A supervised active learning method for identifying critical nodes in Wireless Sensor Network
Authors:
Behnam Ojaghi,
Mohammad Mahdi Dehshibi
Abstract:
Energy Efficiency of a wireless sensor network (WSN) relies on its main characteristics, including hop-number, user's location, allocated power, and relay. Identifying nodes, which have more impact on these characteristics, is, however, subject to a substantial computational overhead and energy consumption. In this paper, we proposed an active learning approach to address the computational overhea…
▽ More
Energy Efficiency of a wireless sensor network (WSN) relies on its main characteristics, including hop-number, user's location, allocated power, and relay. Identifying nodes, which have more impact on these characteristics, is, however, subject to a substantial computational overhead and energy consumption. In this paper, we proposed an active learning approach to address the computational overhead of identifying critical nodes in a WSN. The proposed approach can overcome biasing in identifying non-critical nodes and needs much less effort in fine-tuning to adapt to the dynamic nature of WSN. This method benefits from the cooperation of clustering and classification modules to iteratively decrease the required number of data in a typical supervised learning scenario and to increase the accuracy in the presence of uninformative examples, i.e., non-critical nodes. Experiments show that the proposed method has more flexibility, compared to the state-of-the-art, to be employed in large scale WSN environments, the fifth-generation mobile networks (5G), and massively distributed IoT (i.e., sensor networks), where it can prolong the network lifetime.
△ Less
Submitted 29 April, 2023; v1 submitted 19 April, 2020;
originally announced April 2020.
