CA2023821C - Spread spectrum communications system for networks - Google Patents

Abstract

Spread spectrum chirps (i.e., wideband frequency packets) are propagated on a local area network in a hostile communications environment, such as a powerline or a crowded radio frequency band. chirps are self-synchronizing, data bit (or subdata bit) in length and detectable by all network nodes, to allow the contention resolution and collision detection needed to support carrier-sense based network protocols. A matched filter of the same time length and encoding sequence as the transmitted chirp provides the self-synchronized chirp reception at each receiver.

Description

_, SFP/M°987 PATENT APPLICATION
1 SPREAD SPECTRUM COMMUNICATIONS SfSTEM FOR NETWORKS

James E. Vander Mey Timothy J. Vander Mey y BACKGROUND OF THE INVENTIAN
8 Field of the Invention This invention relates to use of broadband communi-cations for a network. More specifically, the invention 11 relates to use of spread spectrum communications on a noisy 12 network media such as a powerline using carrier sensing 13 protocols.

Description of the Prior Art 16 The use of spread spectrum communications for more 17 reliable and secure communications is well-known. By 18 transmitting an information signal over a frequency spectrum 19 that is broad with respect to the information bandwidth, and in a manner that can be decoded/despread at the receiver, 21 several benefits are realized. First, any particular narrow 22 band frequency impairments due to interference or 23 attenuation will not necessarily impair the received signal 2'~ because of the redundancy of spectrum used for the information signal. Secondly, the encoding/spreading 26 technique can be chosen such that performing the inverse function (decoding/despreading) on the received signal 2$ effectively spreads any received interfering signal or 29 ~anoise" in the process, thereby minimizing the impact of such noise at the receiver.
31 Spread spectrum for data communications, in its well 32 known form, can be achieved in several ways. Typically, the 33 methods used are classified as variations of the "direct 3'4 sequence" or "frequency hop" techniques. (See, fox instance, Spread Spectrum Systems, Second Edition. by Robert 36 C, Dixon, John Wiley & Sons, 1984.) Each of these methods share the requirement of a synchronisation process. This process must take place in order to establish a connection _ 1 -'ti it : 7 ~ ~i ~~ _~.
:'s .
SFP/P9-9S7 PAT~IdT APPLICATION
1 between the transmitter and receiver. The purpose of this synchronization is to allow the transmitter and receiver to follow the same encoding and decoding process in time 4 synchronization - whether it be code transitions of direct sequence type modulation or frequency hop transitions. This 6 synchronization process can be accomplished in a number of well-known ways. Once synchronization is established between a transmitter and a receivers data is generally transferred by modulating the higher frequency encoding/spreading signal with the information signal.
11 As long as the transmitter and receiver stay in 12 synchronization, the data communications capability of the 13 link is enhanced by the spreading function of the encoding 14 signal. Since the receiver is locked in time to the encoding pattern, the receiver averages out over a period of 16 time (correlates) the encoded signal through possible inter-1~ ference, and then demodulates the received information 18 signal from the recovered encoding signal. Interfering 19 signals are first spread by the decoding process, and then filtered by the coirelation/demodulation process. The 21 ability to reject high levels of interfering signals is one 22 of the primary benefits of spread spectrum communications.
23 It is the encoding characteristic that also allows 24 spread spectrum communications to be more secure. ~y selecting an encoding and synchronization process based on a 26 psuedo-random number sequence of long length, which sequence 2~ is known only to the intended receiver, a transmitter can 28 establish communications with the intended receiver that is difficult for any eavesdropping receiver to synchronize to, since the code sequence is unknown to the eavesdropper. In 31 network applications, however, such secure communications 32 are not usually required, at least at the physical level of 33 transmission.
3'~ The synchronized communication characteristic of spread spectrum is also why spread spectrum is believed to be 36 unsuitable fox carrier-sense based networks -- those that 3~ allow multiple access to a media using carrier-sense 3$ techniques for contention resolution and collision detection R ,' .) '.i :)~ ,1 . . ~ i;.

1 (e. g., CSMA/CD networks such as Ethernet, and the proposed 2 EIA CEBus network). The reason that the prior art spread 3 spectrum is unusable in these connectionless packet-oriented 4 carrier-sense based networks, is that spread spectrum is basically a "connected" communications protocol that to the 6 extent possible ignores the rest of what is happening when a 7 "connection" has been established. A transmitter, even with 8 the same encoding technique, will not be received by a receiver that is synchronized to and receiving from another transmitter, if it is at least marginally out of phase in 11 time (out of synchronization). In fact, a common °°network"
12 technique with spread spectrum communications is to use Code 13 Division Multiple Access. Simply put, this is a number of 14 "connected" (synchronized) transmitters and receivers 1' sending and receiving simultaneously, each '°connection"
16 using a different code sequence, unaware of the existence of 17 the others. In fact they can all be using the same encoding 18 sequences as long as any transmitter/receiver pair is 19 synchronized to a different point in the encoding sequence than is any other pair at any point in time.
21 In a network where the media carrier signal is managed 22 and only one transmitter is normally allowed to operate at 23 one time, techniques of contention-resolution and collision-24 detection are used to arbitrate the use of the carrier on the media. These techniques depend on the ability of any 26 receiver to detect the presence of a signal "carrier" on the 2~ media at any instance in time. In spread spectrum this is 28 not practical, since only if synchronization is achieved can 29 the system determine the gresence of a "carrier", and then once synchronization is achieved, na other °°carrier°' 31 (interference-collision) will be detected. Typically, this 32 synchronization process is non-trivial, and requires time to 33 achieve well beyond that allowed for "carrier-detect°' in 3'~ such networks.
The networks referred to here may reside on any media 36 that allows managed multiple accesses. In the case of typical LAN-type networks, such as Ethernet, the media 3$ usually consist of a coaxial cable and connections meeting .... '.:"' ~!.' ~° ri .r. :~ /!
,~, ~~t.., !-J ~.

1 certain stringent requirements. In this ease, the use of 2 spread spectrum communications would typically not be 3 beneficial, since the media is typically well behaved electrically and relatively noise-free. In the case of other media, however, where the environment is not so well 6 behaved or easily controlled, or is potentially noisy or 7 suffers from variable attenuation or other flaws, the benefits of spread spectrum communications could be substantial.
The transmission characteristics of an AC powerline (a 11 typically noisy media) often exhibit narrow band 12 impairments. Most previously known powerline carrier 13 communications systems utilize a single carrier frequency 14 (ASK modulation) or a narrow band of frequencies (FSK
1' modulation), thus making existing systems susceptible to 16 these commonly found narrow band impairments. If these 17 impairments approximately line up with the carrier frequency 18 the system no longer functions properly. The ASK method of 19 modulation, usually at 120 KHz, is by far the most common method in use today especially in the residential (consumer) 21 market.

24 In accordance with the invention, a method is provided that offers the substantial benefits of spread spectrum 26 communications in network environments that use carrier-2~ sense based protocols. The use of spread spectrum 28 technology, for example, on normal 120 volt AC (alternating 29 current) power distribution lines (powerlines) allows more reliable communications than prior art narrow band 31 techniques. Spread spectrum technology has been used 32 previously on AC powerlines successfully in point-to-point 33 (non-network) applications, for example in the commercially 34 available NEC spectrum AC powerline product, to achieve reliable and high rate throughput, but has not been employed 36 heretofore in a managed multiple access (carrier-sense) network type environment on a powerline.
38 The method described here differs from prior art spread ~n ~~ r,..,~ ~~~
L! rl !.:% :i IrJ .:,.

