Rethinking Connectivity Restoration in WSNs Using Feedback From a Low-cost Mobile Sensor Network Testbed

In recent years, a lot of studies proposed to introduce mobility capability to sensor nodes in order to exploit the advantages of mobility in Wireless Sensor Networks (WSNs) applications. In particular, the idea has been exploited heavily to restore connectivity in disjoint WSNs. While the studies demonstrated various capabilities of the proposed connectivity algorithms via simulation, real node and testbed implementations were mostly lacking due to unavailability of proper hardware and/or high robot costs. Since this may hinder the direct applicability of the algorithms in realistic settings, testbeds which can be constructed with low-cost and commercial-off-the-shelf (COTS) hardware are required for realistic evaluations of the connectivity restoration algorithms. To this end, in this paper, we first design a low-cost mobile sensor network (MSN) testbed using iRobot-based mobile nodes. We integrate commercially available iRobot Create platform with IRIS sensor motes to design a new mobile sensor node called iRobotSense. Using a network of iRobotSense nodes, we then implement one of the widely used connectivity restoration algorithms, namely PADRA, on this testbed. The evaluations revealed several factors that need to be taken into consideration in the design and development of connectivity restoration algorithms which were not considered in simulation-based evaluations. We revise several components of PADRA based on the feedback received from the testbed experiments.