Sparse-LBL: a cost efficient and flexible solution for enhanced subsea autonomy

12 February.
2020

Nowadays, various positioning techniques are available for companies and institutes that need valuable navigation information to operate undersea. And while LBL usually remains the preferred solution for the highly accurate positioning of ROVs and AUVs, this method remains costly as it requires the use of many transponders to produce a single position. Operators are thus now looking for new ways that are more efficient, flexible and less costly to conduct their operations. To do so, they can now rely on sparse-LBL, a method that uses the INS equipping subsea vehicles, and that achieves similar or better performance than traditional LBL, while using less transponders. This is indeed made possible by fixing potential INS drift using the measured ranges to the seabed transponders while at the same time filtering acoustic ranges using INS data. This method, that brings increased performance and flexibility, and that reduces deployment costs, has recently been used and evaluated by the Ifremer oceanographic institute, using iXblue’s new Canopus LBL subsea positioning solution.

 

INTRODUCTION TO SPARSE-LBL


Although the Canopus solution enables various techniques such as SLAM (Simultaneous Localization and Mapping) and traditional LBL (Long BaseLine), it has been especially designed for sparse-array applications.

Using fewer transponders than conventional LBL systems, the sparse array technique reaches the same accuracy as traditional LBL systems, while optimizing transponders’ battery power consumption. iXblue’s Canopus transponders can incidentally be deployed on the seabed for multiple years thanks to their extremely low power consumption, both during operation and in standby mode.

“The Canopus solution concept, procedures and performance are all based on one basic principle: the ability to merge precise range measurements to an acoustic transponder with the very precise short-term movements from an INS in order to optimize navigation accuracy.” Explains James Titcomb, Offshore Technical Manager at iXblue. ”Contrary to classical LBL, the optimum real-time fusion is obtained using a Kalman filter, which allows the asynchronous merging of information of various natures, including acoustic range measurements performed by our Ramses transceiver.” Adds James.

 

Classical LBL principle

classical LBL

Sparse LBL principle

sparse-LBL principle

 

The preceding “sparse-LBL principle” graphic illustrates this data fusion:

1. Initially the vehicle’s INS position has a large error ellipse represented by the blue circle surrounding the AUV.

2. The ellipse error is then updated thanks to Ramses transceiver measurement of the first range to the beacon, and its transmission to the INS. The position is now well known in the axis between the vehicle and the transponder.

3. The vehicle moves along the route, the INS providing precise relative movement between acoustic interrogations.

4. As the vehicle moves relative to the beacon, the error ellipse progressively improves on multiple axis, gradually resulting in a more accurate positioning.

Following this principle, each range measurement helps computing a new position, as opposed to classical triangulation algorithms (for which at least three simultaneous range measurements are required in order to compute a position). It is therefore possible to navigate with fewer transponders (i.e. sparse array), without any compromise made on performance.

 

CANOPUS INTEGRATION ON AN IFREMER AUV


boats 1

Ifremer vessel “Europe”

photo 2

Integration of Ramses on Ifremer vehicle “IdefX”

French oceanographic institute Ifremer recently tested the complete Canopus solution and deployed their “IdefX” AUV from the Europe vessel in the Mediterranean Sea. Conducted in water depth between 1,300 and 1,700 m, those sea trials aimed at evaluating the level of accuracy that could be reached for the navigation of the AUV with only two transponders deployed within a 16 km² area.

iXblue’s Canopus solution made use of an INS (Inertial Navigation System), a Ramses transceiver and a DVL (Doppler Velocity Log) mounted within the subsea vehicle. It also made use of the dedicated Canopus transponders, deployed on the seabed and regularly interrogated by the Ramses transceiver.

schema2

Within this system, iXblue’s Phins INS was at the core of the positioning system. Its role was to gather all measurements (mainly speeds from the DVL and ranges from Ramses to fixed calibrated transponders), merge them with its internal sensor (gyroscopes and accelerometers) and deliver the optimal real time navigation information. The Canopus transponders used were the latest generation of iXblue smart seabed transponders, with long lasting listening and pinging capability, embedded environmental sensors and storage, acoustic modem and WIFI features.

 

OPERATING SCENARIO


deployment scenario

 

The Ifremer purpose was to navigate in the largest possible area using a minimum of transponders, while ensuring a high level of positioning accuracy. For these tests, two Canopus transponders were deployed and calibrated, covering a 2 km × 8 km area. Thanks to vertical acoustic propagation, the AUV was able to detect the closest seabed transponder as soon as it started diving and helped the INS navigation. Once the sea bottom reached, the AUV started a preprogrammed survey while maintaining a constant altitude above the seabed. During the whole trajectory, the Ramses transceiver within the AUV detected at least one transponder and a single range aiding navigation was performed.

 

CALIBRATION OF SEABED FIXED TRANSPONDERS AND NAVIGATION IN THE FIELD OF FIXED TRANSPONDERS


In the same way as with the available positioning modes, Canopus offers great flexibility for the method used for transponder calibration. Both SLAM and LMS (Least Mean Square) based techniques are indeed available, and calibration may be conducted using either iXblue’s Gaps USBL system, or Ramses transceiver, as the acoustic interface to the array. If available, inter-beacon ranges are employed and the calibration may be conducted from a surface vessel or a subsea vehicle. For these trials, the Ifremer chose to use Gaps USBL system to perform the box-in operation. The surface vessel thus circled the transponder location in real time and the Ramses transceiver measured the range to the Canopus transponder and used a LMS, reporting the results and giving an estimate of quality (standard deviation and residuals).(Figure 1.)

 

Once calibration of the transponder was done, the AUV was deployed and started its mission. During the survey, the AUV was furthermore tracked from the surface vessel using a Gaps USBL system. After the dive, the direct comparison between USBL tracking and embedded INS/sparse-LBL/ DVL navigation could be performed and showed extremely good results. The positionings estimated in the AUV being much more precise than USBL and overall absolute positioning accuracy was estimated around 1 m.(Figure 2.)

schema1

Figure 1.

schema 2

Figure 2.

 

Finally, bathymetry data was extracted and geo referenced using the INS/sparse-LBL/DVL navigation. Below graphic shows a correct continuity between the different isobathymetry curves and confirms the quality of the positioning.

 

lbl ifremer isobathymetry

 

CONCLUSION


This simple and straightforward real grid survey example demonstrates the efficiency of iXblue’s Canopus solution used in Sparse Array navigation mode:

  • To gain the full benefit of Canopus sparse array navigation, it is only necessary to add a Ramses transceiver to a ROV already equipped with an iXblue INS & any DVL.
  • The system is made extremely redundant and tolerant to data outages (transponder out of range, DVL bottom tracking loss, etc…).
  • While the Canopus supervision software can be used for configuring and monitoring the system, it is also possible to directly connect the Ramses transceiver to the existing INS & ROV power, the use of a topside system not being mandatory.

The above test conducted by the Ifremer thus shows that excellent positioning can be achieved, even at extreme ranges outside of the conventional LBL array. This leads to the possibility of greatly improving the positioning information for significant distances along field routes, and outside of conventional coverage.