CA1117603A - Automatic channel selection - Google Patents

Abstract

AUTOMATIC CHANNEL SELECTION

Abstract of the Disclosure HF and VHF radio channels which are subject to time-varying propagation anomalies and to interference are arranged in groups according to frequency band. Within each group are several channels, each spaced sufficiently close so that they experience essentially the same propagation. The several channels in each group are continuously measured to find which channels have the least ambient energy levels, then one of those channels is selected in a random manner to transmit data.

Description

!l ~

7~i03 round of the Invention (a) Field of the Invention . Broadly speaking, this invention reiates to teleco~muni--cations. More particularly, in a preferred embodiment, this invention relates to methods and apparatus for automatically selecting the best one of n telecommunications channels.
(b) Discussion of t _ Prior Art It is frequently necessary to establish a highly reliable telecommunications link o~er channels which are charac terized by time-varying channel occupancy or uncertain communica-tions support mechanisms. Heretofore, no really successful solu-~
tion to this problem has been found except by the use of human operators located at both ends of the telecommunications link.
However, in many instances, manning both ends of a communications¦
: link is undesirable and in some cases impossible, for example, : when the far end of the link is in some inaccessible or dangerousl : location.
Related prior art apparatus include ~. S. Patent No.
3,160,813, issued December 8, 1964, which utilizes a channel selection system wherein a transmitter and receiver sweep a ; rrequ~ncy Dand in s,~ chronlsm ard -he lecelver -eSisters ~he eld ! ~

~ ;
7~ 3 intensity or siOnal-to-noise ratio to ascertain the frequency for, which the received signal is the highest. The transmitter and receiver local oscillator then operate on that selected frequency.
This requires synchronization between receiver and transmitter and prior communication to start the scans simultaneously and then relies on the strongest signal level to identlfy the best channel. Noise or interference can defeat this technique.
Another system shown in U. S. Patent No. 3,983,492, issued September 28, 1976, also searches for the strongest received signal. A microprocessor and logic circuit, frequency synthesize~, tunable receiver and voltage comparator are used to determine the strongest signal.
The technique disclosed in the present application was I
originally conceived as a solution to the problem of achievin~
extreme reliability in a system which relied on both high fre-quency and very high frequency radio propagation to provide com-munications over difficult terrain. Since spectrum occupancy and : propagation conditions on these frequency bands are extremely tim~
varying, high reliability cannot be achieved by conventional single frequency communications. Of course, one skilled in the art will appreciate that the invention is not limited to radio communications but may be used with equal success on any trans-mission facility which is subject to varying transmission quality The technique disclosed herein provides a solution to the two main problems which are characteristic of the HF and VHF ¦
I radio bands. Communications via HF skywave, necessary to achieve ¦ coverage over long distances or in mountainous terrain, is subjec~
I to propagation anomalies which are severely time varying and '! dependent on .he time of day, the month, the season of the year and current sunspot actlvitv. These variatlors prevent HF radlo 111 ~6iO3 I from providing reliable signal suppor~ on any one preselected frequency, since dif~erent frequencies are affected to varying degrees and at varying times by these phenomena. For example, frequencies which are suitable for night use are usually unsuit- I
able during daylight, and vice versa. Communications at or abovel VHF are generally more stable in ~hat they are subject to fewer propagation variations and, therefore, provide inherently more reliable communications. They are, however, subject to terrain masking. On the other hand~ spectrum occupancy in both of these bands is severe and interference caused by noise and other users also prevents reliable links from being continuously available on any one frequency.
It is therefore acknowledged that a choice of alternate operating frequencies is not only desirable, but essential to highly reliable communication links at HF and VHF, as well as in other crowded portions of the spectrum. As discussed, since manned operation at both ends of the link is in many instances not only undesirable, but impossible, a means of automatically selecting satisfac~ory frequencies is required.
It is the intent of this disclosure to define a general system which is applicable to any communication medium which pro-~
vides the user wlth a number of alternative channels for communic~-tion in an effort to automatically assure the availability of a high quality link. The discussion is generalized so as to apply equally well to HF and VHF, as well as to other rrequency ranges or channels of interest.
Summary of the Invention ¦ The key to reliable operation on any cro~ded radio ¦ frequency band is often the identification of a potentiallv "~sable" requency (channel) Due to the ~arkedly dif~erent L7~
characteristics of HF and ~HF radio channels, identification of a ~ sable" frequency is somewhat difficult. ror the time being, 2 ¦l "usable" channel should be interpreted as bein~ a channel relatively free of interference, either friendly or unfriendly, and of suf~icient quality to support error-rree transmission of data. Therefore, althou~h the deter~ination of a "usable" fre-quency can be made beforehand, an actual measure of channel quality is also required before data transmission is attempted.
The system disclosed and claimed herein therefore utilizes in all cases a known channel test message which is trans~
mitted prior to the data transmission to assure a sufficient leve~
of quality. The aùtomatic channel ordering before transmission utilizes information which is available before transmission, and I
therefore, provides a better basis for trials, assuring signifi- ¦
cantly shorter transmission time, on the average,than random "trial and error" transmissions. Reductions in transmission time~
present advantages in terms of the immunity of an enemy to inter- ¦
cept, as well as savings in power. Thus, another important advan tage of this sytem is that it is nor~ally in a "quiet" state, requiring very little power and emitting no radiation. In addi-. tion, the disclosed technique permits unmanned operation at bothends of the link, since all channel ordering and test transmission is automatic. The remote site requires no operating personnel after the initial installation. System operation is a,ynchronous, I
with no prior knowledge required at the remote site, other than al programmed list of possible operating frequencies and a knowledge of the signal formats to enable the use of digital correlation techniques for both the test ~essage and actual data.
More specifically, the system described and claimed he ein relles on the choice o- not one oDe-atino ireauenc~- cbanneL

