16 August 2018 in IEEE Transactions on Control Systems Technology
This paper presents an optimal tracking control of engine valve timing for an electro-hydraulic variable valve actuator utilizing the adaptively estimated supply pressure that is the main source for valve timing variation. The system supply pressure is online estimated through a model-guided adaptive estimator with one-step prediction capability. The estimation model consists of a static guidance model, an in-event valve model, and a one-step prediction model. Using the estimation model, the adaptive estimator generates a nominal estimation with accelerated convergence, estimates the pressure residual error using in-event dynamics, and provides one-step forward pressure prediction along with the resulting pressure fluctuation. The linear quadratic tracking (LQT) control is used to track the desired valve timing based on an event-by-event valve model, where the adaptively estimated pressure variation is used as the feedforward control in the LQT control to improve the tracking accuracy and a Kalman filter is used to estimate the system states for the LQT state feedback control. The parameter estimation and control algorithms are validated through both the simulation and test bench studies by comparing different estimation and control schemes. Both the steady-state and transient operational conditions are investigated, and the effectiveness of the proposed adaptive LQT valve timing control is demonstrated.
1 September 2020 in Journal of Physics: Conference Series
Inherent errors and random errors are consisting in four-quadrant detector when used for azimuth measurement. Many spot fitting algorithms are proposed to solve the inherent errors caused by the structure of the four-quadrant detector itself, which can reduce the errors in a way. In order to further improve the spot positioning accuracy and eliminate the influence caused by random errors such as background light and thermal noise, we first perform Kalman filtering on the output of the four quadrants when we perform experiments to calibrate parameters. Build an experimental platform, and experimentally verify that the Kalman filtering of the output of the four quadrants before bringing in the spot positioning algorithm can improve the accuracy of the composite fitting algorithm, thereby improving the positioning accuracy.