Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
WO 96/10300 22oo7 I I PCT/US95/12150
Title: Enhanced Frequency Agile Radio
BACKGROUND OF THE IlvVIIVTION
This invention relates to robust methods of frequency re-use
and frequency sharing which prevent independent radio
systems from co-interference due to frequency crowding of available
radio bands, and, more particularly, this invention relates to a method of
frequency agility which can provide a low cost solution to many burst
mode and continuous data communications applications, including security
systems, fire systems, access control, energy management, remote control
of model planes, remote process control, traffic light control, remote
power meter reading, voice communication, radio location, or local area
networks.
DESCRIPTION OF THE REI.EVANT ART
The use of spread spectrum communications and techniques for
diverse commercial and civilian applications has increased in recent
years. By utilizing such techniques to minimize mutual interference and
to provide anti-jamming advantages to multiple-access communications,
as well as aiding in extremely accurate position location using
satellites in synchronous and asynchronous orbits, spread spectrum
techniques are known to offer the advantage of improved reliability
of transmission in frequency-selective fading and multipath environments.
U.S. Pat. No. 4,799,062 to Sanderford, Jr. et al. teaches that
multipath in urban areas poses a problem for accurate position location,
which my be overcome by using a method of synchronization of
transmissions and unique identification codes to derive relative
ranging times for determining position. Compensation for multipath
may include spread spectrum techniques.
U.S. Pat. No. 4,977,577 to Arthur et al, has a wireless alarm
system using spread spectrum transmitters and fast frequency shift
keying for achieving a coarse lock and a fine lock to the spread
spectrum signal. By using spread spectrum techniques, such wireless
alarm systems are highly reliable and provide a safety margin against
jamming and undesirable interference. Other applications of spread
spectrum techniques to commercial uses promise similar
advantages in reliability in communications.
Methods for the serial search and acquisition of utilized spread
spectrum frequencies are well known in the prior art, as shown in M.
~
WO 96/10300 22007 11 PCTIUS95/12150
K. Simon et al., Spread Spectrum Communications, vol. 3, pp. 208-279,
Rockville, Md.: Computer Science Press, 1985. In addition, M. K. Simon et
al., supra. at pp. 346-407 teach spread spectrum multiple access
techniques such as utilizing ALOHA random access schemes.
OBJECTS OF THE INVENTION
A general object of the invention is to achieve superior
jamming resistance compared to other spread spectrum means.
Another object of the invention is to allow multiple systems to
co-exist without undesirable co-interference.
Another object of the invention is to minimize the effects of
data collisions when a system supports numerous non-synchronized ALOHA
protocol transmitters.
An additional object of the invention is to operate within the radio
band allowed by the FCC with minimal cost and minimal frequency
setting components.
A further object of the invention is to provide techniques suitable
for a high level of monolithic circuit integration.
SUMMARY OF THE INVENTION
According to the present invention, as embodied and broadly
described herein, a frequency agile method is provided which has a low
cost solution to many burst mode and continuous data
communications applications, including security systems, fire
systems, access control, energy management, remote control of model
planes, remote process control, traffic light control, remote power meter
reading, voice communication, radio location, or local area networks.
In remote monitoring applications, the frequency agile radio
system typically includes one or more centrally located data collection
receivers with one or more remotely located transmitters. In control
applications, one or more centrally located transmitters may communicate
with a plurality of remotely located receivers. Further, the system can
provide two-way polled type communications where each data node
requires both a receiver and a transmitter.
The method for providing frequency agility includes
using a frequency-agile transmitter and a frequency-agile radio system
for sending a message-data signal by selecting a single pseudo-random
07.
WO 96/10300 2' 2 0 0 7 1 1 PCT/US95/12150
frequency on which to transmit, by generating a preamble signal on a
single carrier frequency for modulating message-data, by transmitting the
preamble signal for a pre-set preamble time for allowing an
appropriate frequency-agile receiver to lock-on the preamble signal,
and by modulating the preamble signal with the message-data signal to
produce a modulated signal. The message-data signal is defined herein to
be a signal having message-data.
In addition, the method for providing frequency agility includes
using a frequency-agile receiver in the frequency-agile radio system for
avoiding occupied radio-frequency channels in a radio spectrum by
scanning the radio spectrum, identifying occupied portions of the radio
spectrum, updating information identifying the occupied portions,
storing the updated information in memory means, associating a time-out
period with the stored occupied portions, and skipping over the
occupied portions of the radio spectrum during the time-out period in
response to the information and while receiving with the frequency-agile
receiver.
Additional objects and advantages of the invention are set forth in
part in the description which follows, and in part are obvious
from the description, or may be learned by practice of the invention.
The objects and advantages of the invention also may be realized and
attained by means of the instrumentalities and combinations particularly
pointed out in the appended claims.
BRIEF DESCRIPTION OF THE DRAWINGS
The accompanying drawings, which are incorporated in and
constitute a part of the specification, illustrate preferred embodiments
of the invention, and together with the description serve to explain
the principles of the invention.
FIG. 1 shows a configuration for a frequency agile
transmitter/receiver system;
FIG. 2A and 2B is a block diagram of a frequency agile transmitter;
FIG. 3 is a block diagram of a frequency agile receiver;
FIGS. 4A-4C shows the frequency sweep over the available
spectrum by a receiver to identify noise, interference or data;
FIG. 5 shows a shunt table;
FIG. 6 iIlustrates a frequency scan algorithm; and
3
WO 96/10300 o PCT/US95/12150
FIG. 7A-7B illustrate algorithms for a trip handler and a time base
shunt decay, respectively.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
Reference will now be made in detail to the present preferred
embodiments of the invention, examples of which are illustrated in the
accompanying drawings, wherein like reference numerals indicate like
elements throughout the several views.
In the exemplary arrangement shown in FIG. 1, the
frequency-agile transmitter is designed to send one full message-data
signal on one single carrier frequency, with the carrier frequency
pseudo-randomly selected. To avoid loss of data throughput, in case the
selected frequency is jammed, the frequency-agile transmitter, after
delaying for a pseudo-random time interval, selects a new pseudo-random
frequency as a next frequency, which is widely separated from the
previous frequency, and re-transmits the entire message-data signal on
that frequency.
The number of redundant re-transmissions and a preselected
average time interval between re-transmissions is programmable, which
allows optimization to a particular installation. For example, if the
number of transmissions required to overcome other sources of data
loss, such as ALOHA collisions, bit error rate, impulse noise, etc., were
5, then by doubling the redundant transmissions to 10 would allow
continued system operation if as much as 1/2 of the available radio band
were jammed.
That is an extreme example, but consider in that example the
time overhead, two to one, compared to the bandwidth overhead, ten to
one or 200 to one, required by conventional spread spectrum techniques
which use all of available radio spectrum simultaneously. The present
invention is much more frequency efficient for equivalent anti-jam
performance. Further, the extra time required to re-transmit with
even 100% redundancy can be overcome by doubling the data rate
which causes only a 3 dB penalty in receiver sensitivity.
The compensating advantage is large for anti-jamming and
frequency re-use. For example, consider two adjacent but independent
frequency-agile radio systems, a first system A and a second system B,
each with single frequency-agile receiver and multiple transmitters. In a
Y
WO 96/10300 2 200 7 PCTIUS95/12150
direct sequence system or a fast frequency hopper, each bit of a
message-data signal is represented on every available frequency. For
example, if a transmitter of the first system A were ON, and were set to
use 63 frequencies, then the data messages signals received by the second
system B are subjected to the multiple access suppression available,
which is 18 dB or less. The frequency-agile radio system of the
present invention responds differently.
If a transmitter of the first system A were ON, and were set to
use 63 hopping channels, then a data message signal received by the
second system B intended for second system B's own transmitter(s)
has only a 1 in 63 temporal chance for being interfered. Since the
receiver is actually occupying only one frequency at a time, a very
small increase to data message signal re-transmissions overcomes the 1
in 63 failure rate due to collisions from the first system A.
