the system structure could depict the character of simple impedance control system. The control system and dynamic model were developed using C code and evaluated using MATLAB/Simulink (2019a).

中文翻译：

---

走向脚踝假肢通用控制系统的仿真研究

在动力型脚踝假肢中，通常使用具有预定参数（在假体试验期间进行调整）的多层控制系统。因此，控制系统无法自适应地与地形变化交互作用，在这种情况下，控制系统对地面行走最为有效，而对楼梯/斜坡的上升或下降效果较差。为了解决不断变化的地形中控制系统的性能，应在控制系统结构中包含自适应机制。在这里，我们进行了一项初步研究，以说明基于遗传算法的自适应模糊逻辑控制系统的适用性。设计方法可以分为两个阶段：基于对生物踝足行为的分析，初始知识库和遗传库的隶属函数。另外，遗传优化机制的构建规则和约束条件（健身功能，突变大鼠，替代率等）。选择Takagi–Sugeno–Kang模糊（TSK-fuzzy）推理系统是因为其系统结构可以描述简单阻抗控制系统的特性。控制系统和动态模型是使用C代码开发的，并使用MATLAB / Simulink（2019a）进行了评估。