This study is mainly to provide an overview of human walking induced energy harvest. Focusing on the proportion of all energy sources provided by daily activity, the available human walking induced energy is divided with respect to the generation principle. The extensive research on harvesting energy results from body vibration, inertial element, and foot press to convert into electricity is overviewed. Over the past decades, various smart materials have been employed to achieve energy conversion. Generators based on electromagnetic induction or the triboelectric effect were developed and integrated. Small captured power and low overall efficiency are criticized. The concept of human walking energy harvest is extended into the wearable walking robotics using other mediums, such as fluid, to transmit power instead of electricity. By comparison, it is indicated that less energy conversion links are involved in energy regeneration of such applications and expected to guarantee less loss and higher efficiency. Meanwhile, in order to overcome the shortage of relatively low power output, comments are made that the harvester should be capable of adaptation under the condition that the mechanical energy of lower limb and feet is subject to change in different gait phases so as to maximize the collected energy.

Accurate and fast prediction of tool conditions is fundamental to improve the machining accuracy and consistency in smart machining systems. The current tool condition monitoring methods, i.e., physics-based and data-driven approaches, have either low prediction accuracy or model generalization. To solve the shortcomings and utilize the benefits of both sides, a novel physics-guided deep learning model is developed in this study. The physics of cutting mechanics and tool wear model is applied to guide the model construction and regulate the network learning process. The model first generates large labeled simulation dataset for the pretraining of the deep network, and solves the problem of labeled-data insufficiency for network training in practice. It then fine-tunes the model through the monitored signal to optimize the network weight and alleviate the deviation between the physical model and the actual machining process, which improves the physical consistency and generalization of the model. Additionally, with the introduction of attention mechanism to the deep residual network, discriminant features can be extracted to distinguish wear values and working conditions. The experimental results show that by learning the physics, the physics-guided deep network can accurately predict the tool wear even under limited training sets and varied working conditions, and it outperforms the original data-driven models and the physical models.

Robot-assisted minimally invasive surgery (MIS) has received increasing attention, both in the academic field and clinical operation. Master/slave control is the most widely adopted manipulation mode for surgical robots. Thus, sensing the force of the surgical instruments located at the end of the slave manipulator through the main manipulator is critical to the operation. This study mainly addressed the force detection of the surgical instrument and force feedback control of the serial surgical robotic arm. A measurement device was developed to record the tool end force from the slave manipulator. An elastic element with an orthogonal beam structure was designed to sense the strain induced by force interactions. The relationship between the acting force and the output voltage was obtained through experiment, and the three-dimensional force output was decomposed using an extreme learning machine algorithm while considering the nonlinearity. The control of the force from the slave manipulator end was achieved. An impedance control strategy was adopted to restrict the force interaction amplitude. Modeling, simulation, and experimental verification were completed on the serial robotic manipulator platform along with virtual control in the MATLAB/Simulink software environment. The experimental results show that the measured force from the slave manipulator can provide feedback for impedance control with a delay of 0.15 s.