1 spectrum in that it does not require a specific synchroniza-2 tion state or process prior to transmission of data, and 3 thus provides the equivalent of a spread spectrum carrier.
This is accomplished in one embodiment by using a specific S encoding sequence (e. g. frequency hop or shift or direct code modulation) to send each data bit on the media as a y separately encoded entity. ~y employing a matched filter in S the receiver of the same time length and encoding sequence as that of the transmitted data bit, the data bit can be received directly in one bit time. No separate 11 synchronization is required, as the same encoding pattern is 12 used with successive data bits. The data stream is then 13 sent as a continuous series of these encoded sequences, each 14 having a change of frequency over a time interval, which are 1S called data ehirps (i.e., pulsed frequency modulated 16 signals). In other embodiments, the data bit may include 1~ multiple data chirps. If all receivers on the network 18 employ the same decoding filters, the presence of a carrier 19 (data chirp) is always detectable.
This technique is analogous to that used in some radar 21 systems where a spread spectrum radio frequency chirp 22 (typically a linearly swept frequency). is transmitted and 23 the echo is received back and processed through a matched 2'~ filter to detect its presence. These filters are typically 2S Surface Acoustic Wave (SAW) devises. The presence of a 26 spread spectrum chirp is detected at the output of the 2~ matched filter, with a time-compression equivalent to the 28 gain of the filter. In this application the filter is used 29 to positively detect the presence of the radar echo and provide a time resolution gain to establish a more accurate 31 distance of the object reflecting the chirp.
32 As in the above described radar system, one embodiment 33 of the present invention also employs individual chirps and 34 a matched filter receiver. These chirps preferably are sent in continuous sequence and need not be of the swept 36 frequency variety. In other embodiments, the chirps~are of the iog sweep, psuedo-random, or time hopping type. Also 3$ the matched filter used is preferably a low cost implementation of a delay line matched filter instead of a SAW
device.
One object of the present invention is to use a wide band signal where the spreading function is independent of the data (spread spectrum) to eliminate or considerably reduce the effects of the narrow band impairments.
All prior art ASK systems such as those sold by X10, Black and Decker, and the new CEBus Standard under development by the EIA (all of which use 120 KHz ASK modulation) can be used in conjunction with the present invention by substituting the wide band swept chirp for the single carrier frequency of the prior art. In one embodiment, during each chirp the frequency is swept from 50 KHz to 450 KHz during the carrier ON
interval of the ASK modulated system. This effectively eliminates the narrow band impairment problem of the prior art since a substantial percentage of the frequencies covered by the wide band sweep (chirp) can be lost without losing the data present in the chirp, and thus the existing systems and the CEBus Standard can be easily retrofitted to operate in accordance with the invention, resulting in a considerable improvement in performance. This performance improvement is desirable in both residential and commercial powerline communicaitons systems.
One of the most expensive performance requirements of prior art spread spectrum signals is the synchronization requirement. In most spread spectrum systems, considerable receiver complexity is needed to properly synchronize with the incoming signal. In addition to receiver complexity, the stability of both the transmitter and receiver clocks is important for long packets, since synchronization once obtained must be.maintained for the duration of the packet. If the differential clock drift over the packet length of both the chirp generating transmitter clock and receiver clock is greater than half the bit period over the packet length, errors will result. In accordance with the present invention, the system instead synchronizes on each bit (chirp). There is no separate synchronization field or area of the packet or chirp and, therefore, clock generators can be low cost since the problematic drift interval is reduced to the data bit time rather than the packet time. Also, the receiver needs no special synchronization circuitry, which provides a considerable cost savings.
Other hostile or uncontrolled media in addition to powerlines, such as radio frequency and infrared local networks, are suitable for application of the present invention. In the case of multiple media networks, such as the proposed EIA CEBus home control network, spread spectrum techniques may be the best choice for a number of media, such as the powerline, radio frequency, infrared media, and twisted pair due to crosstalk considerations with other twisted pairs in the same bundle (e. g. telephone lines with noisy modems/facsimile machines, etc.). The desirability of commonality also suggests application of the invention to well-behaved media such as coaxial cable and fiber optics. The use of wideband communication in accordance with the invention advantageously allows transmission of data at a higher data rate in various media than does the prior art.
According to one aspect the present invention provides a method of transmitting sequences of information symbols on a communication channel between any of a plurality of transmitters and at least one receiver connected to the communications channel, comprising the steps of: each transmitter transmitting a sequence of predetermined wide band signals, each transmitter resolving contention for use of the communications channel by sensing, prior to or during transmission of the sequence, for the presence on the communications channel of a said predetermined wide band signal transmitted by another transmitter, wherein the step of sensing for the presence of a predetermined wide band signal transmitted by another transmitter comprises using a filter that, without prior synchronization, produces an output indicative of receipt of a wide band signal.
According to another aspect the present invention provides a method of transmitting a sequence of information symbols on a communications channel between a plurality of transmitters and at least one receiver, comprising the steps of: representing the sequence of information symbols as a sequence of predetermined spread spectrum chirps; performing a contention resolution process in which a transmitter senses whether other transmitters are transmitting; following a successful contention resolution transmitting the sequence of chirps from the transmitter to the communications channel;
receiving signals from the communications channel at the receiver and processing the received signals using a filter that, without prior synchronization, produces a filter output indicative of the degree of correlation between a said received signal and a said chirp; successively repeating the step of receiving signals to produce a sequence of filter outputs; and processing the sequence of filter outputs to reconstruct the sequence of information symbols.
According to another aspect the present invention provides a method of transmitting data on a communications channel to a receiver comprising the steps of: generating a plurality of signals at a transmitter, each signal being a chirp and being generated at a plurality of frequencies over a - 7a -time interval; providing the signals to a carrier-sense based communications channel; sensing a presence on the communications channel of a carrier signal prior to initiating a generation of signals, said carrier signal including a sequence of at least one chirp, and having a timing which is asynchronous relative to that of at least one other transmitter associated with communications channel; receiving the signals at the receiver from the communications channel; and determining a synchronization for each signal solely from the signal.
According to another aspect the present invention provides a method of transmitting data on a communications channel to a receiver comprising the steps of: generating a plurality of signals, each signal being a chirp and being generated at a plurality of frequencies over a time interval, wherein chirps are transmitted at a predetermined rate for carrier emulation; providing the signals to a carrier-sense based communication channel; receiving the signals at the receiver from the communications channel; and determining a synchronization for each signal solely from the signal thereby synchronizing independently of a source of the plurality of signals.
According to another aspect the present invention provides apparatus for transmitting sequences of information symbols on a communications channel between any of a plurality of transmitters and at least one receiver connected to the communications channel, comprising: means at each transmitter for transmitting a sequence of predetermined wide band signals, the timing of the sequence of signals transmitted by at least one transmitter being asynchronous relative to that of at least one other transmitter, means at each transmitter for resolving contention for use of the communications channel by sensing, - 7b -prior to or during transmission of the sequence, for the presence on the communications channel of a said predetermined wide band signal transmitted by another transmitter, including a said signal with timing asynchronous to that of the sequence signals generated by that transmitter, wherein sensing for the presence of a predetermined wide band signal transmitted by another transmitter comprises using a filter that, without prior synchronization, produces an output indicative of the degree of correlation between a received signal and the predetermined wide band signal.
According to another aspect the present invention provides apparatus for transmitting a sequence of information symbols on a communications channel between a plurality of transmitters and at least one receiver, comprising: a transmitter configured to represent the sequence of information symbols as a sequence of predetermined spread spectrum chirps, to contend for access to the communications channel by sensing whether other transmitters are transmitting, and following a successful contention resolution to transmit the sequence of chirps on the communications channel; a receiver configured to receive signals from the communications channel and to process the received signals using a filter that, without prior synchronization, produces a filter output indicative of the degree of correlation between a said received signal and a said chirp; the receiver being configured to successively process received signal to produce a sequence of filter outputs, and to process the sequence of filter outputs to reconstruct the sequence of information symbols.
According to another aspect the present invention provides a communications system comprising: a plurality of transmitters each generating a plurality of signals, each signal being a chirp and being generated at a plurality of - 7c -frequencies over a time interval; a communications channel for carrying the signals; each said transmitter including means for sensing a presence on the communications channel of a carrier signal prior to initiating a generation of signals, said carrier signal including a sequence of at least one chirp, and having timing which is asynchronous relative to that of at least one other of the plurality of transmitters; and means for receiving the signals from the communications channel, wherein the means for receiving includes means for self synchronizing on each signal of the plurality of signals.
According to another aspect the present invention provides a receiver for a network comprising: means connected to the network for receiving an analog signal including a stream of chirps including a plurality of frequencies from the communications channel and converting the analog signal to a digital signal; means for receiving the digital signal and storing the digital signal; means for comparing the stored digital signal to a predetermined value and providing an output signal in response thereto; and means for self synchronizing and locking into the stream of chirps at a predetermined rate, wherein the receiver synchronizes itself independent of a transmitter of the analog signal.
According to a final aspect a communication system comprising: a plurality of transmitters each generating a plurality of signals, each signal being a chirp and being generated at plurality of frequencies over a time interval, wherein the chirps are transmitted at a predetermined rate for emulation of a carrier signal; a communications channel for carrying the signals; means for sensing a presence of the carrier signal on the communications channel and thereby inhibiting a generation of signals by at least one transmitter, wherein each transmitter generates signals asynchronously to - 7d -signals generated by at least one other transmitter; and means for receiving the signals from the communications channel, wherein the means for receiving includes means for self synchronizing on each signal of the plurality of signals.
BRIEF DESCRIPTION OF THE FIGURES
Figure 1 shows a system including one embodiment of the present invention.
Figure 2 shows one embodiment of the present invention in block diagram form.
Figure 3 shows timing diagrams for the receiver stage of Figure 2.
Figure 4 shows timing diagrams for the transmitter stage of Figure 2.
DETAILED DESCRIPTION OF THE INVENTION
Figure 1 shows a typical household system for control of an electric light in accordance with the invention. The controlling portion of the system at the transmitting end includes conventional product specific logic 10 and an on-off switch 12 controlled by the product specific logic - 7e -r .-..~ S-J ~~n~ i J rJ ~~i f-J
5FP/M°987 PATENT APPLICATIOP1 1 10, and conventional product address circuitry 14 which provides a product address to the product specific logic 10.
3 The product specific logic 10 provides data (i.e., 4 turning the light on or off) to the transmit stage 16. The transmit stage 16 is conventionally coupled by a line coupler circuit 18 to the alternating current (AC) powerline 7 20 in the house.