An additional method is used to enhance resistance to wide
bandwidth jamming and to selective radio channel fading. A hop list or
hop algorithm is designed to provide a minimum hop distance of 500
kHz or more per frequency step, so that, if the coherence bandwidth of the
radio channel is 2 MHz, then several steps overcome the fade.
Each data transmission is proceeded by a pre-determined carrier
preamble signal with a pre-set preamble period which provides the time
needed by the receiver to sweep the entire radio spectrum available for the
system to use and to lock-on to the preamble signal. Such a sweep can
take from one to three miIliseconds, depending on system parameters.
Additionally, the preamble signal can be modulated with a repetitive
data code, called a PREAMBLE SYSTEM CODE, that identifies the
transmitter as belonging to a particular system, so that, if an associated
receiver locks on an energy source as the receiver sweeps, the
receiver can rapidly demodulate the PREAMBLE SYSTEM CODE to determine
if the energy source or data packet were intended for that receiver. If
not, the receiver can continue to sweep its available spectrum to search
for valid incoming messages. The goal is to minimize the dwell time on
any piece of impulse noise, jammer, or data packets which are intended for
an unrelated system receiver.
Prior to system operation the frequency-agile receiver performs a
sweep of the radio spectrum available for system operation. The receiver
logs all channels with higher than expected received energy levels.
cS''
WO 96/10300 22009 11 PCT/US95/12150
Either signal strength or a quieting detector or a phased-locked-loop
"lock detect" circuit or the like is capable of supplying that information.
Each usable frequency channel is associated with its own unique position
in a memory device. The status of that channel is also associated with
that memory location. The status information includes: 1) if the channel is
"clear" or "jammed"; and 2) how many time counts must elapse prior to
re-sapling to determine if that channel has become "clear".
The jammed/clear indication can also be equipped with an additional
associated counter such that more than one occurrence of channel jamming
is required to set the JAMMED flag. That feature makes the system more
robust against impulse noise. Further, the receiver does not consider
a PREAMBLE SYSTEM CODE of an unrelated system to be JAMMING.
Such occasional receptions are expected and quickly discarded prior
to message "data" demodulation. Sufficient extra time is built into the
preamble signal to allow for several collisions with unrelated
PREAMBLE SYSTEM CODES, and for successful decode of a data
message signal.
Upon initial set-up, the receiver determines the frequencies to avoid
and a table containing information of such frequencies is stored in
the receiver's memory. The table is then transferred via electrical
means to the other associated transmitters or transceivers. The
receiver can determine which PREAMBLE SYSTEM CODES are unused
and available during power-up and initialization. The receiver can
transfer the information on unused and available PREAMBLE SYSTEM
CODES as a list of frequencies on which to transmit to other system
transmitters as well.
Such information transference can be magnetic, electrical,
optical, acoustic, or via a display and entered through a keypad connected
to the transmitter. Once the list is so loaded, the transmitter will not
transmit on the frequencies marked as 'jammed' and will use only the
PREAMBLE SYSTEM CODE assigned. The PREAMBLE SYSTEM CODE can be
data modulated via any appropriate means including
frequency-shift-keying (FSK), binary-phase-shift-keying (BPSK), or
amplitude-shift-keying (ASK) modulation. ASK modulation is less
desirable, however, because ASK modulation requires time spent without full
carrier presence.
One separate method no accomplish co-existence with adjacent
~
z20 7 11
WO 96/10300 PCT/US95/12150
systems requires that both the frequency-agile transmitter and the
frequency-agile receiver be highly frequency stable. Such stability must
be greater than the sum of transmitter and receiver frequency uncertainty,
frequency drift, and data bandwidth. For example, the
frequency-agile transmitter can transmit on any one of 50 frequencies,
each separated by 500 kHz. If the required data bandwidth were 25 kHz,
then 10-20 separate channels could fit in each 500 kHz hop slot.
Therefore, 10-20 co-located systems could co-exist with no
co-interference, and TDMA would not be needed. To attain frequency
stabilization, the frequency-agile transmitter and frequency-agile
receiver would simply offset their normal hop by a pre-determined
number of 25 kHz slots. Each system could determine which slot is
unused and then assign each unused slot to all associated system
elements. This method of transmitting on any one of 50 frequencies does
not work, however, if the accuracy or drift of the carrier frequency
were greater than +13.5 kHz. At 915 MHz, the drift equals 15 ppm.
Absent such accuracy, adjacent channels may cause interference.
Therefore, the offsetting of the carrier frequency of each frequency-agile
transmitter by at least the required data bandwidth plus
compensation for frequency inaccuracies allows multiple systems to
co-exist, which utilize this technique.
Previous attempts in the prior art to integrate
voltage-controlled-oscillator (VCO) components with divider and phase
comparison analog and digital circuits have failed, since the phase noise
generated by the digital divider was induced into the VCO and
phase comparator, but such phase noise ultimately reduces the sensitivity
of the receiver. The preferred embodiment runs the digital divider for
a brief period prior to transmission. The divided down signal is then
used to produce a constant frequency error offset term to the VCO.
The digital divider is then disabled after the frequency error offset is
measured so that the VCO can run open-loop without being subjected to
phase noise of the divider harmonics. The receiver is able to
compensate for any short term frequency drift during transmission.
FIG. 1 depicts a frequency-agile receiver 101 which can be used to
accept transmissions from a plurality of frequency-agile transmitters
103. The frequency-agile transmitters are non-synchronized and must
depend on ALOHA type communicating redundancy to ensure data
7
WO 96/10300 I PCT/US95/12150
receptions. Additional optional frequency-agile receivers 102 may be
included to expand system coverage or add spacial diversity to reduce
fading.
FIG. 1 depicts a frequency-agile transmitter 104 sending
message-data signals or command signals to a plurality of frequency-agile
receivers 105. Each frequency-agile receiver is equipped with a unique
address by which to accept commands and data intended for that unit.
Once again, transmitter data redundancy is required to ensure reliability.
FIG. 1 depicts a system whereby a frequency-agile transceiver 106
may be linked to a single remote transceiver or by a plurality of
remote transceivers 107. In such a system, each system element is
associated with a unique address or identification number.
Any number of appropriate two way communications protocols
may be utilized on such a system including poll-response, reservation
request, report by exception, etc.
As illustratively shown in FIGs. 2A and 2B, processor means, which
may be embodied as a microprocessor or custom digital integrated circuit,
is used to generate all transmitter timing. In order to initiate a
transmission, the microprocessor 201 first selects the pseudo-random
frequency to be used, either by table look-up or by an algorithm.
Once chosen, the microprocessor 201 selects the appropriate
digital-to-analog hop setting and outputs the hop setting to the
digital-to-analog 203. The digital-to-analog converter 203 in turn sets a
new voltage level onto the voltage input 218 of the voltage control
oscillator (VCO) 207. The microprocessor 201 then enables 213 the VCO
207.
The VCO 207 then begins to oscillate at the frequency selected by
the voltage control input and the frequency setting resonant element 208.
The frequency setting element can be any appropriate resonant or phase
delay device(s). The quality factor Q and temperature stability of the
resonant element must be great enough to prevent drift outside of allowed
FCC bands. The maximum drift and frequency uncertainty must be
determined. Then a "guard band" for larger than the maximum
uncertainty must be provided on both sides of the intended
transmitted bandwidth so that inaccuracies in the frequency setting
elements do not cause transmissions in disallowed frequency bands.
If the frequency setting element 208 were not stable enough to meet
~
22007 11
WO 96/10300 PCT/US95/12150
these FCC requirements with a reasonably small "guard band", then an
optional divider 209 can be used to lower the carrier frequency to rates
which can be digitally inputted and counted by the microprocessor 201
to determine the frequency of VCO operation. The microprocessor 201 can
then generate a frequency error offsetting term to the digital-to-analog
converter 203 to generate an offsetting voltage. A voltage input of the VCO
207 receives the offsetting voltage so that the VCO 207 is then adjusted
by the offsetting voltage to within a required tolerance range.
