## Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

**Polyphase Pulse Compression Codes with Optimal Peak and Integrated Sidelobes
**

Carroll J. Nunn Gregory E. Coxson Technology Service Corporation Suite 800, 962 Wayne Avenue, Silver Spring, Maryland 20910

Abstract—Near-optimal or globally optimal integrated sidelobe level (ISL) polyphase codes are found for lengths 46 through 80 by using a stochastic optimization technique. Polyphase Barker codes are found for lengths 64 to 70, and 72, 76 and 77 using constrained optimization starting at optimal-ISL codes. 1

2/N and 3/N are probably length 28 and 53 respectively (refer, e.g., to [2] and [3]). Codes achieving 4/N sidelobe levels have been found at lengths up to and including 82 [4].

Keywords: Barker sequence, peak sidelobe level, polyphase Barker, stochastic optimization. I. I NTRODUCTION Pulse compression coding is used in radar applications to obey the same PSL-to-peak ratio maximum, and have complex gain the signal-to-noise (SNR) beneﬁts of a long pulse along elements not necessarily of uniform size. Here we will focus with the range resolution of a short pulse. An important on codes that have elements of uniform size. We refer to them ﬁgure of merit used to describe pulse compression codes is as ”complex uniform”, or polyphase codes. It has become the peak sidelobe level (PSL). Codes with good PSL can be a challenge to extend the list of lengths for which such used to discriminate returns of interest from unwanted close-in sequences are known. Zhang and Golomb found polyphase discrete returns. Barkers up to length 19 in 1989 [6]. Friese and Zottman [7] Perhaps the most well-known phase codes used in radar extended the list to length 31 in 1994. Brenner [8] found codes pulse compression are the binary Barker codes. These are up to length 45 in 1998. Borwein and Ferguson [9] found codes with elements either 1 or −1 for which the PSL-to-peak codes of length up to 63 in 2005. Also in 2005, Nunn [10] voltage ratio is 1/N , where N is the length of the code. Barker found four polyphase Barker codes of length 64. Borwein and codes exist for several lengths up to 13, but no Barker codes Ferguson [11] subsequently presented a polyphase Barker of exist for odd lengths over 13 [1]. It is generally accepted that length 65 in 2007. there are no such biphase codes of even length greater than 4. Furthermore, exhaustive searches have been conducted to establish that the longest codes with PSL-to-peak ratios of

work was funded by PEO IWS 1.0 and 2.0 under BAE prime contract N00024-01-D-7014

1 This

In 1965 Golomb and Scholtz [5] started to look at more general sequences called generalized Barker sequences which

Whereas one goal of this paper is to ﬁnd yet longer polyphase Barker sequences, much of the effort is concentrated on ﬁnding codes with very good integrated sidelobe levels

.g. with 99 percent certainty. The ﬁrst part was to establish the length of the longest polyphase Barker code attainable with the routines and processing power available to the project. It would be difﬁcult to completely describe the methodology of these tools. . a pair of 2. in the background for many months. These methods were used in conjunction with a stochastic search to ﬁnd large numbers of 32-element and 64-element codes with locally optimal ISL. a 3. the use of a local optimization strategy combined with a stochastic. The methodology used in this paper is similar to the method of Borwein and Ferguson [9]. This is done because long codes cannot have Barkerlevel sidelobes without also having good ISL. and then used as starting points for local searches for low-PSL codes.. D ISCUSSION OF M ETHODS AND R ESULTS ISL = 10 ∗ log10 i=0 . Computers that have been used in these searches include a computer with a 2. These routines evolved over a number of years. With this in mind good ISL codes are found. They also used a variation of the Inverse Coupon Collectors problem to establish. N − 1. e. deﬁned as Ri R0 2 II. the authors of this paper used an existing set of routines which were designed to ﬁnd large numbers of good ISL codes. . and have been used effectively to ﬁnd many codes of various lengths both binary and polyphase.2 GHz Opteron processors.4 GHz Intel Pentium D processor. These searches were performed off and on. one of the authors of this paper indicated methods for ﬁnding local minima for objective functions related to pulse compression ﬁgures of merits such as matched ﬁlter ISL and PSL.e. but the basic methodology is as described in Borwein and Ferguson’s paper i. the equivalent of minimum possible ISL values (refer. To ﬁnd polyphase Barker codes. The second objective which was consistent with. and accomplished concurrently with.2 (ISL). .8 GHz single core processors. The work in this paper consisted of three tasks or objectives.4 GHZ Intel quad-core processor. . or global optimization strategy. the ﬁrst was to apply sophisticated searches and signiﬁcant computational effort to seek low. and R0 is the autocorrelation peak [12]. ISL values for codes from length 46 through 80. and a Beowulf cluster combining 18 single core 2. to [13] and [14]). These codes were then used as initial conditions for length32 and length-64 Barker searches. i = 1 − N. For each number of chips from 10 through 80 the program was run for a substantial amount of time on a a set of computers containing one or more Intel or AMD processors. The third part was to partially understand the quantity and where the Ri are the elements of the autocorrelation sequence. In a previous publication [10]. and possibly optimally low. In addition this paper will attempt to give a feel for the overall number of available Barkers. It is one purpose of this paper to use the local optimization/stochastic search strategy to extend the list of known lengths for which polyphase Barker codes may be found. They used a different local optimization strategy with a stochastic search to ﬁnd polyphase Barker codes for lengths from 46 to 63.

