Note: Descriptions are shown in the official language in which they were submitted.
WO 96/37060 2 ~ 9 5 4 5 3 PCTIUS96/07023
METHOD FOF SPREAD SPECTRUM COMMUNICATIONS
r ~c~. uu~ I of the lu.. ~--n
Field of the l~,~. ~-
The present invention generaily relates to spread spectnum, digital
communications, and pdli' I' 1~' to a method for l~ " , Ig frequencies for
frequency hopping lldll_,, ' ' .1:7.
S- ~ - ll of the Prlor Art
Various ~lg i ""s exist for col~ 9 digital pulse, spread spectnum
10 communications. The principal goal is that of encoding frequency selection for
providing communications which are effective in a fsding multipath e.,r;.ur""t",l
and which are secure from both jamming and e -/~,sd~u~.~ ;"9. Spread spectrum
t,al, ,:, typically tai-<e narrow ban 'v.' ill, symbols and spread them out over
a much wider ban : ilh tlhan is neu,:.:.aly for the rate of data lldll~ O n.
15 Unfortunately, "ospl,efic multipath effects are frequency d~,an~ t and thus
are acc~" ~ i by any spread spectnum approach.
In frequency hopping, spread spectnum systems, the carrier frequency is
changed in acco,.ltl"..~ with a p,~ ' ,. ,ed sequence controlled by one of a
variety of codes. The ~dll- Il.'-' ' III~U'~ iS thereby contained in eitherthe
2 0 1l an~l 1 l" ' ' frequency or in the frequency difference between the lldl _ "'
pulses and a reference frequency. The spread spectnum signal may be
received, decoded and demodulated by a receiver constnucted to employ the
same frequency hopping code, once the same p,. , . ,ed code is
sy~ lul ll ul liLed at the receiver with the ll al l~ code. In one prior art method
25 described in US Patent No. 4,612,652, a matched pair of pseudo random
number ~ene, :. are used to '~: ,e the maric and space frequencies of
both a transmitter snd a receiver. The direct generator output and a delayed
version thereof are used to provide separate random frequency sequences for
each of the mari< and space data signals, I--r ~ .31y. The separate gene,
30 are :~nl~:lllulliLud between llall~lll and recelver, rssulting in sy~n,ll,ul);~ l of
both the mark and space frequencies. Unfortunately, the required _,.,-,hlull;...ll
adds Col l I~ .Aity to the 1l dn' -.. ."' - and receiver and their i" lk n and can be
the cause of lldns...:~.s;Jn failures.
WO 96/37060 !t f~ h ,~ PCT/US96/07023
2 1 9 ~ ~ 5 3
S ~y of the Invention
AC~ ly, it is an object of ;he present inventiGn to provide a method of
frequancy hopping, spread spectrum communications.
It is another object of the present invention to provide such a method
5 which is capable of co"".anl - ,9 for multipath and fadins lldll~ i
env;, u, " "e1"ts.
It is a further object of the present invention to provide such a method
which enables error correction and recovary of lost data.
It is a still further object of the present invention to provide such a method
10 which does not require v,, lvl "u"i~", or 5~ IUI ,i~ution between the lldllslllill
and receiver.
In one form ths present invention provides a method for frequency shift
keying l,d"s",;__;ans of the type including choosing different frequencies for
~I dl)SI I " ~- 19 a series of pulses of energy clt,l.encle, It upon the digital data content
15 to be lldl 1~1111-- ' d, wherein the improvement cu, l~ bes the step of . ",i";"g
the frequency choices for a C~hserll~ent pulse l,d"~",i~ion l~pende"l upon
digital data content of at least one previously l,d":,n, ' pulsa.
Brief Dc~ r o~ the D._: _
The present invention is illustratively described in reference to the
20 appended drawings in which:
Fg. 1 is a diagram of an encoding method configurad in ac- ("dance with
one e"ll,udi",~,lt of the present invention; and
Fg. 2 is a diagram of a received lldll~l~lis:,io" decoded in accor.ld"ce with
another e" IbG~ l 11 of the present invention.
