CN102461247B

CN102461247B - Sensor-based wireless communication systems using compressive sampling

Info

Publication number: CN102461247B
Application number: CN201080026620.2A
Authority: CN
Inventors: 托马斯·阿洛伊修斯·塞克斯顿; 西恩·巴塞洛缪·塞门斯; 保罗·詹姆斯·卢斯那; 詹姆斯·厄尔·沃马克
Original assignee: BlackBerry Ltd
Current assignee: Maliki Innovation Co ltd
Priority date: 2009-04-15
Filing date: 2010-04-15
Publication date: 2014-09-03
Anticipated expiration: 2030-04-15
Also published as: EP2420001A2; CA2758937C; CA2758937A1; WO2010121050A3; CN102461247A; WO2010121050A2

Abstract

Methods, devices and systems for sensor-based wireless communication systems using compressive sampling are provided. In one embodiment, the method for sampling signals comprises receiving, over a wireless channel, a user equipment transmission based on an S- sparse combination of a set of vectors; down converting and discretizing the received transmission to create a discretized signal; correlating the discretized signal with a set of sense waveforms to create a set of samples, wherein a total number of samples in the set is equal to a total number of sense waveforms in the set, wherein the set of sense waveforms does not match the set of vectors, and wherein the total number of sense waveforms in the set of sense waveforms is fewer than a total number of vectors in the set of vectors; and transmitting at least one sample of the set of samples to a remote central processor.

Description

Use sensor-based wireless communication system of compression sampling

Technical field

The disclosure is usually directed to wireless communication system, and more specifically, relates to for use the method, apparatus and system of compression sampling at sensor-based wireless communication system.

Background technology

Wireless communication system is disposed to provide for example various voice-and-data related services widely.Typical wireless communication system comprises the multi-access communication network that allows user to share common network resource.The example of these networks is time division multiple access (" TDMA ") system, code division multiple access (" CDMA ") system, single-carrier frequency division multiple access (" SC-FDMA ") system, OFDM (" OFDMA ") system or other similar systems.Utilize various technical standards to support OFDMA system, such as evolved universal terrestrial wireless access (" E-UTRA "), Wi-Fi, World Interoperability for Microwave Access, WiMax (" WiMAX "), Ultra-Mobile Broadband (" UMB ") and other similar systems.In addition, utilize various normal structures (such as third generation partner program (" 3GPP ") and 3GPP2) standard of exploitation the realization of these systems is described.

Along with the evolution of wireless communication system, introduced the more senior network equipment that Enhanced feature, function and performance are provided.This high-level network equipment can also be called Long Term Evolution (" LTE ") equipment or Long Term Evolution senior (" LTE-A ") equipment.LTE is structured on the success of high-speed packet access (" HSPA "), and described high-speed packet access has higher average and peak-data throughput rate, lower delay and better user experiences, particularly in high demand geographic area.LTE, by using wider spectral bandwidth, OFDMA and SC-FDMA air interface and senior antenna method, has realized this higher performance.

Can use single-input single-output system (" SISO "), single input multiple output system (" SIMO ") and multi-input multi-output system (" MIMO ") to set up the communication between subscriber equipment and base station, wherein, in described SISO system, only an antenna is simultaneously for Receiver And Transmitter; In described SIMO system, at receiver place, use many antennas, and at transmitter place, only use an antenna; In described mimo system, at Receiver And Transmitter place, all use many antennas.Compare with SISO system, if use many transmitting antennas, many reception antennas or the two, SIMO can provide the coverage of increase, and mimo system can provide the spectrum efficiency of increase and the data throughout of Geng Gao.

In these wireless communication systems, the input in noise and estimation are general.Sampling thheorem provides the ability that continuous time signal is converted to discrete-time signal, thereby allows input and algorithm for estimating carries out efficiently and effectively realize.The so-called Shannon theorem of well-known sampling thheorem, and the necessary condition relevant to frequency bandwidth is provided, to allow the Exact recovery to arbitrary signal.Described necessary condition is to this signal, to sample with the minimum twice of the peak frequency of signal, and this sampling rate is also defined as Nyquist speed.The sampling of Nyquist speed has following defect: require expensive high-quality assembly, this assembly requires a large amount of power and becomes the sampling that is originally supported in large frequency place.In addition, the sampling of Nyquist speed changes according to the peak frequency of signal, and does not require the knowledge of any other attribute of this signal.