35 -31.28 -32.77 -32.58 77 -18.63 -27.08 -32.61 -31.22 -30.38 -16.13 -17.46 78 -17.41 69 -17.00 -30.66 -33.30 47 -17.48 -16. 76 and 77.49 -16.90 -30.44 .53 -17.83 73 -18.80 -32. and in most cases exceeded.05 54 -17.78 -30.32 -31.57 51 -18. and in the case of the codes from length 10 through 52.93 -33.08 -30. The effort was continued and extended in order to search for better ISL codes and further polyphase Barker codes.92 -16.19 -29.53 46 -17. The amount of effort used at each of these lengths was consistent with.10 -16.06 80 -17.3 nature of available polyphase Barkers. This continued search ultimately yielded a polyphase Barker code at each of the lengths 72. there are no comparison results.00 57 -18.44 -31. existing routines were run for a substantial amount of time for each length from 10 through 80 to ﬁnd large numbers of excellent ISL codes.69 -34. For lengths 64 and above.37 -32. it should be noted that the point of this table is to ﬁnd the lowest ISL values ever found. the level of effort Borwein and Ferguson used to establish the equivalent of best ISL values for lengths up to 45 with 99 percent certainty.84 -17.22 56 -17. and for lengths 72.63 58 -17.78 -29. 76.08 -31.67 -16.57 -31. as described in [10].57 44 -16. These results were always at least as good as.43 -16.46 -30.64 62 -17.50 -16. ISL and PSL of Lowest-ISL Sequences Found N ISL (dB) PSL (dB) Min ISL (dB) in [9] 41 -17.99 -16. with the objective of ﬁnding polyphase Barker codes.44 45 -16.37 -16. The codes which eventually yielded the polyphase Barker codes of lengths 65 through 70 in this paper were found in a relatively short time on a pair of desktop computers.25 -17.19 42 -16. The best ISL values found for lengths 41 through 80 along with the PSL values of the codes with the best ISL are shown in Table 1.11 -29.48 67 -18.53 -33.63 -31.44 72 -17.99 -29. This attempt was successful for all lengths below 71.94 -17. and 77.74 -17.72 -16. 76 and 77.61 52 -16. To accomplish the ﬁrst and second task.70 49 -17.70 -30.85 -17.54 79 -17.45 -16. to ensure that as many polyphase Barkers as possible were found. and often better than.39 -16.93 -16. the comparison results. Though some of the improvements exhibited in Table 1 appear small.87 64 -17.47 -32.74 76 -17. 72.91 68 -18.36 65 -18. The results for lengths less than 64 are compared to the results reported in [9].78 71 -17.53 -29.48 53 -17.70 -29. Table 2 gives the ISL values found for sequences used as starting points for searches that resulted in Barker sequences of length 70.62 -32.14 43 -16.77 -16.60 74 -17.02 -29.06 -28. Table 1. The codes resulting from this search were used as starting points for a constrained optimization search.88 61 -17.25 75 -17.50 55 -18.59 -17.69 60 -17.28 59 -17.28 -29.25 -26.97 50 -16.14 -29.97 -30.98 48 -16.98 -29.45 -31.22 66 -17.54 -16.67 70 -18.83 63 -17.

