WO2008117311A1 - Powerline communication system at vhf/uhf - Google Patents

Abstract

Initial channel modeling and simulation of the Medium Voltage (MV) power lines (11-66 KV) in India have indicated very high bandwidths, comparable to Metro fiber. This communication channel has some static variations and (fading) dynamics. As a consequence a wide variety of wireline offerings are available - Gigabit scale backbone through broadband access, at a variety of repeater spacings. A unique characteristic of the MV powerline at VHF/UHF is that it can radiate, and transmit, simultaneously, resulting in a hybrid wireless-wireline channel, serving both high speed broadband as well as ubiquitous mobility applications, including 3G/4G. Transmission rates close to the Shannon bound can be achieved in these applications, using classical transceiver techniques.

Description

POWERLINE COMMUNICATION SYSTEM AT VHF/UHF

FIELD OF THE INVENTION

This invention relates to a powerline communication system at VHF/UHF. This invention further relates to a hybrid broadband wireless-wireline channel.

DISCUSSION OF PRIOR ART

The authors in [[1][2]] have discussed the theoretical and experimental characteristics (from India), of the MV power line channel at UHF/VHF. Gigabit scale capabilities are possible using repeatered systems. Deployment of this technology requires re-allocation of some/all VHF/UHF TV channels (8 MHz each) for BPL - discussions with authorities in India have indicated that this can be considered.

Figure 1 summarizes the results from [[1][2]]. Figure 1 shows the measured attenuation -.< (dark blue), attenuation from an ideal theoretical model without per pole mismatch loss (yellow), and attenuation from the ideal model with a per pole mismatch loss of about 1 dB (blue-green), linearly normalized for a 750 meter section, with about 20 poles spaced about 35 meters apart. Also shown is the free-space loss (violet) assuming an optimistic inverse-square law propagation over this distance. The span distance of 750 meters was selected to illustrate the capabilities (Gigabits/s) of a repeated wireline communication system based on their measurements. While this measured channel forms the basis of discussion with respect to the present invention, the qualitative insights carry over to other channels encountered in the highly variable MV grid. Moreover, the theoretical curves in Figure 1 (yellow, blue-green) as well as NEC simulations, indicate that the frequency selective behaviour of the MV channel is a fundamental feature. BRIEF DESCRIPTION OF THE INVENTION

The present invention discloses a hybrid broadband wireless/wireline channel, which has static variation with mildly fading characteristics and is frequency selective but amenable to classical equalization techniques In addition, the channel has the unique characteristic of both a radiative mode and a transmission mode, whose relative strengths depend on the frequencies used. For short distances, wireline signals are preferred, the balance shifting to wireless modes as distance is increased, beyond a frequency dependent threshold. The medium can be used simultaneously as a wireless/wireline hybrid, with broadband trunking being done using wireline mode and mobility using wireless mode. This invention primarily uses twin-pair lines and MIMO channels are similar with some additional constraints.

The specific examples of the capabilities of the hybrid channel of the present invention include:

1. Over 200MHz of bandwidth (at 200 MHz and 500 MHz), and using 500 mW per 8 MHz video channel, the Shannon bound is beyond STM- 16.. Achieved transceiver rates are around 1.4 Gbps, suitable for many broadband metro backhaul applications.

2. Over 200MHz of bandwidth, at ImW per video channel, the Shannon bound, and achieved rates, are both beyond GBE rates. 3. At 1 μW/channel, our transceivers operate beyond 2bps/Hz, and using both the

200 MHz and 500 MHz band, get past the STM-I barrier. 4. In the wireless mode, with 1 Watt of injected power on the line, the system can get past 100 Mbps Ethernet in a 8MHz video channel, over distances of 1.5 Km or more, using cell-phone size receive antennas, while simultaneously achieving the wireline rates above.

BRIEF DESCRIPTION OF DRAWINGS

Figure 1 describes the Measured Channel Attenuation, dB per 750 meter span. Figure 2 shows the MV powerline at VHF/UHF. Figure 3 (a) shows Channel Variability w.r.t Conductor Spacing 0.75m to 1.25m.

Figure 4 shows channel variability over different diameters of wires (from 1 cm to 3 cm).

Figure 5 (a) shows the rate of change of attenuation w.r.t frequency, for a range of conductor diameters 0.5 cm to 1.5 cm. Figure 5 (b) shows the fading Trajectory, 20 N/s@2-5Hz.

