IEEE - Aerospace and Electronic Systems - April 2021 - 12

Flight Trial Demonstration of Secure GBAS via the L-band Digital Aeronautical Communications System (LDACS)
proofs to Bob (or any other party receiving the broadcast
message), that the message mj in interval i with MAC j
was actually sent by Alice, since no one else knew the key
ki at that time Ti . The content of the different packets pj
with the respective key in the respective interval contained
in the packet is again depicted in Figure 4.
For Bob to partake in the TESLA protocol, Alice and Bob
need to synchronize their clocks within a margin of acceptable
error [11]. Then, Alice sends TESLA parameters such as functions F , F 0 , and F 00 , the time interval schedule consisting of
interval duration Tint , start time T0 , index of interval i and the
length of the one-way chain, the key disclosure delay d, and a
key commitment to the key chain, allowing Bob to verify that
the received keys are actually part of the key-chain. These
parameters need to be distributed in an authenticated manner.
In the experiment, public keys and certificates of the ground
station and aircraft station were bilaterally exchange via
LDACS. In an operational deployment of LDACS, the distribution of public keys and certificates will likely be realized
via an LDACS specific public key infrastructure, as described
in [14]-[17]. Knowing Alice's public key, Bob can verify the
authenticity of the TESLA parameters and start buffering
messages sent by Alice until he receives the correct key to verify their authenticity.
Note that TESLA authentication requires the buffering of
received messages until Alice's authentication key has been
received. This introduces a key disclosure delay increasing
the communication latency between Alice and Bob.
Time synchronization between aircraft station and
ground station was implemented as described in [11]. The
exchange of TESLA parameters was signed via an Ed25519
digital signature of the ground station. Our implementation
used python3 and the nacl [18] crypto-library with F being
the SHA-512 hash function for key stream generation and
two variations F 0 , F 00 of the blake2b hash function for MAC
key derivation and MAC generation.

LIMITATIONS OF THE EXPERIMENTAL SETUP
Our experiments used a single GBAS ground receiver with
limited ground monitoring, thus without ionosphere and
ephemeris monitoring or B-value checks. However, this has
no influence on the purpose of the experiment; thus, it characterizes GBAS over LDACS similar as proposed in [8].
GBAS corrections and integrity parameters were generated and broadcast for GPS, Galileo, and GLONASS for
L1 100 s and L5 100 s processing modes. As we used the
LDACS message format, we did not transmit final
approach segment (FAS) data and did not use the VDB
message format. Transmission of all data was combined
in one correction/integrity message per epoch. Producing
corrections for all visible satellites our experimental setup
generated data rates as predicted in [8]: Approximately
3500 B/s, distributed over several communication packets
per second. To allow for better characterization of LDACS
12

Table 1.

Flight Trajectories and Experiments.
Exp.

Trajectory

Duration
(s)

Security

Parameters

01

Flight 1

762.160

TESLA

Tint ¼ 1s, d ¼ 1

02

Flight 1

3755.519

TESLA

Tint ¼ 1s, d ¼ 2

03

Flight 1

1459.296

TESLA

04

Flight 2

1213.501

05

Flight 2

2765.582

unsecured

improved msg.
format

06

Flight 2

3028.761

TESLA

Tint ¼ 1s, d ¼ 2
improved msg.
format

unsecured

Tint ¼ 1s, d ¼ 2
improved msg.
format

performance, each set of corrections (every 0.5 s) was sent
twice with the redundant second message 0.2 s delayed.
The position of the antenna on the aircraft was in an
unfavorable place between the wings under the belly of
the aircraft. This was due to the availability of port-holes
in the experimental aircraft (c.f. Figure 1) and resulted in
short outages discussed ahead.

FLIGHT TRAJECTORIES AND EXPERIMENTS
We performed six experiments in two flights. We chose two
different flight trajectories to demonstrate secure GBAS via
LDACS with different pitch and roll alignments of the aircraft-station and ground-station antennas. During the first
flight, we transmitted only TESLA-secured GBAS via
LDACS. We varied the key verification delay to compare
different TESLA parameters for GBAS performance. In the
second flight, the first two experiments were conducted
with unsecured GBAS via LDACS, whereas the last experiment used TESLA again. Flight 2 used a more efficient
GBAS message format. The experiments are summarized in
Table I. Both flights included considerable taxiing times
and preparation times on the apron not included in the table.
The first flight took place on March 26, 2019, had
takeoff at 08:53 UTC, touches down at 10:50 UTC, and
was chosen as dedicated test flight to demonstrate secure
GBAS. Its total airtime was 7000 s with a distance of
1048 km covered. Our goal was to climb, remain at a constant altitude of 6000 m for as long as possible to have different pitch and roll configurations toward the groundstation antenna at constant altitude, and then descend and
land, which we achieved as plotted in Figure 5.
The second flight trajectory was used to directly compare
TESLA-secured GBAS via LDACS and unsecured GBAS
via LDACS in several experiments during the same flight. We
chose a greater distance to our ground station, two different
flight altitudes, steeper and longer curves, and missed
approaches provoking and resulting in more antenna

IEEE A&E SYSTEMS MAGAZINE

APRIL 2021
IEEE - Aerospace and Electronic Systems - April 2021

Table of Contents for the Digital Edition of IEEE - Aerospace and Electronic Systems - April 2021
