چكيده به لاتين
The IEEE 802.15.4 standard was published in 2003, defining a medium access control (MAC) and physical (PHY) layer for low power and low datarate communications such as WSANs. The 802.15.4 cannot provide any bound on the maximum delay experienced by data to reach the final destination and provides a very low delivery ratio and the energy consumption of this standard is high. 802.15.4e extends the previous 802.15.4 standard.
IEEE 802.15.4e standard amendment was published with three new MAC modes, the timeslotted access provides a bounded maximal delay and the channel hopping increases the reliability, Therefore, if the channel used in the current timeslot suffers from interference or multipath fading, the chance of transmission failure next timeslot is decreased since the channel will be switched. The reliability of TSCH has been reported to achieve 99.9995%.
The TSCH uses a fixed procedure to makes the hopping operation between the Channels. band. The default hopping list in TSCH contains all 16 channels and since most of them are occupied by the high power WiFi signal, reliability will go down.
Not only WiFi interference is an issue, also multipath fading can degrade reliability. Hence it is necessary for a clever decision maker to decide whether or not to hop on the other channel given the presence or absence of these interactions, to this end, a blacklist mechanism has been introduced for the adaptive hopping [5], which is a list to isolate undesirable channels. Blacklisting solutions have mainly used learning algorithms assuming the presence or absence of 802.11 interactions as stationary. That is, the statistical parameters of the probability of presence or absence of these interactions are constant and do not change with time, that is an unrealistic assumption and many of these algorithms have been able to report high packet delivery ratio with unrealistic assumption.
In our thesis, we extend the probabilistic model of the presence or absence of an interfering 802.11 to a non-stationary model, such as an environment that can be switched between probabilistic parameters as a Markovian switching.
we were able to achieve a higher packet delivery rate by about 15% compared to one using a kind of learning algorithm called MAB, using an adaptive approach of frequency channel hopping.