Part in these difficulties is difficult, compression sampling provides the new frame for signal sensing and compression, wherein, utilize the particular community " sparse property " of input signal to reduce and represent that reliably signal does not lose the number of the needed value of expectation information simultaneously.

Accompanying drawing explanation

In order to facilitate those of ordinary skills to understanding of the present disclosure and to be put into effect, now to quoting with reference to the illustrated example embodiment of accompanying drawing.In whole accompanying drawings, similar reference numeral refers to unit similar in identical or function.According to the disclosure, accompanying drawing and embodiment are merged in specification, form a part for specification, and for further illustrated example embodiment and various principle and advantages are described, in the accompanying drawings:

Fig. 1 is exemplified with according to an embodiment of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.

Fig. 2 is exemplified with according to another embodiment of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.

Fig. 3 is exemplified with according to another embodiment of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.

Fig. 4 is exemplified with according to an embodiment of the compression sampling system of the various aspects of setting forth herein.

Fig. 5 is according to the flow chart of an embodiment of the compressive sampling method of the various aspects of setting forth herein.

Fig. 6 is exemplified with according to another embodiment of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.

Fig. 7 is exemplified with according to an embodiment of the cut-in method in sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.

Fig. 8 is exemplified with according to another embodiment of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.

Fig. 9 is exemplified with according to an embodiment of the quantization method of the detector in sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.

Figure 10 is the chart being illustrated according to the example of the type of the rarefaction representation matrix using in sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein and sensing matrix.

Figure 11 is exemplified with can be according to an embodiment of the wireless device using in sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.

Figure 12 is exemplified with can be according to an embodiment of the transducer using in sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.

Figure 13 is exemplified with can be according to an embodiment of the base station of using in sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.

Figure 14 is exemplified with according to the simulation result of an embodiment of detected wireless devices in sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.

Figure 15 is exemplified with according to the simulation result of the performance of an embodiment of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.

Figure 16 is exemplified with according to the simulation result of the performance of an embodiment of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.

Figure 17 is exemplified with according to the simulation result of the performance of an embodiment of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.

Figure 18 is exemplified with according to the simulation result of the performance of an embodiment of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.

Figure 19 is exemplified with according to the simulation result of the performance of an embodiment of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.

Figure 20 is according to the example of the certainty matrix using in an embodiment of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.

Figure 21 is according to the example of the random matrix using in an embodiment of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.

Figure 22 is exemplified with the example of the incoherent sampling system in noise-free environment.

Figure 23 is exemplified with according to another embodiment of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.

Figure 24 is exemplified with the example of the harmless sampling system of prior art.

Figure 25 is exemplified with according to another embodiment of the sensor-based wireless communication system that uses compression sampling in noisy environment of the various aspects of setting forth herein.

Figure 26 is exemplified with according to another embodiment of the cut-in method in sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.

Figure 27 is exemplified with according to another embodiment of the sensor-based wireless communication system that uses compression sampling in noisy environment of the various aspects of setting forth herein.

Figure 28 shows according to another embodiment of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.

Figure 29 shows according to another embodiment of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.

Figure 30 show proposition according to the object run region of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.

Figure 31 shows according to another embodiment of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.

Figure 32 shows according to the embodiment of the frequency domain sample of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.

Figure 33 is according to the block diagram of the long-range sampler of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.

It will be appreciated by one of skill in the art that, for clarity, simply, and further help to strengthen the understanding to embodiment, the unit in accompanying drawing is carried out to illustration, but these unit needn't proportionally be drawn.

Embodiment

Although the following discloses exemplary method, equipment and system for using at sensor-based wireless communication system, those skilled in the art will appreciate that, in no case instruction of the present disclosure should be restricted to example shown.On the contrary, expection can realize instruction of the present disclosure in replacing configuration and environment.For example, although describe exemplary method as herein described, equipment and system in conjunction with the configuration of aforesaid sensor-based wireless communication system, but those skilled in the art will easily recognize and can in other system, use this exemplary method, equipment and system, and this exemplary method, equipment and system can be configured to as required corresponding to these other systems.Correspondingly, although exemplary method, equipment and the system used in sensor-based wireless communication system are described below, those of ordinary skills will understand is that, disclosed example is not to realize the sole mode of this method, apparatus and system, and accompanying drawing and describe to be considered to be illustrative and nonrestrictive in itself.