Detalled De.. v.i, of the D,_. ~ s
In one el,ll,oli",e"lt of the present invention, a wide frequency bdll :llh
is divided into a p,. '~ -",; ,ad number of fraquency channels or bins which areused to distribute sequentially lldll:-lll'-- ' pulses of a signal. The frequency
selection or hopping pattern is partially p,t:d~lt~ .""ed and includes alternate3 o sl Ihseql ~4nt frequency paths d~,al ,ui"g upon the data content of each
llul)~lllill~d pulse. Normally the lldll:,lllitltld frequency ~u~ ,u"ts a logical data
value such as "1 " or "O". In the s~ Ihseql IQnt fraquency hop there ara two pairs of
frequencies with each pair cc~ uùl,di, ,g to one of the possible data values of its
i""" p,~de~e ~ol hop. Which pair of frequencies are used to transmit the
............. .. . ..... _ . . .. .. _ .. . .... _ ............. ... .... _ ... ... _ .... .... _ .. _ . _ .. _ .. _ ... ..
WO 96/37060 ' ~ , PCTIUS96/07023
21 ~5~53
s, IhSQr;nQnt data bit depends upon the data content of the previous hop. A ~1
sent on the previous hop wiil dictate use of one of the pairs and a O will dictate
use of the other of the pairs. The individual channels in each frequency pair ofy the s~hseq~ ~snt hop represent the same logical "1 and O and the data content
5 of the subsequent hop deten " ,es which channel of the d~ ~.. " ~ ed pair is used.
The present method is l,~F ~ to the l,d"a",is5ion of more than one
data bit at a time. Trdnal, ,9 two data bits at once requires four possible pulse
frequencies to represent the four possible .;on iilions of two data bits. The
5llh5erjllent pulse lldnsllli~si(", uses foursets of fourfrequencies, orsixteen in
10 all, and d~ '~ . ",i";.,g which set of four is used in response to the four possible data
cor ti~ ~s of the previous bit.
Fg 1 is a diagram showing a simplified encoding method in ac(~or id"~
with one ~",i,odi",t",l of the present invention. Each row of the diagram
se";~ a sequential pulse 1, dl ,~",is:.ion intervai. The diagram Itl,ul tlsel ,.s a
15 single bit lldnsllli:,:,ion rate meaning that only one logic bit is l,tu~sr" ~ 3d in each
interval. The diagram is repetitive so that the four lldllslllia~ion intervals are
repeated in sequential passes.
Each logic value shown ,t"J,t,:,e"ts a separate frequency bin within a full
lldnSIIIi:~.iOII bdn 'I ,. Logical pairs of bins are shown in pd,ar,~i,t3ses,( ). The
2 0 example shows a total of thirty different frequency bins. Within each of thelldll'llli:~aiull intervals, rows, only one of the co,,~,uol,di,,g frequency bins is
used to transmit a pulse. Which frequency bin is used depends upon the data
content of the current bit and the data content of the bits l~d";.~;lled in the
previous intervais. The specific pair of frequency bins which is avaiiable for the
25 .cllhseqll-Qrlt lldllSIIIia:,iOn interval from each frequency bin is co""e..tt,d thereto
by dotted lines. From the fourth lldn~lllia ,iUn interval, each pair of frequency bins
is treated as if it cor,t,:",ol,"~ to the first lldll~llli~aioll interval and the sequence
repeats from the second lldll~ ion interval.
Thus in the first intervai, a puise is lldll:,lllillttd in either of the bins shown
30 l~ sell ,9 the ,t,:",e~.ti~o logic vaiue. in the second lldnallli:,~ion interval
shown, which pair of bins is available depends upon the data content of the first
interval. Which bin of the availabie pair is used depends upon the logic vaiue to
be lldnallF i in that second interval. This scheme continues to the l~dn~ aion
of a puise in the fourth intervai. The ,t ~,e~.live iogic vaiue of the pulse actually
_ _ _ _ _ _ _ . .
