5.3 PDF of L-values for 2D Constellations

The analysis shown in Section 5.2 is slightly more involved in the case of 2D constellations. In what follows, we will first describe how to find the tessellation regions c05-math-0317, and then, we will find the PDF of the L-values for arbitrary 2D constellations.

5.3.1 Space Tessellation

Finding the tessellation regions c05-math-0318 for 1D constellations is relatively simple and can be done by inspection, as we did, e.g., in Example 5.5. A more structured approach is required for 2D constellations, because as we will see, even with very regular constellations such as c05-math-0319QAM, the regions c05-math-0320 may significantly change in a function of the binary labeling.

We start by rewriting (5.27) as

5.67 equation
5.68 equation

where c05-math-0324, c05-math-0325, and c05-math-0326 is given by (3.54).

These tessellation regions c05-math-0327 are mutually disjoint

5.70 equation

and span the whole space

5.71 equation

Moreover, because c05-math-0330 in (5.69) is defined by linear inequalities, the tessellation region may be (i) a (convex) polygon, (ii) a region that extends to infinity (i.e., an “infinite” polygon), or (iii) an empty set (if the inequalities in (5.69) are contradictory).

It should be rather clear that the regions c05-math-0331 cannot be deduced from the so-called Voronoi regions of the constellation (see (3.100)). The Voronoi regions depend solely on the constellation c05-math-0332, while c05-math-0333 depends also the on the labeling (via c05-math-0334 and c05-math-0335), or equivalently, on the subconstellations c05-math-0336 and c05-math-0337. However, the tessellation regions c05-math-0338 can be obtained as the intersection of the Voronoi regions of the subconstellations c05-math-0339 and c05-math-0340, as explained in the following example.

c05f013

Figure 5.13 Tessellation regions c05-math-0341 in (5.69) for a 16QAM constellation labeled by the BRGC (see Fig. 2.14): (a) c05-math-0342, (b) c05-math-0343, (c) c05-math-0344, and (d) c05-math-0345. For clarity, only selected regions c05-math-0346 are identified with labels. The subconstellations c05-math-0347 and c05-math-0348 are indicated by white and black circles, respectively

c05f014

Figure 5.14 Tessellation regions c05-math-0396, for an 8PSK constellation labeled by the BRGC. The subconstellations c05-math-0397 and c05-math-0398 are indicated by white and black circles, respectively

Finding the tessellation regions in Examples 5.6 and 5.7 is relatively simple, however, this is not always the case. For example, when a 16PAM constellation labeled by the M16 (shown in Fig. 2.15) is considered, the labeling cannot be factorized as the ordered direct product of two labelings, so the tessellation regions are difficult to find by inspection. As we will see in Example 5.8, the regions in this case are very irregular, and thus, to find them, we require an algorithmic support.

In order to find c05-math-0366, we need to remove redundant constraints from (5.69). The c05-math-0367 linear inequalities c05-math-0368 and c05-math-0369 in (5.69) can be reformulated in a matrix form6 as

where

5.73 equation

is an c05-math-0372 matrix,7 and where c05-math-0375 and c05-math-0376 are c05-math-0377 matrices whose columns are taken from the respective sets

5.74 equation
5.75 equation

Similarly, the c05-math-0380 vector c05-math-0381 in (5.72) is

5.76 equation

where c05-math-0383 and c05-math-0384 are c05-math-0385 column vectors whose elements are taken from the sets

5.77 equation
5.78 equation

respectively. To make (5.72) correspond to (5.69), we also assume that the matrix c05-math-0388 and the vector c05-math-0389 are ordered in the same way. In other words, if the column of c05-math-0390 is defined by a particular pair of indices c05-math-0391, the same pair defines the corresponding element of c05-math-0392. The same applies to c05-math-0393 and c05-math-0394 whose entries are defined by the pairs c05-math-0395.

In general, (5.72) has redundant inequalities, i.e., the number of linear forms defining the region c05-math-0399 is smaller than c05-math-0400. Removing the redundant inequalities in (5.72) is a “classical” problem in computational geometry, and to solve it, we first require the region c05-math-0401 to be closed, i.e., infinite polygons (which indeed exists, as shown in Fig. 5.13) are artificially closed by adding constraints forming a square with arbitrarily large dimensions c05-math-0402

where c05-math-0404 is a c05-math-0405 vector of ones. We can keep track of these artificial constraints and, if they appear in the final solution, we should eliminate them, which would produce an infinite polygon. However, for practical purposes, retaining the artificial inequalities has negligible effect when c05-math-0406 is sufficiently large.

