Pathfinder DVL Guide
April
2018
EAR-Controlled Technology Subject to Restrictions Contained on the Cover Page.
Page 31
DVL Performance and Influencing Factors
Long Term Performance
TRDI specifies the long-term velocity error of its DVL products as a combination of a percentage of speed
(known as “scale factor error”) and an offset (e.g., ±0.3% ±0.2 cm/s for the Pathfinder 600 kHz bottom
track measurement). The water- and bottom-track versions have the same form but may have somewhat
different values. For water track, the first term scales with the speed through the water (the magnitude of
the vector difference between current and boat velocity). For bottom track, it scales with the magnitude of
the boat velocity over the bottom.
Almost all known bias mechanisms (other than rounding error) in DVLs approach zero as the speed ap-
proaches zero, although it is difficult to demonstrate this empirically. (Note that when the speed is exactly
zero, frozen short-term error can look like bias.) The offset specification term should not be interpreted as
the standard deviation of the bias at zero velocity, but rather as a way of accommodating non-linear be-
havior such as biases that oscillate with velocity or that give an increased scale factor bias at low velocity.
The long-term error is also known as systematic error or bias. If we assume that the predictable compo-
nent of the long-term error has been subtracted out, then the long-term error can be considered to have
zero mean, although it may be non-stationary, in which case the mean may not exist. An example of a bias
that can be subtracted out is the beam pointing error, which can be removed using a calibrated beam-to-
instrument transformation matrix. A calibration always leaves some small residual uncertainty, which can
be considered to be another source of long-term error.
When measuring velocity, the bias represents a small error. When using these velocities for navigation the
bias errors will accumulate and the total error will grow over time. This bias error can be mitigated using
various approaches; TRDI can offer application support in this area.
Bottom Track
The most important source of long-term bottom track error is the beam angle error. TRDI measures this
error and provides a beam correction matrix that reduces this error to within the system’s specified accu-
racy. Improvements in both scale factor and azimuth errors can be made by doing a more extensive cali-
bration on the platform.
The bottom track velocity measurement is proportional to the speed of sound in water. Any error in the
sound speed used in the DVL’s internal signal processing (for example errors in the temperature reading
or salinity setting) propagates directly into scale factor error. Therefore, it is important to program an ac-
curate salinity into the unit (the DVL has a temperature sensor). Another approach to minimizing sound
speed is to incorporate a speed-of-sound sensor. This external reading of the speed of sound can be sent to
the DVL, or the DVL’s sound speed can be set to a constant value and then post-processed to correct the
data for the actual sound speed in the navigation system software. For most applications, setting the salin-
ity accurately and using the DVL’s integrated temperature sensor is sufficient to yield highly accurate
readings.
Water Track
Water Track accuracy is susceptible to the same error sources outlined above. The first term for long-term
accuracy is determined only by the uncertainty of the beam angles, which is the same for water profile,
water track, and bottom track. Thus, one can use the same Long Term accuracy as presented for Bottom
Track above for water profile and water track. When using the Water Layer mode one must also be aware
of the water movement, i.e. currents will be an additional source of error.