Various technology as herein described can be for various sensor-based wireless communication systems.Various aspects as herein described are rendered as the method, apparatus and system that can comprise a plurality of assemblies, unit, member, module, node, ancillary equipment etc.In addition, these method, apparatus and system can comprise or can not comprise add-on assemble, unit, member, module, node, ancillary equipment etc.In addition, can realize various aspects as herein described with hardware, firmware, software or their any combination.Be important to note that term " network " and " system " can exchange use.Relational terms as herein described, such as " ... on " and " ... under ", " left side " and " right side ", " first " and " second " etc. can be only for an entity or action separate with another entity or active region, and needn't require or imply these entities or move between this relation or sequentially of any reality.Term "or" is intended to mean the "or" of inclusive, rather than the "or" of exclusiveness.In addition, term " a " and " an " are intended to mean one or more, can clearly be intended to singulative unless otherwise specified or based on context.

Described wireless communication system can comprise a plurality of subscriber equipmenies and foundation structure.This foundation structure comprises in this wireless communication system not being the part of subscriber equipment, such as the combination of transducer, base station, core net, downlink transmitter, other unit and unit.Described core net can access other networks.Described core net (being also known as central brain (central brain) or remote central processor) can comprise high power infrastructure component, and it can be with acceptable financial cost, and the large function of calculating strength is carried out on two-forty ground.Described core net can comprise infrastructure element, this infrastructure element can with base station communication, thereby physical layer function also can be carried out by described core net.Described base station can be to downlink transmitter communicating control information, to overcome the communication barrier being for example associated with channel fading.Channel fading comprises how radio frequency (" RF ") signal can be flicked from a lot of reflectors, and resulting reflection and attribute.Described core net and base station can be for example identical infrastructure element; Share a part for same infrastructure unit; Or different infrastructure element.

Base station can be called Node B (" NodeB "), base station transmitting-receiving station (" BTS "), access point (" AP "), satellite, router or certain other technology of equal value.Base station can comprise RF transmitter, RF receiver or both, and they and antenna-coupled are communicating with subscriber equipment.

Transducer can be called long-range sampler, remote conversion equipment, distance sensor or other similar terms.Transducer can comprise for example antenna, receiving element, sampler, controller, memory and transmitter.Transducer can be mutual with for example base station.In addition, transducer can be deployed in the wireless communication system that comprises core net, and this core net can access another network.

The subscriber equipment using in wireless communication system can be called mobile radio station (" MS "), terminal, cell phone, cellular handset, personal digital assistant (" PDA "), smart phone, handheld computer, desktop computer, laptop computer, flat computer, net book, printer, Set Top Box, television set, wireless device or certain other equivalent technique.Subscriber equipment can comprise RF transmitter, RF receiver or the two, and they and antenna-coupled are to communicate with base station.In addition, subscriber equipment can be that fix or mobile, and can have ability mobile in wireless communication system.In addition, uplink communication refer to from subscriber equipment to base station, transducer or the communication of the two.Downlink communication refers to from base station, downlink transmitter or the two communication to subscriber equipment.

Fig. 1 is exemplified with according to an embodiment of sensor-based wireless communication system 100 of the use compression sampling of various aspects as herein described.In this embodiment, the system 100 robust high bandwidth real-time radio communication of high user density that can provide support.System 100 can comprise subscriber equipment 106, transducer 110 to 113, base station 102, core net 103 and other networks 104.Subscriber equipment 106 can be low-cost low-power equipment for example.Base station 102 can be used for example a plurality of low-cost low power sensors 110 to 113 to communicate by letter with subscriber equipment 106.

In Fig. 1, system 100 comprise with base station 102 coupling transducer 110 to 113, for from subscriber equipment 106 received communications.Base station 102 can be coupled with core net 103, and core net 103 can access other networks 104.In one embodiment, transducer 110 to 113 can interval for example about 10 meters to hundreds of rice.In another embodiment, can use single-sensor 110 to 113.It will be appreciated by those skilled in the art that when disposing sensor-based wireless communication system, between power consumption, lower deployment cost, power system capacity, other factors and the blocking factor of transducer, have balance.For example, along with the interval of transducer 110 to 113 becomes more approaching, the power consumption of transducer 110 to 113 may reduce, and lower deployment cost and power system capacity may increase simultaneously.In addition,, when approaching with transducer 110 to 113, subscriber equipment 106 can operate by the different RF band of the frequency range of using from bottom wireless network.

