Data Driven Spatial Filtering Can Enhance Abstract Myoelectric Control in Amputees.
Data Driven Spatial Filtering Can Enhance Abstract Myoelectric Control in Amputees.
复制标题
数据驱动的空间过滤可以增强截肢者的抽象肌电控制。
DOI：
10.1109/embc.2018.8513075
复制
发表时间：
2018
期刊：
影响因子：
--
通讯作者：
Dyson M
中科院分区：
文献类型：
--
作者：
Dyson M
基于多传感器技术的肌电控制可以提供增强的信噪比，但增加了硬件成本和复杂性。传感器阵列也是有吸引力的假肢上下文中，当确切的肌肉位置是未知的，如可能是肢体丧失后的情况。我们目前获得的初步数据，而四名截肢参与者从事一个抽象的肌电解码任务。解码器由前臂或上臂的肌肉控制，这取决于肢体损失的程度。我们比较性能使用一对表面肌电图传感器，同时使用数据驱动的加权八个传感器。表现率表明，截肢者参与者能够学习肌电任务。使用多个空间加权传感器时，结果强烈倾向于增强性能。需要进一步的研究来测试在抽象解码中使用额外的肌电传感硬件是否会导致有效的假肢控制。
Myoelectric control based on multi-sensor techniques can provide an enhanced signal to noise ratio but increases hardware cost and complexity. Sensor arrays are also attractive in a prosthetics context when exact muscle positions are unknown, such as may be the case after limb loss. We present preliminary data obtained while four amputee participants engaged in an abstract myoelectric decoding task. The decoder was controlled by muscles of the forearm or upper arm depending on the level of limb loss. We compare performance using a pair of surface electromyography sensors and while using a data driven weighting of eight sensors. Performance rates demonstrate that amputee participants are able to learn the myoelectric task. Results trend strongly toward enhanced performance when using multiple spatially weighted sensors. Further studies are required to test whether the use of additional myoelectric sensing hardware in abstract decoding would lead to effective prosthesis control.