Multi-sensor Data Fusion Algorithm Based on Adaptive Trust Estimation and Neural Network

Multi-sensor data fusion technique plays a key role in the agricultural services such as data collection and processing. However, the collected data usually is featured by redundancies and errors, which deteriorate the reliability of network. In this paper, based on adaptive trust estimation, we pro...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:2020 IEEE/CIC International Conference on Communications in China (ICCC) s. 582 - 587
Hlavní autoři: Zhao, Xuexin, Wu, Junhua, Wang, Maoli, Li, Guangshun, Yu, Haili, Feng, Wenzhen
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: IEEE 09.08.2020
Témata:
Adaptation models
adaptive trust estimation
BP neural network
data fusion
Data integration
Data models
Estimation
Green products
Neural networks
Sensors
WSNs
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Multi-sensor data fusion technique plays a key role in the agricultural services such as data collection and processing. However, the collected data usually is featured by redundancies and errors, which deteriorate the reliability of network. In this paper, based on adaptive trust estimation, we propose a multisensor data fusion algorithm in Trust Neural Network (T-NN), aiming to solve the problem of low accuracy and poor stability of multi-sensor data fusion. In particular, the original data collected by the sensors first are pre-processed by exponential smoothing. Then, trust estimation model is applied to calculate the value of trust among the sensing nodes and optimize the data, and the performance of redundancy and reliability are enhanced. Furthermore, the data optimized is introduced into BP neural network for training and fusion. Extensive simulations show that the algorithm proposed in this paper greatly outperforms adaptive weighted average model and traditional BPNN model, in terms of the accuracy of data fusion.
DOI:10.1109/ICCC49849.2020.9238952