Distributed partition detection and recovery using UAV in wireless sensor and actor networks

Abstract EN

Wireless Sensor and Actor Networks (WSANs) have been extensively employed in various domains ranging from elementary data collection to real-time control and monitoring for critical applications.

network connectivity is a vital robustness measure for overall network performance.

different network functions such as routing, scheduling, and QoS provisioning depends on network connectivity.

the failure of articulation points in the network disassociates the network into disjoint segments.

we proposed distributed partition Detection and recovery using unmanned aerial vehicle (UAV) (DPDRU) algorithm, as an optimal solution to recover the partitioned network.

It consists of three steps : Initialization, operational and detection, and recovery.

in the Initialization phase sink node collects all the information about the network.

in the operational and detection phase, network nodes communicate regularly by exchanging HEARTBEATS, detects failure if some nodes do not get a message from the neighbor node and send failure reports, and sink node identifies network partition.

in the recovery phase, the sink node sends UAV at the positional coordinates of the failed node and examines network recovery after UAV reaches the desired location.

our approach primarily focuses on reducing message overhead by sending few update messages to sink node and energy consumption by engaging network nodes only for communication.

the requirements of the recovery process (physical movement and communication) are fulfilled by UAV.

the algorithm is tested according to the following parameters : detection time, recovery time, message overhead, and distance traveled by UAV.

simulation results validate the efficacy of the proposed algorithm based on these parameters to provide reliable results.

the minimum and the maximum number of messages transmitted are 11 for 10 nodes and 24 for 100 nodes respectively.

hence these results demonstrate that the message overhead in our proposed solution is less as compared to other techniques when the number of nodes increases.