_ Ij I
but on a set of possible operating frequencies distrlbuted across the bands of interest. Before message transmission is attempted, .
a survey is automatical]y made of the entire channel set, and 2 determination of the "best" channel is made. Once this determina-tion is made, operation, when desired, proceeds on the "best"
channel. Choice of the "best" channel however can be made in a number of ways, and is discussed in more detail below. I
The invention and its mode of operation will be more fully comprehended from the following detailed description when 10 taken with the appended drawings, in which: .
. Description of the Drawings FIG. 1 is a spectrum diagram illustrating the manner in which radio frequency channels are grouped and tested for pOssiblq use according to the invention;
FIG. 2a is a block schematic diagram of an illustrative channel selection apparatus according to the invention;
FIG. 2b is a block schematic of an illustrative micro-processor for use in the apparatus of FIG. 2a;
FIG. 2c is a flow chart indicating the operation of thel microprocessor channel ordering and switching; , FIG. 3a is a block schematic of an illustrative adaptive threshold detector for use in the apparatus of FIG. 2a;
FIG. 3b is a block schematic of an illustrative valida-¦
tion circuitry for use in the apparatus of FIG. 2a;
FIG. 4 is a flow chart indicating the operation of the system when data is transmitted from the remote location to the command location; and FIG. 5 is a flow chart indicating the operation of the l system when data is transmitted fro~ the command location to the 301 remote location.

~, , a1'76~3 Detailed De_cription ~f ~he ~ '~rred Embodiments l~en a number of channels are available for the trans-mission of a message, the total time for message delivery can be reduced considerably if the "best" channels can be selected and ¦I tried first, rather than by using a random trial and error proced~
ure. Both embodiments of the invention to be discussed below have provisions for monitoring the steady-state or ambient energy level in each channel. What is needed is a criterion for orderi.n~
¦ the channels to determine the sequence of trials. On time invarient channels typical of above-VHF radio, it would appear that the quietest or lowest ambient energy channel would be the best, and the channels should be ordered on this basis. This criterion leads, however, to a possible ambiguity at HF since thel quietest channel may merely indicate a channel with no signal support mechanism.
Another criterion for determining the "best" channel is,¦
therefore, disclosed herein and this criterion is effective at ; both VXF and HF. More specifically, if channels at HF are groupe~
as shown in Figure 1, with several channels located within severa~
hundred kilohertz of each other, these channels would each receive simil~r ionospheric support. Thus, a set of channels (at HF) conl sisting of the quietest channel from each group would be certain j to include the "best" channel. If these channels are tried in an !
arbitrary order, one would be assured oI trying the best channel I
in a fewer number of tries than by routinely testing every channell in each group. Since the ordering of those channels tried is ¦ arbitrary or random, one may assume the best channel is equally likely to be chosen on each try.
,, Assuming an HF repertoire of 12 frequencies arranged in four groups of three, as in Figure 1, the quietes, of each of the . I .

