You could look at the different attacks in WSN/IoT listed in https://tetcos.com/file-exchange.html. There is some content in https://tetcos.com/machine-learning-netsim.html on detecting these attacks using machine learning. You could probably adapt one of them to a Poisson distribution.

Dear Mehdi Najafi,

Information below should be helpful:

The Poisson distribution is quite useful in detecting attacks in wireless sensor networks (WSNs) due to its ability to model the occurrence of rare events over a fixed period or space. Here are a few ways it can be applied:

Event Detection: In WSNs, the Poisson distribution can model the arrival of packets or events. If the rate of packet arrival deviates significantly from the expected Poisson distribution, it may indicate an attack, such as a Denial of Service (DoS) attack.

Node Deployment: The spatial distribution of sensor nodes can be modeled using a Poisson point process. This helps in evaluating the coverage and connectivity of the network, which is crucial for detecting and mitigating attacks.

Intrusion Detection: By analyzing the frequency and pattern of network traffic, the Poisson distribution can help identify anomalies that suggest malicious activities. For example, a sudden spike in traffic could indicate a jamming attack.

Trust Models: Poisson distribution-based trust models can be used to assess the reliability of nodes in the network. Nodes that deviate from the expected behavior can be flagged as potentially compromised.

https://link.springer.com/article/10.1007/s11761-024-00396-w

https://www.mdpi.com/2079-9292/11/20/3367