The pcregistericp function only estimates a rigid transformation (rotation and translation only, no scale). That means the incoming point clouds are expected to correspond to the same scale. There are variants of ICP like SICP that also estimate scale but pcregistericp function implements the algorithm from the following papers as mentioned in its documentation.
[1] Chen, Y. and G. Medioni. “Object Modelling by Registration of Multiple Range Images.” Image Vision Computing. Butterworth-Heinemann . Vol. 10, Issue 3, April 1992, pp. 145-155.
[2] Besl, Paul J., N. D. McKay. “A Method for Registration of 3-D Shapes.” IEEE Transactions on Pattern Analysis and Machine Intelligence. Los Alamitos, CA: IEEE Computer Society. Vol. 14, Issue 2, 1992, pp. 239-256.