695 -1.4 Four Barker sequences were found for length 70.986 0.854 0.514 13 -3.194 -1.238 -1.24 76 -17. The resulting polyphase Barker codes are shown in Tables 3a.280 25 3.633 41 1.624 -1.637 -1.250 4 0.094 35 -1.542 -2.866 2.041 19 2.431 8 -0.068 -2.098 -1.748 16 1.782 1.130 -3.191 2. Starting Point ISL and PSL N ISL (dB) PSL (dB) 70 -18.549 0.712 -1.149 27 0.41 -31.590 -0.485 -1.586 21 2.044 2.644 -0.91 70 -17.393 1.033 -2.672 37 1.166 2.234 18 2.169 34 -1. 3b.78 70 -17.634 1.042 2.361 5 1.478 -0. Table 3a.501 -0.722 10 -1.023 17 -3.507 .742 7 -2.947 29 -2. Barkers were also found for lengths 65 through 69.231 0.074 -1.694 28 -2. In order to save space.928 -1.016 1. 74. and 3c.805 12 0.034 2.967 15 -0.381 1.989 14 -1.094 -0.843 0.990 0. This is a standard form for polyphase codes which can be achieved for a given code by applying some combination of four PSL-preserving transformations [15].402 -0.442 0.929 38 -2.372 33 -0.27 -33.230 -2.796 2.359 3.705 1.669 0. and 77 through 80.129 2.171 9 -0.760 -1. 3b.93 -33.641 24 -2.605 2.370 -0.043 32 1.276 0.944 -1.620 1.998 30 -1.564 1.377 0.927 0.48 Each of the nine codes in Tables 3a.849 -0. New Polyphase Barkers Length i 65 66 67 3 0.582 1.647 -2. 73.326 22 1.671 -2.578 20 -3.651 0.957 -2.103 -1.725 -0.84 -33.853 42 0.06 72 -17.384 1.745 1.745 40 1.279 1.060 0. these zero phase elements were not displayed.365 -2.531 6 2.487 39 1. Table 2.840 -1.454 26 1.547 -1.873 -0.730 -0.095 0.08 -31.787 2.742 31 2.513 -2.131 1.44 70 -17.889 -1.166 0.68 -31.918 -0. No polyphase Barker codes were found for lengths 71.131 -0.35 -32. and 3c have zero phase in the the ﬁrst two elements.788 0.130 11 -0.828 0.940 36 2.58 77 -17.287 23 -2.

