Identification of dynamic systems using support vector regression neural networks

A novel adaptive support vector regression neural network （SVR-NN） is proposed, which combines respectively merits of support vector machines and a neural network. First, a support vector regression approach is applied to determine the initial structure and initial weights of the SVR-NN so that the network architecture is easily determined and the hidden nodes can adaptively be constructed based on support vectors. Furthermore, an annealing robust learning algorithm is presented to adjust these hidden node parameters as well as the weights of the SVR-NN. To test the validity of the proposed method, it is demonstrated that the adaptive SVR-NN can be used effectively for the identification of nonlinear dynamic systems. Simulation results show that the identification schemes based on the SVR-NN give considerably better performance and show faster learning in comparison to the previous neural network method.

Nonlinear Time-Varying Systems Identification Using Basis Sequence Expansions Combined with Neural Networks

A new method for identifying nonlinear time varying systems with unknown structure is presented. The method extends the application area of basis sequence identification. The essential idea is to utilize the learning and nonlinear approximating ability of neural networks to model the non linearity of the system, characterize time varying dynamics of the system by the time varying parametric vector of the network, then the parametric vector of the network is approximated by a weighted sum of known basis sequences. Because of black box modeling ability of neural networks, the presented method can identify nonlinear time varying systems with unknown structure. In order to improve the real time capability of the algorithm, the neural network is trained by a simple fast learning algorithm based on local least squares presented by the authors. The effectiveness and the performance of the method are demonstrated by some simulation results.

Opportunities and challenges for developing closed-loop bioelectronic medicines

The peripheral nervous system plays a major role in the maintenance of our physiology. Several peripheral nerves intimately regulate the state of the brain, spinal cord, and visceral systems. A new class of therapeutics, called bioelectronic medicines, are being developed to precisely regulate physiology and treat dysfunction using peripheral nerve stimulation. In this review, we first discuss new work using closed-loop bioelectronic medicine to treat upper limb paralysis. In contrast to open-loop bioelectronic medicines, closed-loop approaches trigger ‘on demand' peripheral nerve stimulation due to a change in function(e.g., during an upper limb movement or a change in cardiopulmonary state). We also outline our perspective on timing rules for closedloop bioelectronic stimulation, interface features for non-invasively stimulating peripheral nerves, and machine learning algorithms to recognize disease events for closed-loop stimulation control. Although there will be several challenges for this emerging field, we look forward to future bioelectronic medicines that can autonomously sense changes in the body, to provide closed-loop peripheral nerve stimulation and treat disease.

Multi-agent reinforcement learning using modular neural network Q-learning algorithms

Reinforcement learning is an excellent approach which is used in artificial intelligence,automatic control, etc. However, ordinary reinforcement learning algorithm, such as Q-learning with lookup table cannot cope with extremely complex and dynamic environment due to the huge state space. To reduce the state space, modular neural network Q-learning algorithm is proposed, which combines Q-learning algorithm with neural network and module method. Forward feedback neural network, Elman neural network and radius-basis neural network are separately employed to construct such algorithm. It is revealed that Elman neural network Q-learning algorithm has the best performance under the condition that the same neural network training method, i.e. gradient descent error back-propagation algorithm is applied.

A special hierarchical fuzzy neural-networks based reinforcement learning for multi-variables system

Proposes a reinforcement learning scheme based on a special Hierarchical Fuzzy Neural-Networks (HFNN)for solving complicated learning tasks in a continuous multi-variables environment. The output of the previous layer in the HFNN is no longer used as if-part of the next layer, but used only in then-part. Thus it can deal with the difficulty when the output of the previous layer is meaningless or its meaning is uncertain. The proposed HFNN has a minimal number of fuzzy rules and can successfully solve the problem of rules combination explosion and decrease the quantity of computation and memory requirement. In the learning process, two HFNN with the same structure perform fuzzy action composition and evaluation function approximation simultaneously where the parameters of neural-networks are tuned and updated on line by using gradient descent algorithm. The reinforcement learning method is proved to be correct and feasible by simulation of a double inverted pendulum system.

High-Performance of Power System Based upon ANFIS （Adaptive Neuro-Fuzzy Inference System） Controller

The proposed controller incorporates FL （fuzzy logic） algorithm with ANN （artificial neural network）. ANFIS replaces the conventional PI controller, tuning the fuzzy inference system with a hybrid learning algorithm. A tuning method is proposed for training of the neuro-fuzzy controller. The best rule base and the best training algorithm chosen produced high performance in the ANFIS controller. Simulation was done on Matlab Ver. 2010a. A case study was chopper-fed DC motor drive, in continuous and discrete modes. Satisfactory results show the ANFIS controller is able to control dynamic highly-nonlinear systems. Tuning it further improved the results.

A biologically inspired model for pattern recognition

In this paper, a novel bionic model and its performance in pattern recognition are presented and discussed. The model is constructed from a bulb model and a three-layered cortical model, mimicking the main features of the olfactory system. The olfactory bulb and cortex models are connected by feedforward and feedback fibers with distributed delays. The Breast Cancer Wisconsin dataset consisting of data from 683 patients divided into benign and malignant classes is used to demonstrate the capacity of the model to learn and recognize patterns, even when these are deformed versions of the originally learned patterns. The performance of the novel model was compared with three artificial neural networks (ANNs), a back-propagation network, a support vector machine classifier, and a radial basis function classifier. All the ANNs and the olfactory bionic model were tested in a benchmark study of a standard dataset. Experimental results show that the bionic olfactory system model can learn and classify patterns based on a small training set and a few learning trials to reflect biological intelligence to some extent.