~1'7~g~3 four groups, tried randomly, will on the average result in the 1l~ best overall frequency being among the first two tried. Average !. transmission time is, therefore, reduced by a factor of three.
Operation of the proposed system will be described from the point of view of a symmetrical two-way radio link, each end being capable of operation on a number of frequencies, n. It ¦ should be stressed that the resultant reliabillty of this technique depends both on the number and the frequency of the channel choices. These parameters must be selected on the basis of expected propagation conditions or channel occupancy, in such a . way as to provide sufficient overall reliability.
FIG. 2a depicts an embodiment of the basic system. As shown, n channel inputs are connected to a channel ordering and switching block 10 which includes a microprocessor and interface logic device 12, a frequency synthesizer 14 and a tunable receiver¦
16. The output of the tunable receiver is connected to an envelope detector 17, thence to an analog-to-digital converter 19 ¦
via a low-pass filter 18. The output of the A/D converter, in turn, is connected to the input of an adaptive threshold detector 20, to be discussed in more detail below. Control leads 21 and 22 connect the threshold detector to the microprocessor and inter-, face logic device. The outputs of the tunable receiver 16 and control lead 22 are connected to the inputs of an AN~-gate 23 which is connected to signal processing circuit 24 representing a utilization device or output means whlch processes the output signal for a particular application.
The channel ordering and switching operation OI the microprocessor produces a sequential set of channel samples, ln order of increasing ambient energy levels 2S me2sured bv Lhe 30 1 adaptive threshold detector 20. Before the channels can be orderec the energy of each channel is measured. This ls acco~?lished bv ll a sequential scan of all channels under control of the microprocessor to load a full set of energy measurements into the adaptive threshold detector.
Selection of a suitable microprocessor to accomplish the necessary operations is a matter of design to obtain proper timing, sequencing of a number of programming steps and association of particular memory locations within the microprocessor with particular variables. A representative microprocessor that may be used is shown in FIG. 2b which illustrates a Motorola~ M6800 type unit. The term ROM signifies-read only memory, RAM is-random access memory, and PIA represents-peripheral interface adapter. These are standard designations for conventional commercially available items. Three types of memories are utilized within the microprocessor including a lookup table memory A sufficient to contain n code words, one code word for each of n channels; a measurement memory B sufficient for n digi-tal measurements, each representing the ambient energy level of one channel; and an order memory C for n digital code words representing the n channels to be tried in order of increasing ambient energy level.

The microprocessor accepts as an input signal level measure-ments from line 2] in addressed locations for each of the channels measured.
These values are already stored in shift register 26 of detector 20, FIG.
3a, and are readily transferred to the microprocessor. A one by one comparison of Lhe channel measurements implemented in the microprocessor results in an output digital word representing the next channel to be tried which is formatted by the interface logic to control the frequency synthesizer. In another variation, where n channels represent a number of multichannel circuits arriving at a single point, the microprocessor would simply control an n position programmable digital switch, instead of a frequency syn~hesizer-receiver ~ ~RADEMARK

MP~

, combination.
Following the flow chart of FIG. 2c, the first step of the channel ordering and switching operation is to read the code for channel 1 from memory A and tune the synthesizer to channel 1. I~hen tuned to channel 1, output 21 of the adaptive threshold detector 20 is loaded into memory B. Since all channels haven't yet been measured, the microprocessor selects the next code word ¦
from memory A and loads the next measured value into memory B.
It then proceeds to load measurements of succeeding channels into locations in memory B. When finished, the measurement memory B
contains measured energy levels of all channels. The micro-processor now proceeds to order these stored values.
T~e first step in the ordering subroutine is to compare the first and second measured values in memory B. The smaller of the two is selected and compared in a successive pairwise fashion until the smallest (Bm) of all the measurements in memory 3 is known. Since the memory location of this smallest value also indicates the channel number (m) from which it was taken, the microprocessor looks up the code for that channel from lookup table memory A and stores that code word in the first location of~
order memory C. That will be the first channel to be used for test transmission. Since ordering has not yet been completed, pairwise comparison of the remaining measurements resumes to obtain the next smallest value.
Finally, after all measurement memory B locations have been ordered, order memory C contains code words for the n channels to be tried, in the order of increaslng ambient noise ¦ level.

