Improved wheel speed processing in the RT systems

Product News

December 8, 2010

200_pegasem_wheel_encoderIf you are using a wheel speed input with your RT system, then the latest RT software release will be very useful for you. We have made significant improvements to the wheel speed algorithm in this software release which makes the position accuracy of the RT even better in adverse GPS conditions. The new, improved wheel speed processing means lower drift rates and increased positional accuracy when obstructions such as bridges, tunnels or dense urban canyons occur.

To make this improvement the wheel speed state has been added to the Kalman filter in the RT . Previously the RT only used the wheel speed to update the Kalman filter when the external GNSS was not available. Now the wheel speed is used all the time and this lets the Kalman filter remove more of the poor GNSS measurements. The older processing platform in the original RT products was not fast enough to have the wheel speed state as part of the main Kalman filter but the new, faster processor used in the RT  product does not have a problem with the extra processor load caused by this new state.

Wheel speed is an excellent addition to the RT since it makes a significant difference to the drift. Our tests show that the RT drifts by about 2.5m per kilometre (or minute) in “ideal” conditions when using wheel speed. When there are significant errors in the GNSS position and velocity before an obstruction, like a tunnel, then the drift can be higher. Using the new wheel speed this is reduced so that we should be closer to our specification in non-ideal conditions. Quantifying this exactly is very difficult.

 

Please note: This improvement is only available on the new processor platform (TP500) and cannot be run on older RTs unless they have a processor upgrade. For more information please contact support.

Image: Pegasem wheel speed encoder, see www.pegasem.de.