Robust Control And Applications Of Complex Uncertain Dynamic Systems

In mechanical systems,electronic circuits,robots,and other fields of engineering,there exist many key structured subsystems,which are mainly composed of core compo-nents with uncertainties or nonlinearities.It is of great practical significance to study the robust control problem of this kind of complex uncertain dynamic systems and guarantee the performance against uncertain disturbances.Based on the existing research on com-plex uncertain dynamic systems,robust control and filtering of state-dependent uncertain systems,design and application of a novel vibration sensor using state-dependent nonlin-earities,and robust navigation control and human-following control of mobile robots are comprehensively studied in this thesis.By constructing a new parameter-dependent Lyapunov functional,the problems of robust stability analysis and control of state-dependent uncertain systems are addressed.Based on the parameter-dependent Lyapunov functional,stability conditions of open-loop uncertain systems are given.By resorting to characteristics of state-dependent uncertain parameters,a relaxed stability criterion is derived to reduce conservativeness.In order to take full advantage of the parameter-dependent functional,a model-dependent state feedback stabilization method is proposed,which has more flexibility in controller synthesis.The proposed robust controller design method is applied to synchronization of Chua's oscillator circuit.A robust filter design method for state-dependent uncertain system is proposed by introducing a generalized filtering performance index-extended dissipativity.Some clas-sical filtering problems such as_?,₂-_?,passive and dissipative filtering problems for state-dependent uncertain systems are solved successfully within a unified framework.This can allow us to choose a suitable filtering strategy according to different practical ap-plications or noise levels,which is conveniently implemented by adjusting the weighting matrices in the generalized performance index.The robust filter design method simulta-neously considers disturbances from state-dependent parameter uncertainties and general time-varying parameter uncertainties.The small current estimation problem of a tunnel diode circuit system under uncertain disturbances is solved by using the proposed robust filter design method.A novel vibration sensor using a bio-inspired quasi-zero-stiffness structure with tate-dependent nonlinearities is developed for the real-time measurement of absolute vibration displacement.With this novel bio-inspired vibration sensor,the problems of er-ror accumulation and real-time performance induced by traditional measurement method using accelerometer can be effectively eliminated.The vibration sensor provides a new effective and convenient way for measurement of absolute vibration displacement in mov-ing platforms.Through taking full advantage of the bio-inspired vibration sensor in real-time measurement of absolute vibration motion,a model-based on-line fault detec-tion algorithm is proposed to cope with the real-time detection problem of weak fault with fast time-varying characteristics.Experimental results show that the proposed fault detection method is more accurate and sensitive than other detection methods such as multi-resolution wavelet method.A novel passive suspension system is constructed by adopting the animal-limb-like bio-inspired structure with state-dependent nonlinearities,and it is applied to tracked mobile robot.With the novel bio-inspired suspension,the robot's loading capacity,stability,and obstacle negotiation capability can be significantly enhanced.The tracked mobile robot can be applied to various rough terrain environments.Importantly,robust navigation control using double-layer nonlinear model predictive control(NMPC)method is realized to improve the transient response and trajectory tracking accuracy against unknown slippage disturbances.The upper layer NMPC is performed at a low frequency to optimize the global trajectory tracking performance,and the lower layer NMPC is performed at a high frequency to capture unknown slippage disturbances.Experimental results demonstrate that the novel passive suspension can efficiently absorb strong shock induced by the process of obstacle crossing and hence guarantee smooth motion of the robot,and the robust navigation control method based on double-layer NMPC can effectively improve the transient response and trajectory tracking accuracy with a less computational burden.To realize stable and reliable fast human-robot following,a robust controller with disturbance compensation is proposed for vision-based human following control of this tracked mobile robot.Human-robot following control can be applied to assist transporta-tion of heavy goods or perform cooperative tasks in some unstructured environments con-veniently.The stability and convergence of the robust human-robot following controller are analyzed by using the stability theory of state-dependent uncertain systems proposed in Chapter 2.Through hand gesture recognition and laser radar sensor,human-robot nteraction and collision avoidance during following process are realized respectively.Two examples of human following control including straight line and curved path have illustrated the advantages and effectiveness of the proposed robust human-robot following control method in terms of improving following control performance.