This method has the advantage of not having to run closed-loop, nor
be highly accurate, since this method only needs to achieve the "guard
band" requirements so that inaccuracies in the frequency setting
elements do not cause transmissions in disallowed frequency bands.
Further, if the VCO 207 drifts during the brief message, then the receiver
can track and compensate for the drift after the receiver initially attains a
lock on a transmitter.
Once on frequency, the microprocessor 201 enables subsequent or
final radio frequency amplifiers, embodied as the radio frequency (RF)
output 212, which in turn generates a signal on the antenna
213. The microprocessor 201 then modulates the carrier frequency with
the PREAMBLE SYSTEM CODE, which can be a repetitive five bit code which
contains part of the transmitter's unique larger system code, or
identifier code which is sent during the data portion of the
transmission. The modulation of data can be FSK via resistor R2 206
or BPSK 210 or ASK 212, or any other appropriate modulation means.
The transmitted preamble 214 is sufficiently long enough for the
receiver to a) search the entire radio band available, b) lock-on the
transmitter carrier, and c) validate the PREAMBLE SYSTEM CODE 217.
Further, preamble time is provided so that unrelated PREAMBLE SYSTEM
CODES or impulse noise can be analyzed and rejected with enough time
reserved to recognize and intended PREAMBLE SYSTEM CODE. Once
the preamble is complete, the transmitter then sends its
Message-data signal 215 via one of the above data modulation means. The
message-data signal is then followed by a cyclic-redundancy-check (CRC)
216 error correction/detection code to ensure data integrity by detecting
and correcting error bits appended to the message-data.
If the frequency setting process were highly stable, then the
four least-significant bits of the digital-to-analog converter 204 can be
WO 96/10300 PCT/US95/12150
used for channel selection, providing for system co-existence
on a non-interfering basis. These adjacent channels can be thought of
as the last three or four significant bits of the digital-to-analog input to
the VCO 207. The most-significa.nt-bits are controlled by a pseudo-random
"hop frequency" generator. The least-significant bits stay fixed and
cause a permanent but small frequency offset in the VCO 207.
Providing both the receiver and transmitter use the same offset, the
two are able to communicate. A system with a different offset is
suppressed by the frequency selectivity of the receiver and lignored.
This hopping technique can be readily made hybrid by
additionally modulating the VCO 207 carrier with other direct sequence
methods.
Referring to FIG. 3, a 915 MHz modulated carrier is introduced
into antenna port 300 then filtered and amplified by section 302. If
spatial diversity is desired, an optional section 301 can be added and
controlled by generally known means to lower occurrences of selective
fading.
Mixer 303 receives a local oscillator signal which is generated by
a voltage controlled oscillator (VCO) 304. The frequency of the VCO 304
is set by a frequency setting element 305 and by the voltage preset on
the voltage control input, VIN 315. The voltage control input VIN 316
voltage is generated by a digital-to-analog converter 309 or by
any other appropriate linear means which can produce a controllable
voltage ramp.
If the frequency setting element 305 has a poor absolute
accuracy, or time, or temperature drift, an optional divider 308 can
be used for correcting frequency error. The divided down result is
compared to the crystal 315 of the microprocessor 314 which then
produces a frequency error offsetting voltage via the digital-to-analog
circuit 309.
The down converted output of the mixer 303 can be immediately
demodulated or passed first through an optional second conversion state
306.
In the preferred embodiment, prior to data demodulation, the
signal is first band-limited by a fixed frequency and bandwidth filter
307. The signal is then decoded by the data demodulator. The demodulator
means 316 or 310 must match the transmitter's data modulation utilized.
1o
WO 96/10300 200 7 11 PCTIUS95/12150
In the preferred embodiment, a phase lock-loop (PLL) 310 is
used to detect FSK modulated data. The available frequency spectrum is
initially swept through and monitored via a wider PLL loop intermediate
frequency bandwidth selected by wide/narrow bandwidth select 311, so
that the entire available bandwidth can more quickly swept through. The
energy detector, carrier lock-detect, quieting detector or equivalent means
must have a very rapid response time, which should exceed the impulse
response time of the filter 307 that is, 1/110 kHz=9 microseconds, in this
example.
Once the carrier is initially detected, a more narrow bandwidth
can be selected by 311 which is more representative of the data bandwidth
required for the elected data rate. This narrowing of the bandwidth
improves the carrier to noise ratio. The microprocessor 314 or other
linear means then can be used to close the frequency control loop and
track and compensate for further transmitter or receiver drift components.
If the drift components are minimal, and the transmitted message
brief, a frequency control loop will not be needed. At this point, data
can be demodulated. Alternatively, if the PREAMBLE SYSTEM CODE or
some form of transmitter identifier, does not match, noise and interference
in occupied portions of the radio spectrum can be identified and skipped
over while the transmitter transmits.
If the transmitter is hybrid modulated with ocher spread
spectrum modulation, such as direct sequence, then direct sequence
demodulation must also be added after the hopping frequency is locked
on. This would be an opportunity to include in the combination a
simple parallel correlator means to decode data. Such a combination
would be well suited to time-of-flight radio location applications.
The hop sequence would be optimized for anti-jam and the parallel
correlator optimized to "time stamp" the incoming message.
As an alternative to controlling the voltage input of a voltage
controlled oscillator 304, a numerically controlled oscillator or a frequency
synthesizer may be used, or any method as is known in the art which can
select a frequency based on a control input. If AGC blanking is employed
by 'turning off' the front-end of the receiver, then the rise and fall times
of the amplifier 302 and/or circuit 301 can be controlled in order to reduce
unwanted bandwidth expansion from the effective amplitude modulation
resulting from the rapid blanking control signal. Alternately, the narrowest
WO 96/10300 200" 11 PCT/US95/12150
IF, or baseband, filter can be dumped during the AGC transition time.
As an alternative to the D/A converter 309, dedicated hardware can
be implemented which provides a monotonic increasing or decreasing ramp
or saw tooth waveform. Such implementations are well known in the art
and can be fashioned from operational amplifier integrators.
The frequency agile receiver voltage controlled oscillator 304 also may
be designed such that the narrowest IF filter is not the last one in the
receiver's IF line up. It is further possible to use the voltage controlled
oscillator 304, or an equivalent, to sweep a second local oscillator or any
other intermediate frequency conversion stage in the receiver's IF line-up.
The "wide/narrow bandwidth select" signal 311 could alternatively be
used to select a "chirp/conventional" type filter. The phase-locked-loop
circuit 310 could be replaced with any of the IF/baseband filter approaches
described herein. The D/A converter 309 could additionally provide the AGC
control signal with appropriate sample and hold circuits, or an additional
D/A converter could be provided for that purpose.
FIGS. 4A-4C illustratively shows the frequency sweep operation of
the present invention. In normal operation, the receiver sweeps over
the available spectrum to identify cases of noise, interference or data, and
to store such identified portions of the spectrum in a jammed channel
list.
When no data and no new interference since the last memory
update of the jammed channel list is detected during an associated
time-out period and after a plurality of failed attempts to detect new
data or interference have occurred, the VIN 316 input appears as
depicted by FIG. 4A. The VIN will linearly sweep 401 until the
microprocessor 314 determines, via the microprocessor's list stored in
memory, a frequency to skip over 402 while the receiver is receiving. Once
the maximum frequency point is reached, the VIN 315 input reverses the
direction of sweep 403, in order to minimize the required impulse response
of the VCO 304.
If new impulse noise or jamming were detected 404, such noise or
jamming causes the CPU's 314 algorithm to temporarily stay on that
frequency. The algorithm then attempts to decode a 5 bit PREAMBLE
SYSTEM CODE. If that were not possible, then the VIN 316 sweep
resumes its normal path. The PREAMBLE SYSTEM CODE can also be Grey
coded or Manchester encoded so that impulse noise could be more
/a
22007 11
WO 96/10300 PCT/US95/12150
rapidly detected as an illegal sub-sequence without having to monitor all 5
bits.
