FIG. 9 shows a relationship of a synchronous coordinate system (dr, qr), an estimated synchronous coordinate system (d{circumflex over (r)}, q{circumflex over (r)}) and an injection coordinate system (di, qi) indicating an actual rotor position. Referring to FIG. 9, the injected d-axis is defined as being ahead of the estimated coordinate system d-axis by the injection angle (?_i) (this may be inverted in another embodiment). In this case, the relationship of the first angular error ({tilde over (θ)}_r), the injection angle (?_i) and a second angular error (a difference between the synchronous d-axis and the injection d-axis, {tilde over (θ)}_i) may be as in Equation 8 below.
{tilde over (θ)}_r={tilde over (θ)}_i+?_i ??[Equation 8]

For clear explanation, the terms are arranged such that {tilde over (θ)}_r is the first angular error, {tilde over (θ)}_i the second angular error, and ?_i is the injection angle.

The current signal (i_sig) obtained by dividing the current ripple of the injected q-axis by the clock signal when the high frequency voltage is injected to the injected d-axis is expressed as a function of the second angular error ({tilde over (θ)}_i) as in Equation 9 below. As mentioned above, it is found that the current signal (i_sig) is expressed as a function of the torque command, the rotator angle and the first angular error. In this embodiment, since a new variable of the injection angle (?_i) is applied, the current signal is expressed as a function further including the injection angle as in Equation 9 below.