i The switching operation is then ~ecun, in which the 30 1 microprocessor reads code words from the memory locations in !
~-me~ory C, providing to the frequency synthesizer the code words necessary to select specific channels.
In the arrangement shown in FIG. 2a, system operation depends on the availability of independent samples of each channel. The channels represent n possible frequency slots in a radio communication system wherein the synthesizer, under control of the microprocessor, tunes the receiver to each channel. The synthesizer provides a local oscillator signal for each channel to a mixer in the receiver. The receiver includes appropriate RF
and IF stages that translate all frequency slots to a common IF
frequency. Once this IF is established, the n channel~s are equivalent to any n channels, and could equally originate from a ¦
wire multichannel switch or any other radio propagation mechanism.
In operation, once the channel samples are available at a co~mon intermediate requency, the envelope detector l7 and low-pass filter 18 measure the quiescent or ambient energy level ¦
on each channel. The outputs of low-pass filter 18 are then AID
converted, and further processing of the control signals is I digital. Digital energy levels are next entered into adaptive i ¦ threshold detector 20 which is shown in greater detail in FIG. 3a.
The adaptive threshold detector uses a time-multiplexed (n-word) digital filter with a much longer time constant than low-pass filter 18 and is therefore termed a very-low pass rilter.
As shown in FIG. 3a, the detector 20 comprises an n-word shift register 26. The output of A/D converter 19 is su~tracted from the output of register 26, in a subtractor 27. The output of subtractor 27, in turn, is connected to an amplifier 28 of gain K2, thence ~o a second subtractor 29 which also receives the output of A/D converter 19.
30 ¦ The output of subtractor 27 also .orms one input to a i I' , li, 'I
third subtractor 31 whose other input is an adjustable threshold potential Kl. The output of subtractor 31 is connected to a threshold exceedance decision circuit 32 whose outpu~, via contro~
lead 22, controls the microprocessor switching mechanism. Cir- I
cuit 32 represents a comparator which provides an output when the ¦
signal from 27 exceeds the threshold. The output of the first stage of shift register 26 forms the signal on control lead 21 to control the microprocessor channel ordering.
The short time constant of the low-pass filter 18 pro-vides a short term average energy measure on each of the n channels. The very-low-pass filter (detector 20) has a signifi-cantly longer time constant and provides an ambient, long term average energy measure for each channel which is stored separatel~
in each stage of the shift register. The different time constants of the two filters enable the system to detect the transient caused by the arrival of a-signal in one of the n channels. Each ¦
of the n channels will, therefore, have its own adaptive threshol~
maintained in one stage of the very-low-pass filter. The arrange-ment of the two filters enable adaption of the digital signal level measurement to compensate automatically for time varying interference on each separate channel,to permit detection of sig-nal arrival on any one channel. This avoids the need for receiver/transmitter synchronism.
As discussed, two control signals are provided by the adaptive threshold detector 20. One, on lead 22, is a binary signal which indicates threshold exceedances for each channel.
The other, on lead 21, is the actual digitized energy level of each chanrel. This information is used to order the channels, as discussed previously.
The microprocessor channel switching mechanism, upon 1, .
.

