Distributed and morphological operation-based data collection algorithm

Abstract

When monitoring the environment with wireless sensor networks, the data sensed by the nodes within event backbone regions can adequately represent the events. As a result, identifying event backbone regions is a key issue for wireless sensor networks. With this aim, we propose a distributed and morphological operation-based data collection algorithm. Inspired by the use of morphological erosion and dilation on binary images, the proposed distributed and morphological operation-based data collection algorithm calculates the structuring neighbors of each node based on the structuring element, and it produces an event-monitoring map of structuring neighbors with less cost and then determines whether to erode or not. The remaining nodes that are not eroded become the event backbone nodes and send their sensing data. Moreover, according to the event backbone regions, the sink can approximately recover the complete event regions by the dilation operation. The algorithm analysis and experimental results show that the proposed algorithm can lead to lower overhead, decrease the amount of transmitted data, prolong the network lifetime, and rapidly recover event regions.

CS-FCDA: A Compressed Sensing-Based on Fault-Tolerant Data Aggregation in Sensor Networks

“…A basic Compressed Sensing (CS) algorithm was proposed in paper [ 11 ] targeted for the compression and reconstruction of the urban traffic data and it was shown that the amount of communication in the network can be significantly reduced. A Distributed and Morphological Operation-based Data Collection Algorithm (DMOA) was proposed in paper [ 12 ]. This scheme was based on the non-parametric data construction of the lagged covariance matrix for the reconstruction of the partially lost geological data information.…”

“…For the sake of evaluating the performance of the proposed CS-FCDA scheme, we further compare our proposed CS-FCDA mechanism with the Compressed Sensing (CS) [ 11 ], DMOA [ 12 ] and MTTA [ 13 ] with different evaluation criterions. In order to further evaluate the performance of the proposed CS-FCDA, we have run simulations under the following five different environments.…”

“…In this simulation, we deployed the sensor nodes randomly within a target area which is 300 m × 300 m. The numbers of nodes deployed are 100, 300 and 500, respectively, representing different deployment density. The node energy consumption model in paper [ 12 ] is adopted for this paper and the equations for the data transmission and reception energy consumption are listed in (31) and (32).…”

Edge Computing in Internet of Things: A Novel Sensing-Data Reconstruction Algorithm Under Intelligent-Migratoin Stragegy

“…However, such algorithm is only applicable to the lost data recovery of the sink end database without considering the packet loss problem recovery during collection of the data. Paper [28] proposes the Distributed and Morphological Operation-based Data Collection Algorithm (DMOA), where it randomly selects some nodes in each round in order to participate in data collection, sparse observation matrix extremely is constructed just based on the node number of the received data in order to reconstruct the raw data of the entire network at the sink end. Though this method aims to solve the CS data collection problem which is under unreliable links, as the extremely sparse assumption is adopted, it is not applicable for environment with weak spatial data correlation across the whole network.…”

MR-DFM: A multi-path routing algorithm based on data fusion mechanism in sensor networks

“…al [10] first defined the link cost based on node distance, the information entropy of the nodes, the union entropy and the data amount and then obtained High-dimensional Data Aggregation Control (HDAC) in the network through the dynamic scheduling. Nie et.al [11] proposed three different algorithms, i.e., MSTbased, DMDC-based and COM algorithms to construct data aggregation trees with better energy consumption performances and delay performances with different data growth rates. The location of the one-hop neighbors or the distance information is employed by the L-PEDAPs [12] to construct the local minimum spanning tree and correlated neighbor graph.…”