W096/37060 21~4~ PCT/US96/07023
l,d"a", ' in the fourth interval is treated as the logical equivalent of the first
intorval and the pairs of frequency bins used in the actual fifth Ll~ blll;abiuninterval ars selected from the second ~,dnb",iabion interval shown.
By the method described, an actual lldl IblllibbiOn signal is identitied by
5 d~ r",i" ,9 the length of the sequence of lldllbllliabiUn intervals in which pulses
are detected in the proper frequency bins. Sequences of pulses which are not
sufficiently sustained are discarded as non-signals. Further, an actual pulse
lldns",illt,d in the fourth interval may be used to reconstnuct pulses missing from
the second and third intervals by simply following the dotted lines back through10 the intervals.
The actual frequencies which represent the frequency bins shown in Fig. 1
may be perio.lically l~dbbiylled to further prevent decoding of an illttll~ d
message.
Fig. 2 is a diagram ,~,ur~be, ,9 an example of potential signal pulses
15 received overa portion of a message lldllblllilled in ac~iuldd"ce with one
e",~u.li",er,l of the present invention. The horizontal axis It~,ulubellts a portion of
the total bdl ' ;dll~ of the lldlls",ission with different horizontal positions
,~.,t,se" ,9 different frequency bins or channels within that total ban ' :-ll,. The
vertical axis of the diagram is labeled with successive pulse l,d"b",ibbion intervals
20 to represent the successive frequency hops of the l,d"s",ibsiun. Each circle
shown on the chart ~ bel IS energy detected in the ll~b,UUUti~rO frequency bin
and l,d"a",ibbiùn interval, by a wideband receiver ,,w,, i"g at least the portion
of the bdr, ' :dll~ shown. The lldl IblllibbiUn shown in Fig. 2 is a two bit
Il dl Ibl llibbil)n rate for each frequency hop or 1, dnS~ I ~ission interval. That is, two
25 binary bits of datamtl~ belllillg a total of four possible logical states are lldllblllilltld in each interval.
The first lldllblllibbiUn interval shows energy detected in bins 10 and 12.
The second l,d"s",i~bion interval shows energy detected in the nine frequency
bins for which there are circles shown. Those bins which are logical successors
3 0 to the bins 10 and 12 are respectfully conneultld thereto by dotted lines and are
labeled in quotes (7 with their logical value, i.e. ~01". Because energy,
,t,l.,~se, ~ ~ by the circles, is shown to be present in all of the logical successors
in the second interval as well as bins 10 and 12 of the first interval, s~ ~hss~llJsnt
lldnblllibbiun intervals are neu~bbdly to detennine the correct data sisnal tor the
WO 96/37060 ~2 1 9 5 4 5 3 PCT/US96/07023
~;: 5
first and second intervals. In the third interval, ensrgy is detected in the five
frequency bins shown. Because the detected energy in the third interval can onlybe a logical successor to the energy in bin 14 of the second interval, it is known
that bin 14 is the corred data signal for the second l,d"a",iaaion interval. All of
r 5 the remaining energy detected in the second interval can be discounted because
none of the detected frequency bins have a logical successor in the third interval.
Likewise the proper signal pulses may be traced back to bin 10 of the firs
interval. Also, the correct signal path continues to develop through s~hse~ snt
lld"s",isaion intervals in frequency bins 16-22.
Simultaneously, another data path is shown to be received in ener~qy
detection circles 23-26. Because energy is continually detected in the proper
frequency bins of 5llhceqllent intervals, it is ~l~t~r",i"ecl to be a valid signal. In
this manner simultaneous signals may be l, dr,s",;tl~d in a multi-user network.
This second signal also d~llluilall - that the encoding is not time def,ende"l
15 and, therefore, does not require_)"~il"~., ~ 1.
