ABSTRACT

Low rate wireless personal networks are vulnerable to several attacks focused on the Media Access Layer. Malicious or faulty nodes can subvert CSMA/CA and GTS allocation algorithms, achieving MAC unfairness in order to obtain higher medium access priority or to interrupt legitimate communication, ultimately leading to Denial-of-Service. While there exist secure protocol frameworks for LR-WPANs and diverse countermeasures against these attacks, most of them do not fit the usual computational and power supply constraints of LRWPAN nodes, which require less computationally intensive methods such as trust models. We extend a bayesian trust model designed to detect such attacks by introducing context-dependent parameters and a flexible ageing factor which enable adaptive control of the proposed trust model based on specific context attributes and on changing network conditions.