Figure 5. Example of an Elman network
The Artificial Neural Network simulator used in our experiments was the Stuttgart Neural Network Simulator - SNNS2, it is a free software, and a quite complete neural network simulator that have several additional tools that allow us to create scripts and execute learning and simulation tasks in batch mode. The SNNS facilities also simplify the analysis of the obtained results and creation of graphic plots.
The Table 4 shows the main ANN parameters used in our simulations. For a complete description of these pa- rameters, see the SNNS manual. The input layer is fully connected to the hidden layer, and the hidden layer is fully connected to the output layer. In addition, the hidden layer is fully, recurrently connected (Elman architecture). At each time step of the simulation of a robot’s behavior, the eight sensor signals are scaled to floating point values in [−1.0; 1.0] interval, and supplied to the input layer. The values are propagated to the hidden and output neurons. The hidden neurons use the hyperbolic tangent activation function.
Figure 5. Example of an Elman networkThe Artificial Neural Network simulator used in our experiments was the Stuttgart Neural Network Simulator - SNNS2, it is a free software, and a quite complete neural network simulator that have several additional tools that allow us to create scripts and execute learning and simulation tasks in batch mode. The SNNS facilities also simplify the analysis of the obtained results and creation of graphic plots.The Table 4 shows the main ANN parameters used in our simulations. For a complete description of these pa- rameters, see the SNNS manual. The input layer is fully connected to the hidden layer, and the hidden layer is fully connected to the output layer. In addition, the hidden layer is fully, recurrently connected (Elman architecture). At each time step of the simulation of a robot’s behavior, the eight sensor signals are scaled to floating point values in [−1.0; 1.0] interval, and supplied to the input layer. The values are propagated to the hidden and output neurons. The hidden neurons use the hyperbolic tangent activation function.
การแปล กรุณารอสักครู่..