8 At the receiving and power control end of the system, the receive stage 24 is coupled by a second line coupler circuit 22 to the AC powerline 20. The receive stage 24 11 recovers data provided by the transmit stage 24 from the 12 Signals on the powerline 20, and provides the data to the 13 product specific logic 26. A second product address is 14 provided to the product specific logic by product address circuitry 28. Using the data, the product specific logic 26 16 controls a power switch 30 (shown as including a relay) to turn power on or off to a light 32 cahich in turn is 18 conventionally connected to powerline 20.
1~ The above described system is thus conventional in several respects and the well known aspects of such a system 21 are disclosed in, for instance, U.~. Patent No. 4,755,792, 22 issued July 5. 1988 to Pezzolo et al. In accordance with 23 the present invention,~the transmit stage 16 and receive 24 stage 24 are both different from that in the prior art systems, and are described in further detail below.
26 In the network application of the present invention, 2y the purpose of the chirp is to provide a minimal unit time 28 for carrier emulation (usually a data bit time ar 29 submultiple thereof) and provide the spread spectrum processing gain and interference spreading inherent in the 31 use of a delay line matched filter. The processing gain of 32 the matched filter defines the ratio of the chirp transmit 33 time to the valid filter correlator output time. If a 34 single chirp represents a data bit, ,i~t is known that the gain ratio of the matched filter is equivalent to the ratio 36 of the number of chirps per bit in direct sequence type 3~ spread spectrum systems.
3$ The chirp to be transmitted can be in any number of g . . 4 i (-l ..~ ('J ;~ ~~
S~P/a~_98~ PATENT r.~P~,acATZ~N
1 wideband signal formats. For example, swept frequency 2 chirps are used in one embodiment. A chirp in which the frequencies are swept linearly (i.e., the frequency changes 4 linearly over time) results in a chirp that has energy evenly distributed over the swept frequency spectrum.
6 Alternately, a non°linear sweep is used to tailor the energy at various frequencies, as is best suited for the media used 8 as the transmission channel.
Direct sequence code generated frequency chirps are also used in another embodiment. As is common in many 11 spread spectrum applications, these wideband spectrum chirps 12 are generated using psuedo_random maximal linear codes and a 13 modulation technique such as binary or quadrature phase 14 shift keying (EPSK/QPSK). Each chirp that is separately decodable has a unique code sequence.
16 The modulation of the encoding sequence for each bit 1~ can be accomplished in several ways in accordance with the 18 invention. One implementation for a two_state data stream lg is to send a linearly swept frequency sequence in opposite polarity phases (180 degree phase shift) depending upon 21 which state (e. g. "O°' or "1'°) of data is to be repre-22 rented. This will result in the receiver matched filter 23 using the same correlation network to generate either a 24 positive or negative correlation output, reflecting which of the two states has been received. Alternately, the matched 26 filter can have more than one correlation network GOnneCted 2T to its internal delay line to receive multiple symbol states 2$ if desired. These symbols may include for example data bits ZS and control field delimiters. Another embodiment includes multiple sub-symbol length chirps to encode the symbols.
31 Network nodes not immediately involved in a data 32 transfer but seeking access, can continuously monitor the 33 media, as required by the network contention rules. and 34 would be able to detect the presence of "carrier" (data) being sent by any other node. In the event of simultaneous 36 contention for the media by more than node, the network rules normally allow for the detection of collisions -°
3$ either during a packet preamble time, during a packet .. _ g 'l~~ :~ ~:.'~ E~ -~.
SFP/M°987 PATFI~T APPLICATION
1 transfer, or upon receipt of an (invalid) packet. In the 2 latter case, the network protocol rules provide for the appropriate retransmission. In the former cases. the 4 network protocol rules may or may not require the receiving and transmitting nodes that currently are transferring a packet of data to monitor the media for carrier collisions (i.e. simultaneous interfering carrier from another node).
8 If such monitoring is required, during the portion of time that nodes are to monitor for collisions, the receiver (or listening transmitter) can regard any correlated output 11 that is detected, other than its own, as a carrier 12 collision. The discrimination of a carrier collision 13 correlation output from a correctly received/transmitted 14 data bit can be accomplished by synchronizing the receiver matched filter to the transmitter°s bit data rate. This 16 requires a constant rate and continuous data stream from the 1~ transmitter at a rate known to the receiver. The valid 18 output of the delay line matched filter is then synchronized 19 to the data rate as a narrow period (window) of time. which is nominally equal to the data bit time divided by the gain 21 of the filter. If a correlation signal occurs outside of 22 this window, the signal is presumably some other colliding 23 signal.
24 A, colliding signal is not necessarily harmful to the successful transmission of the data packet in progress.
26 This collision condition is similar to the Code Division 2~ Multiple Access (CDMA) case of multiple transmitters/
28 receivers~using the same code sequence as described above.
29 Simply ignoring all correlating signals at other than the correct time window eliminates all colliding signals except 31 those that are in exact time phase (synchronization) as the 32 transfer in progress.
33 When the network rules allow, either because they do not require the nodes transferring data at some (or all) times to monitor the media for carrier collisions, and/or 36 use invalid packet transfer as collision indications, the 3' data transfer integrity can be enhanced by data stream 38 synchronization. This is accomplished as described above by ° 10 -limiting the valid receive filter correlation output to the anticipated correlation time window in succeeding data bits.
In order for this data stream synchronization to occur, the transmitter must send data chirps at a constant frequency and the receiver must receive them at the same rate. When this data stream synchronization is accomplished, it can be shown that the resultant data transfer robustness is equivalent to that of a similarly encoded prior art spread spectrum system of the same gain.
Slight variations in data stream frequency can be tolerated by appropriate receiver design. For example, the receiver could open up the correlation window by the anticipated maximum time offset error for any single data chirp and then reset the time to the next window on each data chirp received. This method would allow the receiver to track any transmitter within the design error tolerances.
An element of the cost-effective implementation of this described method which allows use in low-cost home-type networks (in addition to less cost-sensitive commercial networks) is the availability of an inexpensive delay line matched filter, such as that described below.
The spread spectrum poweline communication system in accordance with the invention in one embodiment uses discrete frequency sweeps (chirps) to transmit information. Two kinds of sweeps are used, one corresponding to a logical "1", and one corresponding to a logical "0". The "0" sweep is the amplitude inversion of the "1" sweep. Synchronization occurs after the reception of three contiguous "1's", and is terminated by the loss of continuous data reception.