If the PREAMBLE SYSTEM CODE 217 matches, then the VIN stays
constant 405 or in-track with a drifting input frequency 300 if a
frequency locked loop (FLL) is used, so that transmitter and
receiver frequency drift is compensated. The processor 314 then
attempts to decode a data message 215 and CRC 216. After completion
of the message decode, the normal search algorithm resumes and the
previous path of VIN 316 continues.
Performance can be further enhanced by not skipping over the
occupied portion or portions of the band, but rather by dynamically
predicting future noise and jamming signal strengths and then only
attempting to decode future signals which are higher than those predicted
levels. If a future desired signal were lower in power than a jamming
level, then the signal-to-noise ratio is too low for the data to be decoded
and, therefore, the frequency agile receiver should avoid dwelling on those
frequency bins. If, on the other hand, a future desired signal were higher
in power than a jammer or a background noise level, and were higher by
an amount at least equal to the carrier to noise ratio required to decode
data at an acceptable bit error rate (BER), then the frequency agile
receiver should dwell on that frequency and attempt to decode a PREAMBLE
SYSTEM CODE and/or voice or data.
The frequency agile receiver seeks to eliminate dwells on frequency
bins which contain unrecoverable messages. The frequency-agile receiver
should not waste time during the scanning portion of the algorithm, as
such waste may prevent the receiver from finding an otherwise recoverable
desired signal residing in an unoccupied portion of the spectrum later in
the frequency sweep. Therefore, if the frequency agile receiver had a
'perfect' prediction of the background noise, jamming and interference
contained within the frequency band in use, then the frequency agile
receiver algorithm would be able to optimally either dwell on a new signal
which exceeded that prediction, or minimize wasted dwell time and continue
the sweep into other potentially useable areas of the spectrum. The goal of
this technique is therefore accurate prediction. The receiver's algorithm
should seek to use information gathered both in the current frequency
sweep, as well as information collected from previous frequency sweeps, to
make the best prediction of what the background noise, jamming and
13
WO 96/10300 22007 11 PCT/US95/12150
interference environment will look like in the next successive frequency
sweep, which may be only one or several millfseconds in the future.
If the types of interferers could be identified, then prediction could
be made easier or more accurate. There is a wide breadth of potential
jamming and interfering sources which may appear in the unlicensed
portions of the radio spectrum which are attractive for operation due to
FCC regulatory considerations. The types of jammers which may appear in
these bands include on/off keyed devices, and fast and slow frequency
hopping devices, including data sets and portable telephones. Portable
phones include those which find an initially unoccupied frequency and
remain on it for the duration of a conversation, as well as those which are
continuously frequency-hopped according to the FCC's definition. Such
frequency hopping devices are Lfkely to dwell on any single frequency in
the range of 1 ms to 400 ms.
Other potential jammers include direct sequence devices operating at
either low or high baud rates. The low baud rate devices typically occupy
1-3 MHz of bandwidth with dense spectral line spacing. The high baud
rate devices require high chipping rates in order to achieve the required
FCC process gain, If the data is not being richly modulated, these high
chip rates lead to sparsely spaced spectral line peaks. Depending on data
modulation type, such high chip rates can lead to sparsely populated peaks
filled in by energy valleys over the entire spectrum used, which is likely
to be 20-60 MHz.
Amateur video transmission may also be found in these bands, as well
as continuous wave (CW) , continuous transmitting jammers such as pilot
tones, or unintentional harmonics caused by lower frequency licensed or
unlicensed devices or FCC Part 15.249 devices. Fast fading effects from
mobile user transmissions and slow fading effects, such as those
experienced in a building at night, will also need to be considered. Taken
as a whole, there are wide areas of bandwidth which are either unused, of
low intensity, or which are unused during those periods in between burst
type transmissions.
The instant invention can also be used in a two-way transceiving
system. Such a system takes advantage of dynamic reallocation of hop
channels, providing that the number of hop channels exceeds that number
required by the FCC's regulations. For example, a receiving device can
transmit a map of its background noise, jamming and interference
!'~
WO 96/10300 PCTIUS95/12150
environment to a distant transceiver such that the transmitter associated
with that distant transceiver can reallocate transmitting frequencies to
concentrate on those frequencies of the receiver that are least occupied.
These considerations taken, the frequency agile receiver can use an
algorithm which sweeps in a linearly increasing frequency and then resets
to an initial frequency, sweeps in a linearly decreasing frequency and then
resets to the initial value, sweeps in a ramp-up/ramp-down fashion, or
searches over a pseudo random order of the available frequency bins. A
frequency 'bin' is set by the most narrow intermediate frequency (IF) or
baseband filter, or a portion thereof, which is being analyzed by the
receiver in one unit time. In addition, the receiver can skip over portions
of the band which are too densely occupied to be of value and seek to
extract useable information from two separate portions of the available
band, which may be removed in frequency from one another by many
megahertz.
In a preferred embodiment of the instant invention based upon
prediction, a present measured signal strength or quieting or signal lock
quality reading from an individual frequency bin is compared to the last
best prediction of the noise, jamming and interference in that bin. The
receiver does not dwell on that particular frequency bin unless the
presently measured signal strength or quieting or signal lock quality
reading is higher than the predicted environment level by an amount
roughly equal to the carrier-to-noise ratio required to decode a valid
message. Once this threshold is met, the receiver will 'TRIP' and dwell on
that channel for the amount of time necessary to determine if valid data
can be extracted from that frequency bin.
The algorithm then falls into three modes: NORMAL MODE, TRIP MODE
and RECOVERY MODE. In NORMAL MODE, the radio spectrum is scanned and
radio energy received from background noise and from manmade jamming
sources which is at approximately static power output is processed. Small
variations in received power do not exceed the dynamic tracking available
in the frequency agile receiver's algorithm and therefore do not cause an
entry into the TRIP MODE, i.e., associated fading changes of known jammers
are either sufficiently small or slow so that the modulation form on the
jammer will either have slow variations in time or small variations in signal
power. The goal during these occurrences of slowly varying signals in
NORMAL MODE is for the frequency agile receiver algorithm to apply
/,f
WO 96/10300 0 0 PCT/US95/12150
averaging to these sources of radio energy which, in fact, tend to average
to a roughly static level.
The next mode of the frequency agile receiver's algorithm, TRIP
MODE, is triggered when a signal exceeds the predicted level plus the
required carrier to noise ratio, signaling the potential for decoding a valid
message. The receiver's algorithm enters the TRIP MODE when subjected to
rapidly changing background noise or a manmade jamming source which is
impulsive in nature, from other hoppers either fast or slow, from a new
transmitting source which has just been turned on, from fast constructive
fading from an existing transmission, from a larger modulation variation of
a transmitter which has already been on, or as the result of multi-tone IMD
products in receiver. The last source of a TRIP is that of a VALID signal.
A VALID signal may Contain a correct PREAMBLE SYSTEM CODE, in
which case useable data may be recovered. A VALID signal may also result
from a PREAMBLE SYSTEM CODE from a similarly coded frequency agile
transmitter but one which was not intended or desired by the receiving
system. This is possible since the PREAMBLE SYSTEM CODE is used in a
fashion similar to that employed with Code Division Multiple Access. A
VALID but unwanted signal message would be discarded prior to data
decode. In either VALID case, the receiver's algorithm typically does not
seek to alter the prediction variables. Alternatively, if VALID messages
from similarly coded systems are anticipated to be long in duration, the
receiver's algorithm can be implemented to rapidly alter the prediction
variable.
Lastly, there is RECOVERY MODE. when a jammer goes away, or is
reduced in unwanted signal energy, the receiver's algorithm seeks to make
that portion of the radio band available as rapidly as possible.