We then focus on the augmented problem

where c05-math-0408 is an c05-math-0409 matrix and c05-math-0410 is an c05-math-0411 column vector. To solve (5.80), we translate the coordinates c05-math-0412 via c05-math-0413 to obtain

5.81 equation

so that after the translation, c05-math-0416 satisfies all the inequalities in (5.82). The problem then boils down to finding a solution of

Various methods may be used to find a feasible solution of the set of inequalities in (5.83). Here, we have solved the following quadratic optimization problem:

Note that the objective function c05-math-0419 is defined arbitrarily as we are interested in finding any translation point c05-math-0420 satisfying the constraints. If the solution of (5.84) does not exist, it means that the inequalities (5.72) are contradictory and the tessellation region c05-math-0421 is empty.

After c05-math-0422 has been found, (5.82) can be expressed as

where c05-math-0424. The next step is to normalize each inequality in (5.85) to obtain

where the normalization is such that the c05-math-0426th row of c05-math-0427 is equal to the c05-math-0428th row of c05-math-0429 divided by the (positive by definition) c05-math-0430th entry of c05-math-0431.

The original problem in (5.72) expressed as in (5.86) is now in a standard form and can be solved using the duality between the line and points description. This duality states that removing redundant inequalities from (5.86) is equivalent to finding the minimum convex region enclosing the c05-math-0440 rows of c05-math-0441, where each row is treated as a point in 2D space. We thus have to find a convex hull of these c05-math-0442 points, which can be done efficiently using numerical routines available, e.g., in MATLAB® or Mathematica®.

As we will see in the next section, to compute the PDF of the L-values, we need not only the linear pieces defining the tessellation regions c05-math-0443 but also the vertices of the polygons. Finding these vertices is again relatively simple and corresponds to a vertex enumeration problem treated in computational geometry.

c05f015

Figure 5.15 Tessellation regions c05-math-0432 in (5.69) for a 16QAM constellation labeled by the M16: (see Fig. 2.15): (a) c05-math-0433, (b) c05-math-0434, (c) c05-math-0435, and (d) c05-math-0436. For clarity, only selected regions c05-math-0437 are identified with labels. The subconstellations c05-math-0438 and c05-math-0439 are indicated by white and black circles, respectively

5.3.2 Finding the PDF

In the previous section, we have shown how to find the tessellation regions for arbitrary 2D constellations and labelings. The next step toward finding the PDF of the L-values is to find the CDFs c05-math-0447 as required by (5.31). We start by rewriting (5.69) as

where we use

to denote each of the c05-math-0450 nonredundant linear forms c05-math-0451 defining the tessellation regions c05-math-0452, i.e., for each c05-math-0453 there exists a mapping between c05-math-0454 and the indices c05-math-0455 defining c05-math-0456. These nonredundant linear forms are the ones we have already shown, e.g., in Fig. 5.13: if c05-math-0457, c05-math-0458, and c05-math-0459, the region c05-math-0460 is an infinite polygon defined by c05-math-0461 linear forms; if c05-math-0462, c05-math-0463, and c05-math-0464, we obtain c05-math-0465; and for the region c05-math-0466 (the closed polygon containing c05-math-0467 and c05-math-0468), we have c05-math-0469. We also define the vertices of the tessellation regions as c05-math-0470.

The CDF c05-math-0471 (5.32) can be expressed as

where

contains observations c05-math-0474 which satisfy c05-math-0475 (this region is shown shaded in Fig. 5.16) and

where c05-math-0477 given by (5.43). The equation c05-math-0478 defines the line perpendicular to c05-math-0479, as schematically shown in Fig. 5.16.

c05f016

Figure 5.16 The tessellation region c05-math-0480 is a polygon delimited by the lines c05-math-0481, where c05-math-0482 is given by (5.88). The values of c05-math-0483 defining the intervals c05-math-0484 in (5.93) for c05-math-0485 together with the line c05-math-0486 and the two active linear pieces are also shown. The shaded area corresponds to c05-math-0487

The region c05-math-0488 in (5.90) is a subset of (5.87) and depends on c05-math-0489. Further, to simplify the presentation, we assume that the tessellation region c05-math-0490 is a closed polygon defined by c05-math-0491 nonredundant inequalities. Then the line c05-math-0492 either (i) intersects8 two sides of polygon c05-math-0494 and these sides are then active, or (ii) does not have a common point with the polygon. The first case is critical to the calculation of the CDF. We note that while the indices of the linear forms c05-math-0495 defining the active sides change with c05-math-0496, they remain constant as long as, varying c05-math-0497, the line c05-math-0498 does not cross one of the vertices c05-math-0499 of the polygon; i.e., the indices change when c05-math-0500 satisfies c05-math-0501, which may occur for c05-math-0502 distinct values of c05-math-0503 defined by

Thus, denoting by c05-math-0505 the sorted values of c05-math-0506 in (5.92), we find the c05-math-0507 intervals

such that

5.94 equation

For a given value of c05-math-0510, we will use the indices c05-math-0511 and c05-math-0512 to denote the two sides of the polygon which remain active for all c05-math-0513.