The sweep (or chirp) consists of a sine wave (or triangle wave) which is linearly frequency swept through the transmission band. A log sweep is used in a second embodiment.
The reception of the waveform is accomplished by continuously measuring and storing the incoming signal's slope direction.
When this stored pattern sufficiently matches the ideal waveform pattern, correlation is declared.
- lla -:; , . . , ,:.., l i~,..<:) ., ~

1 Signal pattern matching is done by first storing the received signal's slope directions in a digital FIFO (first 3 in, first out) circuit. The appropriate inverting and non inverting taps of the FIFO which correspond to the ideal signal pattern are then summed through resistor summing 6 networks. When the summed voltage reaches sufficient ? potential, correlation is declared. To illustrate this, suppose the signal received exactly corresponds to the ideal "1" signal. Upon complete reception of this signal, all the inputs to the summing network would be "high", thus raising il the output of the summing network above the correlation 12 threshold.
13 One embodiment of the invention is shown in Figure 2.
14 In this embodiment, both the transmit stage and the receive Stage of Figure 1 are included in one receive-transmit power 16 line modem for one node on the network.
i? The following describes the RECEIVE stage 44.
18 An advantage of the sign of slope method of matching a 19 received signal to a ideal signal in accordance with the invention is that it is independent of received signal 21 amplitude. That is, the receiver is looking only to the 22 direction of the signal. This allows for the introduction 23 of signal processing prior to the matched filter to enhance 24 its performance. A non-linear signal compressing circuit such as a logarithmic amplifier improves the signal-to-noise 26 performance in an environment where noise is dominant, since 2? the ratio of the dominant noise to signal is reduced. In an 2S environment where the signal to noise ratio is positive 29 (more signal than noise) this logarithmic amplifier is also usable, since the logarithmic amplifier would in the worst 31 case cause the signal-to°noise ratio to approach zero, and 32 this filter is designed to operate reliably even in the 33 negative signal-to-noise area.
The Following elements are shown in the RECEIVE stage 44.
36 R1-ISOLATION TRANSFORMER: The isolation transformer (i.e., line coupler Rl) performs impedance matching and 3$ current gain functions as well as high frecyuency AC coupling la _ ;~ .- o.~, ; 3 ,~i ... j Fn 'y, :-:~ !.) :..~ r ~pp~p~_gg7 PATENT APPLICATION
1 the circuit to the AC powerline 40 via conventional plug 2 40P.
3 R2-BANDPASS FILTER: The band pass filter R2 limits 4 unwanted signals outside the pass band. This filter R2 preferably has an approximately constant group delay for the 6 frequencies in the pass band.
y R3-LOGARITHMIC AMPLIFIER: The logarithmic amplifier R3 8 is a nonlinear amplifier used to improve the signal-to-noise g ratio when that ratio is less than Odb. The amplifier also provides drive current for the SLOFE DETECT COMPARATORS R4, 11 R5.
12 Logarithmic or similar non-linear signal compression 13 preprocessing also improves the dynamic range of the 14 receiver 44, since it compresses larger signals more than smaller signals. In many cases, this may eliminate the need 16 for automatic gain control devices.
ly R4-SLOPE DETECT COMPARATOR ~1: The slope detect 1S comparator R4 compares the current signal voltage with that lg same signal delayed in time. The result of this comparison will be a voltage proportional to the direction of the slope 21 of the signal. Slope detect comparator ~l R4 is used for 22 comparing the frequency components of the.signal which lie 23 in the upper half of the full signal frequency band. The 24 delay time is approximately equal to one quarter of the period of the average frequency in this upper half of the 26 band.
27 R5°SLOPE DETECT COMPARATOR #2: This comparator R5 is 28 identical in operation to the above comparator, but is used 29 fox comparing the frequency components of the signal which lie in the lower half of the full signal frequency band.
31 Its delay time is approximately equal to one quarter of the 32 period of the average frequency in this lower half of the 33 band.
The two slope detect comparators function as an analog-to-digital converter "front end" to the receiver 44. The 36 analog-to-digital front-end is a "sign of slope" detector, and determines the amplitude direction and not the actual 38 amplitude of the received signal. Its output at any point - ~. ; ' ~'~ ~:': : ~ =.

1 in time is one of two states, indicating a rising or falling 2 amplitude of the signal being received.
Such a detector can be implemented with a single 4 comparator using an RC (i.e., resistor-capacitor) filter circuit including resistor RF and capacitor CF (as shown for R4) on one comparator COMPF input. The input signal is fed ? to one input of the comparator COMPF and the RC filter 8 delayed signal to the other input of comparator COMPF. The g output of the comparator COMPF then continuously indicates the rising or falling direction (sign of slope) of the 11 amplitude of the signal being received.
12 The value of the time-constant of the RC filter used at 13 the input to the comparator COMPF determines the amount of 14 time over which the slope of the input signal is measured to determine its direction. An appropriate slope time is 16 selected to match the desired frequency being sampled. For 1? optimal results. the RC time-constant value chosen is such 18 that the slope measurement time is equal to 1/2 the period 19 of the frequency being sampled.
R6-HIGH FREQUENCY SHIFT REGISTER: The output of SLOPE
21 DETECT COMPARATOR ~1 R4 is clocked into the high frequency 22 shift register R6 by clock R6C at a frequency at feast foul 23 times the maximum frequency of the swept signal. The number 24 of stages is sufficient to hold enough data to cover half of the signal sweep time. Since the signal is swept from low 26 to high frequency, at the termination of the received sweep 2? register R6 contains the slopes of the high frequency half 2$ of the signal.
29 R?-LOW FREQUENCY SHIFT REGISTER: The output of SLOPE
DETECT COMPARATOR #2 RS is clocked into the low frequency 31 Shift register R? at one half the clock rate of the high 32 frequency register by clock R6C and divider R6D. The low 33 frequency shift register R? is sufficiently long to hold 34 enough slope data to cover the entire sweep time. The two Shift registers R6, R? both function as digital delay lines.
36 The output of each comparator R4, R5 can be sampled 3? respectively into the internal delay line digital shift 3$ registers R6, R? at the clock rate of the delay line. The -y;~~~~'.)!: /1 y Y~.f ~~% ~~:r ~.;-i s F~.J w.

