Ove Kent Hagen

A toolbox of aiding techniques for the HUGIN AUV integrated inertial navigation system

Modern AUV designs must handle submerged autonomous operation for long periods of time. The state of the art solution embedded in the HUGIN AUVs is a Doppler velocity log (DVL) aided inertial navigation system (INS) that can integrate various forms of position measurement updates. In autonomous operations, position updates are only available in limited periods of time or space, thus the core velocity aided inertial navigation system must exhibit high accuracy. However, position uncertainty of a DVL aided inertial navigation system will eventually drift off, compromising either mission operation or requirements for accurate positioning of payload data. To meet the requirements for a range of military and civilian AUV applications, the HUGIN vehicles come with at flexible and powerful set of navigation techniques. Methods for position updates include GPS surface fix, DGPS-USBL, underwater transponder positioning (UTP) and bathymetric terrain navigation. Based on synthetic aperture sonar technology, a potentially revolutionary accurate velocity measurement is under development. HUGIN also comes with a navigation post-processing system (NavLab), which can be applied to increase navigational integrity and maximize position accuracy.

The HUGIN real-time terrain navigation system

Submerged long endurance autonomous missions are a real challenge for the navigation systems of autonomous underwater vehicles (AUV). Terrain navigation is a promising technique for obtaining submerged position updates for the navigation system. This paper describes a real-time terrain navigation system developed for the HUGIN AUV, and reports of sea trials, where HUGIN 1000 HUS was navigating accurately in real-time with terrain navigation as the only source for position updates during long transit legs.

An analysis of real-time terrain aided navigation results from a HUGIN AUV

Terrain aided navigation techniques are attractive for obtaining submerged position fixes for an underwater vehicle. This paper describes results from a real-time terrain navigation system developed for the HUGIN AUV family. The system is a result of previous work on terrain navigation at FFI and has been tested in sea trials on several occasions. In the test described herein, DVL altitude data was used in the terrain navigation system, together with a high quality terrain database. After 5 hours of operation with terrain navigation as the only external position sensor, the horizontal navigation accuracy remained within 5 meters. Offline computations on the same data set show that the terrain navigation algorithms are robust to initial position errors, with the expense of longer convergence times. A comparison of the DVL and map data also revealed a bias that should be investigated further.

Using Terrain Navigation to Improve Marine Vessel Navigation Systems

Terrain navigation has been used extensively by underwater vehicles in the last decade. By comparing bathymetric measurements with a digital terrain model (DTM), one can estimate a global position of a vehicle underwater, where global positioning system (GPS) signals are unavailable. With the increasing threat of GPS signal jammers and spoofers to marine vessels, GPS-independent techniques are becoming more interesting for surface vessels as well. This paper explores the idea and challenges of using terrain navigation to detect GPS spoofing and as a substitute position source during GPS jamming. A robust navigation system is suggested based on a GPS-aided inertial navigation system (INS) augmented by terrain navigation. The INS and terrain navigation system of the HUGIN autonomous underwater vehicle (AUV) was adapted to a surface vessel and tested in two experiments on the coast of Norway in scenarios simulating GPS jamming. The results from the experiments clearly indicate the feasibility of such a system, if a DTM of the area is available and the terrain is well suited for terrain navigation.

TerrLab - a generic simulation and post-processing tool for terrain referenced navigation

One of the challenges in underwater navigation for autonomous vehicles is to get position updates for the navigation system. For deep water applications or operations requiring a high degree of covertness, surfacing for GPS is not a solution. The conventional submerged position update techniques, like DGPS-USBL or LBL, requires external physical infrastructure in the operation area of the vehicle. Terrain referenced navigation is a promising technique for submerged position updates. It requires only a bathymetric map of the operation area and the use of a bathymetric sensor during the operation. Different algorithms for terrain referenced navigation have already successfully been tested for underwater applications. These algorithms include variants of the original TERCOM, different Kalman filter based techniques and non-linear Bayesian estimators like the point mass filter and particle filters. To asses the performance of the different algorithms and their robustness on both real and simulated data, FFI has developed a tool called TerrLab. The tool was originally used to qualify algorithms for the real-time terrain navigation system for the HUGIN AUVs. It is now also used to test different sensor and DTM error models for the algorithms.

