在下肢外骨架中异常的步行阶段识别和四肢角度预测
Sheng Wang1,2, Chunjie Chen2, Xiaojun Wu1
1School of Mechanical and Electrical Engineering, Xi'an University of Architecture and Technology, Xi'an 710055, China.
Biomimetics (Basel, Switzerland)
|September 26, 2025
概括
这项研究改善了异常步行阶段检测和下肢角度预测,以控制外骨架. 结合脚角度和步态阶段的新型输入方案提高了预测准确性,CNN-LSTM网络显示出最佳结果.
科学领域:
- 机器人技术 机器人技术 机器人技术
- 生物力学 生物力学
- 机器学习 机器学习
背景情况:
- 控制下肢外骨架在检测异常步行阶段和预测下肢角度方面面临挑战.
- 异常的步态,如剪刀,落脚和摇摇欲的步态,需要精确的相位识别才能提供有效的辅助.
研究的目的:
- 开发和验证一种可靠的方法来检测异常的步态阶段和下肢角度预测.
- 通过提高运动预测精度来增强下肢外骨的控制能力.
主要方法:
- 模拟了三个异常的步态:剪刀,落脚和摇摆.
- 提出了四个离散阶段的划分 (摇摆前,摇摆,摇摆终止,姿势) 以用于单腿步行阶段识别.
- 利用卷积神经网络 (CNN) 和支持矢量机器 (SVM) 进行离散相识别.
- 使用自适应频率振荡器进行连续相位估计.
- 开发了一种创新的输入方案,整合了三轴脚关节角度和连续步态阶段,用于运动角度预测.
- 使用CNN-长期短期记忆 (LSTM) 网络评估预测准确度.
主要成果:
- 成功地使用CNN和SVM识别了离散的步态阶段.
- 通过自适应频率振荡器实现了连续步行阶段估计.
- 建议的信息融合方案显著提高了下肢角度预测的准确性.
- CNN-LSTM网络在预测四肢运动角度方面表现出卓越的表现.
结论:
- 拟议的四个离散阶段的划分增强了异常步行的识别.
- 整合脚关节的角度和连续的步态阶段是有效的四肢角度预测.
- 开发的方法有望改善下肢外骨架的控制和功能.
相关概念视频
Bones of the Lower Limb: Femur and Patella
The femur is the body's longest and strongest bone spanning the thigh region. Its head articulates with the acetabulum of the hip bone to form the hip joint. A minor indentation on the medial side of the femoral head, called the fovea capitis, serves as the site of attachment for the ligament of the head of the femur. This weak ligament spans the femur and acetabulum and supports the hip joint. The narrowed region below the head is the neck of the femur. The inclination angle between the neck...
Ankle Joint
The ankle is formed by the talocrural joint (crural = leg). It consists of the articulations between the talus bone of the foot and the distal ends of the tibia and fibula of the leg. The superior aspect of the talus bone is square-shaped and has three areas of articulation. The top of the talus articulates with the inferior tibia. This is the portion of the ankle joint that carries the body weight between the leg and foot. The sides of the talus are firmly held in position by the articulations...