In current embodiment, transducer 110 to 113 can use respectively communication link 114 to 117 and base station 102 couplings, and communication link 114 to 117 can be supported for example optical fiber cable cable connection, coaxial cable connection, other connections or their combination in any.In addition, sensor-based information can transmitted each other in a plurality of base stations 102, to support various functions.Utilize for example combination of antenna, RF front end, baseband circuit, interface circuit, controller, memory, other unit or unit, transducer 110 to 113 can be designed to low cost.A plurality of transducers 110 to 113 can be for supporting the combination of for example aerial array operation, SIMO operation, MIMO operation, beam forming operation, other operation or operation.Persons of ordinary skill in the art will recognize that aforementioned operation can allow subscriber equipment 106 to send with lower power level, thereby cause for example lower power consumption.

In system 100, subscriber equipment 106 and base station 102 can be used for example procotol to communicate.Described procotol can be for example combination of cellular network protocols, Bluetooth agreement, wireless local loop (" WLAN ") agreement or any other agreement or agreement.Persons of ordinary skill in the art will recognize that cellular network protocols can be any one in a lot of standardized cellular network protocols of using in such as LTE, UMTS, GSM and other system.The part of the procotol that transducer 110 to 113 is performed can comprise a for example part for physical layer function.Persons of ordinary skill in the art will recognize that the performed reduction function of transducer 110 to 113 can cause the combination of power consumption, other advantages or the advantage of lower cost, less size, reduction.

Transducer 110 to 113 can be powered by for example combination of battery power source, interchange (" AC ") electric power source or other power sources or power source.Can use for example automatic repeat request (" ARQ ") agreement, be supported in the communication that comprises real time communication between transducer 110 to 113, subscriber equipment 106, base station 102, core net 103, other networks 104 or their arbitrary combination.

In current embodiment, transducer 110 to 113 can compress the uplink signal receiving (" f ") sending from subscriber equipment 106, to form sensing signal (" y ").Transducer 110 to 113 can use respectively communication link 114-117, and to base station, 102 provide this sensing signal (" y ").Then, this sensing signal (" y ") can be processed in base station 102.Base station 102 can be to transducer 110 to 113 move instructions, and wherein, described instruction can relate to for example combination of wave beam control, other instructions or the instruction of data transaction, oscillator tuning, use phase sample.In addition, subscriber equipment 106, transducer 110 to 113, base station 102, core net 103, other networks 104 or their arbitrary combination can be used the combination of for example media access control (" MAC ") hybrid ARQ agreement, other similar agreements or agreement to comprise the communication of real time communication.Equally, subscriber equipment 106, transducer 110 to 113, base station 102, core net 103, other networks 104 or their arbitrary combination can be used for example combination of existence signaling code (presence signaling code), empty time-code, fountain codes (fountain code), other communications codes or communications codes to communicate, described existence signaling code can be in the situation that do not need the cooperation of transducer 110 to 113 to operate, described empty time-code may require channel knowledge, and described fountain codes can be for registration and real-time Transmission.Some in these communications codes may require for example to apply the advantage that various signal processing technologies are utilized any inherent attribute of these yards.

In Fig. 1, following functions can be carried out in base station 102, such as transmitting system Overhead; Use the existence of transducer 110 to 113 detection subscriber equipmenies 106; Bidirectional real-time with subscriber equipment 106; The combination of other functions or function.Persons of ordinary skill in the art will recognize that transducer 110 to 113 can be very much more cheap than base station 102 and core net 103.

By with cycle rate, aperiodic speed or value that the two measures continuous time signal form discrete-time signal, carry out sampling.In current embodiment, the effective sampling rate of transducer 110 to 113 may be less than the actual samples speed of transducer 110 to 113 uses.Described actual samples speed is the sampling rate of analog-digital converter (" ADC ") for example.Effective sampling rate is measured in output at transducer 110 to 113, and it is corresponding to the bandwidth of sensing signal (" y ").By lower effective sampling rate is provided, transducer 110 to 113 can consumption rate with other transducers of actual samples speed operation power still less, and without any need for compression.Can be by Redundancy Design in the deployment of system, thus make the loss of transducer minimally to affect the performance of this system.For very eurypalynous signal, can carry out the reconstruct to sort signal by base station 102, core net 103, other networks 104 or their arbitrary combination.