lllq~iO3 co~nand of the threshold e~ceedance decision circuitry 32, selects one of the n channels for further processlng by sign21 detector ¦ circultr~7 This depends totally on the signal modulation and formats and is application dependent. Should subsequent process-ing and authentication determine that the signal was not a valid one, the channel switching mechanism would resume its "scanning"
operation. A validation circuit 25 is connected to the micro-processor and to output signal processing circuit 24 for the pur-pose of analyzing a received digital channel test message which is compared with a locally stored version of that message to con-firm its authenticity. FIG. 3b shows a representative circuit that can be used for this purpose. The N-bit received digital word can be stored in register R and cycled upon command C into a one bit comparator NAND gate simultaneously with the N-bit stored message from register S. Depending upon a predetermined disagree-ment count, a validity signal is then provided.
Equipment deployment is envisioned as comprising two units, a manned "command" unit and an unmanned "remote" unit.
The invention does not, however, require operators at either site, since all functions are automatic, and could readily be applied to a situation where message transmission was initiated auto-matically at the command site. Installation procedures and the . types of message transmitted depend on the specific application .~ and are not discussed. This channel selection technique is applicable whether data is generated at the co~mand or remote units, and therefore each case is described separately CASE 1: Data Source at Command Unit _ I
This application requires the transmission of a aata message from the command to the remote unit with ~.igh reliability.
~ After le remote unit has been installed, lC lies quietlv lD t:~e I

7~ 3 field, monitoring each of the possible operating frequencies.
This requires very little ?ower and, therefore, is consistent ~iit,=
portable battery constraints. The command unit, si~ilarly, is left in a quiet condition, continuously making a determination of the usefulness of each of the possible operating frequencies.
This determination results in a continuously updated order of use fulness which dictates which frequency is currently considered the "best".
When message transmission is desired at the command unit, action is initiated which results in a series of events, as described in the flow diagram of FI&. 4. More specifically, the adaptive ordering is frozen, and a channel test message is trans-mitted on the "best" channel. The remote unit, unaware of ~Jhich ¦
frequency the command unit has selected, monitors all frequencies and looks for a sudden increase in signal levels on one of its channels. When this energy increase is detected, timing acquisi-tion and message detection and authentication are attempted. In the event that the message is not authentic~ continuous channel monitoring is resumed. If the received message is verified, the i remote unit locks on the frequency on which it received the test message, and transmits a reply message. It then awaits receptionj of the real command message on the same channel. Upon receipt of the remote unit's reply message, indicating a channel of suffi-cient quality, the command unit proceeds to transmit the actual command message on the same channel. Should a reply not be received, the command unit assumes its "best" channel is no longer suitable, and proceeds to try the next in its order.
CASE 2: Command Unit Interrogates Remote Unit, ~nich Must Trans-mit Actual Data Applications such as this are useful in the re~.ote jl interrogations of passive sensors which have been i~planted in I the field to collect data. The ~echnique provides the useful capability of interrogating sensors in inaccessible locations, andi would prove extremely useful in areas with high tactical VHF
activity.
Message transmission is in this case also initiated at the command unit, with remote unit operations shown in FIG. 5.
Here, the only message transmitted by the command unit is a channel test message to measure quality, which is sent on the "best" channel. Reception and authentication of this message at the remote unit, in this case, elicits not a reply message, but a transmission of the actual data message on the same channel.
Should this not be successfully received at the command unit, the ¦
command unit assumes the channel quality was poor, and a channel test message is transmitted on the next "best" frequency, until reception is achieved. The remote unit would not, of course, transmit its data until a test message was received and authenti cated, indicating a valid request and identifying a good channel.
In both the above-discussed cases, proper selection of 20 the format of the channel test message is essential, in order to provide a valid measure of channel quality. There is no point in locating a channel of sufficient quality for the test message if that quality is not good enough for error-free reception of the actual data. In most cases, this dictates that the signal energy per bit for the test message must be essentially equal to that of the actual data.
One skilled in the art may make various changes and sub-stitutions in the configuratlon shown without departing from he spirit and scope of the invention as set ~orth in the appended claims.

Il.
.-- .

Claims (13)

THE EMBODIMENTS OF THE INVENTION IN WHICH AN EXCLUSIVE PROPERTY OR PRIVILEGE
IS CLAIMED ARE DEFINED AS FOLLOWS:

1. In a communication system of a type that includes n channels each of which is subject to propagation anomalies and/or interference, said channels being arranged into m groups of P channels each, wherein n = m x p, a method of selecting a channel for the transmission of a data signal between near and far ends of a communications link, said channel having character-istics which maximize the probability of successful transmission of data, comprising the steps of:
(a) at the near end of said transmission system, continuously measuring the ambient energy levels of each of the P channels in the first one of said m groups to determine which of said P
channels has the least ambient energy level;
(b) re-iterating said measuring step for each of the remaining (m-l) groups;
(c) selecting a given one of said channels with the least ambient energy levels as the most likely channel to maximize the probability of successful data transmission; and then (d) transmitting a test message to the far end of said system over said selected channel.