In view of the three modes and the different needs of each, the
present invention proposes a dual algorithmic prediction approach. One
portion of the algorithm attempts to apply averaging to those signals which
are anticipated to remain within a relatively static range. The other
portion of the algorithm tries to rapidly adjust to a signal which is likely
to have large fluctuations in signal strength. This dual approach yields a
type of hysterisis, such that if the signal stayed within a predetermined
window, or within a dynamic TRIP LEVEL range, then the algorithm
performs averaging within that window.
If the received signal fell outside of the predetermined window, or
/4
22.007 11
WO 96/10300 PCTIUS95/12150
the dynamic TRIP LEVEL range, then the receiver's algorithm seeks to
rapidly match that new reading. This can be accomplished with classic
hysterisis or windowing techniques. Alternatively, rate of signal change can
be tracked. Rate of signal change methods which are common in the art
include Type II servo loops which seek to match or anticipate second or
third order, rate of change phenomenon.
It is important to provide means in the receiver's algorithm such that
'one time' events have a minimal skewing effect on the prediction algorithm.
One time events can come from very rapid fading, frequency hoppers, rapid
impulse noise, or large but infrequent modulation deltas. In addition,
statistics can be maintained on each frequency bin measured to show the
percentage of successful transmissions received, or the percentage of
successful data throughput over a unit of time. Since frequency hopping
transmissions will be randomized over the frequency band available, if
there is no jamming present in the frequency band in use, then all of
these statistical successes will be roughly equivalent. Statistics can also
be maintained on the number of false TRIPS experienced in each frequency
bin. The receiver can then take advantage of this information to either
initially, or dynamically, reallocate the channels by either ignoring those
channels exhibiting excessive false TRIPS and thereby preventing wasted
dwell time, or sending information to the associated transmitting devices so
that the -transmitting devices avoid utilizing those frequency bins in
future transmissions. If frequency avoidance is used, the transmitter still
transmits on a specified number 0~ channels sufficient to meet FCC
requirements.
A goal of the frequency agile receiver is to sweep the available band
as rapidly as practicable while providing as much resolution per frequency
bin as possible. A further goal is to yield an accurate representation of
the radio environment, while minimizing signal strength loss. The receiver
may therefore take one or more signal strength or quieting detection or
signal lock quality samples per transmitted bandwidth. The receiver also
sweeps as rapidly as possible in order to minimize the required transmitter
preamble time and make a finer time resolution prediction. The last IF
filter, or the one with the narrowest bandwidth, is the filter that is most
affected by these requirements and is, therefore, the filter that sets the
limit of performance.
Typically, in a conventional filter, as a signal is swept through it at,
17
WO 96/10300 PCT/US95/12150
or approaching, a time proportional to the reciprocal of bandwidth, two
limiting effects occur. First, the filter loss increases, thereby reducing the
effective sensitivity of the resulting receiver. Second, the energy from a
previous frequency bin dwells in the filter while the next frequency bin is
being analyzed. The energy dwelling from frequency bin to frequency bin
is a form of ringing, and is due to the infinite response time of the filter.
The dwelling energy from frequency bin to frequency bin makes
measurements appear to smear into one another. Thus, the energy from a
strong jammer will dwell in the filter one or more sample times delayed and
can be greater than an otherwise decodeable desired signal having a
weaker signal strength.
There are several ways to minimize these two effects on the
performance of a rapidly sweeping frequency agile receiver. The method
previously described using two different IF filter bandwidths can be
beneficial. The step using a wide bandwidth filter is used initially during
the sweep time. The wide bandwidth filter has a faster response time than
a narrow bandwidth filter The wide bandwidth, however, allows additional
noise to be placed within the filter, thereby reducing the effective
sensitivity of the receiver and equivalent process gain. Once the desired
signal is acquired, a narrow bandwidth filter is used when decoding the
data portion of the message. while the receiver is no longer slewing the
vco.
A method which overcomes this previous shortcoming is to use a
chirp filter matched to the sweep speed of the receiver's slewing VCO. The
effective bandwidth of the chirp filter can be matched to the desired
incoming signal. The chirp filter reduces the effects of ringing or
smearing from the filter, and at the same time achieves full receiver
sensitivity. Once the initial signal is acquired during the sweep mode, the
receiver stops sweeping the VCO and switches to a conventional narrowband
filter matched to the data bandwidth, prior to attempting to decode data.
An alternative to using the chirp type filter during the acquisition
phase is to use an IF filter with the narrowest bandwidth or baseband
filter, match that filter bandwidth as closely as possible to the bandwidth
of the transmitted signal, and then provide means for dumping the filter
after each sample. This technique eliminates smearing due to signal energy
dwelling in a filter during subsequent samples. Another approach is to use
a filter designed with a finite impulse response (FIR) so that unwanted
/8
2 2 ~ 0 7 1. 1.
WO 96/10300 PCT/US95/12150
energy cannot linger in the filter. when using the foregoing approaches,
the chosen filter should have a fast response time, or signal energy will be
lost in fast sweep/fast sample implementations, effectively reducing the
sensitivity of the receiver.
A further approach is to use a digital signal processor (DSP), or the
equivalent, to provide a fast Fourier transform of a section of bandwidth
equalling one or more frequency bins simultaneously. A DSP can also
simultaneously implement multiple narrowband filters, thereby increasing the
effective sweep speed of the radio spectrum utilized. A DSP algorithm can
be flexible, and may implement a chirp type filter during the sweep portion
of the receiver algorithm, and/or a narrow band conventional type filter
once the algorithm locks to a single frequency.
Another approach is to use a single IF filter which has excellent
impulse response time as well as desirable ringing/smearing characteristics.
Attractive results have been achieved by both guassian and synchronous
filters. The shortcoming of these filters is that the skirt roll-off degrades
with high signal levels. These filters work well in combination with AGC
strategies, as described below, to minimize the peak energy which may be
introduced.
The receiver 310 may use any of several filter implementations, with
the chosen filter providing desirable results, which may be enhanced by
the foregoing approaches. The filter implementations include use of a
phase lock loop design, a DSP filter, an appropriate surface acoustic wave
( SAW ) implementation, a switched capacitor bank, or any of the classic
active filter techniques having filters modified for dumping their energy
storing elements, e.g., capacitors. In these filter implementations, the
leading edge of the filter should be as sharp as practicable during the
condition when the filter is being swept at a rapid rate. The benefits of
the sharp leading edge include narrowing the effective bandwidth,
improving the co-channel performance, and improving resistance to the
effects of nearby jamming signals.
Prior to anticipated higher levels of jamming, the sweep speed of the
VCO can be slowed down so that the filters can settle, or the sweep speed
of the VCO can be significantly increased such that very little energy is
collected by the narrowest IF or baseband filter.
A further alternative, which can be used with the foregoing methods,
is to provide automatic gain control (AGC) at the front-end stage, or at an
ip
WO 96/10300 PCTIUS95/12150
intermediate frequency stage, of the receiver. The AGC can be in the form
of on/off signal blanking, e.g., gain dumping, or may be in the form of
smaller incremental steps operating over a wide dynamic range. If the AGO
uses smaller incremental steps, then the receiver's algorithm can attempt to
match the AGC signal to that of the incoming signal such that the resulting
RF output of the front-end stage of the receiver yields a nearly level
signal strength envelope as the local oscillator is swept. using an AGC
with smaller incremental steps reduces undesirable effects from the AGC
control signal, which would tend to expand the bandwidth of an incoming
RE signal and could cause the equivalent of smearing. This AGC strategy
can also serve to reduce the effects of front-end signal compression and of
undesired intermodulation components which may be created by non-linear
operation of subsequent receiver stages.
If the AGC used 'on/off' blanking, then undesirable amplitude
modulation artifacts may result from a rapidly changing AGC control signal.
It is therefore desirable to moderate the ramp speed of the 'on/off' control
signal. Such moderation techniques are well known in the art.
