**7. Simulation of the humaniod manipulator based upon virtual reality**

The humanoid manipulator 27-DOF model is built by using VR environment. It is shown in Fig. 15. The simulation of this model is achieved by solution the IKP with analytical model firstly. The data for analytical solution was using to learning FRNN that is shown in previous section with FRNN structure (15-33-27). The inputs are ( I=27 ) six for each two arms and two lags ( three position of end-effecter), one for waist and two for neck. The output of FRNN has dimension (N=27). The outputs are six angles of joint for each the limbs, one for waist and two for neck.

The initial posture form of humanoid manipulator is shown in Fig. 15. After solution of IKP by FRNN and get the joint angles. The values of joint angle were implemented by forward kinematic solution to get the posture of humanoid manipulator. The interaction between Matlab/Simulink and VR model will used the calculation of IKP by FRNN to implement the posture. The overall simulation design is shown in Fig. 16.

Fig. 15. Humanoid manipulator in VR.

Recurrent Neural Network with Human Simulator Based Virtual Reality 113

From the description of the all types of RNN, some of special modified in the main types to get special RNN structure related to special application such as the calculation of system parameters by identification for the states system, recognition of fault with experience system and From the description of main types of training RNN can be seen that the RNN is very flexible structure to apply many mathematical algorithms and implement to solving system problem. The implement of feedback technique with memory in RNN gives the ability to dealing with many problems that needs high calculation with iteration to get the solution.

The solution of IKP is achieved by different method but the powerful method with practice application is the analytical solution. The calculation of posture for human body or manipulator which has the similar kinematic structure with human can be achieved by analytical solution, but this solution is very difficult because of high mathematic and very length calculation. This difficulty is increasing when we need calculated the posture with time to get the movie of manipulator. This problem is solving in this chapter by using FRNN

The results of calculation are achieved for each limb separately at beginning such as for arm, lag and neck. After checking the accuracy of results by applying the forward kinematic to get the same coordinate of end- effecter for each part. The overall calculation of IKP is achieved to get the posture of manipulator. The data base for IKP is used to identify the FRNN for all point of envelop for movement where the internal memory and feedback in FRNN assistance this

The link between Matlab / Simulink and VR environment and the suitable of Matlab to execute any algorithm with high calculation are assistance to implement the posture of manipulator with high accuracy for two cases of running (movies and once posture). *The* 

The calculation of IKP by RNN can be used as data base for implement the manipulator

The VR model for manipulator can be used as human – robot interactive system based

Implement the algorithm of IKP solution based FRNN as microprocessor to able used

Abraham Ajith, (2005), 129: Artificial Neural Networks, *Handbook of Measuring System* 

Baruch Ieroham, Gortcheva Elena, Thomas Federico & and Ruben Garrido, (1999), A

Cheron G., etl, 2007, Toward an Integrative Dynamic Recurrent Neural Network for

*Design*, edited by Peter H. Sydenham and Richard Thorn. 2005 John Wiley & Sons, Ltd. ISBN: 0-470-02143-8. *Oklahoma State University, Stillwater, OK, USA*, From Web:

neuro-fuzzy model for nonlinear plants identification, Proceedings of the IASTED International Conference Modeling and Simulation (MS '99), May 5-8, 1999,

Sensorimotor Coordination Dynamics, Recurrent Neural Networks for Temporal

structure to overcame the problems of high calculation in the IKP solution.

**8. Conclusion** 

with back propagation training algorithm.

*future* works for this design are:

**9. References** 

as practice system with human robot 27 DOF.

with human robot manufacturing for many applications.

http://www.softcomputing.net/ann\_chapter.pdf

Philadelphia, Pennsylvania – USA, 291-021 – 326

on pc computer or microcomputer chip.

Data Processing, 2007

Fig. 16. The overall simulation of Humanoid Manipulator 27-DOF with Virtual Reality

The results of the simulation can be display by two methods. The first by running the design as one iteration by input one set of position and orientation of the limbs where the results is the posture form of humanoid manipulator in VR same as results shown in Fig. 17.a,b,c,d. The second method is achieved by enter many sets of coordinates to design. The results were movie of model in side the VR environment according to the path of inputs coordinate.

Fig. 17. a, b, c, d. The simulation results for posture of humanoid manipulator 27-DOF. Based upon FRNN with VR model.
