This paper is related to the problem of low computational resources available on actual processing modules used in the nodes of wireless sensor networks, in the context of continuously increasing complexity of optimization algorithms, communication protocols and implemented services. The increase in complexity is driven by the efforts to conserve the energy and to enhance the network throughput but it is also caused by the evolution of these technologies toward the integration into the Internet of Things. The results obtained give an overview on the computational resources (computing time and memory size) necessary to perform the usual arithmetic operations using different representations of the operands on multiple platforms. These results are aimed to support the optimal selection of WSN platforms relatively to the computational effort required by specific application as well as the adaptation of the operations performed and of the related operands to a specified platform.
展开▼