36066 2.64371 2.868 2.348 -0.05258 -2.523 3.27012 -0.3947 2.10839 0.746 2.7585 -0.424 0.69751 20 2.703 -2.8895 -1.931 0.053 0.33779 0.61902 -1.49634 -1.83961 -2.2964 0.4734 2.4344 2.5633 -1.4287 0.1708 1.14618 15 1.078 -2.82098 7 -1.455 0.0771 -0.26739 -1.00855 8 2.933 -1.99762 -2.01951 0.5770 1.290 2.09872 0.176 2.63622 -1. New Polyphase Length i 68 69 3 0.520 0.306 0.00357 2.01812 -2.91657 -3.80808 0.0375 1.0834 -0.85139 -0.797 1.65086 1.7318 -1.25934 -1.354 -2.4187 -1.532 0.7560 1.67005 0.04328 -2. continued Length 68 69 70 -2.17547 2.94564 -0.02962 2.094 -0.2603 2.43274 -2.5 i 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 Table 3a.32073 0.1562 -1.091 Table 3b.5716 -2.97088 -1.09061 1.49084 -2.0912 1.24951 -0.2813 -3.8551 Barkers 70 1.1533 11 1.08902 -0.523 2.6002 1.5046 0.7483 1.107 -1.3980 2.81485 6 -2.815 2.38315 2.0508 1.09963 -0.01598 -1.80092 0.6224 2.61698 1.13284 1.7057 -1.77191 -2.00861 0.1933 -1.6898 -1.1530 -0.28353 -2.47052 -2.30729 2.217 -2.34463 0.141 0.9203 -1.7844 0.177 -2.4276 -1.692 -1.25682 -2.075 -2.618 -2.6239 -1.66465 2.90333 -1.850 1.26534 0.694 1.50078 18 1.766 -2.35055 0.163 -0.16998 -0.15889 -1.387 1.872 1.5365 0.2799 2.9506 -1.6279 2.004 -0.6485 -2.16039 0. continued Length 65 66 67 -0.37481 0.056 2.50286 3.1368 0.382 -1.87771 -1.54634 0.10378 -0.08816 -3.1282 1.5801 0.5165 -0.26634 2.65743 -2.15376 -0.07472 -2.1169 -3.69489 13 1.20297 16 2.749 -2.3322 i 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 Table 3b.1478 2.85718 3.376 0.853 1.40911 2.48012 3.1541 -2.12196 2.127 0.942 -2.949 -1.30113 -2.00515 2.44501 -1.58296 5 -2.306 2.458 2.10051 1.64288 12 1.16378 0.688 1.510 -2.18334 2.69645 17 1.105 0.13499 2.824 0.66255 2.38929 0.43504 14 2.139 -2.418 -0.5290 -2.4964 1.4076 2.5231 -2.226 0.22836 2.0920 19 1.4148 0.2047 -0.351 1.04429 -0.36299 -1.06233 -0.379 2.67878 1.38185 2.4067 0.345 -1.93941 0.1069 0.1235 2.19691 -2.546 -3.14218 -0.29756 10 2.381 2.22248 2.13029 2.454 -3.54268 0.5654 -2.6271 1.056 1.07976 4 -0.263 1.98574 -2.1759 -1.0478 -2.8677 0.967 0.7039 1.32013 2.840 0.50735 -0.1702 1.20069 -2.43624 -0.60719 0.659 -2.659 -3.41804 0.8912 -0.784 -2.546 -1.46227 -2.75635 -2.15192 -0.48749 -2.289 -0.9324 -0.40224 1.89255 9 1.9185 -1.331 -2.46549 -0.13762 .9010 -1.01319 0.0184 -0.58034 -1.9425 -2.42523 2.130 2.9365 -0.0150 0.024 -2.031 -1.45249 -2.6488 2.3617 -2.0949 2.80148 0.79157 2.

or they were not found because the methods employed were not capable of ﬁnding them in the time alloted.1116 0.2282 1.5051 0.0806 -0.3015 -1.1418 -1.2759 0.3693 1.6826 1.1339 -0. As with previous efforts to ﬁnd polyphase Barkers.9933 0.3058 1.7022 -1. continued Length 72 76 77 -2.278 -0.0835 -1.527 -2.1562 -0.9669 0.4911 -0.021 8 2.2363 1.7175 -1.5185 2.6545 2. A RE T HE F ISH A LL G ONE ? It is natural to ask why there is a dearth of larger polyphase Barkers in our results.3944 1.1234 1.7501 -0.4643 22 1.8421 -0.6148 2.5668 2.751 2.0082 -2.9381 2.2223 1.1876 -1.1046 19 -2.3613 2.0985 -1.616 Barkers 77 0.6868 -2.667 12 1.1198 -2.1384 -0.2771 2.0373 15 2.1035 -1.1159 III. New Polyphase Length i 72 76 3 -0.8166 3.5054 3.5211 -0.6055 2.6511 18 -1.7053 0.9632 30 0.7288 -2.1631 34 2.3695 2.7383 -0.2246 -2.8598 0.6102 25 -2.7583 36 -1.3892 2.7578 39 1.2108 2.1442 -2.2318 0.4875 -0.7357 -1.297 3.4648 1.9859 2.9202 -0.6 Table 3c.3781 -1.0393 1.0878 -0.7238 -2.3623 -2.0706 2.8063 -0.0662 20 -0.3907 3.2543 0.6939 0.3614 2.3549 7 2.3141 -2.5966 10 1.6986 -2.4486 -2.1393 1.5553 -0.436 -2.5594 -0.7337 0. Although this is .4045 -1.9272 0.1991 -1.5654 17 2.2405 -2.3881 -1.7354 -1.0962 -1.0609 3.9737 2.8675 6 -2.0565 2.5396 i 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 Table 3c.2648 21 -1.1971 2.3064 -0.3394 2.7153 1.8183 2.3757 3.1547 16 2.5782 1. I F M Y N ET C OMES BACK E MPTY.1757 1.8288 2.5173 37 -2.8669 -0.878 0.7485 -1.7974 1.3578 1.249 -2.8018 -2.2554 4 -1.7094 -0.8844 35 -0.6193 23 0.5553 -0.3654 1.6182 -1.4842 1.3423 -0.1582 -2.8147 0.6961 0.5286 2.5541 0.7773 -2.0986 -1.3513 -0.8031 0.8297 29 2.3551 2.5616 -2.5124 -2. it is not possible to determine for certain whether they were not found because they do not exist.8485 -0.7884 1.1287 38 2.3336 0.4647 1.8144 1.151 1.0338 13 1.4187 0.5457 -2.3479 -1.2493 -0.5406 -0.5818 -0.6827 0.7057 2.7853 0.9032 1.8723 -2.3097 -2.3605 1.7806 1.3462 -2.97 33 -2.3807 0.8379 27 -0.7471 24 -0.0729 -1.9241 9 2.9883 -3.3461 -2.6034 1.847 31 -2.6224 0.5133 1.9605 -0.7774 26 2.4539 -0.1198 -0.0655 -0.7293 2.5864 -0.1221 0.9594 5 -1.8794 32 -2.3649 28 -1.0035 1.8776 2.4183 -0.7964 -2.0665 -2.6578 2.0903 1.3454 -1.0554 0.7581 -0.444 -0.2644 2.482 -1.4087 2.3757 14 2.3057 2.9635 -1.2897 2.1797 -0.2608 11 0.4022 0.0879 0.0393 -1.

