Aerospace and Electronic Systems Magazine February 2018 - 24

Feature Article:

DOI. No. 10.1109/MAES.2018.170048

MH370 Burst Frequency Offset Analysis and
Implications on Descent Rate at End of Flight
Ian D. Holland, Defence Science and Technology Group, Edinburgh, Australia

INTRODUCTION
On 7 March 2014 at 16:41:43Z,1 Malaysian Airlines flight MH370
departed Kuala Lumpur (KL) International Airport bound for
Beijing. Less than an hour later, following the last recorded radio transmission from MH370 at 17:19:30Z, the plane's secondary
radar transponder went offline. As evidenced by Malaysian military radar, the plane (registration number 9M-MRO) then veered
off course unexpectedly, backtracked across the Malaysian Peninsula, and was tracked heading northwest from Penang through
the Malacca Straits. After disappearing from radar at 18:22:12Z,
it reestablished a satellite communications (SATCOM) link with
the Inmarsat satellite I-3F1 at 18:25:27Z. By analyzing a series of
automated messages exchanged via that satellite between the plane
and an Inmarsat ground station in Perth, Australia, it was determined that the plane continued to fly for 6 h before finally ceasing message exchange with the ground station at 00:19:37Z on 8
March 2014. This article discusses specifically the analysis of burst
frequency offset (BFO) metadata from the SATCOM messages. It
is shown that the BFOs corresponding to the last two SATCOM
messages from the plane at 00:19:29Z and 00:19:37Z suggest that
flight MH370 was rapidly descending and accelerating downward
when message exchange with the ground station ceased.
An initial analysis by Inmarsat of the SATCOM metadata for
MH370 in the last 6 h of flight suggested that MH370 had flown
into the Southern Indian Ocean before SATCOM was lost (see [1]
for further details). As summarized in [2], an intensive aerial and
surface search was undertaken in the Southern Indian Ocean by an
international search team during March and April 2014, with no
MH370-related debris found. On 28 April 2014, the aerial search
concluded, and the search transitioned to an underwater phase [2].
The Australian Transport Safety Bureau (ATSB) took responsibility for the definition of the underwater search zone. It convened an
international flight path prediction working group, bringing togeth1

All times given in this article are in UTC or Zulu time, denoted by Z.

Author's current address: Defence Science and Technology
Group, Cyber and Electronic Warfare Division, P.O. Box 1500,
Edinburgh, SA 5111, Australia, E-mail: (ian.holland2@dst.
defence.gov.au).
Manuscript received February 6, 2017, revised December 14,
2017, and ready for publication December 18, 2017.
Review handled by P. Willett.
0885/8985/17/$26.00 © Commonwealth of Australia 2018
24

er experts in SATCOM, statistical data processing, and aviation, to
estimate the most likely final location of flight MH370. The group
consisted of representatives from the Australian Defence Science
and Technology (DST) Group and the other organizations listed in
the Acknowledgments section of this article.
New methods of analyzing the Inmarsat data were developed
by the group, resulting in the release of reports concerning the
likely final location of flight MH370 from the ATSB in August
2014 [3], October 2014 [4], and December 2015 [5]. Inmarsat also
published an article regarding its contribution to the flight path reconstruction effort [1]. The DST Group contribution that assisted
in the definition of an extended priority search area in December
2015 [5] has been detailed in [2]. This demonstrated how Bayesian
analysis was used to identify a high-probability region of where
the plane was believed to be at the time of last SATCOM transmission (00:19:37Z, 8 March 2014). The DST Group Bayesian method used a prior probability distribution defined by the Malaysian
military radar, a likelihood function describing the relationship between SATCOM measurements and aircraft position and velocity
during the flight, and a model of the aircraft dynamics.
This article does not cover the Bayesian method used in defining the underwater search area. For details on that method, refer to
[2]. Instead, this article focuses on the following:
1. A review of the statistical analysis of BFOs for several previous flights of 9M-MRO
2. Examination of the effects of the plane's vertical velocity on
the BFO
3. An analysis of the effect of the track angle of the plane on the
BFO toward the end of flight
4. The establishment of a BFO trend throughout the last 6 h of
flight so as to determine an expected BFO at end of flight
5. An analysis of the behavior of the frequency oscillator in the
plane's satellite data unit (SDU) after power outage events,
such as those believed to have occurred twice during flight
MH370, and how this affects the BFO
6. An analysis that shows that the final two BFOs are consistent
with MH370 being in a rapid descent and accelerating downward
This article first presents a timeline of key events during the
MH370 accident flight. Then a review of the SATCOM model is
provided, along with a review of the BFO statistics. This serves as a

IEEE A&E SYSTEMS MAGAZINE

FEBRUARY 2018

Table of Contents for the Digital Edition of Aerospace and Electronic Systems Magazine February 2018