The drawback with a full AGC approach is that the receiver's
algorithm primarily uses time cues in order to decrease the level of
predictions of future noise, jamming or interference sources. With a full
AGC approach, the AGC circuits eliminate signals which have fallen below
immediate predicted levels and therefore make these signals unmeasurabie.
A compromise approach to full AGC, is to allow a portion of the RF signal
energy to be above the noise floor and therefore remain measurable. This
compromise approach allows the receiver algorithm to take advantage of
those measurements to more rapidly predict decreasing levels of noise,
jamming and interference. The compromise approach allows most of the
unwanted anticipated signal energy to be eliminated in the very first stage,
or early stage, of the receiver and still provides very robust performance.
The architecture of the frequency agile receiver can be designed
with a digital-to-analog converter, driving the AGC circuit with sufficient
resolution so that the resulting received signal strength indication (RSSI)
level is compared against a single TRIP LEVEL. With such an architecture,
the resulting RSSI level, or quieting output level, does not have to be
analog-to-digital converted, thus reducing the cost and complexity of the
architecture.
In the typical AGC approach, the receiver's first section 301, and/or
'a o
WO 96/10300 2 ? a o 1 1 1 PCT/US95/12150
section 302 is either blanked or replaced with a circuit capable of variable
gain and/or variable attenuation. Further, the AGC or blanking circuits can
be distributed in the receiver's IF strip for reducing cost and improving
performance. It is, on the one hand, desirable to reduce unwanted signal
power immediately at the front-end of the receiver. It also is desirable to
place a variable attenuator after an IF filter or baseband filter which is as
narrow as practicable, but which does not create the limiting factor of
ringing or smearing. The bandwidth of the filter 307 would likewise be
increased such that the impulse response time and signal loss of filter 307
would not be a factor to the following stage.
A further enhancement to the receiver's prediction capabilities is to
provide an adaptive TRIP LEVEL. The adaptive TRIP threshold has several
areas of value. As the receiver sweeps through the available band at
increasingly rapid rates, the narrowest bandwidth RF, IF, or baseband
filter will increasingly attenuate the received signal, due to response time.
It is, therefore, desirable to TRIP the algorithm in the most sensitive
manner possible while at the same time not creating so many false dwells
that the receiver is likely to entirely miss a preamble of a valid
transmission on a later frequency. This can be accomplished with the
support of the software filters which seek to predict background noise,
jamming, and interference. The TRIP threshold can be effective even at
extremely low values, such as 3 dB.
The system of the instant invention seeks to balance sensitivity with
robustness against a hostile environment. In order to be consistent with
this approach, the PREAMBLE SYSTEM CODE can be designed to allow for a
single bit error within its repetitive sequence and still uniquely decode the
preamble's ID. This approach typically expands a 5 bit code to more than
five bits. Designing this facility into the coding of the PREAMBLE SYSTEM
CODE effectively provides for single bit error correction, making the
PREAMBLE SYSTEM CODE more robust, while at the same time lowering the
needed EB/No to successfully decode the PREAMBLE SYSTEM CODE bits. This
approach is consistent with very low TRIP levels. For example, if a 5 bit
PREAMBLE SYSTEM CODE were expanded to 10 bits, to allow for one bit of
error tolerance, then a bit error rate as low as 1 x 10 -1 would provide
usable results. The E B/No required for such a low bit error rate,
depending on modulation type, can be as low as 1 to 3 dB.
It is usually desirable to rapidly discard a transmitted message which
a~
WO 96/10300 1 22007 11 PCT/US95/12150
is not intended for, or desired by, the receiving device. This may be
accomplished by comparing further repetitions of a PREAMBLE SYSTEM CODE
in the transmitted preamble and, if a subsequent code does not match, to
abort the message. Additionally, a 'full' system code in the transmitted
message, representing hundreds or even millions of unique systems, can be
matched against a full system code which is stored within the frequency
agile receiver's memory means.
The frequency agile receiver can store a table of desired transmitter
identification's (ID's) in memory means. The transmitter ID, or full system
code, can be located near a 'data start' marker at the beginning of the
data message. If the ID code or full system code does not match with one
of the desired codes stored in the memory means, then that message can be
aborted without forcing the receiver to dwell for an entire message time.
The 'data start' pattern can be provided by either embedding an illegal bit
sequence, such as three or four sequential 1's or 0's in a row, or the data
start pattern can be a reserved orthogonal PREAMBLE SYSTEM CODE. Since
the PREAMBLE SYSTEM CODE can be so robust, the following portion of the
data message would benefit from an interleaved forward error correction
scheme which approximates the robustness of the PREAMBLE SYSTEM CODE.
A wide range of modulation types may be used with good results in
the instant invention. In general, the modulation type should use the least
amount of bandwidth possible while maintaining a very rapid skirt roll-off.
Desirable modulation forms include, but are not limited to, low index FSK,
SFSK, FFSK, MSW, and GMSK.
A narrow bandwidth utilized at the transmitter allows for a narrower
bandwidth to be used at the receiver, serving to increase sensitivity and
to better reject jamming and interference. Conversely, as the bandwidth is
widened to allow for increasing data modulation rates, the overall message
dwell time is reduced. Brief messages are desirable to avoid the effects of
a rapidly changing RF environment and of undesirable interferers such as
frequency hoppers. Minimum bandwidth is desirable for co-channel
isolation and to minimize the amount of undesirable signal energy which can
be introduced into the narrowest IF filter or baseband filter. Specific
applications of the instant invention can be evaluated to make this trade-
off.
The frequency agile receiver can be applied to transmitters which do
not transmit on a multiplicity of pseudo random frequencies. In such a
az
2200 7 11
WO 96/10300 PCT/US95/12150
case, the receiver provides the benefit of automatically eliminating
frequency uncertainty caused by an unknown transmitter frequency. This
reduces the cost of frequency setting elements while optimizing receiver
sensitivity by permitting bandwidths approximately equal to the bandwidth
of transmitted voice or data. This provides an attractive approach for
cordless phones and PCS applications.
FIG. 5 shows a shunt table. The shunt table is stored in memory
means and is updated either during system setup and/or dynamically
through ongoing system operation. The purpose of the shunt table is to
associate frequency bins, relating to received frequencies, i.e., a current or
voltage input of the receiver's VCO or synthesizer divider inputs, to other
information concerning the past history of that frequency bin. Past
history information can enhance predictions of future anticipated noise,
jamming level or interference level within that frequency bin. Each
frequency bin may be either approximately equal to the bandwidth of the
transmitted signal or may represent a fraction of the bandwidth of the
transmitted signal. It is also possible to make the frequency bin wider
than the transmitted bandwidth, which can improve sweep speed but does
not optimize sensitivity.
The frequency bin number 500 can be stored in the shunt table as a
variable. Alternately, the frequency bin number 500 can be associated with
a physical memory address in a table whereby an offset number can be
used at a start 517 address to index into the table. The end 518 of the
table represents 'n' entries which equal the number of hops available or
practically 2-4 times the number of available hops utilized by the
transmitter to provide finer spectrum resolution at the receiver.
Optionally, the D/A value controlling the receiver's VCO 516 can also be
stored in the shunt table. This value is the equivalent of the selected
receive frequency. The advantage of storing the frequency bin number
500, or of storing the value of the D/A converter which drives the voltage
input of the receiver's VCO 516, is that the receiver algorithm can then
skip through a wide variety of frequencies which may not be adjacent to
one another. This further helps facilitate systems which take advantage of
dynamic reallocation of frequencies.
Each frequency bin is associated with a SHUNT LEVEL 501 which
represents the prediction from the receiver algorithm's previous frequency
sweep(s). The next piece of information, which is optional, is a single bit
a.3
22007 11
WO 96/10300 PCT/US95/12150
which indicates that the last reading taken from that frequency bin was a
false TRIP 515. This bit can be replaced or augmented by the following
optional variables. Attack 502 may be stored to provide a second order
representation of the rate of rise of undesired signal energy. It can be
used as part of a servo loop to track an increasing level of jamming
energy in this frequency bin. Decay 503 provides the same functionality as
attack, but is used for jamming signals which are reducing in signal
energy.