1 result is that the delay line shift registers contain the sign (e. g. 0- - ,1= +) of the elope of the signal at that 3 point in time, thus indicating whether the received signal amplitude is rising or falling.
The clock rate of each internal delay line shift register R6, R7 is determined by the highest frequency to be 7 detected by the filter. An oversample rate is required with 8 a higher oversample rate giving better results. An oversample rate of four times the highest frequency gives satisfactory results and an oversample rate greater that 11 eight does not seem to contribute significant further 12 improvement.
13 The amount of storage required for the delay line is a l4 function of the oversample rate and the highest frequency component of the chirp. If no storage compression 16 techniques are used (see below), the number of storage bits 17 required can be calculated as the oversample rate times the 18 highest frequency times the time duration of the signal 19 chirp to be received. For example, for a 100 microsecond chirp pulse with a highest frequency component of 450KHz, 21 and an oversample rate of four, the required number of bits 22 for delay line storage would be 180.
23 Rg,Rg-SUMMING NETWORKS: The outputs of the HIGH and LOW
24 r9REQUENCY SHIFT REGISTERS R6, R7 are summed respectively through two parallel resistor networks R8, R9. Non-26 inverting and inverting taps are used which correspond to the slopes of the ideal received signal. If, for example, 28 the ideal signal was received all outputs to the summing 2g network would be °'high". Since only half of the bits in the I,pW FREQUENCY SHIFT REGISTER (LFSR) R7 will contain data on 31 the low frequency half of the signal, only this half of the 32 shift register is used for summing. Furthermore, since the 33 Clock frequency of the LFSR R7 is half that of the clock 34 frequency of the HFSR R6, the outputs of the LFSR R7 are multiplied by a weighting factor of 2, so that they have 3S equal weight when considered in the time domain. This can be accomplished by allowing the value of the summing resistors of the LFSR R7 to equal one half the value of the ~.Y 'v ~! ~~ly . {. ;,' =1 .% l'..r -SFP/M-9~7 PATEiJT APPLICATIOI9 resistors of the HFSR R6. It is not necessary to include all the valid taps of the registers into the summing network 3 in order to obtain satisfactory performance.
4 This method of matching the received and stored signal values to that of the desired ideal signal preferably uses a 6 summing network of the stored signs of slopes and comparing the summing network output value to an expected value. A
simple resistor summing network is preferably used, where each point of interest in the delay line is tied to an appropriate value resistor. The ideal signal determines 11 which polarity of the storage bit in the delay line is tie 12 to the resistor.
13 Different configurations of the summing network are 14 used in various embodiments. A simple embodiment is to tie each and every storage element to a single resistor, with 16 the polarity being a function of the ideal signal. All 1~ resistors in this case would have the same value, and if the 18 signal received is exactly the same as the ideal signal, the 19 summing network would be at maximum value when the received signal sign of slope values are shifted to the corresponding 21 delay line storage elements. A signal of opposite amplitude 22 to the ideal signal would similarly result in the minimum 23 summing network value, and with an appropriately chosen 24 ideal signal a random signal should result in a midpoint value.
26 Another embodiment uses less than all the storage 2~ elements in the summing network, in some cases for improved 2$ performance. Hy eliminating those points that the ideal 2g signal suggests are near transition points (i.e., the slope 3p values would start to approach zero and transition to the 31 opposite sign), better discrimination is possible. For the 32 frequency swept chirp, the best discrimination appears to be 33 achieved by using only those points that are peaks where the slopes would reach their maximum values (nodes). This occurs twice (once positive/once negative) per cycle of the 36 ideal waveform.
As an example of the above delay line storage and 3$ summing network requirements, assume a (non-baseband) ~.: :.' ~n !3 ~" t.W

1 linearly frequency swept chirp signal of 150Khz to 459Khz 2 over a period of 100 usec. The clock frequency required for 3 a four times oversample rate would be l.8Mhz, giving a 4 nominal delay line storage requirement of 180 bits (l.BMhz times 100 usecj. The number of cycles would be 30~ giving a total of 60 cycles. The summing network is built using one resistor per node. The values of each resistor in this case would be proportional to the distance in time between the adjacent nodes to give the correct weighting to the summing network.
11 In one alternative embodiment. instead of blocks R4 to 12 R9, a commercially available matched filter is used, such as 13 a charge-coupled transversal (matched] filter model 9602 14 from EG&G-Reticon.
R10-CORRELATION DETECTORe The carrelation detector R10 16 compares the voltage generated by the SUMMING NETWORKS R8, 17 R9 to two predefined voltages. If these voltages are 13 exceeded, a signal detected output is generated. The 19 reception of the ideal signal and the reception of the amplitude inverted ideal signal will produce output voltages 21 equal in magnitude but opposite in direction. Because of 22 this, the correlation detector must be able to detect and 23 distinguish a "high°' or °'low" correlation.
24 The correlation detector determines that the value of the summing network exceeds the threshold established to 26 determine that a signal matching the ideal signal has been 2~ received. In the absence of noise, the output of the 28 summing network would reach its maximum when a signal 29 matching the ideal is received. With noise present, the summing network will reach some value between its maximum 31 for minimum for opposite amplitude] and its midpoint when 32 the received signal containing the ideal signal is 33 received. Other signals may also cause the summing network 34 output to vary beyond its midpoint. It is the function of the chirp design code sequence, bandwidth, etc.j to provide 36 the maximum discrimination possible. Various threshold values are appropriate for particular applications.
3$ The correlation detector for the summing resistor _...~ r~'-- ,,._ .i 1 network preferably includes two comparators (not shown) 2 conventionally set to the thresholds for the maximum and 3 minimum desired. The output of the comparators is then the indication of a received ideal signal of the corresponding amplitude phase.
6 R11-RAW DATA/CARRIER SENSE OUT: The output of the CORRELATION DETECTOR R10 is fed directly out to the user (i.e., product specific logic 42) through the saw data out line R11. The signal is conditioned, however, so that it ld has a minimum pulse width, and is used to indicate the 11 presence of a "carrier" for the product specific logic 42 12 needed to handle network protocols.
13 R12-SYNCHRONIZATION AND LOCK: The output of the 14 CORRELATION DETECTOR R10 is fed into the synchronization and lock circuitry R12. This circuitry R12 looks for three 16 consecutive "high" correlations each at a predefined 1~ interval. When this occurs, the circuit R12 places itself 18 in "lock" mode. When in lock mode the circuit R12 will 19 ignore any correlations until it opens its correlation window. If no correlations are seen by the time the 21 correlation window is closed, the circuit R12 will fall out 22 of lock mode and again begin looking for the synchronization 23 pattern. If the correlation output is one clock cycle long, 24 the correlation window will open after the elapsed time equals the sweep period minus one clock cycle, and will 26 close after the elapsed time equals the sweep period plus 27 one cloak cycle. If the correlation output is two or three 2$ clock cycles long, then the correlation window will open 29 after the elapsed time from the first correlation equals the sweep period and will close after the elapsed time from the 31 first correlation equals the sweep period plus two ClOCk 32 cycles. This feature allows the circuit R12 to reject 33 spurious correlations outside the correlation window, while enabling the circuit to accommodate slightly varying clock frequencies.
36 R13-SYNCHRONIZATION STATUS OUTPUT: The "lock"/°'n0 lock"
status of the synchronization circuitry R12 is fed out to 3$ the user 42 through the synchronization status output line - 16 - ..

