A complete methodology for an unmanned coaxial rotor helicopter with unstructured uncertainties was proposed to achieve high-accuracy tracking performance from modelling to robust control. An integrative approach was introduced to systematically construct a whole dynamic model. The key parameters were selected carefully after iteratively being checked by empirical coefficients to decrease the budget and risk of programme. Moreover, a new control scheme is proposed to simultaneously incorporate six inputs to control six states based on the investment of singularity value responses and the general rule of relative gain array. Coprime factor uncertainty model is considered to represent a class of unstructured uncertainties, such as unmolded actuator dynamics and unpredicted interferences between two rotors. Furthermore, the $H_{\infty }$ loop-shaping control was proposed to apply the control design of the coaxial rotor helicopter to manage complicated uncertainties and multivariable coupling. Finally, simulation results show the effectiveness of the proposed controller design in the step response of the closed loop. The stable closed-loop plant is achieved and the tolerant size of unstructured uncertainty is up to 36.09%. Good step responses and satisfied decoupling were also investigated in detail.

This study proposes a new method for the immediate fault warning and fault root tracing of CNC lathes. Here, the information acquisition scheme was formulated based on the analysis of the coupling relationship between the mechanical parts of CNC lathes. Once the collected status signals were de-noised and coarse-grained, transfer entropy theory was introduced to calculate the net entropy of information transfer between the mechanical parts, after which the information transfer model was constructed. The sliding window method was used to determine the probability threshold interval of the net information transfer entropy between the lathe mechanical parts under different processing modes. Therefore, the transition critical point was determined according to the information entropy, and the fault development process was clarified. By analyzing the information transfer changes between the parts, fault early warning and fault root tracking on the CNC lathe were realized. The proposed method realizes the digitalization and intelligentization of fault diagnosis and has the advantages of timely and efficient diagnosis. Finally, the effectiveness of the proposed method is verified by a numerical control lathe tool processing experiment.