Reliable Shortest Paths in Wireless Sensor Networks: Refocusing on Link Failure Scenarios from Applications

Abstract

Mission-critical applications of wireless sensor networks (WSNs) require collecting all data from sensors without any loss. Existing hop-by-hop and end-to-end retransmissions still face challenges in data collection reliably over shortest-paths in WSNs due to unreliable links and resource-constraints (energy, bandwidth). Such paths easily break when operating environments are harsh and vary from time to time and from location to location. In this paper, we propose rSP, an approach to preserve reliable shortest-paths in a WSN considering those environments. We assume that link failures are stochastic and independent. We use an algorithm to calculate the steady-state unreliability and availability of links over shortest-paths in order to optimize the extra energy consumption for a shortest-path failure. We then propose an algorithm to find local routing path reliability (LRPR) from each sensor to its upstream sensors to preserve a shortest-path reliable. If some links around some locations at some point of time appear more vulnerable than other links, the best reliable link for that time is chosen. Simulation results based on empirical dataset show that rSP improves the reliability over 70% and the energy-efficiency in WSNs by 50% compared to well-known approaches.