Percent yield 504, which is optional, may maintain a figure of merit
relating to the success of the receiver algorithm in extracting usable
desired messages from that frequency bin. The average false trip 519 table
entry, which is optional, tracks the number of false dwells that occurred on
that frequency bin which may have been caused by impulsive jammers.
This information can then be used to either avoid spending receiver dwell
time on that bin or as a method to transfer this receiver's algorithm to
TRIP and dwell on that frequency bin.
A desired signal 511 is shown which is below the SHUNT LEVEL plus
TRIP LEVEL 510, and is therefore rejected as unrecoverable. Such a signal
would not provide a sufficiently high carrier-to- noise ratio to reliably
decode a message. A desired signal 512 is shown above the SHUNT LEVEL
plus TRIP LEVEL and causes a TRIP of the receiver's sweep algorithm such
that the receiver dwells on that frequency bin and correctly decodes the
desired signal. Alternately, both the SHUNT LEVEL and the TRIP LEVEL can
be summed prior to storage in the shunt table. The hysterisis range
around the SHUNT LEVEL is shown as 519.
Referring to FIG. 6, the receiver software includes an algorithm to
sweep or pseudorandomly step through a list of frequencies contained in
the shunt table. The TRIP handler seeks to identify if the TRIP is a result
of a desired message or undesired noise, jamming or interference. If the
TRIP was due to undesired noise, jamming or interference, the TRIP handler
readjusts the SHUNT LEVEL 501 an appropriate amount.
Once the receiver software is initialized, it jumps to point 'A' 621 in
order to enter the frequency sweeping algorithm. Next, if a chirp type
filter architecture is optionally utilized, then the algorithm selects the
chirp
IF or baseband filter which is architecturally prior to the hardware in the
receiver which measures either signal strength, quieting or signal lock
quality. Alternatively, the sweep algorithm can be re-entered through node
3~
WO 96/10300 PCT/US95/12150
'B' 622 once a pass of a single frequency bin has been completed.
If the instant invention uses an implementation whereby the voltage
controlled oscilla.tor 304 runs predominantly open loop, then occasionally an
additional step may be used to calibrate the voltage controlled oscillator 304
to a known reference such as a crystal 315. This is repeated in a time
interval which is less than the maximum acceptable receiver frequency
drift. This frequency measurement computation can be accomplished
through the optional digital divider 308 in conjunction with timer counter
hardware and/or software within CPU 314 or by phase locked means.
Alternately, if the voltage controlled oscillator 304 is part of a
synthesizer, once settled on an initial frequency, the loop can be opened
and the voltage into the voltage controlled oscillator 304 can be integrated,
or ramped, by any means as is known in the art. As a further alternative,
the voltage controlled oscillator 304 can sweep through an in- band
'intentional jammer' which is purposely generated by the receiver's
hardware. Such an 'intentional harmonic' would also be modulated in a
manner such that the receiver's algorithm could uniquely identify the
harmonic.
The 'intentional harmonic' would be generated by a stable reference
such as a crystal source so that it would appear at a known position in the
band over which the receiver was sweeping. The algorithm within CPU 314
would then be able to use this as a known reference to correct the actual
scan position of the voltage controlled oscillator 304. Additionally, more
than one harmonic can be placed in-band so that a two point correction or
a piecewise linear curve fit could be implemented by algorithmic means.
The frequency bin pointer 508 is incremented 601 to point to the next
shunt table entry. Alternatively, the frequency bin pointer 508 can be
stepped through a pseudo random sequence, or other randomizing means.
Next, if step 602 decides the frequency bin pointer SOS has reached the
end 518 of the table, then control of the algorithm passes to reset 603 the
shunt table pointer. Next, the voltage controlled oscillator 304 is slewed
604 to the start frequency so that the voltage controlled oscillator 304 has
adequate settling time prior to taking the next read. Optionally, the SHUNT
LEVEL 506 is dynamically recalculated starting with a comparison of the
false TRIP count 507 to the number of additional PREAMBLE SYSTEM CODE
repetitions designed into the transmitted preamble 605. If the false TRIP
count 507 is lower than the number of additional repetitions, then the TRIP
as_
WO 96/10300 PCT/US95/12150
LEVEL 506 is reduced 606 by one half of one dB, or by some other
preselected amount. If the false TRIP count 507 is not less than the
number of additional repetitions, then the TRIP LEVEL 506 is increased 607
by one half of one dB, or by some other preselected amount. To prevent
the TRIP LEVEL 506 from increasing or decreasing into unreasonable
ranges, the TRIP LEVEL is constrained 608 to a TRIP LEVEL which is
greater than 3 dB but less than 10 dB. These numbers can be preselected
to any desirable bounds. Alternatively, the dynamic TRIP point can be
updated on every Mth pass. Lastly, the false TRIP count 507 is reset 609.
Control is then passed to reading 611 the SHUNT LEVEL from the
shunt table. The step of writing 612 the SHUNT LEVEL is optional, and
may be used alternatively to augment the rest of the algorithm. The
SHUNT LEVEL 501 is used either in part or in whole to control the AGC
input of the front-end of the receiver. The AGC input would be in block
301 and/or block 302. Alternatively, variable gain or variable attenuation
may be distributed through any part of the receiver. The step of writing
612 the shunt level can be used to eliminate or reduce undesirable signal
energy which is predicted by the shunt table. The step of writing 612,
therefore, reduces the amount of energy which may cause subsequent RF
stages to either overload, ring or create undesirable non-linear signal
components. The step of writing 612 can also be used to increase the
effective dynamic range of the receiver system. A form of blanking can be
implemented by simply disabling the amplifier stage in 302 and/or 301,
thereby effectively reducing the input signal by approximately 30 dB. The
AGC control signal also may be synchronized with subsequent filter
dumping to eliminate unwanted Fourier components which may be measured
on the receiver' s output.
Next, the algorithm increments steps, ramps or pseudo randomly
jumps 613 to the next VCO frequency. This can also be accomplished by
reading a D/A, or frequency, value 516 from the shunt table. Next, the
algorithm optionally may dump 614 the last narrowest IF or baseband filter.
The algorithm then allows a delay 615 which is appropriate for the IF or
baseband filter to settle. Once the filter settles, the A/D converter 313
reads 616 a level representing signal strength, quieting or signal lock
quality. The new reading is then compared 617 to the TRIP LEVEL 506 plus
the SHUNT LEVEL 501. If the algorithm stores a value in the shunt table
that is equal to the SHUNT LEVEL pluLs the TRIP LEVEL, then the new
v~~
WO 96/10300 PCTIUS95/12150
reading is compared only to that combined stored level. If the new reading
exceeds this level then control is passed to C 719 for TRIP handling. If
this new reading is not greater, then the algorithm then seeks to find if
the new reading is 618 less than or equal to the SHUNT LEVEL 501 minus
the TRIP LEVEL 506. Effectively, this creates a hysterisis band which is
plus or minus the TRIP LEVEL 506 selected. If the reading is within a
hysterisis band then control passes to set 619 the SHUNT LEVEL in order
to provide ordinary averaging of the SHUNT LEVEL to the newly read data.
If the new reading is outside of the hysterisis band, then the algorithm
seeks to very rapidly move the SHUNT LEVEL to the newly read input 620.
Alternatively, attack 502 and decay 503 variables may be used from the
SHUNT LEVEL to augment the hysterisis algorithm and provide means to
create averaging about levels which tend to be static while simultaneously
rapidly acquiring jamming signals which are changing quickly. Lastly,
control is directed to B 622 whereby the algorithm repeats itself.
