This paper presents a modified training method for Recurrent Neural Networks. This method depends on the Non linear Auto Regressive (NARX) model with Modified Wavelet Function as activation function (MSLOG) in the hidden layer. The modified model is known as Modified Recurrent Neural (MRN). It is used for identification Forward dynamics of four Degrees of Freedom (4-DOF) Selective Compliance Assembly Robot Arm (SCARA) manipulator robot. This model is also used in the design of Direct Inverse Control (DIC). This method is compared with Recurrent Neural Networks that used Sigmoid activation function (RS) in the hidden layer and Recurrent Neural Networks with Wavelet activation function (RW). Simulation results shows that the MRN model is better than RS and RW in identification the forward dynamics and provides good results in the Direct Inverse Neuro- Controller (DINC).
In this research the results of applying Artificial Neural Networks with modified activation function to
perform the online and offline identification of four Degrees of Freedom (4-DOF) Selective Compliance
Assembly Robot Arm (SCARA) manipulator robot will be described. The proposed model of
identification strategy consists of a feed-forward neural network with a modified activation function that
operates in parallel with the SCARA robot model. Feed-Forward Neural Networks (FFNN) which have
been trained online and offline have been used, without requiring any previous knowledge about the
system to be identified. The activation function that is used in the hidden layer in FFNN is a modified
version of the wavelet func
In this research the results of applying Artificial Neural Networks with modified activation function to perform the online and offline identification of four Degrees of Freedom (4-DOF) Selective Compliance Assembly Robot Arm (SCARA) manipulator robot will be described. The proposed model of identification strategy consists of a feed-forward neural network with a modified activation function that operates in parallel with the SCARA robot model. Feed-Forward Neural Networks (FFNN) which have been trained online and offline have been used, without requiring any previous knowledge about the system to be identified. The activation function that is used in the hidden layer in FFNN is a modified version of the wavelet function. This approach ha
... Show More
(24)
(21)
The inverse kinematics of redundant manipulators has infinite solutions by using conventional methods, so that, this work presents applicability of intelligent tool (artificial neural network ANN) for finding one desired solution from these solutions. The inverse analysis and trajectory planning of a three link redundant planar robot have been studied in this work using a proposed dual neural networks model (DNNM), which shows a predictable time decreasing in the training session. The effect of the number of the training sets on the DNNM output and the number of NN layers have been studied. Several trajectories have been implemented using point to point trajectory planning algorithm with DNNM and the result shows good accuracy of the end
... Show MoreThis paper includes an experimental study of hydrogen mass flow rate and inlet hydrogen pressure effect on the fuel cell performance. Depending on the experimental results, a model of fuel cell based on artificial neural networks is proposed. A back propagation learning rule with the log-sigmoid activation function is adopted to construct neural networks model. Experimental data resulting from 36 fuel cell tests are used as a learning data. The hydrogen mass flow rate, applied load and inlet hydrogen pressure are inputs to fuel cell model, while the current and voltage are outputs. Proposed model could successfully predict the fuel cell performance in good agreement with actual data. This work is extended to developed fuel cell feedback
... Show MoreAbstract
This paper presents mechanical and electrical design, and implementation process of industrial robot, 3-DoF type SCARA (selective compliment assembly robot arm),with two rotations and one translation used for welding applications.The design process also included the controller design which was based on PLC(programmable logic controller) as well as selection of mechanical and electrical components.The challenge was to use the available components in Iraq with reasonable costs. The robot mentioned is fully automated using programmable logic controller PLC(Zelio type SR3-B261BD),with 16inputs and 10 outputs. The PLC was implemented in FBD logic to obtain three different automatic motions with hi
... Show More
(2)
(9)
(6)
(50)
(43)
This paper presents a proposed neural network algorithm to solve the shortest path problem (SPP) for communication routing. The solution extends the traditional recurrent Hopfield architecture introducing the optimal routing for any request by choosing single and multi link path node-to-node traffic to minimize the loss. This suggested neural network algorithm implemented by using 20-nodes network example. The result shows that a clear convergence can be achieved by 95% valid convergence (about 361 optimal routes from 380-pairs). Additionally computation performance is also mentioned at the expense of slightly worse results.
This paper adapted the neural network for the estimating of the direction of arrival (DOA). It uses an unsupervised adaptive neural network with GHA algorithm to extract the principal components that in turn, are used by Capon method to estimate the DOA, where by the PCA neural network we take signal subspace only and use it in Capon (i.e. we will ignore the noise subspace, and take the signal subspace only).