Figure 6 shows the radiation energy of 2 conductors perpendicular to the length of the wire and ground.

Figure 7 shows wireless versus wireline coverage (a) versus Frequency (b) Versus

Distance. Figure 8 shows channel capacities, over 175-275 MHz and 475-575 MHz band.

Figure 9 shows spectral efficiencies with total input power of 1 Watt over 175-275 MHz and 475-575 MHz band.

Figure 10 shows the impulse response of the channel.

Figure 11 shows a simulation model in MATLAB. Figure 12 shows the achieved data rates in simulation, w.r.t total power over a 100 MHz channel (12 video channels).

Figure 13 shows the Data Rate Versus Distance Curve.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

The basic feature of the MV powerOne channel at VHF/UHF is that it radiates in evanescent modes in addition to transmitting energy in the guided wireline mode. The channel at VHF/UHF (Figure 2) has frequency selective fading, CCITT antenna noise, NEXT/FEXT from adjacent channels and self pickup. The geometry changes from section to section, with mismatches in between. Major aspects of this variability are analyzed below.

A. ATTENUATION, FREQUENCY SELECTIVITY:

Channel attenuation is due to ohmic losses, radiation (though limited), and frequency selectivity is due to reflections, impedance and radiative resonances, etc. Attenuation for an ideal N-conductor MV line under certain excitation conditions, can be bounded using the equation from [2]:

Where a. is the attenuation, R_ohm is the ohmic resistance, the characteristic impedence Zo is given

z_θ %-**&* α>

^-4 - ftJΪ

where b is the distance between the wires and r is the radius of the wire. F(β] is given from [2] as

Where A(φ) is the array factor, which is frequency dependent. Using the formulae for

A(φ) from [2], F(β)for a twin-pair excited with current phasors [l,α,- ], can be written as

IsF (/?) = Jl+ 0 α π" +2β, cos(βώ cas{φ))dφ (4)

Results from this equation can be shown to be consistent with radiative perturbation to eigenmodes from a classical Quasi-TEM analysis ([4] - details omitted). Frequency selectivity is further enhanced due to mismatches causing reflections geometry changes etc. Theoretical and experimental results based on Equation (1) have been plotted in Figure 1 (taken from [2]).

B. CHANNEL VARIATION

The geometry of the MV grid, while much better controlled than the LT grid, is far from ideal, with changes in conductor diameter (weasel, squirrel), spacing, configuration (horizontal /triangular/vertical), splices, etc occurring frequently. The channel magnitude and phase response, poles, zeros, etc all vary from instance to instance. Transceivers for the MV channel have to be robust to these variations. The impact of changing conductor diameter and spacing can be estimated from Equation (1,2), by changing b, and r, thereby changing the characteristic impedance Zo and the factor F(β). Variations w.r.t conductor spacing are plotted in Figure 3 (a) and (b), from 100 MHz to 1 GHz. Figure 3 (b) it shows that small changes in the conductor spacing (+/- 2 cm possibly due to wind loads) can cause significant fades (upto 5-10 dB), near notches corresponding to high attenuation (e.g. between 200-400 MHz). Figure 4 further indicates that the depth of these notches is dependent on the conductor diameter, i.e. weasel/squirrel, etc. A change in diameter from 2.5 cm to 3 cm causes a 50 dB drop in channel response at 320 MHz. High quality training and adaptation algorithms have to be used to track the MV channel at VHFAJHF. This variation/fading is not due to multipath, but due to changes in the geometric parameters changing the impedance and radiation characteristics of the single channel. Configuration changes, multipath reflections, etc will cause further modulations over and above these effects. Additionally, a section is not of uniform geometry - the channel responses of individual portions of the section will vary. The end-to-end response is composed of the cascade of individual responses, and will show some averaging behaviour. However, these irregularities are less significant in MV lines compared to LT in general. Long straight regular sections of MV, running for kilometers in length are frequently encountered. Degradations due to major configuration changes can be eliminated by completely regenerating the signal there.

