Limit cycle attractor
simulation :
- Implementation using Coupled
Map Lattice (CML) :
Shown on the right plot the 3D projection of the 11-D attractor
of the transmitter state. Both authorized and unauthorized receiver can
not observe the transmitter state and need to reconstruct the attractor in
a reconstructed embedding space using delays of the coupling signal y(n)
which is transmitted from the transmitter to the receiver (Left figure).
- 3D projection of the
transmitter 11D attractor
:
The system trajectory converges to an attractor which takes the
form of a limit cycle :
- Changing the receiver dynamics
changes the position of the attractor :
In order to enhance security of DDE the dynamics of the receiver
is altered, preferably at the beginning of each transmitted bit. Changing
the transmitter dynamics results in a change in position of the attractor,
and prevents an attempt to reconstruct the position of the attractor by
monitoring a large number of trajectories. The attractor shown in
the above figure has different dynamics and therefore different position
than the previous attractor
.
- Changing the receiver dynamics
changes the position of the attractor :
At the beginning of each transmitted bit, the transmitter state
is set to a random initial value. Shown below is the superposition of multiple
trajectories that starts at a random initial state and converge to the
system attractor
. The trajectories of the same transmitted bit converge
to the sytsem's attractor
, however the transient depends on the initial state.
- Reconstructed embedding state
space of the transmitter :
Since neither authorized nor unauthorized receiver can observe the
transmitter secret state, both need to reconstruct the converging trajectories
and the attractor in a reconstructed embedding space. Shown in the above
figure is a reconstruction of the
attractor
in a embedding space, E[n], created using
delays of the coupling signal y : E[n] =
[ y[n], y[n-1], y[n-2] ] .