Research On Modeling And Control Methods For Multi-modal Unmanned Aerial Vehicle Switching Systems

Autonomous UAVs have been widely applied in military and civilian fields such as unmanned reconnaissance/surveillance,high-altitude fire strikes,land and resources detection,emergency rescue and disaster relief,in which the multi-modal phenomena widely exist.In terms of the changes of the controlled object,the heavy-load UAVs or the composite aircrafts release large payload-weight ratio materials,weapons,parasite aircrafts and other loads,which shows the typical multi-modal characteristics.In addition,during the pickup and delivery process of aerial manipulations,the changes of robot arms will affect the system parameters and the dynamics models.In terms of environment changes,the hybrid aerial/terrestrial vehicles experience different forces in different environments(the ground provides continuous support in the terrestrial mode).Besides,the different wind resistance and the intermittent measurement of the target will also lead to multi-modal phenomena.Focusing on the path planning and switching control methods for multi-modal unmanned aerial vehicle systems,this thesis proposes a new method for the stability analysis and control synthesis of multi-mode systems under asynchronous switching,which is extended to nonlinear fuzzy systems.Also,the imitation learning method is introduced into the reinforcement learning framework,which ensures the policy performance while improving the training efficiency.In addition,this thesis also proposes a class of multi-modal autonomous unmanned vehicle called Syta B(Smooth-Transition Hybrid Terrestrial/Aerial Bicopters).Specifically,the main work of this thesis is summarized as follows.First,the thesis is concerned with the problems of mapless navigation for unmanned aerial vehicles in the scenarios with limited sensor accuracy and computing capability.A novel learning-based algorithm called Soft Actor-Critic from Demonstrations(SACf D),integrating reinforcement learning with imitation learning,is proposed.Specifically,the maximum entropy reinforcement learning framework is introduced to enhance the exploration capability of the algorithm,upon which the thesis investigates how to significantly accelerate the convergence rate while improving policy performance reliably.Further,the proposed algorithm enables an implementation of mapless navigation for unmanned aerial vehicles and experimental results show that it outperforms the existing algorithms.The thesis investigates the problem of control synthesis for a class of discrete-time semi-Markov jump linear systems,in which the sojourn time of each mode is bi-boundary(with upper and lower bound).The system is subject to modal asynchrony,which means that the switchings of the mode-dependent controller to be designed lag behind the ones of the controlled plant,and the lag is mode-dependent.In contrast with the traditional mode-independent lag commonly assumed in the existing studies,not only is the modal lag more practical and general,but also it yields less conservatism of the controller design.By virtue of the semi-Markov kernel approach,the conditions on the existence of the anticipated stabilizing controller capable of overcoming the modal asynchrony are derived.Illustrative examples including a class of vertical take-off and landing(VTOL)helicopter models are presented to demonstrate the necessity and the validity of the designed antimodal-asynchrony controllers.Afterwards,the thesis is concerned with a class of discrete-time hybrid fuzzy systems subject to semi-Markov switching,in which the sojourn time of each mode is with upper and lower bounds.A practical scenario of transitional asynchrony is taken into account for the first time,where the switchings of controllers to be designed lag behind the ones of the controlled plant,and the lags depend on the transition between adjacent modes.By means of the semi-Markov kernel approach,numerically testable stability criteria are obtained,based on which existence conditions of the anticipated stabilizing controller capable of overcoming the transitional asynchrony are derived.Compared with the previous studies assuming the mode-independent or mode-dependent lags,the derived results are less conservative.Two illustrative examples including a class of bicopters are given to demonstrate the effectiveness and potential of the designed anti-transitional-asynchrony controllers.The thesis details the design,modeling and control of Syta B,a vehicle capable of hybrid terrestrial/aerial mobility with smooth transition,where the structure embedding a bicopter is adopted for the first time.In contrast to previous hybrid terrestrial/aerial vehicles with a quadrotor embedded,Syta B not only requires less energy for the same takeoff weight,but also regulates its attitude less frequently to restrain the vibration of sensors.Three modes,the terrestrial/aerial/transitional modes,are considered for the vehicle,and the dynamics modeling and controller design of each mode are carried out.The transition between terrestrial and aerial locomotions is smooth compared with the case of non-inclusion of transitional mode such that the bounce and shake of the vehicle are alleviated.The energy efficiency is compared between the terrestrial and aerial modes,and between the energy-saving and high-maneuverability paradigms(a choice enabled in the terrestrial mode of Syta B).Experimental results are presented to demonstrate the potential of Syta B,the effectiveness of the designed controllers,and the necessity of considering the transition process.Finally,the thesis is concerned with the issue of smooth control for a class of discretetime switched linear systems subject to a semi-Markov chain,in which the switchings of controllers lag behind the ones of the controlled plant.A new concept of smoothness performance for stochastic switched systems is put forward in a new form including the statistical information.On this basis,the piecewise control scheme with transitiondependent smooth duration is introduced for the first time,which brings less conservatism and reconciles the transient and steady-state performance in contrast to the previous studies considering the single-duration or fixed-duration ones.Besides,the receding solution methodology of linear matrix inequalities is proposed to obtain the stability criterion and the existence condition of the proposed smooth controllers.Illustrative examples including a class of hybrid terrestrial/aerial bicopters are presented to demonstrate the validity and applicability of the developed theoretical results.