C. CHANNEL NOISE

Very few measurements of noise on the MV grid at VHF/UHF have been carried out. The channel noise of the system of the present invention are based on extrapolations of the Korean results [5] to beyond 30 MHz. Noise rapidly decreases and approaches levels about 20-30 dB above CCIR even at 30 MHz, and is expected to approach CCER. antenna noise at higher frequencies (see [2] for more details). In addition, NEXT/ FEXT are both present through near field induction as well as far field radiative pickup (possibly from other transmissions on the same grid), governed by the classical Friis transmission equation [3], and powerline harmonics. Powerline harmonics are much less at VHF/UHF compared to the standard 0-30 MHz BPL band, since they are blocked by the transformers, and have low energy in the VHF/UHF band to begin with. The CCIR antenna noise is used for channel capacity analysis

D. CHANNEL DYNAMICS

The MV lines are a relatively stable channel compared to low voltage (LT) lines, due to the high tension in the conductors. However, they exhibit a unique characteristic of wireline fading. Fading is primarily due to changes in geometry (spacing, sag, etc.) due to wind and/or similar dynamic forces. Exact analysis of geometry changes requires a computationally expensive 3-D full-wave calculation. A simplified approach is adopted and changes in wire-distance due to changes in wind-loads are considered in isolation, leading to changes in the conductor spacing b, and hence changes in the characteristic impedance Zo, and the array factor F(β). Also, a quasi-steady state analysis is performed assuming that fields are always at their static values corresponding to the instantaneous geometry (which changes with time). Hence taking the time derivative of Equation (1) is a valid approach.

First, the change in characteristic impedance can be related to change in distance between the wires as

dt b dt

Next, the change in the array factor can be expressed as:

^Α ₌ ²J _ 2a_tβ— cos(φ)sm{βdi cosfø))# (6) dt o dt Where ώ£ is the rate of change of distance between the wires.

This integral does not have a convenient closed-form solution (other than a power series), and has to be numerically evaluated. Finally, the caternary sag in the wires is given as ^wlA_τ • Thus the rate of change of distance between the wires can be related to change in tension on the wire as

wl² dT

(7) dt 8F² dt

Where w is the weight per unit length of the wire, 1 is the span length of the wire, and T is the tension in the wire, assumed directly modulated by the wind force F (dT/dt = dF/dt). To find the rate of change of attenuation due to change in wind force, differentiating equ. 1 we get (Rohm is constant)

This results in the governing nonlinear differential equation (dα/dt depends on both dF/dt, dZ/dt, and α itself)

where dF/dt, dZ/dt are related by equations (5), (6), and (7) to the changing wind force. The nonlinearity of these differential equations indicates that the channel characteristics can change non-harmonically w.r.t the wind-load spectrum. Exact solution for a time varying wind-force is complex -only present fade rates w.r.t a constant rate of change of wind-load are presented.

Figure 5 (a) shows the rate of change of attenuation w.r.t frequency, for a range of conductor diameters 0.5 cm to 1.5 cm. The wind force change has been assumed to be 20 N/s, which can be encountered in high wind situations. The fading shows a 300 MHz resonant behaviour (with non-fading frequencies due to cancellation between increased radiation and larger characteristic impedance) for the ideally straight section MV section model, real lines would exhibit more complex frequency dependency.

Figure 5 (a) indicates that the change in attenuation is less than 4dB/Km/sec below 500 MHz, rising to about 10 dB/Km/sec at IGHz. Indeed, the' actual fade rates will be less, because the wind direction cannot be assumed uniform across a 1 Km section - the fading would scale as a random walk in the a-domain. Figure 5 (b) shows a simulation from 100

MHz to IGHz, of the governing differential equation (9), with a time varying wind load

(average 20 N/s), with a temporal spectrum of upto 2-5 Hz. The total fading over a few minutes is limited to within 20 dB except at IGHz (< 10 dB in the 200/500 MHz passbands). This relatively slow dynamics indicates that very fast tracking equalizers, or complex power control algorithms, are not required for this channel. In the present invention, equalizers converged in 2000 samples (20 microseconds) at the symbol rate

(100 Msamples/s), much faster than these dynamics. Cross talk due to radiation from other portions of the grid exhibits fast fade dynamics, but is limited in power for the ranges of antenna gains encountered in the MV grid.

E. WIRELESS MODE

The MV grid at VHF/UHF radiates energy, which can be exploited as a channel also, when permitted by the radio licensing authority. Since wireless attenuation follows a power law (1/R² , 1/R³ ), whereas wireline attenuation is exponential, wireline signals are well matched to broadband trunking with short repeater spans, while the radiated signal is excellent for mobility applications. It can be shown that for an exponentially attenuated current distribution along the z-axis, with propagation constant α+jβ, the electric field decays as

Aπr (a + jβ- jβcos(θ)) ' ' r

leading to a 1/R² reduction in power density, with distance. The presence of ground, multiple conductors, etc generally increases the rate of decay, but exact answers have to be obtained numerically.

