When determining speed, as with any measurement, a level of uncertainty exists regarding the determined value. To counteract this, a range of different sensor systems are combined in modern train control systems. In his article, Florian Exner briefly and concisely explains the topic of "sensor diversity".
Only by measuring speed very precisely can rail vehicles be closely synchronised while travelling at high speeds. For example, if a measurement indicates the speed to be between 28 km/h and 32 km/h and a set of points can only be traversed at 30 km/h, the train must brake, even if its true speed is 28 km/h. A train that is able to determine the speed more accurately may pass through the points at a higher speed.
In addition to measurement precision, availability is also a major issue in terms of speed and position determination.
The fact is that no sensor functions optimally in every potential scenario. For example, the measurement accuracy of axle-mounted sensors suffers in the event of sliding or spinning wheels, which may occur during braking or when there are wet leaves on the rails.
In the same manner, GPS system accuracy, for example, is impaired if there are insufficient satellites within the line of sight and camera systems become unreliable if visibility is restricted by fog or heavy rain.
Sensor diversity for ETCS and CBTC
Modern systems therefore combine different sensors in order to leverage the systems’ strengths and compensate for their weaknesses. Standards such as the European Train Control System (ETCS) and the Communication-Based Train Control System (CBTC) explicitly require sensor diversity. An established combination consists of Doppler radar sensors and axle-mounted sensors. The measurement uncertainties of these two sensor types do not stem from the same cause, which means that at least one of the two sensor types is almost always available. More sensors can improve the precision and safety of speed measurement at the expense of greater complexity and component costs.