Increasing Connectivity in Wireless Sensor-Actuator Networks Using Dynamic Actuator Cooperation

Distributed systems based on networked sensors and actuators with embedded computation capabilities are commonly used to monitor and control the physical world. To provide a meaningful service such as disaster and emergency surveillance, meeting real-time-and-energy constraints and the stability of transmit queues are the basic requirements of communication protocols in such networks. In settings with sparse distribution of actuator nodes, multi-hop routing is traditionally used to relay information to a remote sink. A problem with this approach is that the loss of connectivity of actuator nodes may lead to partitioning of the network. In this paper, we address the problem of minimizing power consumption at each actuator node and minimizing assignment overhead at each sensor node while ensuring network connectivity. We propose that each actuator is enabled with two wireless interfaces: one to communicate with its assigned sensor network, and the other to communicate with the network of neighboring actuators. At any instant of time, an actuator can adjust its transmit power level to ensure connectivity and pro-actively inform its attached sensor network in case of mobility. These strategies have an associated cost. We show in this paper that at any instant of time, the strategy chosen by an actuator to adjust power and control-overhead due to mobility are optimal subject to constraints. The proposal is validated by means of analysis and simulations.