Figure 6 shows the fields for a twin-wire line including ground, computed using the Numerical Electromagnetic Code (NEC), up to 50 meters perpendicular to the direction of the length of wire, at a longitudinal distance of 30 meters from the transmitter operating at IW. The electric field of 40 mV/meter is well above wireless threshold, even when extrapolated to distances beyond 1 Km. This radiative coverage can be estimated using the power picked up by a small 5cm x 5cm antenna. Figure 7 shows the received power (from NEC) assuming a 0 dBW source for both the wireless and wireline modes (wireline power available is that available by directly contacting the line). Figure 7 (a) shows the relative power as a function of frequency, for a fixed span distance of 750 meters - wireless mode power is high (at 250 MHz) when wireline mode power is low and vice versa. The tradeoff as a function of distance at 175 MHz in Figure 7 (b) shows that wireless mode starts dominating wireline after about IKm.

An analysis based on eigenmodes corresponding to zeros of Hankel functions can be performed, and shows that the wireless mode is due to excitation of evanescent modes fed from a guided wireline TEM mode at pole mismatches. As such there are multiple radiators, and the system has MIMO behaviour. Details are disclosed in other applications. F. CHANNEL CAPACITY: IDEAL LINES

A rigorous analysis of the capacity of the MV grid at VHF/UHF has to take both the wireline and wireless capacity, and require concepts from the classical MIMO log-Det results of Foschini and others. From Shannon's Information Capacity Theorem, the capacity of an AWGN wireline channel is defined to be

C= B 1Og₂(I + SNR) bits/sec,

where B is the bandwidth utilized for transmission of signal and SNR is the Signal to Channel Noise ratio, adjusted for system margin/gap (8.8 dB for QAM), with waterfilling being used for frequency selective channels. Capacity estimates have been made for the MV channel in Figure 1, using CCITT noise temperatures (100-1000 K for VHF/UHF), - impulse and other powerline specific noise is limited at VHF/UHF. Figure 8 and Figure 9 show the Shannon capacities and spectral efficiencies over the channel passbands at 175- 275 MHz and 475-575 MHz, given an input signal between lmW-500 mW (27 dBm) over each 8 MHz channel (12 of them in each 100 MHz band). Over each band, spectral efficiencies peak at about 15bps/Hz, resulting in aggregate system capacity exceeding 2.5Gbps (STM- 16). Even with just 100 mW of power per channel, aggregate Shannon capacity exceeds 2Gbps over the two bands. At just ImW per channel, the two bands together can give more than lGpbs, going past GBE rates.

G. TRANSCEIVER SPEEDS: SIMULATION RESULTS

QAM Transceivers were designed for the MV powerline channel, and simulated in MATLAB/SIMULINK using Communication Blockset and Signal Processing Blockset. The simulation model (Figure 11) is a baseband model of the communication system with the channel modeled as a long FIR to match the measured channel characteristics (which includes impedance and radiative resonance, multipath, etc). The system model consists of a random data generator, M-ary QAM modulator, a raised cosine pulse shaping filter followed by the channel model. The channel noise is modeled as AWGN — powerline specific impulse noise is less significant at VHF/UHF. The receiver model consists of an equalizer block, a raised cosine receive filter followed by a Gardner symbol timing recovery block and an M-ary QAM demodulator block.

Figure 10 shows the impulse response of the channel as used for the simulations. The response is very long (about 1000 samples in length at a 100 Msymbol/second sample rate), corresponding to 10-100 microsecond tails due to various resonances, reflections, etc.

