Research On Key Technologies For Reprogramming Wireless Sensor Network Applications

Wireless sensor networks (WSNs) can be deployed in complex environments for large-scale, long-term, and real-time monitoring of the physical world. Therefore, they have been successfully applied in a variety of applications. In recent years, WSN applica-tions may update executable code on sensors due to changes in application require-ments. Researchers refer to this problem as reprogramming of WSN applications, and reprogramming for short. Despite its importance, implementing reprogramming is not a trivial task. Reprogramming needs to disseminate application code via wireless communications, which require a large amount of data transmission. This makes re-programming energy and time consuming, and consequently shortens the lifetime and reduces the availability of WSN applications. Therefore, reprogramming efficiency has been an important research theme in the community for years.To address this problem, existing efforts try to improve the reprogramming ef-ficiency by reducing the size of the script (i.e., data for updating executable code on sensors) and decreasing the number of packets for disseminating the script. In partic-ular, these efforts use the delta to the binary code or introduce the intermediate code to reduce the size of the script. Meanwhile, they leverage on less data relaying sensors and decrease communication collisions among these sensors to decrease the number of packets for disseminating the script. However, they may not fully control changes on executable code caused by changes in application requirements. In addition, they may not efficiently take advantage of application-specific information to further reduce the number of packets for disseminating the script.In this paper, we propose an efficient reprogramming approach for WSN appli-cations. Different from existing efforts, the proposed approach introduces a specially designed programming language to control changes from application requirements to programs, and leverages on small-scale and loosely-coupled intermediate code gen-erated by the language's processing system to control changes from programs to ex-ecutable code. With these efforts, the proposed approach can further reduce the size of the script. Meanwhile, the proposed approach leverages on additional information of sensors such as locations and battery power to decrease the number of data relaying sensors, and optimizes the multi-channel communication to eliminate collisions among these sensors. These efforts can further decrease the number of packets for disseminat-ing the script.In particular, this paper makes the following contributions.Provide a technical framework for efficient reprogramming of WSN applications. By analyzing factors and their relationship which determine the efficiency of the reprogramming process, this paper provides a technical framework which combines a programming language, a processing system of the language, and a bulk data dissemination protocol.Propose a reprogramming-oriented logical programming language. The lan-guage considers a WSN application as the combination of data processing and system operations. It uses constructs of fact and rule from the traditional logi-cal programming language to deal with data processing, and extends with con-structs of event and action to handle system operations. Meanwhile, constructs of the language shield system-level implementation details. The organization of these constructs inherits the loose coupling nature of the logical programming language. Experimental results show that the proposed language can make ap-plication programs concise and easy to modify.Propose a language processing system which aims to support efficient repro-gramming of WSN applications. The processing system consists of a compiler and an interpreter. The compiler endeavors to reduce the size of the script. In particular, it first reduces the size of intermediate code by generating the inter-mediate code with no control logic of the code execution. In addition, it also diminishes differences between intermediate code by keeping the intermediate code of unchanged parts of old and new application programs the same. The interpreter focuses on reducing the overhead of the code execution. It first opti-mizes the intermediate code according to features of the code executing process. In addition, it also optimizes the execution process according to computational and communication requirements of specific applications. Experimental results show that the proposed language processing system can efficiently reduce the size of the script and the overhead of the interpretive execution.Propose a bulk data dissemination protocol to support efficient reprogramming of WSN applications. The protocol aims to decrease the number of packets for dis-seminating the script. In particular, it decreases the number of data relaying sen-sors by optimizing the dissemination route and increasing communication range of sensors according to locations and battery power of sensors, respectively. In addition, it eliminates communication collisions among data relaying sensors by optimizing the multi-channel communication based on complete information of collisions.Design and implement a reprogramming service management system. The sys-tem consists of programs of web site, server site, and WSN site. It can provide ef-ficient reprogramming services on multiple WSN applications for multiple users.