| Wireless sensor network is a new technology that has very broad market prospects and has a profound impact in various fields of human life. In context-aware system based on wireless sensor network, developers face a major question is: changing of the business rules allow system upgrade and maintenance become increasingly difficult. Rule engine can be a very good solution to this problem. Rule engine can separate business rules from the code. It use natural language syntax to compose rules, to receive data input, and in accordance with the rules to make business decisions.The algorithm of most rule engine is Rete and its variants. The core idea of Rete algorithm is state saving. Clearly, this algorithm is not applicable to the scene that facts change frequently. However, in context-aware system based on wireless sensor network, to deal with a large number of real-time information, the rule engine's efficiency has become the bottleneck of the whole system.In order to solve the above-mentioned performance issues, the paper design of a context-aware architecture based on wireless sensor network called MamboSenser. MamboSenser is a distributed architecture. It composed of a set of agents who represent real-world entities. We call it context-aware agent. In context-aware agent, Drools are used as smart components. Compare to a centralized manner, MamboSenser distribute the work of reasoning to each agent. Agents collaborate to complete the specific objectives, from a certain extent, make performance arising.To simplify the application development based on MamboSenser, this paper is designed and implemented the SmartJADE framework. SmartJADE extends JADE (Multi-Agent System Development Platform) API, integrates the basic functions of Drools into these agents so that agents are born with a "smart." And, an agent can control another agent's rules engine. Use SmartJADE can greatly enhanced the efficiency of development of context-aware system. |
PDF Full Text Request