To clarify the above discussion, consider the example shown in Fig. 5.16. In this case, we have c05-math-0514 and c05-math-0515, c05-math-0516, and c05-math-0517. For the particular value of c05-math-0518 in Fig. 5.16, c05-math-0519 so c05-math-0520 and c05-math-0521. Furthermore, in this figure we schematically show the values of c05-math-0522 as a projection of the vertices on the line connecting c05-math-0523 and c05-math-0524 which becomes the “axis” for the arguments c05-math-0525 of the PDF (see also Fig. 3.2). The origin of this axis corresponds to c05-math-0526. Note that the origin does not necessarily belong to c05-math-0527.

The calculation of (5.89) requires now a 2D-equivalent of the integration shown in (5.48). This is practically the same problem we need to solve when evaluating the bit- and symbol-error rates, so the approach we have adopted here is also explained in detail in Section 6.1.2. More particularly, we decompose the observation space c05-math-0528 into elementary wedges (see also (6.34)) defined as

5.95 equation

Integration over such wedge can be expressed as9

5.97 equation

where c05-math-0538 is the bivariate Q-function defined in (2.11).

As shown in Fig. 5.16, the union of the wedges and the polygon over which the integration must be carried out yields the entire observation space. In the particular case shown in Fig. 5.16 we can write

where we used shorthand notation c05-math-0540 to denote the parameters of the linear form (5.91). We note that the linear form c05-math-0541, being redundant for the definition of c05-math-0542 (for the particular value of c05-math-0543 used in this example), does not affect the integration result.

We are now ready to find the conditional PDF of the L-values c05-math-0567.

Analyzing Theorem 5.10, we note that the indicator function in (5.106) takes into account the variation of the indices c05-math-0587 and c05-math-0588 because of the change of c05-math-0589 moving throughout the intervals c05-math-0590. As we see, the form of the PDF does not depend uniquely on c05-math-0591, c05-math-0592, and c05-math-0593, but also on other symbols from the constellation which define the polygon's sides. We also note that the Gaussian piece c05-math-0594 we have already obtained in the case of 1D constellations, is still present, but it is now multiplied by the “correction” terms (5.107).

5.3.3 Case Study: 8PSK Labeled by the BRGC

In this section, we consider an 8PSK constellation, which is probably the simplest nontrivial example of a 2D constellation. Furthermore, because of its practical relevance, we consider the BRGC. The constellation and labeling are shown in Fig. 5.14.

The tessellation regions are simple to find as we have already shown it in Example 5.7: they are always either (i) the Voronoi regions of the constellation symbols (e.g., c05-math-0595, c05-math-0596 in Fig. 5.14) or (ii) a union of two Voronoi regions (e.g., c05-math-0597, or any c05-math-0598, see again Fig. 5.14). This property spares us the need to apply the algorithmic approach of Section 5.3.1, which means we do not need to set the artificial constraints (5.79). Second, because of symmetry of the BRGC, the regions also have a notable symmetry, which we will exploit to simplify the analysis.

As for c05-math-0599PSK constellations c05-math-0600 c05-math-0601, we obtain

and thus,

5.113 equation

The expression in (5.112) shows that all the lines delimiting the regions pass through the origin and so does the line c05-math-0604.

Let us focus on the case c05-math-0611 which also gives us the results for c05-math-0612 (as explained in Example 5.7). This case is shown in Fig. 5.17. The piece of the PDF c05-math-0613 depends on the form of the tessellation region c05-math-0614 as well as on the position of the symbol c05-math-0615 with respect to this region. An inspection of the constellation in Fig. 5.17 lets us conclude the following:

c05f017

Figure 5.17 Some tessellation regions for 8PSK labeled by the BRGC and c05-math-0605. To obtain c05-math-0606, c05-math-0607, c05-math-0608, or c05-math-0609 for c05-math-0610, integration should be carried over the shaded regions

  • Because of the symmetry of the constellation and the tessellation regions, we may write
    5.114 equation
    5.115 equation

    thus, it is sufficient to find c05-math-0618 for c05-math-0619, conditioned on c05-math-0620.

