Testbed design and localization in MiNT-2: A miniaturized robotic platform for wireless protocol development and emulation

A significant challenge in the development of robust wireless networking protocols is often the need to prototype and test these protocols in a small-scale setting before they can be widely deployed. Two contrasting prototyping and testing methods are currently used, requiring a choice between convenience and accuracy. The first involves simulating a wireless network solely in software. Although convenient, this option fails to accurately account for real-world factors such as realistic radio propagation models and their interaction with node mobility and obstacles. The second relies on setting up a large-scale physical testbed that, although accurate, represents a single design point and tends to be expensive to reconfigure and manage. The MiNT project at Stony Brook University was one of the first to propose an accurate and inexpensive small-scale physical testbed built using commercially-available robots coupled with a version of NS2 built to work cooperatively on multiple nodes. MiNT combines the best features of the two popular performance evaluation methods, achieving network accuracy comparable to that of large-scale physical testbeds without abandoning the convenience and flexibility of software simulation. In this paper, we describe our initial experiences in developing MiNT-2, the next generation of MiNT. MiNT-2 represents a fresh redesign of MiNT that at once simplifies and improves the original design, and extends it with a range of new features. The paper describes a number of these improvements including a new, simplified, node design, an improved node localization using RFIDs, node position calibration, and automated layout configuration. We also demonstrate the accuracy of the new localization approach and outline planned testbed improvements.