Case Overview of Missing Flight – MH370

For today’s blog post we will look at a case study where AUV technology has been implemented; the investigative search of the plane that went missing in 2014, flight MH370. This was a large operation lead by the Australian government department, Australian Transport and Safety Bureau (ATSB) who were publishing ongoing reports on their website. The end of flight search area analysis was conducted by the Australian Defence and Science Technology group (DST). The drift analysis and landfall of debris involved collaboration with an oceanographer from the Commonwealth Scientific and Industrial Research Organisation (CSIRO). The survey operations was contracted out to ocean survey company Fugro, with a base in Perth, on the West Coast of Australia.

The story of this aircraft is a standard Malaysian airlines flight on 7th March 2014 heading from Kuala Lumpur to Beijing. There were communications between air traffic control from flight departure with final radar return 100 minutes into the flight after already deviating from the planned flight path, the plane continued to fly for 6 hours. During this time there was a series of ‘handshakes’ between the Ground Earth Station (GES) and the plane via satellite.  These crucial pieces of information included satellite communication (SATCOM) data and their timing and frequency metadata. Burst timing offset(BTO) and burst frequency offset(BFO) could then be used to calculate the plane location.

A search area was later refined after analysis. The methods used included Bayesian filtering on information such as plane dynamics and drift analysis and performing backward calculations to determine the probable final accident location. The initial 60,000 square kilometres search area was expected to take 12 months. The Bayesian method performed by DST included SATCOM data, aircraft dynamics and environmental effects, produced a probability density function of the location of the aircraft. This is visualised in the image below by DST.


The drift analysis measured aircraft parts relative to ‘oceanographic drifters’. This used the existing archive from the Global Drifter Program by NOAA of satellite tracked buoys. The drift study did an analysis on a flaperon, confirmed from the plane, which was found on the coast of La Reunion Island July 2015. The analysis used the the oceanographic drifter models, as well as the lack of debris on the west coast of Australia, to lower the search area of the sea surface in 2014.

Maps can be found through here.

The initial search area had a first priority of 20 nautical miles either side of the 7th arc and around 700 kilometres along it. The 7th arc is a certain distance from which the plane was in communication via satellite to ground control. It is inferred from burst timing offset and is the 7th communication since losing contact. The enormity of the search area means that the technology of AUVs have helped make it realistic. Previously, the search would have taken a decade. The initial bathymetry survey needed to be done as this remote area is lacking in detailed ocean mapping. The depths that were to be reached could be at least 6000 metres, the information obtained would inform on what is involved and planning for the underwater search and which sensor types that would be used.

While the search was a joint agreement between Australia, Malaysia and China, Australia was the leading investigator and here we will focus on their deployed equipment. Fugro was contracted by ATSB for the initial bathymetry survey in mid May and later contracted for the extensive underwater search in August 2014. There were three survey vessels that were used over the search period. The vessels included deep tow systems, EM 2040 multibeam sonar that had the ability to reach 6000 metres in depth and a Kongsberg Hugin AUV. The three vessels were the Fugro Equator, Fugro Discovery, Fugro Supporter.

The initial bathymetric survey commencing in June 2014 to December 2014, would build on low resolution seafloor topology information mainly gained from satellite altimetry data, which can estimate water depth. (A link to an article on satellite altimetry and an example of its use.) The survey vessel for this was the Fugro Equator that was equipped with a multibeam system and sub-bottom profiler, which found some significant differences to the existing data.

Following the initial bathymetry was the extensive area search for the final accident location. This commenced in October 2014 with the Fugro Discovery. The search area later expanded to 120000 square kilometres. The AUV had several missions during the search operation and could be deployed in areas that had more complex terrain than where the deep tow could survey. The initial proposal had the AUV involved once the debris field had been found and it could then do the detailed mapping and optical imaging to be provided to aircraft engineers. As the terrain became too complex for the deep tows, an AUV was then deployed, being self propelled with an endurance of around 24 hours before having to re-surface and recharge. This was done from the Fugro Supporter commencing January 2015. Outlining this equipment shows the extensive resources taken up by this operation, lasting over 2 years with the search being suspended just in January, 2017. It is not surprising then, that this is the most expensive aircraft search in history. This video from Fugro summarizes features utilized by the AUV.

This is an impressive project with documentation from ATSH, CSIRO, DST and Fugro with the survey focus. ATSB providing updates and search area analysis with some media interest. While this is a non commercial project and therefore being government funded, such large scale projects can test capabilities of equipment. It also has a human interest, exploring a mystery and mapping largely unmapped underwater areas, but also affects the people impacted by the tragedy.


