A state estimator with reduced sensitivity to sensor quantization

The authors present a simple method to reduce measurement quantization effects in state estimation. The work addresses a practical problem in industrial electronics, where many modern electronic sensors produce quantized outputs (e.g. incremental encoder), or are sampled with finite quantization in digital control applications (e.g. N-bit analog-to-digital converter). Classical methods often describe quantization error as a random process, but the authors' approach is motivated by stability considerations from nonlinear systems analysis. Lyapunov stability analysis shows that the effect of quantization is greatly reduced by embedding the nonlinear quantization characteristic within the state estimator in a deterministic fashion. Several simulation examples complement the theoretical analysis. To the authors' knowledge, the contribution is unique among recent works for its simplicity in addressing the case of large quantization error.