Microsoft Research posted its paper from CHI'12 conference "Shake’n’Sense: Reducing Interference for Overlapping Structured Light Depth Cameras" by Alex Butler, Shahram Izadi, Otmar Hilliges, David Molyneaux, Steve Hodges, and David Kim.
"The key behind Shake‟n‟Sense is to minimally vibrate a Kinect camera unit using an offset-weight vibration motor and thereby artificially introduce motion blur. Both the structured light DOE illuminator and the IR camera of the Kinect will move in harmony, which means that depth sensing works as normal, albeit with a little induced blur. However, even minor almost imperceptible motion of the sensor in this way causes blurring of structured light patterns from other units which serves to eliminate most of the crosstalk."
Update: It looks like University of North Carolina researchers proposed the same technique at about the same time.