Recharging vs. Replacing Sensor Nodes Using Mobile Robots for Network Maintenance

International audienceWireless sensor networks (WSNs) have been of very high interest for the researchcommunity since years, but the quest for deploying a self-sustained network andeffectively prolonging its lifetime has not found a satisfactory answer yet. Two mainapproaches can be identified that target this objective: either ``recharging'' or``replacing'' the sensor nodes that are running out of energy. Of particular interest aresolutions where mobile robots are used to execute the above mentioned tasks toautomatically and autonomously maintain the WSN, thus reducing human intervention.Recently, the progress in wireless power transfer techniques has boosted researchactivities in the direction of battery recharging, with high expectations for itsapplication to WSNs.Similarly, also sensor replacement techniques have been widely studied as a means toprovide service continuity in the network.Objective of this paper is to investigate the limitations and the advantages of these tworesearch directions. Key decision points must be identified for effectively supportingWSN self-maintenance: (i) which sensor nodes have to be recharged/replaced; (ii) inwhich order the mobile robot is serving (i.e., recharging/replacing) the nodes and byfollowing which path; (iii) how much energy is delivered to a sensor when recharged.The influence that a set of parameters, relative to both the sensors and the mobilerobot, on the decisions will be considered. Centralized and distributed solutions arecompared in terms of effectiveness in prolonging the network lifetime and in allowingnetwork self-sustainability. The performance evaluation in a variety of scenarios andnetwork settings offers the opportunity to draw conclusions and to discuss theboundaries for one technique being preferable to the other