The authors present a new approach for estimating the impulse response of indoor radio channels using the signal subspace concept. The technique is based on the use of information theoretic criteria, a high-resolution spectral analysis technique, and principal component analysis. It is shown that the new technique can be used to overcome difficulties experienced by conventional methods and, as a result, can produce more accurate estimates of the channel parameters. The use of this technique is demonstrated by data analysis using both simulated and measured indoor radio channel data.