Virtual sensors or soft sensors have been applied to paper machines since the early 1990's. The majority of these applications are based purely on statistical methods, while the physics of the papermaking process are not considered. Successful control applications that are based on virtual sensor signals are rarely encountered. Statistical and mathematical methods are needed to filter measurement errors and to tackle phenomena such as collinearity. Combining mathematical techniques with technological process know-how is the key element in creating virtual sensor applications, which are robust and accurate enough for paper machine control. Voith Paper has adopted this strategy and is now able to integrate its OnV VirtualSensors into OnQ GradeControl (Voith's MD control). This innovative virtual sensor application is used to control the MD dry weight of the sheet during breaks and start-up periods. Voith Paper is currently extending this control strategy to include the filler content of the sheet. Furthermore, OnV VirtualSensors can provide real-time paper quality measurement signals, which otherwise could only be measured in the laboratory. Examples of this recent development of OnV VirtualSensors are shown for paper strength and porosity.