., r r SFP/hl-987 PATENT APPLICATION
R13.
The output of the comparator circuitry is therefore 3 further processed by the above described data stream synchronizer circuitry. The data stream synchronizer limits the recognition of correlator R10 output to specific windows of time corresponding to expected correlation times as determined by the received data stream.
8 Because this type of above described matched filter can be used to match arbitrary frequency ranges, it can be used in applications which involve modulated carriers, either by 11 processing the received signal through a mixer and using the 12 matched filter at baseband, or alternately by skipping the 13 mixer stage entirely and matching the received signal 14 directly at its modulated frequency. The trade-off is the cost of the mixing stage versus the additional number of 16 bits in the delay line shift register to handle the higher 17 frequency of the modulated frequency.
18 R14-DATA OUTPUT: Correlations occurring in the 19 correlation window are provided to the user (i.e., the product specific logic 42) by the data output line R14.
21 Figure 3'shows waveforms (signal amplitude versus time) 22 for the operation of the receive stage 44 of Figure 2. As 23 shown in Figure 3, waveform A shows the signal chirp 24 received at point A designated in Figure 2 (the log amplifier R3 output). Waveform B shows the signal at point 26 g~ which is the output of the slope detector comparators R4, R5. Waveform T is an indication of time. Waveform C shows 28 the signals at point C, which is the output of the resistor 29 networks R8, R9. As shown, waveform C is the analog correlation signal which has a lower limit of 0 volts, and 31 an upper threshold voltage of VAC. Waveform D is the 32 digital signal at the output of the correlation detector R10 33 at point D; waveform D is either a "zero°' chirp or a "one"
34 chirp.
The following describes the TRANSMIT stage 46.
36 Sl-DATA IN: DATA IN line S1 accepts Strobed data from 3y user 42. If the DATA BUFFER FULL line S1 is set, all 3$ strobed data will be ignored.

~r/J r~o :lJ ~'~ lr ~.~

1 S2-DATA BUFFER: DATA BUFFER S2 is a one-bit "double 2 buffer" allowing the user 42 to queue up the next bit of data while the previous bit is still being transmitted.
4 This allows for continuous data transmission and is necessary fox maintained synchronization. When the COUNTER
S4 has reached its terminal count, data is loaded from the BUFFER S2, the COUNTER S4 is cleared and started, and the DATA BUFFER S2 is cleared. If no data is present in the DATA HUFFER S2 when terminal count is reached, transmission is halted until further data is loaded. One complete 11 counting cycle of the COUNTER S4 corresponds to the 12 transmission of the swept waveform.
13 S3-DATA BUFFER FULL : Lane S3 alerts the user 42 that 14 the DATA BUFFER S2 has been filled. If the DATA BUFFER FULL
line S3 has been set. any further data presented on the DATA
16 IN line S1 is ignored.
1~ S4-COUNTER: Counter S4 is a nine bit counter driven by 18 clock S4C which generates consecutive addresses. These 19 address the ROM S6 which contains a lookup table holding values proportional to the transmit waveform voltages.
21 S5-TERMINAL COUNT DECODE LOGIG: Logic S6 generates an 22 appropriate signal when the terminal count of the counter S4 23 has been reached. This state corresponds to the completion 24 of a sweep transmission.
S6-ROM: The chirp pattern ROM S6 (read only memory) is a6 addressed by the COUNTER S4 and contains a lookup table 27 whose values are proportional to the swept waveform voltages.
29 S7-DATA SELECT & ENABLE: Logic S7 performs amplitude inversions on the binary values present at the ROM S6 31 output. Inversion is selected/deselected by the level of 32 the DATA line.
33 gg-LATCH: Latch S8 holds the output value of the DATA
3~ SELECT & ENABLE S7 while the ROM S6 address is being redefined.
36 Sg-SUMMING NETWORK: Summing network S9 is a binary ' 3' weighted resistor summing n~awork which generates a voltage 38 proportional to the magnitude of the binary value presented ~' ~-"'n 'l' ::. :-' ~~ 't.' r ; ~:
i-1 at its inputs. Thus latch S8 and summing network S9 together are a digital-to-analog converter.
S10-P04~ER A~1P~IFIER: Power amplifier S10 provides the 4 necessary drive current to drive the secondary winding of the ISOLATION TRANSFORMER R1 (i.e., line coupler R1) with a 6 voltage proportional to the output of the SUMt~iING NETWORK
7 S9.
8 Figure 2 also shows transmit/receive selector line 511.
Figure 4 shows waveforms for the operation of the transmitter stage. Waveform E is the data signal supplied 11 from the product specific logic 42 (see Figure 1) at point E
12 of Figure 2. Waveform F is the input clock signal at point 13 F, Waveform G is the output signal from the counter S4 at 14 point G. Waveform i3 is the driver enable signal provided by the terminal count decode logic S5 to the line driving 16 amplifier S10 at point H. Waveform 2 is the driving 17 amplifier S10 output signal (i.e., the wave signal on the AC
18 powerline) at point I. As shown, the signal at point I is 19 floating (i.e., indeterminate) when the driving amplifier S10 is not enabled.
21 The psuedo-random code sequence modulation referred to 22 above conventionally has the frequency spectrum 23 characteristics of a bandspread of twice the modulating code 24 rate between first zero nodes centered on the carrier frequency, corresponding to the two sidebands of the 26 modulating code. The shape of the power spectrum would be (sin x)/x with the 3 db bandwidth of the signal 28 approximately .88 times the code rate.
29 The frequency spectrum produce by a swept frequency chirp is effectively a single sideband signal, as opposed to 31 the dual sideband signal of the psuedo-random code. The 32 shape of the power spectrum will be reat~ngular wer the 33 frequencies being swept and the edges of the spectrum are 34 sharp. Thus the use of the given bandwidth is mush more effective witfi the swept frequency chirp. This may be 36 advantageous in applications where available bandwidth is 3y narrow and the absence of spectral spillover is an important 3~ consideration.

