Abstract
This paper investigates the local linear model tree (LOLIMOT), a typical neural fuzzy model, in the multiple-input-multiple-output model predictive control (MPC). In the conventional LOLIMOT, the structural parameters including centres and variances of its Gaussian kernels are set based on equally dividing the input data space. In this paper, after the structural parameters are initially obtained from the input space partition, they are optimized by the gradient descent search, from which the space partitions are further adjusted. This makes it better for the model structure to fit the input data statistics, leading to improved modelling performance with small model size. The MPC based on the proposed structurally optimized LOLIMOT is then implemented and verified with both numerical and diesel engine plants. Validation results show that the proposed MPC has significantly better controlling performance than the MPC based on the conventional LOLIMOT, making it an attractive solution in practice.
Highlights
M ODEL predictive control (MPC) has been used in many fields including robotics, vehicle, aerospace and power electronics [1]
Examples include the MPC with finite horizon based on the fuzzy discrete systems [7], the output feedback predictive control based on the Takagi–Sugeno (T–S) fuzzy model [8] [9], the nonlinear MPC (NMPC) based on self-feedback fuzzy network [10], and the adaptive T–S fuzzy model-based predictive controller
This paper investigates the local linear model tree (LOLIMOT) [18] [19], a typical neural fuzzy model
Summary
M ODEL predictive control (MPC) has been used in many fields including robotics, vehicle, aerospace and power electronics [1]. The performance of the MPC well depends on modelling the target plant because of the model-based optimization to obtain the control actions [2] Both linear MPC [3] and nonlinear MPC (NMPC) [4] have been proposed. The performance of the LOLIMOT relies on the model structure including the model size (i.e. the neuron number), the centers and variances of the Gaussian kernels. A novel gradient descent search approach is proposed to optimize centers and variances of the Gaussian kernels, based on which the input space partitions are further adjusted. This leads to more efficient Gaussian kernels in the LOLIMOT. The proposed NMPC is verified by both numerical data and experimental data from the turbocharged diesel engine platform, where significantly better control performance is observed in both cases
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