copious amounts of time and effort were spent in ﬁnding the number of polyphase Barkers for short lengths (below 20). Interestingly. Similar tests were done for lengths 44 and 48. Figure 2 shows the results for the 48 codes found during the 305 trials. We do believe however that our search was relatively thorough at those lengths. Each time the routine was run. It is the authors opinion that the numbers of poly-phase Barker codes not found during this search are not sufﬁcient in number to effect the trend shown in Figure 1. 120 100 80 number of codes In an attempt to collect evidence of this kind. Of the 305 times there were four codes found each of the 305 times. One code was found only 4 times and all the others were found 20 or more times. We believe that if we have made the case number of codes found during this set of 305 trials was exactly the same as the number of codes found during the above-described exponentially increasing search. At each of these sizes the probabilities of ﬁnding individual codes followed a similar pattern. In this test the program was run a large number of times over a period of several days on an 18 element Beowulf cluster. Some were found every time. To see if the exponential increase in effort was sufﬁcient to capture the number of polyphase Barker codes for lengths from 10 to 52. and some were found only a small percentage of the time. 1. The amount of effort was then extrapolated from this lower length to lengths up to 52 chips. In fact a careful examination of Figure 2 suggests the possibility of a relatively small number of further polyphase Barkers at length 52. an additional test was done at length 52. As a ﬁrst step. At the end of this time the program had executed 305 times. The effort was seen to be growing at an exponential rate. evidence can be gathered and presented which suggests the answer to this question. Numbers of polyphase Barker codes as a function of code length these lengths were established.7 not a question that can be answered for certain. . they are plotted in Figure 1. in terms of numbers of polyphase Barkers found. The exact number of times the routine was run had no particular signiﬁcance other than the fact that it was a relatively large number. and we found no further codes beyond those found in the exponentially increasing search. This does not show that there are no more polyphase Barkers at that length. The number of local optimizations needed to ﬁnd all of the polyphase Barkers at 60 40 20 0 10 15 20 25 30 35 code size 40 45 50 55 Fig. and the numbers tabulated. That effort was then expended. a number of polyphase Barkers were found. the The shape of graph in Figure 1 appears to show that the number of polyphase Barker codes is rapidly decreasing after its peak at 38 chips. was determined for sizes from 10 to 52. search yield.

