S.S. Ge

Nonlinear adaptive control using neural networks and its application to CSTR systems

Abstract In this paper, adaptive tracking control is considered for a class of general nonlinear systems using multilayer neural networks (MNNs). Firstly, the existence of an ideal implicit feedback linearization control (IFLC) is established based on implicit function theory. Then, MNNs are introduced to reconstruct this ideal IFLC to approximately realize feedback linearization. The proposed adaptive controller ensures that the system output tracks a given bounded reference signal and the tracking error converges to an e -neighborhood of zero with e being a small design parameter, while stability of the closed-loop system is guaranteed. The effectiveness of the proposed controller is illustrated through an application to composition control in a continuously stirred tank reactor (CSTR) system.

Adaptive NN control for a class of discrete-time non-linear systems

In this paper, adaptive neural network (NN) control is investigated for a class of single-input single-output (SISO) discrete-time unknown non-linear systems with general relative degree in the presence of bounded disturbances. Firstly, the systems are transformed into a causal state space description, adaptive state feedback NN control is presented based on Lyapunovs stability theory. Then, by converting the systems into a causal input-output representation, adaptive output feedback NN control is given. Finally, adaptive NN observer design and observer-based adaptive control are presented under the assumption of persistent excitation (PE). All the control schemes avoid the so-called controller singularity problem in adaptive control. By suitably choosing the design parameters, the closed-loop systems are proven to be semi-globally uniformly ultimately bounded (SGUUB). Simulation studies show the effectiveness of the newly proposed schemes.

Sparse-Representation-Based Graph Embedding for Traffic Sign Recognition

Researchers have proposed various machine learning algorithms for traffic sign recognition, which is a supervised multicategory classification problem with unbalanced class frequencies and various appearances. We present a novel graph embedding algorithm that strikes a balance between local manifold structures and global discriminative information. A novel graph structure is designed to depict explicitly the local manifold structures of traffic signs with various appearances and to intuitively model between-class discriminative information. Through this graph structure, our algorithm effectively learns a compact and discriminative subspace. Moreover, by using L2, 1-norm, the proposed algorithm can preserve the sparse representation property in the original space after graph embedding, thereby generating a more accurate projection matrix. Experiments demonstrate that the proposed algorithm exhibits better performance than the recent state-of-the-art methods.

Adaptive neural network control of a wheeled mobile robot violating the pure nonholonomic constraint

In this paper, adaptive neural network control is presented for a wheeled mobile robot violating the pure nonholonomic constraints. The nonholonomic constraint of the vehicle is assumed to be violated by an unknown slippage. Under a restricted assumption of the slippage, the proposed controller is constructed at the dynamical level using backstepping. The neural network (NN) controller deals with the unmodelled dynamics in the robot and eliminates the need for the error prone process in obtaining the LIP form of the system dynamics. In addition, the time-consuming offline training process for the NN is avoided. All the system states are shown to be able to track the desired trajectory. Simulation results are given to show the effectiveness of the proposed controller.

A direct adaptive controller for dynamic systems with a class of nonlinear parametrizations

Abstract In this paper, the adaptive control problem is considered for a class of nonlinearly parametrized systems. By introducing a novel kind of Lyapunov functions, a direct adaptive controller is developed for achieving asymptotic tracking control. The transient performance of the resulting closed-loop system can be guaranteed by suitably choosing the Lyapunov function to construct the controller. The effectiveness of the proposed scheme is illustrated with two examples.

Controller design for a single-link flexible smart materials robot with experimental tests

In this paper, controller design is investigated for a single-link flexible smart materials robot, which combines both the advantages of flexible robots and piezoelectric materials. To avoid the drawbacks resulting from model uncertainties and/or model truncations, model-free controllers (both decentralized and centralized) are proposed for tip regulation and residual vibration suppression. In contrast to traditional model-based methods, the controllers presented here are derived from the basic energy-work relationship and are independent of the system dynamics.

Locally connected graph for visual tracking

Abstract Based on fast feature extraction, the subspace representation model provides a compact notion of the “thing” being tracked rather than treating the target as a sparse feature representation. The main challenges of the subspace representation model can be attributed to the difficulty of handling the appearance variability of a target object. In this paper, we present a subspace learning algorithm based on graph embedding that uses a Locally Connected Graph (LCG). By constructing a supervised graph with several types of labeled target samples, the algorithm can effectively learn the semantic subspace modeling for some appearance variability. Moreover, by using an additional constraint connection among several subgraphs, the algorithm can obtain a more compact subspace model. In comprehensive experiments, our algorithm has demonstrated better performance than alternatives reported in the recent literature.

Terrestrial navigation based on integrated GPS and INS

The Global Positioning System (GPS) and Inertial Navigation System (INS) have complimentary features that can be exploited in an integrated system, thus resulting in improved navigation performance. The INS is able to provide accurate aiding data on short-term vehicle dynamics, while the GPS provides accurate data on long-term vehicle dynamics. In this paper, a complete solution is presented for terrestrial navigation based on integrated GPS and INS using Kalman filtering technique.

Robust adaptive control for a class of nonlinear systems with generalized Prandtl-Ishlinskii hysteresis

In this paper, a robust adaptive control is proposed for a class of nonlinear systems preceded by unknown hysteresis. The generalized Prandtl-Ishlinskii (P-I) model is used to describe the characteristics of unknown hysteresis. The challenge addressed here is to fuse the generalized P-I model with controller design without constructing a hysteresis inverse. The Nussbaum-type function is used to solve the problem of unknown control directions and the high-order neural network approximate method is used to overcome the computational complexity. The global stability of the closed-loop system is achieved, and the effectiveness of the proposed control approach is demonstrated through simulation example.

Direct MNN control of continuous stirred tank reactor based on input-output model

A direct multi-layer neural network control scheme is investigated for a class of continuous stirred tank reactors (CSTR). The CSTR plant under study is discretized to an input-output based /spl tau/-step ahead discrete-time model. By implicit function theorem, the existence of the implicit desired feedback control (IDFC) is proved. Multi-layer neural networks are used as the emulator of the desired feedback control. Projection algorithms are used to guarantee the boundness of the multi-layer neural network weights. Simulation results show the effectiveness of the proposed controller.