Figure 12 shows the achieved raw rates obtained in simulation on baseband communication over each of the 175-275 MHz and 475-575 MHZ bands. These and all succeeding numbers are for a raw uncorrected channel operating around 10-4 BER. Half rate and 2/3 rate convolutional encoders have enabled us to lower the BER below maximal acceptable limits for communication systems (< 10^"12). Figure 12 shows that the channel offers wide opportunities from very high speed transmissions (2 x STM-4) with reasonable power (500 mW per sub channel) to high speed ( 2x STM-I) transmission with extremely low power (1 microwatt per sub channel) . The transceiver design is a conventional QAM receiver, adapted to the MV channel using a front end equalizer. A Zero-Forcing equalizer is acceptable for high SNRs, but the performance degrades rapidly as the SNR is lowered. Normalized LMS performs much better in this scenario, with shorter equalizer lengths at high input symbol rate. A comparison between the ZFE and NMLS under similar performance conditions of input signal strength, SNR and modulation scheme is tabulated in Table 1. DFE structures perform even better.

The computational requirement of 5Gflops for the NMLS (1.2 Gflops using DFE) is achieveable using modern DSP's/ASICS. NMLS convergence was within 2000 samples, corresponding to 20 microsecond adaptation, much faster than the channel dynamics of upto 4dB/Km/second . Based on this design, a variety of transceivers can be configured, and are tabulated in Table 2. These transceivers use the 175-275 MHz and 475-575 MHz band (each of 100 MHz bandwidth). The channel being modeled is by default the 750 meter measured MV line. Double length lines (1.5 Km) can also be used with high front end receiver gains, upto 4-QAM. It is clear from Table 2 that the MV aerial grid offers metro fiber class trunking bandwidth (STM-4/STM-16), very low power broadband access bandwidth (STM-I), as well as ubiquitous broadband wireless mode coverage (100 Mbps+ in an 8MHz channel). The coverage pattern in wireless mode is not circular, but follows the grid. This can be an advantage, since the MV grid tracks habitation. Figure 13 shows a distance-data-rate curve for wireline mode at 500 mW per 8 MHz channel, using 100 MHz bandwidth. A variety of offerings are available, ranging from almost 700 Mbps at 750 meters down to 100 Mbps over 2 Km.

The method and apparatus of the present invention utilizes transceivers well known in the state of art of CATV and OFDM (including an inverse FFT a channel normalizing 'gain and an FFT) alongside frequency hopping well known in the state of art of CATV and OFDM is used- to avoid bad spectral regions of the channel. Further, the invention is capable of finding its use in 3G/4G backhaul and powergrid monitoring.

REFERENCES

1. Prasanna, G.N.S,"The MV aerial power grid at VHFAJHF rivals fiber incapacity, "Optical Fiber Communication Conference, 2006.

2. Prasanna, G.N.S,"Aerial MV lines at VHFAJHF: Quasi-TEM Analysis and Experimental Results," IEEE International Symposium on Power Line

Communications and Its Applications, 2006.

3. Jordan, E.C., and Balmain, K.C., Electromagnetic Waves and Radiating Systems, Prentice-Hall, India, 2nd Ed, 1968

4. Amore, M. D, Sarto, M.S., A New Formulation of Lossy Ground Return Parameters for Transient Analysis of Multi-Conductor Dissipative Lines, IEEE

Transactions on Power Delivery, VoI 12, No 1, Jan 1997, pp 303-314. 5. Lee, JJ, et al, Measurements of the Communications environment in Medium Voltage Power Distribution Lines for Wide-Band Power Line Communication, ISPLC, 2004.

Claims

CLAIMS:

1. A method for transmitting a broadband signal over a set of one or more conductors of a power line network, such as a ubiquitous medium voltage (MV) or low voltage network, transmission occurring at UHF/VHF wherein a transmission channel is divided into a plurality of sub channels each being separated from the other in time, or in carrier frequency, comprising the steps of: a. Allocating unused sub channels which result in using those channels for transmission as per the invention, which aforesaid channels have not been earmarked for broadcast communication using electromagnetic radio wave propagation; b. Generating and Transmitting a waveform over two or more conductors in free space; c. Receiving a waveform with a loss less than free space loss, over two or more conductors and/or one or more antennas. d. Coupling a low power signal to a high power transmission, MV or LV line; e. Signal Processing; f. Regeneration; and g. Error Correction.

2. A method as claimed in claim 1, wherein the step of allocating unused sub channels includes a step of: a. Spectral shaping to preferentially occupy spectral notches in the spectrum of said broadcast, said spectral notches exemplarily caused by horizontal or vertical flyback in the case of TV broadcast, b. Transmission at a power level non-interfering to said broadcast communication.

3. The method as claimed in claim 1, wherein the step of transmission uses a modulation scheme suitable for OFDM implementation.

