Wireless sensor networks (WSN), which consist of numerous energy-constrained sensor nodes, are mostly duty-cycled and rely on multi-hop routing to collect data. Data collection mechanism is thus a building block of WSN systems. The performance of low duty cycle WSN nowadays, however, are far restricted by the limited understanding and the underachieved design of the forwarding mechanisms. This dissertation is based on real-world WSN measurements and mainly addresses three key issues that emerge from real applications, namely bursty-loss aware lazy forwarding, duplicate detectable opportunistic forwarding and end-to-end delay measurement, modeling and optimization. We develop theoretical principles and practical approaches to address the above issues. Through extensive real-world...[ Read more ]

Wireless sensor networks (WSN), which consist of numerous energy-constrained sensor nodes, are mostly duty-cycled and rely on multi-hop routing to collect data. Data collection mechanism is thus a building block of WSN systems. The performance of low duty cycle WSN nowadays, however, are far restricted by the limited understanding and the underachieved design of the forwarding mechanisms. This dissertation is based on real-world WSN measurements and mainly addresses three key issues that emerge from real applications, namely bursty-loss aware lazy forwarding, duplicate detectable opportunistic forwarding and end-to-end delay measurement, modeling and optimization. We develop theoretical principles and practical approaches to address the above issues. Through extensive real-world implementations and trace-driven experiments, we evaluate those proposed methods and verify the efficiency.