2. The method according to Claim 1 comprising the further steps of:
(e) at the far end of the system, verifying that said test message is authentic and has been properly received; and then (f) transmitting a reply signal to the near end of the system which is indicative of the results of said verifying step.

3. The method according to Claim 2 comprising the further steps of:
(g ) monitoring said selected channel at the near end for said reply signal, reception of said reply signal indicating satisfactory reception of said test signal at the far end of the system; and then (h) transmitting said data signal to said far end of the system over said selected channel.

4. The method according to Claim 3 comprising the further steps of:
(i) at the near end of the system, monitoring said selected channel for said reply signal and if not received;
(j) selecting another one of said least ambient energy level channels; and then (k) re-iterating steps (d) through (h) above.

5. The method according to Claim 2 wherein said verify-ing step comprises:
(1) monitoring the energy level received at the far end of the system on all n channels;
(m) detecting which of said n channels is the channel corres-ponding to the selected channel at the near end of the system by virtue of a sudden increase of the energy level monitored on said channel;
(n) comparing the incoming test signal on said channel with a locally stored version of said test signal and, if said comparison results in agreement;
(o) generating said reply signal and applying same to said channel for transmission to the near end of the system.

6. The method according to Claim 1 comprising the further steps of:
(p) transmitting a test message to the far end of said system over said selected channel;
(q) at the far end of the system, verifying that the test message is authentic and has been properly received; and then (r) transmitting said data signal to the near end of the system.

7. The method according to Claim 6 comprising the further steps of:
(s) monitoring said selected channel for said data signal and, if said data signal is not successfully received;
(t) selecting another one of said least ambient energy level channels; and then (u) re-iterating steps (p) through (r) above.

8. The method according to Claim 6 wherein said verify-ing step comprises:
(v) monitoring the energy level received at the far end of the system on all n channels;
(w) detecting which of said n channels is the channel corres-ponding to the selected channel at the near end of the system by virtue of a sudden increase in the energy level monitored on said channel;
(x) comparing the incoming test signal on said channel with a locally stored version of said test signal and, if said comparison is substantially identical;
(y) transmitting said data signal to the near end of the system over said channel.

9. In a transmission system of a kind that includes n channels each of which is subject to propagation anomalies and interference, said channels being arranged into m groups of p channels each, wherein n = m x p, apparatus for selecting a channel for the transmission of a data signal between near and far ends of a telecommunication link, said channel having charac-teristics which maximize the probability of successful data trans-mission, which comprises:
means for supplying input signals to each of said n channels;
means for continuously measuring the ambient energy levels of each of said channels;
means for continuously selecting within each of said m groups, the channel having the least ambient energy level and for randomly selecting one channel from the m channels having the least energy levels; and means for transmitting a test signal to the far end of the system over said one selected channel.

10. The apparatus according to Claim 9 wherein each of said m groups occupies a different segment of the frequency spectrum and each of the p channels in each group is separated from the other channels in the same group by no more than several hundred kilohertz.

11. The apparatus according to Claim 9 wherein said measuring means comprises:
an envelope detector circuit;
a low-pass filter connected to said envelope detector;
an analog-to-digital converter connected to said low-pass filter; and an adaptive threshold detector connected to said analog-to-digital converter for averaging and storing the ambient energy levels in digital form of each said channels.

12. The apparatus according to Claim 11 wherein said adaptive threshold detector comprises a very-low-pass digital filter including:
an n-word shift register for storing the ambient energy measure-ments made on each of the n-channels in said system and supplying said measurements to said channel selecting means; and a threshold exceedance decision circuit connected to said analog-to-digital converter for measuring a signal exceeding a given ambient energy level, the output of said threshold exceedance circuit being connected to said channel selection means.

13. The apparatus according to Claim 12 further includ-ing:
means for validating a data signal received from said far end.