Traditional robotic foot and ankle have received considerable attention for adaptive locomotion on complex terrain and land buffering ability. This study aims to develop a robotic ankle based on series elastic actuator with a parallel spring (SEAPS) by mimicking the two degree-of-freedom human ankle and related muscles. A unique four-SEAPS-based spring mechanism enhances motion adaptation and landing buffer. A dynamic model based on the SEAPS drive is presented, whereas a kinematic solution is realized. The bionic ankle has a wide range of motions with 30° in both the sagittal and frontal planes, which cover most of the human ankle motions. Four experiments were conducted to thoroughly characterize the capabilities. First, the static stiffness and the 2-D/3-D trajectory tracking performance of the proposed ankle were tested. Second, the angle sensing capacity under inclined road surface and the land buffering performance from a specific height were evaluated. The results show that the maximum 3-D motion tracking error is smaller than 2.3%, and the minimum sensing error of inclined road is smaller than 0.5%. The proposed ankle can closely track the 3-D walking trajectory of natural ankle joint with relative root-mean-square error well below 1.0%. Compared with no springs, landing buffer of the ankle with SEAPS can yield remarkable reduction in both peak ground reaction force (52.2%) and joint torque (57.9%). These findings prove that the SEAPS-based bioinspired robotic ankle exhibits high land buffering performance in an unstructured environment, and also well reproduces the human-like movement in terms of complex 3-D ankle motions.
12 October 2022 in IEEE Transactions on Industrial Electronics
This article presents a novel polycentric hybrid knee prosthesis (PHKP) by enabling a multibar linkage to work for high-torque activities in active mode and for low-torque activities in passive mode. The mode transition is implemented the servo-controlled telescopic slide rails. Distinguished from the single-axis hybrid knee prostheses, the PHKP knee provides geometric stability in stance and notably increases toe clearance in swing by using an optimally designed instantaneous center of rotation trajectory. In active mode, the PHKP knee provides high transmission ratio and continuous peak torque offering sufficient support for torque-demanding activities. Impedance control was employed to regulate joint torque as a response to knee motion in the active mode. It was found that due to the inherent geometric stability of the polycentric mechanism in stance, the damping coefficient of the controller can be tuned much smoother than the single-axis knees leading to more compliant and comfortable interactions among the wearer, the prosthesis and the ground. The stair ascent tests showed that the joint angle and torque profiles of the PHKP knee are similar to those of the able-bodied by demonstrating a step-over-step pattern with a peak angle of 73 $^{\circ }$ and a maximum torque of 0.92 N $\cdot$ m/kg.
In the existing research on prosthetic footplates, rehabilitation insoles, and robot feet, the cushioning parts are basically based on simple mechanisms and elastic pads. Most of them are unable to provide adequate impact resistance especially during contact with the ground. This paper developed a bioinspired heel pad by optimizing the inner structures inspired from human heel pad which has great cushioning performance. The distinct structures of the human heel pad were determined through magnetic resonance imaging (MRI) technology and related literatures. Five-layer pads with and without inner structures by using two materials (soft rubber and resin) were obtained, resulting in four bionic heel pads. Three finite element simulations (static, impact, and walking) were conducted to compare the cushioning effects in terms of deformations, ground reactions, and principal stress. The optimal pad with bionic structures and soft rubber material reduced 28.0% peak vertical ground reaction force (GRF) during walking compared with the unstructured resin pad. Human walking tests by a healthy subject wearing the 3D printed bionic pads also showed similar findings, with an almost 20% decrease in peak vertical GRF at normal speed. The soft rubber heel pad with bionic structures has the best cushioning performance, while the unstructured resin pad depicts the poorest. This study proves that with proper design of the inner structures and materials, the bionic pads will demonstrate distinct cushioning properties, which could be applied to the engineering fields, including lower limb prosthesis, robotics, and rehabilitations.
29 April 2024 in IEEE Transactions on Neural Systems and Rehabilitation Engineering
To overcome the challenges posed by the complex structure and large parameter requirements of existing classification models, the authors propose an improved extreme learning machine (ELM) classifier for human locomotion intent recognition in this study, resulting in enhanced classification accuracy. The structure of the ELM algorithm is enhanced using the logistic regression (LR) algorithm, significantly reducing the number of hidden layer nodes. Hence, this algorithm can be adopted for real-time human locomotion intent recognition on portable devices with only 234 parameters to store. Additionally, a hybrid grey wolf optimization and slime mould algorithm (GWO-SMA) is proposed to optimize the hidden layer bias of the improved ELM classifier. Numerical results demonstrate that the proposed model successfully recognizes nine daily motion modes including low-, mid-, and fast-speed level ground walking, ramp ascent/descent, sit/stand, and stair ascent/descent. Specifically, it achieves 96.75% accuracy with 5-fold cross-validation while maintaining a real-time prediction time of only 2 ms. These promising findings highlight the potential of onboard real-time recognition of continuous locomotion modes based on our model for the high-level control of powered knee prostheses.
25 July 2022 in IEEE Robotics and Automation Letters
A hybrid knee prosthesis is proposed in this letter, which consists of a polycentric structure in passive mode for low-torque activities and a single-axis structure in active mode for high-torque activities. A novel mode transition mechanism controls self-holding electromagnets for switching modes between the four-bar linkage and single-axis structure. Compared with the conventional single-axis hybrid knee, the four-bar polycentric mechanism with varying instantaneous center of rotation (ICR) can enhance the geometric stability and increase the toe clearance in passive mode. For active mode, we developed a custom embedded electric system, employed torque control for stance and position control for swing. The results of bench tests indicated that the bandwidth of the controller was suitable for locomotion. The clinical test of level-ground walking without sudden buckling and stumble was validated by three subjects. Regarding climbing stairs, a typical high-torque activity in daily locomotion, all subjects reach the maximum knee torque around 0.95 Nm/kg comparable to the able-bodied.
Similar to biological muscles in nature, artificial muscles have unique advantages for driving bionic robots. However, there is still a large gap between the performance of existing artificial muscles and biological muscles. Twisted polymer actuators (TPAs) convert rotary motion from torsional to linear motion. TPAs are known for their high energy efficiency and large linear strain and stress outputs. A simple, lightweight, low-cost, self-sensing robot powered using a TPA and cooled using a thermoelectric cooler (TEC) was proposed in this study. Because TPA burns easily at high temperatures, traditional soft robots driven by TPAs have low movement frequencies. In this study, a temperature sensor and TEC were combined to develop a closed-loop temperature control system to ensure that the internal temperature of the robot was 5 °C to cool the TPAs quickly. The robot could move at a frequency of 1 Hz. Moreover, a self-sensing soft robot was proposed based on the TPA contraction length and resistance. When the motion frequency was 0.01 Hz, the TPA had good self-sensing ability and the root-mean-square error of the angle of the soft robot was less than 3.89% of the measurement amplitude. This study not only proposed a new cooling method for improving the motion frequency of soft robots but also verified the autokinetic performance of the TPAs.