Implementation of a wireless medium access control protocol for cooperating objects

Cooperating Objects (COs) are mobile and sentient entities that react the real time sensed data coming from large number of sensors embedded into the environment as well as the requests coming from the other COs in the vicinity. Therefore, the delay in conveying the data is an important quality of service (QoS) requirement. Moreover, COs generally run unattended and autonomously for extended period of times which makes the energy consumption another important QoS parameter. To fulfill the requirements related to these parameters is a challenging task especially when COs are mobile. We introduce power controlled and loosely synchronized (PCLS) MAC protocol for COs. PCLS protocol employs loosely synchronized listen-sleep schedules. Therefore the objects do not need to learn the schedules of their neighbors, and shorter preambles, which lower the protocol overhead, become possible. In addition to loose synchronization, PCLS protocol employs open-loop transmission power control. This power control scheme makes use of the periodic HELO messages to determine the transmission power level that satisfies the network connectivity. It does not depend on the neighbor IDs. These features of PCLS protocol reduce the energy consumption and the link layer delay, and also make PCLS resilient against topology changes due to node failures and mobility. Effects of the protocol parameters on the energy consumption and delay are studied analytically, and verified through simulations. Performance evaluation is performed in OMNet++ simulation environment based on the Tactical Underwater Surveillance System in which the objects are assumed to move according to the reference point group mobility (RPGM) model. PCLS protocol is also compared with B-MAC and LMAC protocols. Simulation results reveal better performance than both protocols for energy consumption as well as delay. Keywords: MAC protocols, cooperating objects, underwater surveillance systems, mobility models, OMNet++.