Kalman Filter for enhancement of noisy speech
2016-08-23
2 0 0
no vote
Other
Earn points
Kalman filter, also known as linear quadratic estimation (lqe), is an algorithm that uses a series of observations over time, including noise (random variation) and other inaccuracies, and produces often more accurate estimates of unknown variables than those based on a single measurement of a person. More formally, the Kalman filter recursively operates on noisy input data streams to generate statistical optimal estimates of the underlying system state. The filter is named Rudolph (Rudy) e. Kalman, one of the main developers of its theory. Kalman filter has many applications. A common application is for guidance, navigation and vehicle control, especially for aircraft and spacecraft. In addition, the Kalman filter is widely used in the field of signal processing and metrology, such as the concept of Yintai sequence analysis. The algorithm works in a two-step process. In the prediction step, the estimated values of the current state variables generated by the Kalman filter and their uncertainties. Once the results of the next measurement (a certain amount of damage and a certain amount of error, including random noise) are observed, these estimates are updated by using the weighted average, with more weights given to the estimates with higher certainty. Because of the recursive nature of the algorithm, it can run in real time using only the input measurement and the previously calculated state, and its uncertainty matrix; no additional past information is needed. It is a common misunderstanding that Kalman filter assumes all error terms and measures Gaussian distribution. Kalman's base paper uses orthogonal projection theory to show that the covariance is minimized by the derived filter, and this result does not require any assumptions, for example, the error is Gaussian. [1] Then, he shows that the filter produces the exact conditional probability estimation in special cases, and all the errors are Gaussian distribution.
matlab
语音
滤波
卡尔曼
环境
增强
噪声
Related Source Codes
GMSK Linear Receiver
0
0
no vote
NSGA-II algorithm
0
0
no vote
NSGA-III multi-objective optimization algorithm
0
0
no vote
Compressed sensing example
0
0
no vote
CFAR detector example
0
0
no vote
No comment