The logical sequence of frequencies to be used is cl~tt,r", ,ed in response
to the l,dnsi"i~ion bit rate and the available spectrum to be used, in a manner
with makes full use of the available spectnum prior to allowing any frequency
repetition. Such repetition limits the number of frequency hops which can be
20 reconstnucted if lost. The number of sequential pulses which can be lldns",ill~d
prior to any frequency repetition is the number of hops that can be reconstnucted
from the method of the present invention. The use of frequency bins in this
manner allows discontinuity of the full l,d"~",isaiùn ban ~' : 'Ih. Thus continuous
emitters, located in the middle of the full lldllalll;~iùll spectnum, only block their
25 limited balid~kllll portion of the spectnum.
As ",e"tiol)ed, spread spectrum techniques in the H.F. band (3-30 MHz.)
are readily sl~ ~Fr ~ '. to cornuption from frequency cle~,end~"l ionospheric
con.litio"s. Different frequencies penetrate the ionoaphe,~ to different altitudes
prior to reflection back to earth and the same frequency may also be reflected
3 o from different altitudes. The first condition upsets the sequence of pulses
l,d,,ar,, ~ at different frequencies and the second condition introduces
additional, or multipath, signals which are likewise differently timed and out of
sequence. The practice of the present invention therefore includes the periodic
sounding of the iouua~.l,er~ with a p,~ tt"",;"ed sequence of lldnsllliaaioil
WO 96i37060 t P ~ PCT/U596107023
21954,53
pulses in as many as all of the separate frequency bins used in the full
lldllallliaail:)ll bar, ;JII,. ne . Ii~ ,i"9 each of the received pulses with the
p,. ",i"ad sequence provides delay and multipath cor,~ctions for the
lldllal"iaaiun er,~r;,u"l"e"l. In one ~",bodi",t:"l, this sounding may be executed
5 every second to thereby allow COI I ~I~an- n for the vast majority of varying
k"~o:,ulla~ic .;ur, ~ns. After the sounding, the actual data signal is lldllaulilled
and the received pulses are realigned in time in response to the sounding results
and prior to decoding in acco"~d"ce with the present invention.
The encoding method of the present invention is also colllr with bit
10 dia~J~Iaiùn, which is an o,ll,ùgor,dl or i"d~Jel"~tl"l rnethod of encoding. By
example the data bits of a sisnal to be lldilalllitt~d are first aligned, at laast
~I,eo,tlticall), as sequential rows of a matrix. The data is then read from the
matrix, for encoding in acco" d".~a with the present invention, as sequential
columns of bits. Thus an entire missing column of data may be lost in
15 lldnsllliasiun and it will show up in the recaived and decoded signal as single bits
lost in sequential data strings. Such i".la~,anJ~, It or ol Ihogu, Idl decoding
techniques may be taken dliV... ' ~ of to iteratively decode, correct recode rmd re-decode the signal and thereby reduce gaps in lost data.
Cor,~v; 101~
2 0 The present invention provides a method for encoding spread spectnum
pulse l,d"s",i~aiùns which are cor"uer - d for varying dlllloa~Jllalic conditions.
The present invention further provides for recovery of lost signals which operates
both ill.l~.alldr~"lly and in co"lb;r,dtiun with other known techniques to provide
:,y~ ~uO :ir results when compared to the recovery ability of individual techniques.
25 The present invention does not require any frequency or time sequence
alignment between l,d"a",;tt~r and receiver priorto l,d"a",iaaiun thus reducing
complexity and lldllal"iasion delay and failure. This l,dns",illt"/lt,caiver
i"depel "~t"lce further enables the reception of simultaneous signals by the same
receiver thareby supporting radio network ~ ns.
The r~",l.odi",~"ts described above are intended to be taken in an
illustrative and not a limiting sense. Various ". ~.la and changes may be
made to the above e"lbodi",6rlta by persons skilled in the art without departingfrom the scope of the present invention as defined in the appended clairns.
