Research On Dynamic Quantization Policies Of Control Systems

As the fast development of computer and communication technology, the networked control system(NCS) has been widely used in our life. So more and more researchers have focused on those hot points about NCS. One of them is the dynamic quantization policy under bandwidth constraint.This paper chooses a class of linear systems with bounded noise as the research object, and gives out the corresponding dynamic quantization policy to reduce the quantization error. Later, it raised out an optimal quantization performance problem under dynamic quantization policy. It turns out to be getting the answer of an optimal function. Furthermore, the paper gives out the upper bound and lower bound of the optimal quantization performance. Because the complecity of the computation, we consider the sub-optimal quantization policy instead. The simulation result confirms that the efficiency of dynamic quantization policy, it's performance is close to the upper bound of the optimal quantization performanece, that is to say, it's close to the performance of optimal quantization policy.After that, this paper researches the dynamic quantization policy for a class of nonlinear feedforward systems, and effects to the stability of those systems. In the former results of this field, researchers have raised out the dynamic quantization policies below: one is the n-bit dynamic quantization policy under Locally Lipschitz condition; the other one is the 1-bit dynamic quantization policy under Global Lipschitz condition. The mainly work of this paper is extending the 1-bit dynamic quantization policy from the Global Lipschitz condition into the Locally Lipschitz condition. By modifying the former policy and using saturation control, we can stabilize the nonlinear feedforward system by 1 bit, which is the least bandwidth. This paper has gave out the proof of the result, and verified it by simulation. Also, by adjusting the dynamic quantization policy, we can stabilize the nonlinear feedforward system with bounded noise, which has been proved by simulation.