[Show abstract][Hide abstract] ABSTRACT: When Multi-DSP parallel architecture transfers to distributed memory way from shared memory way, its parallelism with fine-grained become weak, and it's difficult to offer SIFT's complex computing and satisfy the need of real-time. In the paper, a parallel algorithm, based on distributed shared memory architecture with homogeneous Multi-core DSP, referring to the DSM architecture model of parallel processing machines is presented. Firstly, the master processor separates the task into several small tasks by exploiting the coarse-grained parallelism inherent; then, through a high-speed network for data-exchange, each small task transfers to a related subsystem, based on a homogeneous multi-core DSP; finally, the DSP partitions the small task across multiple cores by exploiting the fined-grained parallelism. The experimental result shows that, comparing to the traditional way, the proposed algorithm increases the speedup, calculates 45 frames on 640×480 images, and achieves the real-time application.
[Show abstract][Hide abstract] ABSTRACT: Angle random walk (ARW) is the noise component at the output of a fiber optic gyroscope (FOG) and it affects the FOG short-term accuracy. Using only the noise property described by random walk coefficient, we developed an unsupervised adaptive filter (UAF) for suppressing the noise. Theoretical analysis illuminates that UAF performs as good as the best FIR filter in the sense of means square. Experimental results both in stationary and non-stationary cases show that UAF is more effective than wavelet denoising and that based on fast Fourier transform.
[Show abstract][Hide abstract] ABSTRACT: A model of piezoelectric actuator with hysteresis has been built in this paper with Prandtle-Ishlinskii model. After that, a radial basis function (RBF) neural network based adaptive inverse control scheme for nonlinear systems with unknown hysteresis nonlinearity is developed. A nonlinear filter based on RBF neural networks is used in hysteresis inverse plant modeling. We use the inverse model as the controller to control the piezoelectric actuator model directly. The simulation results show that the method interposed in this paper can restrain the hysteresis effect to lower than 1.25%.
[Show abstract][Hide abstract] ABSTRACT: In this paper, we introduced the primary problem of common DC-DC converters which are used in complex environments and the design of robust controller for a buck converter in the same environment. The results of simulation showed that even in complex environmental conditions, the converter can still have good performance and stability. The robust design of the electronic system can effectively reduce the relying on the quality of electronic components, and also improves the reliability of DC-DC converter in the complex environments.