we use helpers working in UCJ mode to transmit the artificial noise. To make a full use of
the CSIs of the transmitter and helpers, we propose an antennas allocation scheme to
allocate the antennas of them. In PHY security, the Channel State Information of Eaves-
dropper (ECSI) is vital. If the ECSI is known, the transmitter or helpers can target the
eavesdropper and beamform the artificial noise to enhance the interference intensity.
Meanwhile, without the precise location but a range of area where the eavesdropper may
be, i.e., imperfect ECSI, the transmitter can also get a satisfying secrecy capacity [8].
Generally, ECSI is completely unknown to the transmitter or helpers. In this situation, the
eavesdropper cannot be targeted and the artificial noise is broadcasted, which degrades the
secrecy capacity. Enhancing security performance without ECSI is still a difficulty in PHY
security. In this paper, the proposed scheme has a chance to find out the eavesdropper and
improve the secrecy capacity of the transmission without the ECSI. The new double check
scheme can distinguish the eavesdropper among the helpers. To optimize the secrecy
capacity, we use Genetic Algorithm (GA) as our optimization strategy and set the relevant
parameter to suit our model. To the best of our knowledge, this is the first method of
dealing with the eavesdropper in MIMO-OFDM, and both the schemes, including the
antennas allocation scheme and the double check scheme, are the first of its kinds in this
type of network environment.
The following parts of this paper is organized as this: in Sect. 2, the system model of the
multiple multi-antennas helpers scheme in the MIMO-OFDM network is introduced. In
Sect. 3, the antennas allocation scheme is proposed, and the double check scheme is in
Sect. 4. We present the secrecy capacity calculation in Sect. 5, and the parameters of GA is
settled in Sect. 6. The performance evaluation is shown in Sect. 7, following the conclu-
sion of this paper in Sect. 8.
2 System Model
In this paper, we build a network system model with one transmitter, one receiver, one
eavesdropper and multiple helpers. We assume an equitable network environment with
each node has the same properties. In this network, each node has the same opportunity to
be a helper or a eavesdropper. To secure the transmission, the transmitter needs to find
multiple helpers to send the artificial noise, and prevent the eavesdropper from overhearing
the confidential signal.
In Fig. 1, the transmitter sends the confidential signal to the receiver, accompanied by
the artificial noise sent by helpers. We assume that the global CSI is known to all, so the
helpers can prevent the receiver from being interrupted by the noise they send. But they do
not know which one is the eavesdropper. So there is a chance for the transmitter to choose
the eavesdropper to protect the confidential signal. In that case, the transmitter needs at
least two helpers to protect the same subcarrier in the same time. Meanwhile, the trans-
mitter tries to find out the eavesdropper to enhance the secrecy capacity.
The number of antennas for each node is M, and the number of helpers is N. We assume
there is enough subcarriers for the transmitter to allocate its antennas separately. And the
channel between the transmitter and the receiver is H
tr
. The channels between the helpers
and the receiver are H
hri
, where i ¼ 1; 2; ...; N. The channel between the transmitter and
the eavesdropper is H
te
, which is unknown to the transmitter and helpers. The channels
between the helpers and the eavesdropper are H
hei
, which is also unknown, where
i ¼ 1; 2; ...; N.
Preventing Eavesdropper by Artificial Noise in MIMO-OFDM... 1105
123