Abstract
Numerous applications rely on data obtained from a wireless sensor network where application performance is of utmost importance. However, energy usage is also important, and oftentimes, a subset of sensors can be selected to maximize application performance. We cast the problem of sensor selection as a local search optimization problem and solve it using a variant of stochastic hill climbing extended with novel heuristics. This paper introduces sensor network configuration learning, a feedback-based heuristic algorithm that dynamically reconfigures the sensor network to maximize the performance of the target application. The proposed algorithm is described in detail, along with experiments conducted and a scalability study. A quick method for launching the algorithm from a better starting point than random is also detailed. The performance of the algorithm is compared to that of two other well-known algorithms and randomness. Our simulation results obtained from running sensor network configuration learning on a number of scenarios show the effectiveness and scalability of our approach.
Highlights
Sensor applications are deployed by using data obtained from a wireless sensor network (WSN), and this approach is applicable in many domains
This paper introduces sensor network configuration learning (SNCL), a feedback-based learning algorithm that takes feedback from the application performance along with the current network state and dynamically reconfigures the WSN with the goal of learning a configuration that will maximize the performance of the target application
We developed the SNCL Algorithm 1, which dynamically reconfigures the wireless sensor network by using the feedback from application performance to learn a configuration that will maximize the performance of the target application
Summary
Sensor applications are deployed by using data obtained from a wireless sensor network (WSN), and this approach is applicable in many domains. It can be used in smart homes to detect individual movement and classify the task in which the individual is engaged [1]; in this case, the goal of the application is to maximize recognition accuracy. Many times, a subset of sensors must be selected to maximize application performance, allowing the possibility of turning off unnecessary sensors to save the energy needed to power them. A naive approach would be to exhaustively try all possible combinations of sensors; such an approach is infeasible as the state space increases exponentially with the number of sensor nodes, and the time to evaluate each combination in a real-world environment would be prohibitive
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have