15.01.2013 Views

U. Glaeser

U. Glaeser

U. Glaeser

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

FIGURE 34.58 Viterbi algorithm detection on the PR4 trellis.<br />

The Viterbi detector will start to search the trellis accumulating branch distance from sequence y. In<br />

the first recursion of the algorithm, there are two paths of length 1 at the distance<br />

from y. Next, each of the two paths of length 1 are extended in two ways forming four paths of length 2<br />

at squared Euclidean distance from the sequence y<br />

and this accumulated distance of four paths labels the four trellis states. In the next loop of the algorithm<br />

each of the paths are again extended in two ways to form eight paths of length 3, two paths to each node<br />

at level (depth) 3.<br />

Node 00<br />

Node 10<br />

Node 01<br />

© 2002 by CRC Press LLC<br />

d( y,0)<br />

( 0.9 – 0)<br />

2<br />

= = 0.81<br />

d( y,1)<br />

( 0.9 – 1)<br />

2<br />

= = 0.01<br />

d( y, ( 0,0)<br />

) 0.81 ( 0.2 – 0)<br />

2<br />

= + = 0.85<br />

d( y, ( 0,1)<br />

) 0.81 ( 0.2 – 1)<br />

2<br />

= + = 1.45<br />

d( y, ( 1,0)<br />

) 0.01 ( 0.2 – 0)<br />

2<br />

= + = 0.05<br />

d( y, ( 1,1)<br />

) 0.01 ( 0.2 – 1)<br />

2<br />

= + = 0.65<br />

d( y, ( 0,0,0)<br />

) 0.85 ( −0.6 – 0)<br />

2<br />

= + = 1.21<br />

d( y, ( 1,0, – 1)<br />

) 0.05 ( −0.6 + 1)<br />

2<br />

= + = 0.21 surviving path<br />

d( y, ( 0,0,1)<br />

) 0.85 ( −0.6 – 1)<br />

2<br />

= + = 3.41<br />

d( y, ( 1,0,0)<br />

) 0.05 ( −0.6 – 0)<br />

2<br />

= + = 0.41 surviving path<br />

d( y, ( 0,1,0)<br />

) 1.45 ( −0.6 – 0)<br />

2<br />

= + = 1.81<br />

d( y, ( 1,1, – 1)<br />

) 0.65 ( −0.6 + 1)<br />

2<br />

= + =<br />

0.81 surviving path

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!