Energy Saving Using Scenario Based Sensor Selection on Medical Shoes

Sensor Network Configuration Learning for Maximizing Application Performance

“…We performed the experiments by using the ns-3 sensor network simulator ( ) to evaluate our algorithms [ 6 ] and compare them with the probabilistic selection [ 8 ] sensor selection algorithm, genetic algorithm and pure randomness (use of a different random configuration on every iteration) under the scenarios listed in Section 4 . By iteration, we mean a complete evaluation cycle of the algorithm: set the configuration, then wait for the reward obtained using that configuration.…”

“…The work closest to ours, Xu and Potkonjak [ 8 ], proposes iterative selection and probabilistic selection for the sensor selection problem and demonstrates how this works with a medical shoe. As with the previous paper, they find that semantic information is more important than focusing on the accuracy of the raw sensor data.…”

“…The simulation results using a variety of reward Equations show that SNCL performs better than pure randomness, a standard genetic algorithm [ 7 ] and probabilistic selection [ 8 ], a recently published sensor selection algorithm. In particular, SNCL is able to find a configuration that maximizes application performance in all scenarios; whereas pure randomness and genetic algorithm are never able to find the maximum within the given time frame, and probabilistic selection only finds it in two scenarios due to trying all nodes on as a first step.…”