36
A. M. Said et al.
radio transmission TX energy, radio reception RX energy and radio interfered
INT energy. In order to calculate this metrics we used the formula [31] (Eq. 1,
Table 2) as follow :
Energy(mJ) = (transmit ∗ 19.5mA + listen ∗ 21.8mA
+ CP U ∗ 1.8mA + LP M ∗ 0.0545mA)
∗ 3V /4096 ∗ 8
(1)
P ower(mW ) =
Eneregy(mJ)
T ime(s)
Table 2. Equation parameters description
Variables Meaning
LPM
Power consumption parameter that indicates the power used when in sleep condition
CPU
Power parameter that indicates the level of node processing
Transmit Parameter related with node communication while transmitting
Listen
Parameter related with node communication while receiving
5.3 Power Tracking per Mote for Each Simulation
We used data containing 1000 instances of consumed energy values for each node
in the network. Figure 7 depicts the evolution of power tracking of each node in
the four scenarios:
• scenario 1: when we have a normal behavior in the network, all the sensors
show a regular energy consumption in terms of receiving (node 0) and sending
(nodes from 1 to 10). We use this simulation to collect the training data for
the proposed IDS.
• scenario 2, 3 and 4: for those scenarios, we have a high sending values for the
malicious motes. This is explained by the fact that when a malicious mote
joins the network, it asks the other motes to recreate the DODAG tree and
also to send data that they have, in order to steal as much data as it can.
That is why it have a high receiving values too. The other motes do not
distinguish that this is a malicious mote, therefore they recreate the DODAG
tree, and send their information through the malicious node. We used the first
simulation scenario as dataset for our IDS, describing the normal behavior
of the network. This 1 h information was enough to detect the malicious
activities of the rank attack. Meanwhile, each time we add a malicious mote,
the anomaly detection rate increases as shown in Fig. 8.
In each simulation of malicious mote, the proposed IDS indicates the anomaly
detection ratio which increases each time while adding another malicious mote.
This aims to determine the impact of the number malicious motes compared to
normal behavior of the system.
A. M. Said et al.
radio transmission TX energy, radio reception RX energy and radio interfered
INT energy. In order to calculate this metrics we used the formula [31] (Eq. 1,
Table 2) as follow :
Energy(mJ) = (transmit ∗ 19.5mA + listen ∗ 21.8mA
+ CP U ∗ 1.8mA + LP M ∗ 0.0545mA)
∗ 3V /4096 ∗ 8
(1)
P ower(mW ) =
Eneregy(mJ)
T ime(s)
Table 2. Equation parameters description
Variables Meaning
LPM
Power consumption parameter that indicates the power used when in sleep condition
CPU
Power parameter that indicates the level of node processing
Transmit Parameter related with node communication while transmitting
Listen
Parameter related with node communication while receiving
5.3 Power Tracking per Mote for Each Simulation
We used data containing 1000 instances of consumed energy values for each node
in the network. Figure 7 depicts the evolution of power tracking of each node in
the four scenarios:
• scenario 1: when we have a normal behavior in the network, all the sensors
show a regular energy consumption in terms of receiving (node 0) and sending
(nodes from 1 to 10). We use this simulation to collect the training data for
the proposed IDS.
• scenario 2, 3 and 4: for those scenarios, we have a high sending values for the
malicious motes. This is explained by the fact that when a malicious mote
joins the network, it asks the other motes to recreate the DODAG tree and
also to send data that they have, in order to steal as much data as it can.
That is why it have a high receiving values too. The other motes do not
distinguish that this is a malicious mote, therefore they recreate the DODAG
tree, and send their information through the malicious node. We used the first
simulation scenario as dataset for our IDS, describing the normal behavior
of the network. This 1 h information was enough to detect the malicious
activities of the rank attack. Meanwhile, each time we add a malicious mote,
the anomaly detection rate increases as shown in Fig. 8.
In each simulation of malicious mote, the proposed IDS indicates the anomaly
detection ratio which increases each time while adding another malicious mote.
This aims to determine the impact of the number malicious motes compared to
normal behavior of the system.