Linearization by output-injection under sampling

Linearization by output injection has been studied in both continuous-time and discrete-time contexts. The object of this work is to discuss the possible preservation of such a property under sampling. It is shown that, up to an error in 0(δⁿ⁺¹), δ being the sampling period and n the state space dimension, such a property can be maintained under sampling. The pursued approach relies on differential/difference representations of discrete-time systems which provide suitable combinatoric descriptions of the sampled equivalent models.