25 and 0.5 degrees per chip were applied. all the Barker seis in line with the authors’ expectation that longer polyphase quences found for lengths 10 to 52 were examined. A more exhaustive study of this phenomenon is trials in which two independent random phases were drawn given in [9] and [11]. This section will show results from by length 40. the binary bit representation for the sequence Doppler sensitivity was examined by applying a progressive phase ramp across each Barker sequences for length 10 to length 52. In each case the trend is for the proportion of computations aimed at assessing each of these issues. Figure 4 shows the minimum. from a uniform distribution. One phase was used to generate a unit-amplitude complex multiple for the sequence. For each of Barker codes can in general be expected to require larger these lengths. This To assess sensitivity to quantization. Frequency of occurrence of polyphase Barkers of length 52 Fig. 2. Figure 3 shows the logarithm of the fraction of trials for which the sidelobes of these sequences obey the Barker constraint. However. generated was then truncated. Both of these operations preserve sidelobe level. as a function of length for 4-bit and 8-bit binary representation. . each code was subjected to 6000 Monte Carlo alphabets. Phase ramps using increments of 0. 6000 Trials for having found most of the polyphase Barker codes in the 10 to 52 chip region. S ENSITIVITY TO D OPPLER AND P RECISION Figure 3 shows that the fraction decreases more slowly for the 8-bit representation than for the 4-bit representation. it is not unreasonable to believe that the polyphase Barker codes could run out at lengths not far beyond those found in this paper. Any pulse compression code will exhibit some sensitivity to about one quarter by length 52.8 350 0 300 −1 250 −2 number of trials log (Fraction) 200 −3 150 10 −4 100 4−bit representation 8−bit representation −5 50 0 0 5 10 15 20 25 30 code number 35 40 45 50 −6 10 15 20 25 30 35 Sequence length 40 45 50 55 Fig. polyphase Barkers to decrease log-linearly with length. IV. The fraction with 4-bit to Doppler and to the number of bits used to represent representation drops below one out of every ten thousand cases elements of the sequence. 3. Percentage of Sequences Remaining Barker for Random Phase Ramp. and the second phase was used to generate a progressive ramp in phase across the sequence.

