USBL (Ultra Short Base Line), SBL (Short Base Line), and LBL (Long Base Line) are all acoustic positioning systems, designed to help establish relative and absolute position underwater.  Because of their use of acoustic technology, all three systems depend upon an accurate measurement of sound velocity, although the degree of dependence varies from system type to system type.

USBL - Ultra Short Base Line - systems are generally easy to install, and hence are often used on small boats or buoys.  USBL's use a small transducer array to determine the range and bearing to a transponder that is co-located with the target  being positioned.   In the diagram below, the array is placed on a boat and the transponder is placed on a tow-body.  In many applications, multiple transponders are used, typically with a physical separation of several cm between each transponder.    

The sequence of events to determine the location of the target (see picture above, the towbody) is as follows:

1. The USBL system emits a specific acoustic pulse to query the transponders in the vicinity;
2. The pulse travels through the water to the transponder;
3. The transponder detects the USBL signal and responds with a unique transponder acoustic pulse;
4. The transponder pulse returns through the water to the USBL array;
5. The USBL array detects the transponder signal and determines the round trip acoustic travel time and phase delay of the signal to each of the transducers in the USBL array;
6. The sound speed at the USBL array is used to calculate the received bearing of the transponder signal;
7. The average sound speed of the surrounding water is used to calculate the range to the transponder. Or if refraction is being included the sound speed profile of the surrounding water isused to calculate range and adjust the vertical bearing to the transponder.

SBL arrays are typically fitted to ships and platforms.  Transducer spacings range from 10 to 50 meters.  These systems use triangulation to determine the positions of transponders in the vicinity of the array.  The range between each SBL transducer and the transponder is measured.  The intersection of the range arcs from each SBL transducer determines the position of the transponder relative to the SBL array.

If the transponder is attached to the SBL system with an umbilical, the range from the transponder to each SBL transducer can be determined with a single acoustic pulse from the transponder.  If there is no umbilical cable, the range from each SBL transducer to the transponder and back to the SBL transducer must be determined separately.

SBL systems do not require a sound speed reading at the SBL array since no phase measurement is being made.  SBL systems do require knowledge of the sound speed profile of the water column to accurately calculate range and account for refraction.

LBL - Long Base Line - systems use transponder arrays with transponder separations from 100m to several km.  The transponders are fixed to the sea floor and accurately positioned.  The item to be tracked is fitted with a transducer and processor. 

The sequence of events to determine the location of an item is as follows:

1. The target to be positioned emits an acoustic pulse from it's transducer;
2. The pulse travels through the water to each of the LBL transponders;
3. The transponders detect the signal and respond with a unique transponder acoustic pulse;
4. The transponder pulses return through the water to the target's transducer;
5. The target's processor determines the round trip acoustic travel times to each of the transponders in the LBL array;
6. The sound speed of the water is used to calculate the ranges to the transponders. If the target is operating at a similar depth to the LBL transponders a sound speed reading at the target will suffice.  However, if the acoustic pulses are travelling vertically as well as horizontally, then the average sound speed for ray tracing is required for accurate distance calculations.  Average sound speed is normally collected via vertical profile. 
   

An error in sound speed causes a systematic error which will scale the signal travel time and thus the computed range to the detriment of calibration and tracking quality.  To visualise the effect, think of a histogram displaying the measured ranges against frequency of occurrence.  Ideally the dataset should be free from systematic error and subject only to random error.This will give a ‘short and tall’ normal distribution trend line.

If sound speed is not accounted for sufficiently, the histogram will reflect this with a larger measurement spread.  Sometimes this can be seen as two distinct groupings of reciprocal measurements if the sound speed changed during baseline calibration. The significance of this error has been exacerbated due to the repeatable centimetric precision achievable over greater distances using Wideband LBL than was previously possible with tone systems.  Indeed,the greater the distance the greater the significance of sound speed error.To achieve the quoted 0.03m overall accuracy of Wideband LBL, I recommend regularly monitoring sound speeds across each baseline...with sound speed sensors.