Neuromechanical force-based control of a powered prosthetic foot

Abstract

This article presents a novel neuromechanical force-based control strategy called FMCA (force modulated compliant ankle), to control a powered prosthetic foot. FMCA modulates the torque, based on sensory feedback, similar to neuromuscular control approaches. Instead of using a muscle reflex-based approach, FMCA directly exploits the vertical ground reaction force as sensory feedback to modulate the ankle joint impedance. For evaluation, we first demonstrated how FMCA can predict human-like ankle torque for different walking speeds. Second, we implemented the FMCA in a neuromuscular transtibial amputee walking simulation model to validate if the approach can be used to achieve stable walking and to compare the performance to a neuromuscular reflex-based controller that is already used in a powered ankle. Compared to the neuromuscular model-based approach, the FMCA is a simple solution with a sufficient push-off that can provide stable walking. Third, to assess the ability of the FMCA to generate human-like ankle biomechanics during walking at the preferred speed, we implemented this strategy in a powered prosthetic foot and performed experiments with a non-amputee subject. The results confirm that, for this subject, FMCA can be used to mimic the non-amputee reference ankle torque and the reference ankle angle. The findings of this study support the applicability and advantages of a new bioinspired control approach for assisting amputees. Future experiments should investigate the applicability to other walking speeds and the applicability to the target population.

Toe-in Gait, Associated Complications, and Available Conservative Treatments: A Systematic Review of Literature

“…Based on the results of the gait analysis, femoral anteversion influences the kinematic and kinetic parameters while walking. 11, 24,25) However, it should be emphasized that it influences mostly Patients: 18 subjects with toe-in gait (10.5~17.5 years) and 17 normal subjects (10.6~17.3 years) Intervention: None Comparison: Kinematic and kinetic parameters while walking. Principle component analysis (PCA) was performed using MATLAB.…”

“…It seems that increase in femoral anteversion angle is associated with reduced dynamic control of hip and patellofemoral joints in frontal and transverse planes. Naseri el al, 2020 24) Patients Functional walking was assessed using the FAQ scale Participants were assessed 3 times: 1 day before surgery, 6 months after surgery and 1 year after surgery. Range of motion of hip was evaluated in prone position Hip range of rotation improved after surgery.…”

Challenges and solutions for application and wider adoption of wearable robots

“…IMU technology is cheap and can be easily attached to wearers; thus, it is commonly placed on parts of the body not connected to the wearable robot-for example, in a leg exoskeleton, an IMU can be placed on the arm or trunk (Lazzaroni et al, 2020). Alternatively, the IMUs can be embedded in the robot itself, providing an alternative to classic robot sensors (Naseri et al, 2020). As the individual components of the IMU tend to have noisy outputs, obtaining IMU orientation requires the use of techniques such as Kalman filtering (Nazarahari and Rouhani, 2021).…”

“…Existing algorithms range from simple thresholding (e.g., activate assistance if user bends forward) to more complex approaches based on classification and regression using machine learning (Novak and Riener, 2015;Tucker et al, 2015). For example, classifiers can learn to identify desired gestures from EMG-based on a training dataset of previously recorded and manually labeled EMG data (Yun et al, 2020); similarly, regression algorithms can learn to estimate desired robot torque from ground reaction force (Naseri et al, 2020) or EMG (Gui et al, 2019).…”

Data-efficient human walking speed intent identification

“…In contrast to both EMG and EEG, which sense the user directly, intent information can be inferred from the physical connections between the human and robot without direct human measurements. Ground reaction force sensors, torque sensors, motor encoders, motor current draw, and inertial measurement units (IMUs, such as in the motion capture suit in Figure 1) can determine progression through the gait cycle, which helps infer appropriate robot action (Pacini Panebianco et al, 2018; Gambon et al, 2019; Elery et al, 2020; Lazzaroni et al, 2020; Naseri et al, 2020; Liang et al, 2021; Tan et al, 2021).

Figure 1. Subject wearing XSens inertial motion capture suit, making changes of intended walking speed on large research treadmill.

…”