4. The method as claimed in claim 1, where transceivers well known in the state of art of OFDM (including an inverse FFT a channel normalizing gain and an FFT) are used.

5. The method as claimed in claim 1, wherein frequency hopping well known in the state of art of OFDM is used to avoid bad spectral regions of the channel.

6. The method as claimed in claim 1, wherein the step of transmission uses a modulation scheme used by TV or CATV transmission.

7. The method as claimed in claim 1, wherein the step of signal generation, processing, or reception includes the step of using standard TV, CATV and/or FM equipment, exemplarily modulators, amplifiers, and modems.

8. The method as claimed in claim 1, an agile (i.e. frequency variable) or fixed frequency CATV modulator is used for signal generation and transmission.

9. The method as claimed in claim 1 , a CATV receiver is used for reception.

10. The method as claimed in claim 1, wherein a CATV transceiver is used for both transmission and reception.

11. The method as claimed in claim 1 , wherein a CATV amplifier is used for boosting the signal at a point in between transmission and reception.

12. The method as claimed in claim 1, wherein a CATV amplifier is used followed by a signal processor, to boost the signal level, and sharpen the transitions for an appropriate modulation scheme.

13. The method as claimed in claim 1, wherein a CATV transceiver is used with a pre and post equalizer either separately attached or integral with it, for both transmission and reception, to have the channel seen by the said CATV transceiver resemble CATV.

14. The method as claimed in claim 1, wherein the step of transmission includes the step of two or more conductors being excited simultaneously, with an appropriate choice of relative phase and/or amplitude for reducing radiation, and hence reducing channel attenuation.

15. The method as claimed in claim 1, wherein the step of transmission includes the step of two or more conductors being excited simultaneously, with an appropriate choice of relative phase and/or amplitude for increasing radiation, and hence increasing overage by the wireless mode of transmission. •

16. The method as claimed in claim 1, wherein the step of reception further includes signal detection on multiple conductors, and/or antennas located possibly close to the two or more conductors.

17. The method as claimed in claim 1, wherein the step of transmission further comprises the step of adaptive amplitude modulation (power control) and/or phasing of currents injected at multiple conductors, to reduce radiation, and attenuation, the adaptation being done by co-operative signaling between the steps of transmission and reception.

18. The method as claimed in claim 1, wherein the step of transmission further comprises the step of signaling and/or data channel spread over multiple amplitude/phase/frequency excitations at each conductor, for diversity gain.

19. The method as claimed in claim 1, wherein the step of transmission further comprises the step of using multiple frequency channels and/or multiple timeslots for multiple accesses on the MV line.

20. The method as claimed in claim 1, wherein the step of transmission further comprises the step of using the MV backbone for backhauling (i.e. connecting a point with switching infrastructure to the antennas) cellular wireless and/or wire line signals.

21. The method as claimed in claim 17, wherein the MV backbone is used for implementation of 4+G systems, using possibly small antennas located close to each other (micro-cells).

22. The method as claimed in claim 1, wherein the step of signal processing uses Reed-Solomon and/or Turbo coding for improving SNR or reducing emissions.

23. The method as claimed in claim 1, wherein the step of transmission further comprises the step of using multi-conductor and/or multi-antenna excitation and/or reception, possibly with optimal amplitude and/or phase, to improve SNR and/or bit rate and/or reduce emissions and/or other performance metrics.

24. The method as claimed in claim 1, wherein the step of regeneration further comprises the step of placing repeaters and/or regenerators periodically, with a repeater or regenerator handling one or more channels, said repeaters being exemplary amplifiers, and regenerators being exemplary decoders that completely decode the signal, and/or sharpen the transitions.

25. The method as claimed in claim 1, wherein the step of allocating unused sub channels includes the step of frequency planning and/or allocation, to minimize interference between adjacent systems carrying possibly different data. This step exemplarily allocates sub channels whose geographic coverage follows the contours of the MV grid, and which said coverage need not be approximately circular.

26. A method as claimed in claim 1, wherein the step of allocating unused sub channels includes the step of matching the frequency spectrum of the modulation chosen to the audio, chrominance and luminance signals of a TV broadcast system (including frequency notches and tilt).

27. A method as claimed in claim 1, where both wireless and wire line modes are used, exemplarily with both wire line transceivers and wireless transceivers.

28. A method as claimed in claim 16 being used as a multi-access channel connecting all transceivers on the same power line.