''l '; ~'' 'e ~, .a ~i 1 For the sign of slope detector in the case of a psuedo-2 random code sequence chirp, the optimal RC value is 3 constant. The optimal slope time of the RC time constant is determined by the carrier frequency, and is one-half the period of the carrier. The summing network should use all 6 but the element corresponding to the ideal signal carrier phase changes in the delay line storage to get optimal 8 results.
The RC time constant for the swept frequency chirp is ideally not constant. Theoretically it should be different 11 for each frequency. To accommodate the need for a varying 12 RC time constant for optimal results, one of two techniques 13 are preferably used. First, the filter is divided into 14 several sections corresponding to the different optimal slope times. For example, in a filter designed for a 16 150Khz-~SOKhz swept signal, two slope comparators R4, R5 and 1~ two corresponding delay lines R6r R7 are used -- one of each 18 for the higher frequencies and one of each for the lower 19 frequencies. This is the circuitry shown in Figure 2. Two sections provide an adequate performance for this particular 21 chirp. For chirps having more octaves of range (i.e., wider 22 bandwidth), the filter preferably is divided into more 23 sections.
24 If the chirp is swept from low to high frequency, the high frequency delay line R6 need contain only as many bits 26 as there are high frequency sample points (e. g. one-half) 2y since they are received last, and the low frequency delay 28 line R7 can be clocked at a lower rate (e. g. one-half) 2g therefore requiring less storage bits. The result for a filter for the chirp described is the amount of storage 31 required is the same as if there was one delay line clocked 32 at the higher frequency (180 bits).
33 An alternate technique is to convert the received high 34 frequency slopes to longer time slopes in the delay line register. This can be accomplished by combining multiple 36 stored slopes into one slope element to estimate the sign of 3y the longer time slope. The same number of bits can be used 38 in the delay line shift register, simply converting them to - a2 -E.J cl 5.. ~ '_7 ~_, ~J .'.

1 longer slope times at various points in the shift register. Alternately, since the longer slope times reflect lower frequencies, when conversion is made to the longer slope times, they can also be correspondingly compressed in time without losing resolution in the summing network. The 6 result is a potential savings of storage bits in the delay 7 line shift register by using in effect multiple sequential 8 shift registers at decreasing clock rates.
The above description of the invention is illustrative and not limiting. Other embodiments of the invention will 11 be apparent to one of ordinary skill in the art in light of 12 the disclosure.

Claims (81)

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

1. A method of transmitting sequences of information symbols on a communications channel between any of a plurality of transmitters and at least one receiver connected to the communications channel, comprising the steps of:
each transmitter transmitting a sequence of predetermined wide band signals, each transmitter resolving contention for use of the communications channel by sensing, prior to or during transmission of the sequence, for the presence on the communications channel of a said predetermined wide band signal transmitted by another transmitter, wherein the step of sensing for the presence of a predetermined wide band signal transmitted by another transmitter comprises using a filter that, without prior synchronization, produces an output indicative of receipt of a wide band signal.

2. The method of claim 1 wherein the predetermined wide band signals are used for contention resolution and for transmission of the information symbols.

3. The method of claim 2 wherein the receiver reconstructs the sequences of information symbols by a method comprising the steps of:
receiving signals from the communications channel and processing the received signals using a filter that, without prior synchronization, produces a filter output indicative of receipt of a wide band signal;
successively repeating the step of receiving signals to produce a sequence of filter outputs; and processing the sequence of filter outputs to reconstruct the sequence of information symbols.

4. The method of claim 3 wherein the same filter used by the receiver is used by the transmitter in the sensing step.

5. The method of claim 1, 3 or 4 wherein the filter comprises a matched filter.

6. The method of claim 5 wherein the matched filter is of the same time length as that of the transmitted wide band signal.

7. The method of claim 6 wherein the matched filter comprises a delay line.

8. The method of claim 5 further comprising a correlation detector for comparing the signals obtained by the matched filter to a predetermined signal level.

9. The method of claim 5 wherein the matched filter comprises a sign of slope filter.

10. The method of claim 1, 3, or 4 wherein the wide band signal is a spread spectrum chirp.

11. The method of claim 10 wherein the spread spectrum chirp is frequency swept.

12. The method of claim 10 wherein the information symbol is a single bit, and wherein each state of the bit is represented by a chirp identical except for a 180 degree difference in phase.

13. The method of claim 12 wherein the filter uses the same correlation network to generate a correlation output with a polarity reflecting which of the two states of the bit has been received.

14. The method of claim 11 wherein the spread spectrum chirp comprises at least one linearly swept portion.

15. The method of claim 10 wherein the spread spectrum chirp is a direct sequence code generated chirp.

16. The method of claim 10 wherein each information symbol is represented by multiple chirps.

17. The method of claim 1, 3 or 4 wherein the communications channel is a carrier-sense, multi-access network with contention resolution.

18. The method of claim 10 wherein the information sequences are packets of information symbols.

19. The method of claim 1, 3 or 4 wherein the sequence of wide band signals does not contain a synchronization portion.

20. The method of claim 1, 3 or 4 wherein the communications channel comprises a power line.

21. The method of claim 1 wherein the sequence of signals is transmitted at a uniform, predetermined rate and wherein each receiver, following a synchronization with at least one signal, uses the predetermined rate to lock synchronization for at least a plurality of further signals.

22. The method of claim 1, 3 or 4 wherein the sequence of signals is transmitted at a uniform, predetermined rate and wherein each receiver, following a synchronization with at least one signal, goes into a lock mode in which the receiver uses the predetermined rate to examine the output of the filter only during windows separated by a time period and synchronized with the expected times of receipt of successive signals.

23. The method of claim 22 wherein at least one signal comprises an initial sequence of symbols of a data packet.

24. The method of claim 23 wherein the sequence of wide band signals comprises a contention resolution preamble followed by data.

25. The method of claim 1, 3 or 4 wherein the sequence of wide band signals comprises a contention resolution preamble followed by data.

26. The method of claim 22 wherein the time period between windows is adjusted to allow the receiver to track a transmitter transmitting at frequencies within a frequency error range.

27. The method of claim 22 wherein, following a failure to receive a signal within the window, the receiver falls out of the lock mode and begins looking for a synchronizing pattern.

28. The method of claim 1, 3 or 4 wherein each of the information symbols corresponds to a unique wide band signal.

29. The method of claim 26 wherein the information symbols are bits, and each bit corresponds to at least one wide band signal.

30. The method of claim 5 wherein the communications channel is a portion of a carrier-sense, multiple-access network with contention resolution.

31. The method of claim 30 wherein the communications channel comprises a power line.

32. The method of claim 10 wherein the communications channel is a portion of a carrier-sense, multiple-access network with contention resolution.

33. The method of claim 32 wherein the communications channel comprises a power line.

34. The method of claim 19 wherein the wide band signal is a spread spectrum chirp.

35. The method of claim 34 wherein the communications channel comprises a power line.

36. The method of claim 22 wherein the wide band signal is a spread spectrum chirp.

37. The method of claim 36 wherein the communications channel comprises a power line.

38. Apparatus for transmitting sequences of information symbols on a communications channel between any of a plurality of transmitters and at least one receiver connected to the communications channel, comprising:
means at each transmitter for transmitting a sequence of predetermined wide band signals, the timing of the sequence of signals transmitted by at least one transmitter being asynchronous relative to that of at least one other transmitter, means at each transmitter for resolving contention for use of the communications channel by sensing, prior to or during transmission of the sequence, for the presence on the communications channel of a said predetermined wide band signal transmitted by another transmitter, including a said signal with timing asynchronous to that of the sequence signals generated by that transmitter, wherein sensing for the presence of a predetermined wide band signal transmitted by another transmitter comprises using a filter that, without prior synchronization, produces an output indicative of the degree of correlation between a received signal and the predetermined wide band signal.

39. The apparatus of claim 38 wherein the same predetermined wide band signals are used for contention resolution and for transmission of the information symbols.

40. The apparatus of claim 39 wherein the receiver comprises circuitry configured to reconstruct the sequences of information symbols by receiving signals from the communications channel and processing the received signals using a filter that, without prior synchronization, produces a filter output indicative of the degree of correlation between a said received signal and a said chirp signal; wherein the circuitry is configured to successively receive signals to produce a sequence of filter outputs and to process the sequence filter outputs to reconstruct the sequence of information symbols.