M. Jan. Elect. vol. −26 −28 −30 Minimum Mean Maximum No Phase Ramp −32 −34 −36 10 15 20 25 30 35 Sequence Length 40 45 50 55 Fig.. 1986. Letters.. It should be noted that there is signiﬁcant variability in the performance degradation due to Doppler effects. May 31 to June 3. poster session. Proc. D. and Bass. [15] Golomb. Inf. pp. 2000 IEEE Intl. Mar. no. all the lengths for which polyphase Barkers exist will be found. IEEE Trans. Apr. P. vol. Proc. E. IEEE. Figure 5 does the same for an increment of 0. In addition we extended the length for which globally minimum ISL codes are likely to have been found to length −26 −28 −30 −32 80. 51. S UMMARY In this paper we have extended the length for which −20 −22 Peak Sidelobe to Peak Ratio (dB) −24 polyphase Barker codes are known to exist to length 70. no. “Constrained optimization applied to pulse compression codes. Manitoba.. and Coxson. CMSMITACS Joint Conference. Proc. no. [14] Langford.R. G. no. 911-912. Statistician.50 degrees. and ﬁlters”. Winnipeg.. Further polyphase Barker codes were found at lengths 72. [5] Golomb. [4] Nunn. Proc. 23. Scientist. pp. 16. ”Best Known Autocorrelation peak sideobe levels for binary codes of length 71 to 105”. 6 Aug. C. S. Amer. vol. “Polyphase Barker sequences up to length 45 with small alphabets”. Side-lobe levels of polyphase Barker codes after injecting 0. pp. 1998. September 1963. 345-382.. 4. 1930-31. and 77. “Solution of the inverse collector’s problem”. 1576-1577. and Zhang. 533-537. [12] Nunn. 190-194. 2005. Radar Conference. 1. IEEE Trans. “Recent results on polyphase sequences”. NY. We do however believe that sometime in the near future. p. IEEE Trans. 2. H. Inf. Math. P. of the IEEE. R. [13] Dawkins. [8] Brenner. VA. “Siobhan’s Problem: the coupon collector revisited”. R. IEEE Trans. 44. vol. and Ferguson. “Generalized Barker sequences (transformations with correlation function unaltered.. vol. “Barker sequences and difference sets”. R. Alexandria.. and in many of the cases exhibited the degraded sidelobes are better than can be expected for even the best of the match-ﬁltered biphase codes. 1256. 2005 IEEE Intl Radar Conf. pp. B. Washington DC. A. l’Enseignement Math´ ematique. [10] Nunn.. R. 2005. Theory. 76-82. 32-35.R.J. vol. [11] Borwein. changing generalized Barker sequences)”. and Welch. and Zottman. 11. Wiley. 35. “Multi-parameter local optimization for the design of superior matched ﬁlter polyphase pulse compression codes”. pp. 34. [16] Eliahou. −20 −22 Peak Sidelobe to Peak Ratio (dB) −24 R EFERENCES [1] Turyn. no. 38. 10 Nov. 392-395. Theory. C. 817-824. vol. Theory. “Longest binary pulse compression codes with given peak sidelobe levels”. pp.. Side-lobe levels of polyphase Barker codes after injecting 0. We do not claim that we have found all Fig. 4.J. 2002. “On Barker codes of even length”... vol. 1965. 44. 5. [2] Levanon.25 deg. Inf. and Langford. July 1989. M.. A. pp. pp. L. 1999. 435-440. Elect. 1994. 366. “Polyphase Barker sequences up to length 31”. and Scholtz. Mayer. and Kervaire. and Mozeson. We also exhibited evidence that the numbers of polyphase −34 −36 10 Barkers reach a peak in the late 30’s and start decreasing at 15 20 25 30 35 Sequence Length 40 45 50 55 a fairly rapid rate. 1991. “Sixty-phase generalized Barker sequences”. [9] Borwein. R.50 deg. 30.. Letters. N. and Win. 76.. N. C.9 −18 Minimum Mean Maximum No Phase Ramp V. of phase per chip maximum and mean PSL observed for the sequences of each length with a phase ramp using an increment of 0. p. Feb. vol. “Barker sequences”. vol. vol.. 51. S. S. Inf. and Ferguson. of phase per chip −18 lengths for which there are polyphase Barkers. Radar Signals. M. S. 74. pp. pp. 2008. 2007. vol 45. . no. E. 4. Oct. 1564-1567..25 degrees. [7] Friese. Theory. pp. “Polyphase sequences with low autocorrelation”. [6] Golomb. [3] Kerdock. 27.. pp. R.

- Cs High Performance Comp
- The User Guide Chapter One
- 07 Processor Basics(1)
- Hybrid Dwg Technique
- Lecture 01
- Energy Aware Soft
- SFC51[1]
- 2
- Chapter 1 Computer Abstractions and Technology
- Report (CSA)
- microprocessor lecture
- 4915
- Manual Omega 92
- Digital Circuit Testing and Logic Analysis
- UploadedFile_130350129202317180
- ArchitectureR3-12-03
- Desktop 6th Gen Core Family Spec Update
- PSCSUmanualE
- 750431
- Notes
- Intl
- f
- Hardware
- Untitled
- Mhz
- MELJUN CORTES COMPUTER Archetecture & Data Communication
- ARM_MCUL1
- Plc Final Project (Uthm)
- mlaba02ex6
- Pipeline

- MODELING AND SIMULATION OF MOS MIS MATCH.pdf
- Note_20131126_092807
- Base Paper (1)
- rtosconcepts-1232972644787004-2
- lec1
- 09
- fpgas-an-overview-1192695858110009-3
- MOS MISMATCH-2.pdf
- Note_20131126_092807_1
- vlsi
- Undrstandng DSP.pdf
- Verilog AMS Tutorial
- col10553-1.3.pdf
- ADP
- Electron
- MIT6_012S09_lec01
- Mat Lab Commands
- Cadence
- AMS IC Design- [Bath]
- 11 MOSbiasing
- Trans Analysis
- File 5eb1cd9f24a07a58 Recommended Books for ECE G..
- AMS IC Design- [Bath]
- SLC_vs_MLC whitepaper
- Lecture5-MOS
- 465741-FPAA-Solution-for-Sensor-Signal-Conditioning
- Designing Analog Chips
- Archimedes

Sign up to vote on this title

UsefulNot usefulClose Dialog## Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

Loading