29. A method as claimed in claim 25 being used as a multi-access channel, with lower frequencies allocated to further away transceiver pairs, possibly to improve SNR, reduce power and or increase bit rate.

30. A method as claimed in claim 25 being used to perform 3G/4G backhaul and for powergrid monitoring.

31. An apparatus for data communication at UHF/VHF over a power line network wherein a transmission channel is divided into a plurality of sub channels each being separated from each other in time or in carrier frequency, comprising: a. Transmission means (transmitter); b. Reception means (receiver); c. Coupling means (coupler) by using known methods; d. Signal Processing means (signal processing); e. Regeneration means; and f. Error correction means.

32. The apparatus as claimed in claim 31, wherein the signal processing means comprises standard CATV and/or FM modulators and modems.

33. The apparatus as claimed in claim 31, where transceivers well known in the state of art of OFDM (including an inverse FFT a channel normalizing gain and an

FFT) are used.

34. The apparatus as claimed in claim 31, wherein frequency hopping well known in the state of art of OFDM is used to avoid bad spectral regions of the channel.

35. The apparatus as claimed in claim 31, wherein the transmission means performs auxiliary filtering by means of pre-shaping.

36. The apparatus as claimed in claim 31, wherein the reception means performs auxili ary filtering by means of post-shaping.

37. The apparatus as claimed in claim 31, wherein the transmission means comprises means for multiple conductors being excited simultaneously for reducing relative radiation, and hence reducing channel attenuation.

38. The apparatus as claimed in claim 31, wherein the transmission means comprises means for multiple excitations being replaced by signal detection on multiple conductors, and/or antennas located possibly close to the MV line.

39. The apparatus as claimed in claim 31, wherein the transmission means comprises fixed or adaptive amplitude modulation (power control) and/or phasing of currents injected at multiple conductors, to reduce radiation, and attenuation, the adaptation being done by co-operative signaling between the steps of transmission and reception.

40. The apparatus as claimed in claim 31, wherein the transmission means comprises signaling and/or data channel spread over multiple amplitude/phase excitations at each conductor, for diversity gain.

S 41. The apparatus as claimed in claim 31, wherein the transmission means comprises using multiple frequency channels and/or multiple timeslots for multiple access on the MV line.

42. The apparatus as claimed in claim 31, wherein the transmission means comprises0 using the MV backbone for backhauling wireless and/or wire line signals.

43. The apparatus as claimed in claim 31, wherein the transmission means comprises the MV backbone being used for implementation of 4+G systems, using possibly small antennas located close to each other (micro-cells). 5

44. The apparatus as claimed in claim 31, wherein the signal processing means comprises using Reed-Solomon and/or Turbo coding used for improving SNR or reducing emissions. 0

45. The apparatus as claimed in claim 31, wherein the transmission means comprises using multi-conductor and/or multi-antenna excitation and/or reception, possibly with optimal amplitude and/or phase, to improve SNR and/or bit rate and/or reduce emissions and/or other performance metrics. 5

46. The apparatus as claimed in claim 31, wherein the regeneration means comprises placing repeaters and/or regenerators periodically, with a repeater or regenerator handling one or more channels, said repeaters being exemplary amplifiers, and regenerators being exemplary decoders that completely decode the signal, and/or sharpen the transitions. 0

47. The apparatus as claimed in claim 31, wherein the transmission means comprises frequency planning and/or allocation, to minimize interference between adjacent systems carrying possibly different data.

48. The apparatus as claimed in claim 31, wherein the transmission means comprises matching the frequency spectrum of the modulation chosen to_. the audio, chrominance and luminance signals of a TV broadcast system (including frequency notches and tilt).

49. The apparatus as claimed in claim 31, wherein the transmission means comprises choosing the frequency spectrum of the modulation such that it occupies the spectral notches of a vanilla TV signal, and possibly at a level transparent to TV sets.

- 50. The apparatus as claimed in claim 31, being used in both wireless and wire line modes.

51. The apparatus as claimed in claim 31, being used as a multi-access channel connecting all transceivers on the same power bus.

52. The apparatus as claimed in claim 31, being used as a multi-access channel, with lower frequencies allocated to further away transceiver pairs, possibly to improve SNR, reduce power and or increase bit rate.

53. The apparatus as claimed in claim 31 being used to perform 3G/4G backhaul and for powergrid monitoring.