1D localization of highly correlated mobile stochastic EM sources using neural model
2017
In this paper, a possibility to use a multilayer perceptron neural network for the spatial localization of highly correlated mobile stohastic electromagnetic sources is considered. The neural model architecture for 1D DoA estimation of stochastic sources and the way of chosing the input data for the neural model by selecting the appropriate elements from the spatial correlation matrix are presented in the paper. Model accuracy is verified on the example of determining the angular position of two mobile stochastic sources moving along the 1D path and whose level of mutual correlation is within the range [0.8–0.95].
Keywords:
- Covariance matrix
- Spatial correlation
- Angular displacement
- Artificial neural network
- Multilayer perceptron
- Mathematics
- Matrix (mathematics)
- Correlation
- Architecture
- Artificial intelligence
- Pattern recognition
- spatial localization
- mutual correlation
- Algorithm
- model architecture
- Computer science
- multilayer perceptron neural network
- Correction
- Source
- Cite
- Save
- Machine Reading By IdeaReader
5
References
2
Citations
NaN
KQI