Recent developments in the HUGIN AUV terrain navigation system

This paper describes recent developments in the HUGIN AUV terrain navigation system, which uses terrain measurements in order to provide submerged position updates for the main inertial navigation system. In previous versions of the system, a prior bathymetric map database has been required, sometimes restricting the use of the system. To relax this requirement, a real-time map generator has been implemented, such that the vehicle is able to build its own map during the mission. When returning to the mapped area, the system can use this map in order to obtain a terrain navigation fix, thus reducing the uncertainty of the navigation system. Results from a sea trial of the concept are presented, in which the system effectively brings the navigation uncertainty down from more than 50 to about 10 meters. Further results, using the system in the traditional manner with a prior map database, are also included.

Bayesian Terrain-Based Underwater Navigation Using an Improved State-Space Model

This paper focuses on terrain aided underwater navigation as a means of aiding an inertial navigation system. It is assumed that a prior map is present and Bayesian methods are used to estimate the position of the vehicle. Traditionally this has been done using a crude low-dimensional model in the Bayesian filters. An improved state-space model is introduced, implemented in a particle filter/sequential Monte Carlo filter and tested on real AUV (autonomous underwater vehicle) data. Compared to conventional filter models, the new model yields smoother, slightly more accurate results, though problems with overconfidence occur.

Low altitude AUV terrain navigation using an interferometric sidescan sonar

In low altitude operations of the HUGIN 1000-MR autonomous underwater vehicle, the footprint of its multibeam echo sounder (EM 3000) becomes small. Terrain navigation depends on terrain variability to be present within the footprint, and becomes less accurate in these scenarios. The problem is addressed by including interferometric sidescan measurements from the on board synthetic aperture sonar (HISAS 1030). Some preliminary results are presented on the comparison of the position accuracy obtained from terrain navigation using each sensor separately, and for the fusion of the two sensors. The comparison is done on co-registered data collected by HUGIN 1000-MR in variable terrain.

Collaborative indoor navigation for emergency services personnel

First responders and other emergency services personnel must often enter buildings which prevent the use of GPS or other satellite navigation signals for positioning. Loss of navigation capability combined with the fact that the buildings are often unknown to the personnel in question makes it more difficult for individual team members to coordinate with one another, and difficult or impossible for the team leader to monitor and direct the actions of each team member. While inertial navigation or pedestrian dead reckoning provide for some degree of navigation in GPS signal denied environments, these solutions degrade with time and may require prohibitively large and expensive inertial solutions to navigate over extended periods, while also allowing each individual user to accumulate independent positioning errors and thereby appearing to ‘drift away’ from one another. This paper presents an implementation of a collaborative navigation system utilizing each of user-to-user radio links, Global Navigation Satellite Systems (GNSS) when available, inertial navigation, pedestrian dead reckoning, as well as camera based Simultaneous Location and Mapping (SLAM) to provide a team of users with absolute and relative situational awareness for themselves and their team. The application of collaborative navigation to such a team provides the triple benefits of providing improved absolute navigation accuracy, improved relative navigation accuracy, and greatly enhanced situational awareness for all cooperating team members.

Toward autonomous mapping with AUVs - line-to-line terrain navigation

In autonomous mapping surveys the autonomous underwater vehicle (AUV) cannot rely on receiving position aiding from a surface vehicle or any other external infrastructure. The AUV can however use any payload sensor to restrict the position error drift in the navigation system. In this paper we consider using the bathymetric sensor in line-to-line terrain navigation. Terrain navigation will limit position error growth, and also ensure enough overlap between the lines such that the survey area is covered. We show that sparse map coverage from a line of bathymetry leads to a bias when used in a straightforward implementation of the point mass filter terrain navigation algorithm, and we develop a modification to the algorithm that handles this robustly. The algorithm achieves good performance when tested off-line on real data from a HUGIN AUV equipped with the HISAS 1030 interferometric synthetic aperture sonar.