چكيده به لاتين
One of the main challenges in teleoperation systems is the existence of time delay in the communication network. This delay, in addition to the unstability of the system, also affects its transparency. In recent years, many studies have been conducted to use intelligent methods to reduce the time delay effect in teleoperation systems. In this thesis, we intend to introduce a new method for improving the performance of teleoperation systems under variable time delay.
The methods of reducing the impact of network time delay can be divided into three categories: the use of the virtual environment, the prediction of user movement and time delay estimation. In the first method, which is also referred to as the mode-mediated teleoperation, it is predicted by creating a virtual environment on the master's side of with the same impedance to the environment, and in the second the user’s motion will be predicted. Sothat this prediction can compensate the effect of time delay. The main challenge in this method is the existence of a variable and unknown time delay in the real communication channel. Finally, in the third method, the time delay is targeted itselt and the time delay of a network is estimated using intelligent or non-intelligent methods. In this research, the goal is to predict the movement of the user's hands using intelligent methods in the first stage and then estimate the network time delay to improve these methods. By combining these two methods, it is possible to predict the movement of the user in real networks where the time delay is variable and unknown.
In this thesis, a method is first proposed to estimate the time delay in a teleoperation system, and then a neural network is created to predict the movement of the user's hand according to the time delay given to it, and finally a new system combining these two systems is created which can predict the user's movement according to the variable time delay.
The proposed methods are first simulated in Matlab's simulink environment and then implemented on phantom Omni haptic devices. The results show that by adding the time delay estimation method, the performance of predictive methods for user movement in variable and uncertain time delay networks is appropriate and significant