If the step of comparing 617 the new reading is used in conjunction
with the AGC function of setting 612 the shunt level, then the SHUNT
LEVEL 501 is reduced by an amount equal to that written into the AGC
circuit. If gain stage disabling is employed, then the SHUNT LEVEL is
reduced by an amount equal to the attenuation of the disabled amplifier, or
amplifiers, such as the ones in block 302 and/or 301. Further, in AGC
assisted implementations, the 'below shunt' hysterisis provided by the steps
of comparing 618 and setting 620 will only be effective if the AGC value is
only a portion of the total SHUNT LEVEL 501. Lastly, if the instant
invention is implemented with AGC operating over the entire dynamic range
of the receiver, and such that the receiver algorithm uses detected RSSI to
seek a zero point, then the SHUNT LEVEL 501 can be reduced by time
based decay means per FIG 7B.
As an alternative to the above algorithms, an implementation of the
instant invention may post-process TRIPS which are initially identified as
increases in signal level, but whereby the algorithm does not TRIP and
dwell on those frequency bins. In such an implementation, during the
sweep time, the algorithm would utilize an additional column within the
shunt table to indicate that the present reading yielded an increase Over
the prediction relating to the present frequency bin. After a complete pass
of the available radio spectrum, the receiver's algorithm could then choose
the most likely candidate to attempt to decode a PREAMBLE SYSTEM CODE,
Q
7
WO 96/10300 200 7 11 PCT/US95/12150
or some transmitter identifier or data. This post-processing approach
would provide additional system cues to make a decision before spending
time dwelling on any individual frequency bin.
Once the algorithm in FIG. 6 is tripped, control is passed to C 719 of
PIG. 7. Initially, the method deselects 700 the chirp type filter, if one has
been used, and selects 700 the non-chirp alternate filter. A non-chirp
filter is used during data decode since the voltage controlled oscillator 304
does not slew the receiver's local oscillator once a TRIP occurs. Next, the
method may optionally use feedback means, such as a quad circuit or a
frequency error output from a PLL, to provide 701 drift correction between
the transmitting and receiving devices. A frequency error term is
established and is used to compensate the voltage controlled oscillator 304
through the D/A converter 309 if the voltage controlled oscillator 304 is
open loop. Alternatively, if closed loop means such as a synthesizer were
used, then the feedback loop can be opened and the voltage controlled
oscillator 304 can be directly slewed. Alternatively, if a closed loop
synthesizer were used and remained closed, a directly compensating
reference crystal 315 steers the output of the voltage controlled oscillator
304. This last approach has the advantage of automatically correcting the
data sample clock at the same time the voltage controlled oscillator 304
frequency error is corrected. In the drift correction approaches, however,
the tracking range should be governed to prevent the receiver from
locking onto an adjacent, or more powerful, undesired signal.
The method seeks to decode 702 the PREAMBLE SYSTEM CODE which
is repeated in the transmitted signal. This code may be repeated several
extra times as determined by the anticipated number of false TRIPS in the
frequency sweep of FIG. 6. The PREAMBLE SYSTEM CODE can provide
timing information so that the receiver can clock data at an optimal sample
point. The PREAMBLE SYSTEM CODE may be a 5-10 bit code designed with
special characteristics providing cues to the frequency agile receiver so
the frequency agile receiver can rapidly abort attempts to decode
undesired radio energy as a message. One such characteristic cue can be
detected 703 in three or four bit times, such as an illegal bit sequence and,
if not found, the handler aborts to step 710. Additionally, the receiver
hardware or software algorithms may be able to provide information
indicating that the modulation itself is not appropriate or recoverable, in
which case the TRIP handler aborts to step 710. If these tests pass, then
aFs
WO 96/10300 PCT/US95/12150
the method seeks to determine 704 if the PREAMBLE SYSTEM CODE is in one
of the several valid orthogonal groups which may be used by similar
systems but which are not intended for the receiver. If the 5-10 bits
subsequently received cannot be decoded 704 as a valid group, then the
TRIP handler aborts to step 710. If the 5-10 bits received can be
interpreted as a member of a valid group, then the algorithm acknowledges
that the TRIP was not caused by jamming and clears 705 the false TRIP bit
515. Alternately, if the total message transmit time of a valid, but
undesired, transmission is anticipated to last many sweeps of radio
spectrum available, it may be desirable to adjust the SHUNT LEVEL
accordingly, or to mark the channel as occupied or as jammed for some
number of frequency sweeps.
If the PREAMBLE SYSTEM CODE matches 706 the instant receiver's
preset code, then control is passed to establish 707 the data clock sample
position. If the code does not match, then the algorithm passes control to
A 621 so that the receiver can continue to search for possible desired
signals. Steps 702, 703, 704, 706, and 707 can be combined into one
hardware/software block by use of parallel correlation, table look-up or
equivalent algorithmic means.
Once the data clock is established 707 then the algorithm may decode
the first leading bits in the data message which may optionally contain the
full system code. If the full system code does not match 708, then the
transmission is not intended for the receiver and control is passed to A 621
with delay. If the full system code does match 708, then the remainder of
the 1. message is decoded 709 and is used according to the application of
the system. Control is then turned to A 621 in order to search for further
valid signals. In addition, the transmitter's ID may be located in the
leading portion of the following data. The receiver can, in conjunction,
maintain a data base of 'desired' transmitter IDs which are a subset of all
possible IDs. If the new message does not match the 'desired' list, then
the receiver may abort the message to prevent wasted dwell time.
In the case where the TRIP handler aborted, the method determines
710 if the false TRIP is a product of frequency hoppers or other rapid
pulse jammers, which are not likely to affect the next subsequent signal
reading of that frequency bin. If the false TRIP bit 515 was not set 710 in
the last reading of that frequency bin, then control is passed to step 712.
If the last reading of that frequency bin did 710 result in a false TRIP,
a9
CA 02200711 2005-04-01
then it is desirable to set 711 the SHUNT LEVEL 501 to the new jamming level.
As an alternative to this approach, or to augment this approach, the attack
variable
502 may be used to establish a first or second order derivative which seeks to
predict the rate of rise of a rapidly increasing interfering signal level.
The method sets 712 the false TRIP bit 515 in the shunt table
increments 713 the false TRIP count 507 and exits the TRIP handler, returning
to A
621 to seek further desired messages.
The method shown in FIG. 7B is used if the AGC loop seeks to use
the entire dynamic range of the system. The signal level cues, which are lower
than
the zero level of the servo loop, cannot be used to reestablish or reduce the
SHUNT
LEVEL 501. Alternatively, this time-based SHUNT LEVEL decay algorithm can
be used to enhance or augment the other algorithms. The time-based SHUNT
LEVEL decay algorithm runs in a continuous loop beginning with the step of
delaying 714 for some appropriate period of time which can be in the range of
a
portion of a single frequency sweep or as long as many minutes. This method
requires a separate shunt table pointer to be incremented 715 in order to
index
through the list of SHUNT LEVELS 501. An optional step may be used in
conjunction with a higher order loop to calculate 716 a decay. A function
controlled by the decay variable 503 sets the amount W by which, in the
subsequent
steps, the SHUNT LEVEL 501 shall be decremented 717. The method resets 718
the shunt table pointer if the shunt table pointer has reached the end of the
shunt
table. The method then loops back to 714 and repeats.
As an alternative to these Van Newman algorithms, the prediction
can be implemented by fuzzy logic or neural network means. The present signal
strength or quieting reading in any frequency bin can be input, as well as
other
cues, to re-train the network and to yield results which adapt to the radio
environment, background noise, jamming and interference.
It will be apparent to those skilled in the art that various
modifications can be made to the frequency agile spread spectrum system of the
instant invention without departing from the scope or spirit of the invention,
and it
-30-
CA 02200711 2005-04-01
is intended that the present invention cover modifications and variations of
the
frequency agile spread spectrum system provided they come within the scope of
the
appended claims and their equivalents.
-31-
