A new type of wheel–leg deformation mechanism, based on an electromagnetic clutch and gear rack transmission mechanism, is designed for moving on complex terrain. The walking model is established to study the kinematics characteristics and obstacle-surmounting performance of the mechanism. The stability cone method is used to evaluate the rollover stability and anti-interference ability in order to ensure the normal operation of the mechanism.

A contour error prediction model based on LSTM (long short-term memory) is established. By inputting power, vibration signal, and temperature values, the contour error of parts can be predicted to adjust the feed speed in real time. In this way, we can obtain not only higher machining accuracy, but also higher machining efficiency and reduce the scrap rate. The model is compared with other models, and the optimal model is selected to obtain the optimal prediction effect.

The haptic interactive robot based on a cable-driven parallel mechanism studied in this paper is mainly used for interactive training between an astronaut and a virtual environment. This haptic interactive robot can accurately simulate the motion and mechanical characteristics of objects in a microgravity environment and can safely interact with astronauts.

A multi-body dynamic model of a wind turbine power generation system (WTPGS), with a turbine blade model and hydro-mechanical hybrid power transmission (HMHPT), is developed and simulated. In HMHPT, the hydrostatic power transmission (HPT) improves the system controllability and the gear train obtains higher speeds in the electric generator. The HMHPT's benefits are improved system controllability and higher generator speed. Pump and motor leakages are analyzed based on system performance.