  • Griffin, DA, Oke, PR and Jones, EM (2016). The search for MH370 and ocean surface drift. CSIRO Oceans and Atmosphere, Australia. Report number EP167888. 8 December 2016.
  • Davey et al. 2016, Bayesian Methods in the Search for MH370, SpringerBriefs in Electrical and Computer Engineering
  • ATSB (2016) MH370 — First Principles Review. ATSB, 20 December 2016
  • ATSB (2015) MH370 – Definition of underwater search areas. ATSB, 3, 10 December 2015.
  • Millar, D, 2015, The Search for MH370: Challenges in performing an underwater search in a remote area of the deep ocean. US Hydro. March 16-19, Maryland USA.

Breakdown and components of AUV

The concept of AUVs has been in development over the past few decades, however there have been rapid advances in recent years. An early example of an unmanned surface vessel is from 1996-97 of MIT’s AutoCat and the HUGIN project in the early 1990s. Existing and robust technologies are integrated into AUVs to give them functionality and make them appropriate for certain applications such as bathymetry and surveying. These technologies include SONAR and positioning methods long baseline (LBL) and ultra short baseline (USBL). While these are acoustic positioning, another category of positioning technologies are inertial which involve measurements of the vessel’s velocity and directions. Improvements in inertial and acoustic positioning technologies has enabled the further autonomy and development of AUVs.

The drive for the development of AUVs has come from military and research. The US navy has had their master plan for UUV documented in 2000.  There have been projects using USV as supporting navigation for AUVs by LBL in 2007. Building AUVs can be done by adding controls, navigation and telemetry to any small craft. Initial prototypes used old military missiles with a higher dependence on telemetry control that would be considered an ROV. The use of AUVs are promising for defence and science.

The makeup of AUVs can be broken down into certain components. They would have different functions but can also work together as a check. Here we can look at the groups of positioning, data collection and sensors, mapping and navigation as well as some platforms and vessels they operate on. Positioning components that can be used are LBL, USBL, Doppler velocity log (DVL) and IMU. Some of these can be combined to improve performance. Data collection sensors can include acoustic Doppler current profiler(ADCP), side scan sonar, multibeam echo sounders and optical methods and cameras. These are referred to as payload sensors in AUV specifications. Mapping and navigation can include robotics and computer based processes such as SLAM and its different forms. These are computer processes as they are made up of algorithms using input data to produce obstacle avoidance, filtering and real-time mapping. Finally, platforms can be in surface vessels and underwater missile models made by a number of companies. Some of these parts will be explained below.

flow diagram
Figure: Example breakdown of an AUV


LBL systems use a network of at least three acoustic beacons with predetermined known locations. These are sparsely spread along the seafloor to create the long baseline. The AUV can then infer its position by triangulating measurements to each of the beacons. This system can also be deployed using GPS intelligent buoys on which the acoustic beacons are installed. Beacon locations are determined by GPS.

USBL use a single beacon with a single receiver on the hull of the surface ship. The receiver can get its position through the phase differences from the beacon, from which an AUV can get it position relative to the ship.

ADCP uses Doppler shift, which measures the change in pitch of the returned echo from scattering by particles in the current. This allows the system to measure water currents in the water column. DVL can work alongside the ADCP and can log the velocity of the vessel through data from the ADCP. This as an inertial system will track the position of the vessel.

Data collection:

Sidescan sonar (SSS), previously been utilized with the use of  an underwater towfish by a ship, can be placed directly on the AUV. These collect data line by line to produce an acoustic image of the seabed. As the AUV is a more stable platform aids in the progress and use of Synthetic aperture sonar as an improved option to SSS.

Multibeam echo sounder as an extension of the single beam which directs a single beam vertically and receives its response. The multibeam transmits a fan of beams. The high resolution systems would be used for local surveys and suitable because of their small size.


SLAM as mentioned is a concept from above ground robotics. When applied in AUVs these processes can increase autonomy from smarter systems. This can make path planning a more dynamic and responsive and involves partial data processing to be fed back to the vehicle. Two SLAM processes are feature based and view based where an autonomous vehicle can get its location relative to a number of recognised features or takes a whole sweep view and compare to previous views to get an updated location. This is a recent concept applied to AUVs to increase autonomy.

