Abstract:Problem of adaptive state observer synthesis for linear time-varying (LTV) system with unknown time-varying parameter and delayed output measurements is considered. State observation problem has attracted the attention of many researchers $[4]$. In this paper the results proposed in the $[2]$, $[9]$, $[10]$ are developed. It is supposed that the state matrix can be represented as sum of known and unknown parts. Output vector is measured with known constant delay. An adaptive identification algorithm which reconstructs unknown state and unknown time varying parameter is proposed.
What problem does this paper attempt to address?
This paper aims to solve the problem of designing an adaptive state observer for linear time - varying systems (LTV) under the conditions of partially unknown state matrices and delayed measurements. Specifically, the paper focuses on how to design an adaptive observer that can effectively estimate the system state when there are unknown time - varying parameters in the system and the measurement signal has a fixed delay.
### Main research contents:
1. **Problem background**: In control applications, the measurement data provided by actual devices usually has a delay, which makes the design of observers and control laws for dynamic systems more complicated. For linear time - invariant systems (LTI), this problem has been widely studied. However, for linear time - varying systems (LTV), this problem is still an open research topic.
2. **System description**: The paper considers a linear time - varying system, whose state equation can be expressed as:
\[
\dot{x}(t)=A_0(t)x(t)+B(t)u(t)
\]
where \(x(t)\in\mathbb{R}^n\) is the unknown state vector, \(A_0(t)\in\mathbb{R}^{n\times n}\) and \(B(t)\in\mathbb{R}^n\) are known matrices, and \(u(t)\in\mathbb{R}\) is a known input signal. It is assumed that the state matrix \(A_0(t)\) can be decomposed into a known part and an unknown part:
\[
A_0(t)=A(t)+\theta(t)I
\]
where \(I\in\mathbb{R}^{n\times n}\) is the identity matrix and \(\theta(t)\) is an unknown time - varying parameter.
3. **Measurement model**: The measurement of the output vector \(y(t)\) has a known constant delay \(d\):
\[
y(t)=C(t)x(\phi(t))
\]
where \(\phi(t)=t - d\), \(d>0\) is the constant delay, and \(C(t)\in\mathbb{R}^{n\times n}\) is a known matrix.
4. **Objective**: Design an adaptive observer so that it can effectively estimate the system state vector \(x(t)\) in the presence of unknown time - varying parameters and measurement signal delays in the system.
### Solution methods:
1. **Parameter estimation**: First, estimate the unknown parameter \(\omega\), and then use this parameter to estimate the unknown time - varying parameter \(\theta(t)\).
2. **State estimation**: Use the generalized parameter - estimated - based observer (GPEBO) technique to estimate the system state vector \(x(t)\) based on the estimated \(\theta(t)\).
### Simulation results:
The paper verifies the effectiveness of the proposed adaptive observer through numerical simulation. The simulation results show that the observer can accurately estimate the system state and unknown parameters in the presence of measurement delays.
### Future research directions:
The paper points out that future research can be further extended to the situation where the system state is unmeasurable and can only be estimated through the output with a delay.