41. The apparatus of claim 40 wherein the same filter used by the receiver is used by the transmitter for sensing.

42. The apparatus of claim 38, 40 or 41 wherein the filter comprises a matched filter.

43. A method of transmitting a sequence of information symbols on a communications channel between a plurality of transmitters and at least one receiver, comprising the steps of:
representing the sequence of information symbols as a sequence of predetermined spread spectrum chirps;
performing a contention resolution process in which a transmitter senses whether other transmitters are transmitting;
following a successful contention resolution transmitting the sequence of chirps from the transmitter to the communications channel;
receiving signals from the communications channel at the receiver and processing the received signals using a filter that, without prior synchronization, produces a filter output indicative of the degree of correlation between a said received signal and a said chirp;
successively repeating the step of receiving signals to produce a sequence of filter outputs; and processing the sequence of filter outputs to reconstruct the sequence of information symbols.

44. Apparatus for transmitting a sequence of information symbols on a communications channel between a plurality of transmitters and at least one receiver, comprising:
a transmitter configured to represent the sequence of information symbols as a sequence of predetermined spread spectrum chirps, to contend for access to the communications channel by sensing whether other transmitters are transmitting, and following a successful contention resolution to transmit the sequence of chirps on the communications channel;
a receiver configured to receive signals from the communications channel and to process the received signals using a filter that, without prior synchronization, produces a filter output indicative of the degree of correlation between a said received signal and a said chirp;
the receiver being configured to successively process received signals to produce a sequence of filter outputs, and to process the sequence of filter outputs to reconstruct the sequence of information symbols.

45. A communications system comprising:
a plurality of transmitters each generating a plurality of signals, each signal being a chirp and being generated at a plurality of frequencies over a time interval;

a communications channel for carrying the signals;
each said transmitter including means for sensing a presence on the communications channel of a carrier signal prior to initiating a generation of signals, said carrier signal including a sequence of at least one chirp, and having timing which is asynchronous relative to that of at least one other of the plurality of transmitters; and means for receiving the signals from the communications channel, wherein the means for receiving includes means for self synchronizing on each signal of the plurality of signals.

46. The device of claim 45, wherein each chirp is a signal swept through a range of the frequencies.

47. The device of claim 46, wherein a range of the frequencies is from about 50 KHz to about 450 KHz.

48. The device of claim 46, wherein each chirp is swept linearly through the plurality of frequencies.

49. The device of claim 46, wherein each chirp is swept non-linearly through the plurality of frequencies.

50. The device of claim 45, wherein the communications channel is a portion of a carrier-sense multiple access based network.

51. The device of claim 45, wherein each signal contains only one bit of information, and wherein the means for synchronizing individually synchronizes each bit.

52. The device of claim 51, wherein the signals do not include a synchronization portion.

53. The device of claim 45, wherein each chirp is a pulsed frequency modulated signal.

54. The device of claim 53, wherein each pulsed frequency modulated signal represents a fraction of a data bit.

55. The device of claim 45, wherein the means for receiving further includes a matched filter for obtaining the signals from the communications channel.

56. The device of claim 55, wherein the means for receiving further includes a correlation detector for correlating the signals obtained by the matched filter to a predetermined signal level.

57. The device of claim 45, wherein the means for receiving includes data synchronizing means for terminating receipt of a sequence of the signals upon detection of an interruption in the sequence of the signals.

58. The device of claim 45, wherein the communications channel comprises a powerline.

59. The device of claim 45, wherein the communications channel comprises at least one radio frequency transmitter and receiver.

60. The device of claim 45, wherein the communications channel comprises at least one infrared transmitter and receiver.

61. The device of claim 45, wherein each transmitter includes means for modulating each signal by a pseudo-random code.

62. The device of claim 45, further comprising means included in each transmitter for resolving contention for the communication channel by detecting during a portion of each generation of signals a presence of said carrier signal on said communication channel from at least one other of the plurality of transmitters.

63. The device of claim 45, wherein each chirp is modulated in opposite polarity phases.

64. The device of claim 63, wherein each chirp represents a fraction of a data bit.

65. A method of transmitting data on a communications channel to a receiver comprising the steps of:
generating a plurality of signals at a transmitter, each signal being a chirp and being generated at a plurality of frequencies over a time interval;
providing the signals to a carrier-sense based communications channel;
sensing a presence on the communications channel of a carrier signal prior to initiating a generation of signals, said carrier signal including a sequence of at least one chirp, and having a timing which is asynchronous relative to that of at least one other transmitter associated with communications channel;
receiving the signals at the receiver from the communications channel; and determining a synchronization for each signal solely from the signal.

66. The method of claim 65, further comprising the step of sweeping each chirp through a range of the frequencies.

67. The method of claim 65, further comprising the step of providing only one bit of information in each chirp.

68. The method of claim 65, further comprising the step of terminating a receipt of the sequence of chirps upon detection of an interruption in the sequence of chirps.

69. The method of claim 68, wherein the step of terminating includes the step of detecting a predetermined sequence of bits.

70. The method of claim 65, wherein the step of generating includes the step of modulating each signal by a pseudo-random code.

71. The method of claim 65, further comprising the step of resolving contention for the communication channel by detecting during a portion of each generation of signals a presence of said carrier signal on the communications channel from at least one other transmitter.

72. A receiver for a network comprising:
means connected to the network for receiving an analog signal including a stream of chirps including a plurality of frequencies from the communications channel and converting the analog signal to a digital signal;
means for receiving the digital signal and storing the digital signal;
means for comparing the stored digital signal to a predetermined value and providing an output signal in response thereto; and means for self synchronizing and locking into the stream of chirps at a predetermined rate, wherein the receiver synchronizes itself independent of a transmitter of the analog signal.

73. The device of claim 72, further comprising synchronization means for receiving the output signal and synchronizing the output signal to the analog signal.

74. The device of claim 73, wherein the synchronization means includes means for detecting a predetermined output signal and synchronizing the output signal to analog signal in response thereto.

75. The device of claim 72, wherein the conversion means includes means for determining a direction of a slope of a portion of the analog signal and providing the digital signal as a function of the direction of the slope.

76. The device of claim 72, wherein the delay means includes at least one shift register for storing the digital signal.

77. The device of claim 72, wherein the comparing means includes:
summing means for summing the digital signal to a summed value; and correlating means for comparing the summed value to the predetermined value.

78. The device of claim 72, wherein the delay means and comparing means are portions of a matched filter.

79. The device of claim 72, wherein the comparing means includes means for storing the predetermined value, and wherein the predetermined value is a plurality of values.

80. A communication system comprising:
a plurality of transmitters each generating a plurality of signals, each signal being a chirp and being generated at a plurality of frequencies over a time interval, wherein the chirps are transmitted at a predetermined rate for emulation of a carrier signal;
a communications channel for carrying the signals;
means for sensing a presence of the carrier signal on the communications channel and thereby inhibiting a generation of signals by at least one transmitter, wherein each transmitter generates signals asynchronously to signals generated by at least one other transmitter; and means for receiving the signals from the communications channel, wherein the means for receiving includes means for self synchronizing on each signal of the plurality of signals.

81. A method of transmitting data on a communications channel to a receiver comprising the steps of:
generating a plurality of signals, each signal being a chirp and being generated at a plurality of frequencies over a time interval, wherein chirps are transmitted at a predetermined rate for carrier emulation;
providing the signals to a carrier-sense based communications channel;
receiving the signals at the receiver from the communications channel; and determining a synchronization for each signal solely from the signal thereby synchronizing independently of a source of the plurality of signals.