In current embodiment, each can comprise the combination of direct sequence despread unit, fast Fourier transform (" FFT ") unit, other unit or unit transducer 110 to 113.Base station 102 can for example be selected for the direct sequence code of despread unit or the instruction of sub-chip timing (sub-chip timing), select the combination for instruction, other instructions or the instruction of the frequency separation of FFT unit or the number of frequency range to transducer 110 to 113 transmissions.These instructions can transmit with a for example ms interval, and each instruction simultaneously, after receiving, is carried out in 1/10th milliseconds by transducer 110 to 113.In addition, subscriber equipment 106 can send and receive information with the form of time slot, grouping, frame or other similar structures, and these structures can have for example duration of one to five millisecond.Time slot, grouping, frame and other similar structures can comprise the set of the time-domain sampling of catching continuously, maybe can describe continuous real number value or the set of complex values.

In Figure 100, system 100 can be included in the transmission of the overhead information between subscriber equipment 106, base station 102, core net 103, other networks 104, transducer 110 to 113 or their arbitrary combination.Described overhead information can comprise the combination of for example guiding and synchronizing information, wireless wide area network information, WLAN information, other information or information.Persons of ordinary skill in the art will recognize that by 106 pairs of restricting user equipments and monitor bottom-layer network, additional networks or the demand of the two, can reduce its power consumption.

In Fig. 1, if the enough proximity transducers 110 to 113 of subscriber equipment 106, subscriber equipment 106 can send uplink signal with low transmission power level.Transducer 110 to 113 can carry out compression sampling to received uplink signal (" g "), to generate sensing signal (" y ").Transducer 110 to 113 can use respectively communication link 114 to 117, and sensing signal (" y ") is sent to base station 102.For example the 1st layer function can be carried out in base station 102, such as demodulation code; Layer 2 feature, such as packet numbering and ARQ; And higher functions, such as registration, channel allocation and switching.Base station 102 can have sizable computing capability and come in real time, approaches in real time or the two side by side carries out the function that calculating strength is large.

In current embodiment, base station 102 for example can use such as the knowledge of the described communication channel of the antenna related matrix of subscriber equipment 106, with the combination of number, other factors or the factor of the approaching transducer 110 to 113 of subscriber equipment 106, application link is adjusted strategy.This adjustment strategy of knowledge of described communication channel can require the processing at cycle interval, for example, and a ms interval.This strategy can allow for example with optimum spatial multiplexing gain and diversity gain, to operate.Equally, can communicate by letter each other in a plurality of base stations 102, carrying out for example dirty paper code (dirty paper coding) (" DPC "), described dirty paper code is for effectively send the technology of down link signal by having stood certain communication channel for the known interference in base station 102.In order to support these technology, other base stations that receive extra uplink signal from subscriber equipment 106 can provide to the base station 102 being associated with subscriber equipment 106 this uplink signal (" f ").Persons of ordinary skill in the art will recognize that a plurality of subscriber equipmenies 106 can communicate by letter with base station 102.

Fig. 2 is exemplified with according to another embodiment of sensor-based wireless communication system 200 of the use compression sampling of the various aspects of setting forth herein.In the present embodiment, can the provide support robust high bandwidth of high user density of system 200 is implemented radio communication.System 200 comprises subscriber equipment 206, transducer 210 to 213, base station 202, core net 203 and other networks 204.In the present embodiment, transducer 210 to 213 can be carried out a part for the 1st layer function, such as receiving uplink signal and carry out compression sampling.In addition, communication link 214 to 217 can be used respectively in base station 202, to transducer 210 to 213 transmission instructions.These instructions can be for example to use specific multiple access access code (such as direct sequence code or OFDM code) to compress.In addition, base station 202 can be to transducer 210 to 213 transmission instructions, to carry out sampling with for example sampling rate in the sampling rate twice at particular phases place.