  • For c05-math-0621 the lines c05-math-0622 always intersect with the same sides of the tessellation regions, therefore we do not need to take into account intervals changing with c05-math-0623 as in (5.106). This is obvious from Fig. 5.17: the line c05-math-0624 crosses the same two sides of the region c05-math-0625, the line c05-math-0626 crosses only one side of c05-math-0627, and so on. Considering only c05-math-0628 thus simplifies the analysis.
  • Inspecting the regions c05-math-0629 and c05-math-0630 (shown shaded), we conclude that
    equation

    i.e., c05-math-0632.Moreover, for c05-math-0633 we observe the following relationship:

    equation

    i.e., c05-math-0635.

After differentiation and generalization to other regions we obtain the following relationships:

5.117 equation

The advantage of (5.116)–(5.118) is that we can now express the PDF c05-math-0639 using (5.33) as

5.119 equation

The same can be done for the PDF pieces conditioned on c05-math-0642, namely,

5.121 equation
5.122 equation
5.123 equation

which yields

5.124 equation

Thus, instead of varying the regions c05-math-0648 for a constant c05-math-0649 as suggested by (5.33), in (5.120) and (5.125), we limit our considerations to two regions c05-math-0650 and change the symbols c05-math-0651. We emphasize that this is possible in the particular case of an c05-math-0652PSK constellation labeled by the BRGC because of the symmetries of the constellation and the binary labeling. This is very convenient because we only need to find c05-math-0653 and c05-math-0654 for c05-math-0655.

The same can be done for c05-math-0656. In this case, only one symbol needs to be considered, namely, c05-math-0657. It can be shown that in this case

5.126 equation

In order to reuse the results obtained for c05-math-0659, we can further exploit the equivalence (up to a rotation) of the tessellation regions c05-math-0660 and c05-math-0661 (see Fig. 5.13), i.e., c05-math-0662 for c05-math-0663 and c05-math-0664.

Up to now, we have shown the PDF c05-math-0665 can be expressed as a function of c05-math-0666 and c05-math-0667. The next step is to use Theorem 5.10. Before doing this, however, we study some properties of c05-math-0668PSK constellations.

More specifically, using c05-math-0669, where c05-math-0670 (see (2.52)), we find

5.127 equation
5.128 equation
5.129 equation
5.130 equation
5.131 equation

where

5.133 equation

Similarly, we can find

5.134 equation

We can now analyze the contribution of the region c05-math-0680. From (5.106), we obtain

where c05-math-0682, and c05-math-0683. Here, c05-math-0684, c05-math-0685 and c05-math-0686, so (5.107) becomes

5.137 equation

where we used c05-math-0688.

Note that there is only one correction term c05-math-0689 in (5.136) because the region c05-math-0690 is a semi-infinite polygon and the line c05-math-0691 intersects only one of its sides. Equivalently, we might set an artificial constraint to close the polygon, e.g., c05-math-0692, where c05-math-0693 and c05-math-0694. Then c05-math-0695 and c05-math-0696, cf. (5.107).

Similarly, for c05-math-0697 we obtain

5.138 equation

where c05-math-0699, c05-math-0700, and c05-math-0701 and c05-math-0702, so

5.139 equation
5.140 equation

with c05-math-0705, and c05-math-0706.

The form of c05-math-0707 is shown in Fig. 5.18 for c05-math-0708 and c05-math-0709, where we identify the pieces c05-math-0710. Similarly, we show the corresponding results for c05-math-0711 in Fig. 5.19. We can observe that for c05-math-0712, the term c05-math-0713 has a “dominating” contribution to c05-math-0714. In particular, in Fig. 5.18 (a), we see that for c05-math-0715, c05-math-0716 takes on the largest values, while in Fig. 5.18 (b), we see that for c05-math-0717, the piece c05-math-0718 dominates the others. Finally, in Fig. 5.19 we see that for c05-math-0719, c05-math-0720 is again the dominating piece.

c05f018

Figure 5.18 PDF of the L-values for an 8PSK constellation labeled by the BRGC and c05-math-0721: c05-math-0722 (thick solid lines) in (5.120) and the pieces c05-math-0723 (lines with markers) for (a) c05-math-0724 and (b) c05-math-0725

c05f019

Figure 5.19 PDF of the L-values for an 8PSK constellation labeled by the BRGC and c05-math-0726: c05-math-0727 (thick solid line) in (5.120) and the pieces c05-math-0728 (lines with markers)

..................Content has been hidden....................

You can't read the all page of ebook, please click here login for view all page.
Reset
18.221.12.52