This paper considers problems related to output error models for output data estimation and parameter identification in missing output data systems. In this regard, a new sequentially parallel distributed adaptive sig...
详细信息
This paper considers problems related to output error models for output data estimation and parameter identification in missing output data systems. In this regard, a new sequentially parallel distributed adaptive signal processing method with the implementation of a low complexity least squares algorithm is introduced to estimate the parameters of an auxiliary model relative to the original system, as well as handling irregularly missing output data in a stochastic framework. The validation of the proposed distributed architecture using the low complexity least squares algorithm is presented in terms of computational complexity and processing time. Measurement results show that the proposed architecture using the low complexity least squares approach provides fast convergence, parallel linear computational complexity, and significantly reduced processing time compared to the sequentially operated recursive least square (RLS) algorithm for parameter identification in missing output data systems. Finally, the effectiveness of the low complexity least squares algorithm is tested with a system example
In this paper, a low communication parallel distributed adaptive signal processing (LC-PDASP) architecture for a group of computationally incapable and inexpensive small platforms is introduced. The proposed architect...
详细信息
In this paper, a low communication parallel distributed adaptive signal processing (LC-PDASP) architecture for a group of computationally incapable and inexpensive small platforms is introduced. The proposed architecture is capable of running computationally high adaptive filtering algorithms parallely with minimally low communication overhead. A recursive least square (RLS) adaptive algorithm based on the application of multiple-input multiple-output (MIMO) channel estimation is implemented on the proposed LC-PDASP architecture. Complexity and Communication burden of proposed LC-PDASP architecture are compared with that of conventional PDASP architecture. The comparative analysis shows that the proposed LC-PDASP architecture exhibits low computational complexity and provides an improvement more than of 85% reduced communication burden than the conventional PDASP architecture. Moreover, the proposed LC-PDASP architecture provides fast convergence performance in terms of mean square error (MSE) than the PDASP architecture.
暂无评论