Towards Predictable Real-Time Routing for Wireless Networked Sensing and Control

Real-time routing is a basic element of closed-loop, real-time sensing and control, but it is very challenging due to dynamic, uncertain link/path delay. A basis of real-time routing is determining probabilistically guaranteed path delays, but this problem is NP-hard and it may well have to be solved by resource-constrained devices in a distributed manner; the highly varying nature of link/path delay makes it necessary to adapt to in-situ delay conditions in real-time routing, but it has been observed that delay-based routing can lead to instability and low data delivery performance in general. To address these challenges, we propose the minimum-delay (MD) algorithm that computes probabilistic path delay bounds in pseudo-polynomial time via dynamic programming, and the algorithm is amenable to lightweight, distributed implementation. For enabling adaptivity while addressing instability in real-time routing, we propose a multi-timescale-adaptation (MTA) framework that ensures longterm optimality while addressing short-term dynamics at the same time. We evaluate the performance of MDand MTA-based real-time routing through a testbed of 130 TelosB motes, and we find it outperform existing real-time routing protocols by a significant margin, for instance, improving the packet delivery ratio and deadline catching ratio by a factor up to 6.04 and 6.67 respectively.