General Electric Company

Aria Pezeshk

John Anderson Fergus Ross

The present techniques provide systems and methods for decoding an optical data signal in an optical system to retrieve source information while decreasing errors resulting from optical and electronic noise in the optical system. The techniques involve using decoding algorithms to estimate the a posteriori state probabilities and the a posteriori transition probabilities of the data encoding, and estimating bit state probabilities. The probability density function used to estimate bit states is parameterized by the expected optical and electronic noise in the optical system. Different optical and electronic noise variances, or different probability densities, may be stored in registers or look-up tables to be accessed by a decoder while decoding the optical data signal.

Method for decoding under optical and electronic noise

The present techniques provide systems and methods for decoding an optical data signal returned from an optical disc to retrieve source information. The decoding method is based on a 16 state trellis diagram, and may decode an optical data signal encoded through a modulation code where the input-to-output relationship is not convolutional, such as the 17 Parity Preserve/Prohibit (17pp) modulation code. A trellis diagram may enable non-convolutional trellis-modulated data to be more efficiently decoded. Further, the 16 state trellis diagram of the present techniques provides a unique path for each input-to-output bit pair, such that no information about input bits may be lost on parallel paths in a trellis diagram.

Bit probability estimation method for noisy modulation code

The present techniques provide systems and methods for decoding a data signal with a control bit to improve bit estimation. The techniques in one embodiment involve using decoding algorithms to estimate the a posteriori state probabilities and the a posteriori transition probabilities of the data encoding, and estimating bit state probabilities. The techniques further involve using a control bit in the bit stream and comparing the estimation of the control bit state in the segment of the bit stream with a test control bit determined based on an average of bit states from the encoded segment of the bit stream. If the estimation of the control bit and the test control bit are not equal, the state of the bit estimate with the lowest confidence probability will be changed.

Growing community of inventors

State College, PA, United States of America

Aria Pezeshk