Major models and companies:


ASV Ltd.:

Liquid Robotics (Wave Glider):

Hugin(Konsberg maritime):


Lurton, X 2008, An Introduction to Underwater Acoustics, Springer

Paull, L., Saeedi, S., Seto, M., & Li, H. (2014). AUV navigation and localization: A review. IEEE Journal of Oceanic Engineering, 39(1), 131-149.

E. Manley, “Unmanned surface vehicles, 15 years of development,” OCEANS 2008, Quebec City, QC, 2008, pp. 1-4.

US navy Master plan :

Navigation and uncertainty

Navigation for AUV is a major component that has been in continual development. It has been broken down to three types, these being dead reckoning and inertial navigation system (INS), acoustic navigation and geophysical navigation. These different models and methods can then be integrated in different ways to increase the robustness of navigation systems, similar to how multiple methods and measurements are integrated for increased redundancy for traditional  positioning. Given the dynamic nature of positioning for navigation, models are developed with computing algorithms to generate a prediction for the vessel’s position.

Dead reckoning is where with the knowledge of velocity and direction, the new position of the vessel can be deduced from its original position. This will require vessel speed sensors as well as water current velocity information and compass. Hence, ADCP, DVL and IMU are used as part of this method. Errors in this method are cumulative and significant drift can happen over a period of time.

Acoustic navigation are LBL and USBL systems where errors can mainly arise from the assumed beacon geometry and their positions and also from the assumed sound speed profiles which will affect the distance measurements derived from the pings received from the beacons.

Paull, Saeedi, Seto, Li 2014. USBL, LBL systems

Geophysical navigation is identifying features from the sensor data and matching to bottom features in an available initial map or model of the area. 3D forward scanning would be used for these sensors. Therefore, the challenge is in obtaining an accurate and valid initial map as well as having appropriate environmental features. This leads onto algorithms used to match parameters from sensor data and SLAM algorithms.

A combination of these are used for AUVs, some as additional aids. Dead reckoning and INS are more suited to short lines or survey paths with more turns as errors are cumulative relative to the body. Surfacing can give the vessel a fixed position to correct errors or when in deeper conditions LBL and USBL can provide aids. Vertical uncertainty can increase in depth as measurements depend on the climatology model for average temperature as well as a model for water density.

Integrating the different methods and their errors requires uncertainty modelling. Commonly, a Kalman filter is applied which is used for prediction states that was developed in 1960. This algorithm is applied using recursive equations on time updates and measurement updates to have covariances converge or be reduced to an appropriate tolerance. This applied in real-time will only use historical data. Post processing can be done for better results, mostly for vertical uncertainty. There are variations of the algorithm, such as the extended kalman filter, which is used in many SLAM systems. There are other techniques for state estimation including the particle filter which is a stochastic method with random sampling.

Welch,Bishop 1995. Operation of Kalman filter

The Centre for Coastal and Ocean Mapping with the University of New Hampshire has done AUV uncertainty modelling using NOAA vessels. Also, NTNU have published work on particle filters for robust navigation.

Further reading:
Welch, G., & Bishop, G. (1995). An introduction to the Kalman filter.

Leonard, J. J., Bennett, A. A., Smith, C. M., Jacob, H., & Feder, S. (1998). Autonomous Underwater Vehicle Navigation. In MIT Marine Robotics Laboratory Technical Memorandum.

J. S. Byrne and Schmidt, V. E., “Uncertainty Modeling for AUV Acquired Bathymetry”, U.S. Hydrographic Conference (US HYDRO). Gaylord Hotel, National Harbor, Maryland U.S.A., 2015.

Zhao, B., Skjetne, R., Blanke, M., & Dukan, F. (2014). Particle filter for fault diagnosis and robust navigation of underwater robot. IEEE Transactions on Control Systems Technology, 22(6), 2399-2407.


There is one of multiple a research institutions working alongside an Applied Underwater Robotics lab (AUR-lab) for involvement in various projects linking to industries and to further the use and knowledge of AUVs and USVs. The Norwegian University of Science and Technology(NTNU) has their Centre for Autonomous Marine Operations and Systems(AMOS). NTNU-AMOS is a 10 year research program starting in 2013 to 2022 to address the challenges autonomous marine operations and systems. They currently have two research areas that their work can come under. One is unmanned vehicles and robots and the other is safer, smarter and greener marine operations. A link to their webpage on their project with autonomous underwater robotics is  here. I will outline a couple projects and research work relating to AUVs and USVs.

