Research On Reliability Assurance Approaches In Wireless Sensor Network Systems

Wireless sensor networks (WSNs) have been and will continue to be applied to a wide range of civil and military areas, ranging from environmental monitoring, traffic management, to national defense and security, for their low energy consumption and self-organized characteristics. Because of the harsh deployment environment and limited node capability, reliability has become a great challenge in WSN system design. To address this challenge, this dissertation focuses on investigating techniques to ensure the reliable operation of WSN systems at both networking and system levels.At the networking level, firstly, both transmission reliability assurance scheme and application data traffic variation will cause the workload of node to change frequently, which results in heterogeneous convergecast. To alleviate the problem, MAC protocol shall adjust its operational schedule according to the nodes' workload. In this dissertation, MAC scheme supporting cross-layer schedule optimization was studied, and W-MAC, a workload-aware MAC protocol for heterogeneous convergecast was proposed. Collision-free time slice scheduling mechanism based on sub-tree workload collection, and low-overhead high-speed scheduling adjustment mechanism were designed to minimize the power consumption of nodes, and achieve ultra-low data transmission latency in heterogeneous convergecast. After investigating the MAC scheduling, network organization and data transmission reliability assurance schemes were studied. On one hand, existing methods repair network organization errors caused by node or communication failures at the cost of high energy consumption. To address this issue, E-NORS, an energy-efficient network organization repair scheme based on adaptive working cycle, was proposed. E-NORS achieves a good balance between power consumption and repair speed by dynamically adjusting the duty cycle in control packet listening. On the other hand, existing FEC schemes can not ensure the end-to-end data transmission reliability. To achieve reliable transmission, RAH, a data transmission reliability assurance scheme based on hop-by-hop adaptive FEC, was proposed. RAH distributes the end-to-end data transmission reliability requirement to each hop on the data delivery path, and lets each node on the path to optimize its FEC rate by employing a FEC rate adjustment mechanism based on two-level PID control. Therefore, the end-to-end data transmission reliability is ensured, and the overhead of redundant data transmission is minimized.At the system level, in order to reduce the chance of system failures and to restore the system after a failure occurs, many human-node interaction operations are required throughout the full life cycle of a WSN system, which includes testing, deployment and maintenance. Different operations require the node to behave differently, and part of the requirements even conflict with each other. To solve this issue, the conception of working mode, which provides an abstract on the behavior patterns of human-node interaction operations in the life cycle of a WSN system, was introduced. Adopting the conception of working mode, MEAN, a WSN middleware that groups application programs based on behavior patterns, was designed. The proposed middleware manages the running and suspending of application programs in different working modes, and controls the behavior of a node, to meet the requirements of human-node interaction and data collection operations. At the same time, the application program organization and switching strategies, services and modules provided by the middleware will greatly increase the development efficiency and the extensibility of node software.At last, by applying the proposed techniques, including the network organization reliability assurance scheme and the middleware, a wireless sensor system for long-term microclimate monitoring at Mogao Grottoes was built. To meet the requirements of heritage monitoring, a tiered communication architecture and specific hardware and software were developed. The deployed system has been running for more than 2 years, which verifies the effectiveness of the design of the system and proposed techniques.Future directions include: optimizing dynamic working cycle adjustment mechanism in E-NORS; optimizing the data transmission reliability requirement distribution mechanism in RAH; taking routing into consideration to prolong the network lifetime while ensuring data transmission reliability; and evaluating the proposed schemes further in applications.