Mobility-Aware MAC Protocol for Wireless Sensor Network

Mobility-Aware MAC Protocol for Wireless Sensor Network Tang Zhiyong and Waltenegus Dargie TU Dresden Chair of Computer Networks

Outline • MAC protocols in wireless sensor networks • Mobility • MA-MAC • Evaluation • Future work

MAC Protocols • Most contention based MAC protocols in wireless sensor networks define duty cycle to avoid idle listening and overhearing (thereby utilising energy efficiently) • In SMAC nodes synchronise sleeping schedule at the beginning of a listen period in order to be able to communicate with each other SYNC Contention Listen Sleep

MAC Protocols • The problem with SMAC is the cost of synchronisation • Preamble based MAC protocols (such as BMAC) avoid the need for synchronisation, but the cost of preamble itself is high in some situations Preamble Data RX Sleep Listen Preamble Data TX Listen

MAC Protocols • XMAC minimises the cost of preamble by dividing the preamble into several short Strobes • Preamble transmission ceases when a recipient sends acknowledgement Data Sleep Listen RX Data TX Listen

Mobility • Whereas SYNC packets in SMAC-like protocols enable dynamic neighbour discovery, the cost of SYNC packets is often considered to be inefficient. • On the other hand, the problem with preamble-based protocols is their basic assumption that the receiver is already known to the transmitter • However, this can be solved by sacrificing a small cost of neighbour discovery beacons at the beginning of a listen period

Mobility • More importantly, neither schedule-based nor preamble-based protocols support dynamic handover of communication • If a communication link is broken in the middle of a transmission, the transmitting node should wait until the next listen phase • This introduces a delay in message transfer

MA-MAC • Conceptually, when a transmitting node senses that either the receiver or itself senses mobility one of the following takes place: • The node completes transmitting all the data before a link breaks, hence, there is no need to deal with mobility. • The node negotiates for dynamic rate adaptation, so that it can send all the data at a higher rate and complete transmission before the link breaks. • The node initiates and completes a handover before the link breaks.

Handover Mobile node Relay node Relay node Relay node Relay station Base station

Handover Mobile node Relay node th2 th1 Relay node Relay node Relay station Base station

Handover • th1 and th2 can be determined by evaluating the RSSI values of the two relaying nodes • However, these values are not reliable • Moreover, the method requires repeated measurements and comparisons, which will reduce the data transmission rate • At present th1 and th2 are determined on the average packet loss rate set by the application (itself a function of d Mobile node Relay node th2 th1 Relay node Relay node Relay station Base station

State Transition Handover not successful? Sleep.

Modification of XMAC • MA-MAC interleaves data transfer with handover request when it senses an imminent link breakage.

MA-MAC’s Implementation

Experiment • The aim of the experiment was to analyse the performance of MA-MAC in terms of • Handover threshold (20 vs. 40 m) • Sending interval (30 vs. 250 ms) • Speed of mobility (slow vs. fast speed)

Experimental Setting 20 – 40 m 20 m

Latency

Packet Loss

Observations • Latency is much affected by the handover threshold, increases as the threshold distance increases • Packet loss is mainly affected by the sending interval • Estimating mobility has been our serious challenge (unreliable RSSI; advanced filters increase delay) • Relaying nodes should be awake to participate in a handover request; this either increases the energy cost or requires dense deployment

Mobility-Aware MAC Protocol for Wireless Sensor Network