To begin with, the AUR-lab was created in 2009 and work in collaboration with other research groups. An recent example is the Blue Mining project which is a European project where they have so far been mapping deposits of extinct seafloor massive sulphides. While the details of mining is of less interest here the methodologies involved the use of an AUV that will identify areas that will be later studied in higher resolution by an ROV. This is an example of where these platforms are better suited to particular functions directly relational to the level of control or autonomy of the vehicle. The whole project can be followed through this link.

(BlueMining 2016)

Further challenges explored at NTNU-AMOS includes increasing the level of autonomy, multi-vehicle systems and path planning. The multi-vehicle projects have focussed on the communication between AUV and USV or UAV as an aide for navigation and guidance, and this integration of technologies. The research looked to find the effects of the roll and pitch of the UAV and the effects of motions on the AUV due to ocean waves as well as the communication quality. Communication architecture requires investigation as it can prove to have bottlenecks effecting the system efficiency and effective operation. Communication links in this project was through wifi modems. The network scheme is in the image below. Bottlenecks could be found through the capacity of the system.

(Johansen et al. 2014)

A project involving the tracking of an AUV by USV was done in a controlled location in a fjord which was done in collaboration with Maritime Robotics developing the USV in 2015. Acoustic information would be communicated from the AUV to the USV and a tracking algorithm would use this to predict the position of the AUV. The algorithm is based on constant bearing by intercepting a line of sight with a velocity vector. The AUV drifted from predicted positions and could surface to apply corrections. Operators can monitor the vehicles on shore leading to increased autonomy of systems.

Project papers:
T. A. Johansen, A. Zolich, T. Hansen and A. J. S⊘rensen, “Unmanned aerial vehicle as communication relay for autonomous underwater vehicle — Field tests,” 2014 IEEE Globecom Workshops (GC Wkshps), Austin, TX, 2014, pp. 1469-1474.

P. Norgren, M. Ludvigsen, T. Ingebretsen and V. E. Hovstein, “Tracking and remote monitoring of an autonomous underwater vehicle using an unmanned surface vehicle in the Trondheim fjord,” OCEANS 2015 – MTS/IEEE Washington, Washington, DC, 2015, pp. 1-6.


This is the post excerpt.

The area of unmanned systems for surveying and hydrographic purposes is one that is growing and developing with space for innovative ideas. Therefore, this can be followed with much interest and how they can be most valuable and effective is still being explored. This page will focus on Autonomous underwater vehicles (AUVs) through a range of posts.
As this is quite a recent area of development there are many new terms used that would be unfamiliar to most. Some acronyms used when discussing this topic relating to the vehicle are:

  • AUV – autonomous underwater vehicle;
  • ASV – autonomous surface vehicle;
  • UUV – unmanned underwater vehicle;
  • ROV – remotely operated vehicle;
  • UAV – unmanned aerial vehicle;

There are differences between these while UUV is a more general term. Within these terms there are many different designs and sizes and more being developed. They are then integrated with sonars for bathymetry information and high resolution mapping and imagery. There can also be integrated systems, running a number of vehicles cooperatively.
Some further acronyms include:

  • IMU – inertial measurement unit;
  • SLAM – simultaneous localisation and mapping;

where IMU is a sensor unit on the vehicle used to get its orientation and direction, and SLAM is a more computer based process originating from above ground robotics.

My interest in autonomous vehicles grows from being exposed to a scanning survey vehicle in the open cut mining industry. Similar aspects then extends to unmanned underwater vehicles and the development of these types of machines. Development is growing with the industries most vested being defence and military, oil and gas, as well as deep water research. Therefore, it is a booming sector with commercial interests noted by magazine Hydro International and having released three special issues on unmanned systems over three years. It is seen as of particular value as these vessels can access areas unreachable by people or are deemed unsafe and so would be important to many industries.

A further challenge of the development is that autonomous mapping can be seen as a broad area that is interdisciplinary. Expertise in computer science, algorithms, robotics, sensors, positioning is required for these machines. Some issues that can be explored include navigation challenges, underwater communication, autonomous mapping, detection and classification, as well as system and data interoperability.

The potential of this sector may yet to be identified as the new technology is developed and experimented with in different applications and industries. This requires effective collaboration and  is why this is an interesting topic as it is unknown where the sector will go and how far certain technologies will be taken. There will also be discussion whether the use is cost-effective and therefore financially worthwhile for the data and outputs.