A generative wireless sensor network framework for agricultural use

Precision agriculture and agricultural decision making assistance have attracted researchers of engineering and agronomy as well as industries as a promising application of wireless sensor network (WSN) technology. However, agriculture is an industry with severe cost-cutting pressure and there is little advance guarantee that the WSN increases profit of farmers. Compared with other applications of the WSN such as power grid, civil engineering, etc., investment to the WSN cannot be expected so much. Nevertheless, if the problem of the cost is resolved, the WSN can be used with wide variety in the farm. The ready-made system satisfying requirements and constraints fully and generally of different farmers increases size and complexity of the system and finally the required investment. On the other hand, the full custom-made system for the individual farmer is not practical. This paper presents a concept of generative WSN framework enabling generation of semi custom-made WSN systems that satisfies requirements and constraints of individual farmers within a prescribed scope and also its fundamental technologies. The framework introduces a paradigm of software product line and manages core assets which are basis of the WSN system to be generated. Two tiered feature modeling, variation point mechanisms for hardware, and integer programming based optimization of system configuration are additionally developed technologies to realize the generative WSN framework.