The function that calculating strength is large can be carried out in base station 202, for example to detect the existence of subscriber equipment 206 the sensing signal receiving from transducer 210 to 213 (" y ").Once the existence of subscriber equipment 206 be detected, base station 202 can sensors configured 210 to 213, to improve the reception to the uplink signal from subscriber equipment 206 (" f ").This improvement can be associated with the combination of timing, frequency, coding, other characteristics or characteristic.In addition, subscriber equipment 206 can use for example fountain codes to send uplink signal (" f ").For high bandwidth low power communication, subscriber equipment 206 can send and comprise for example uplink signal of real-time voice with fountain codes.Grouping transmission rate for this uplink signal can be for example in 200Hz arrives the scope of 1kHz.Transducer 210 to 213 can have the limited decision-making capability of substantially being controlled by base station 202.

In Fig. 2, transducer 210 to 213 can intensively be disposed, for example: approximately transducer of each hundred meters of spacing distance 210 to 213, the approximately combination of transducer of every ten meters of spacing distances 210 to 213, other configurations or configuration.Transducer 210 to 213 can comprise downlink transmitter, or with the downlink transmitter place that coexists, described downlink transmitter is for supporting to send 202 down link signals that receive from base station.In addition, base station 202 can provide down link signal to long-range downlink transmitter with communication link, and described long-range downlink transmitter is such as being the legacy cellular steel tower with antenna sector, being arranged on the combination of cellular transmitter on building or auroral poles (light pole), low-power unit, other unit or unit in office.The deployment of this long-range downlink transmitter can be the combination of supporting for example building deployment, street lamp deployment, other deployment or deployment.In addition, will be appreciated that a plurality of subscriber equipmenies 206 can communicate by letter with base station 202.

Fig. 3 is exemplified with according to another embodiment of sensor-based wireless communication system 300 of the use compression sampling of the various aspects of setting forth herein.In the present embodiment, system 300 represents multiple access system.System 300 comprises subscriber equipment 306, transducer 301, base station 302 and downlink transmitter 308.In Fig. 3, transducer 310 can comprise for uplink signal being carried out to the receiving element of down-conversion.Those of ordinary skills are by the Design and implementation requirement of understanding for this receiving element.

In Fig. 3, downlink transmitter 308 can be coupled in base station 302, wherein, downlink transmitter 308 can with for example honeycomb steel tower place that coexists.Base station 302 can comprise for example gatherer, for collecting sensing signal from transducer 310; Detector, for detection of the information signal comprising in this sensing signal; Controller, for controlling transducer 310; The combination of other unit or unit.Base station 302 and downlink transmitter 308 place that can coexist.In addition, downlink transmitter 308 can be coupled to base station 302 with communication link 309, and this communication link 309 can be supported for example optical fiber cable connection, microwave link, coaxial cable connection, other connections or their arbitrary combination.The configuration of system 300 can be similar to conventional cellular system, such as the combination of gsm system, UMTS system, LTE system, cdma system, other system or system.The layout of combination of subscriber equipment, base station, downlink transmitter, other unit or unit that persons of ordinary skill in the art will recognize that these system demonstrations.

In current embodiment, subscriber equipment 308 can use procotol to communicate by letter with base station 302, to carry out such as following functions: the combination of random access, paging, initiation, resource distribution, channel allocation, the expense signaling that comprises timing, navigation system identification, the channel that is allowed for access, switching message, training or pilot tone signaling, other functions or function.In addition, subscriber equipment 308 and base station 302 can transfers voice information, the combination of packetized digital information, circuit switched data information, other information or information.

Fig. 4 is exemplified with according to an embodiment of the compression sampling system of the various aspects of setting forth herein.System 400 comprises compression sampling device 431 and detector 452.In Fig. 4, compression sampling device 431 can use the sense waveform (" φ of sensing matrix (" Φ ") _j") input signal (" f ") is carried out to compression sampling, to generate sensing signal (" y "), wherein, φ _jrefer to j the waveform of sensing matrix (" Φ ").Input signal (" f ") can have length N, and sensing matrix (" Φ ") can have the sense waveform (" φ that M length is N _j"), and sensing signal (" y ") can have length M, and wherein, M can be less than N.If (" f ") is enough sparse for input signal, can recover information signal (" x ").Those of ordinary skills are by the characteristic of identification sparse signal.In a definition, it is sparse that the signal that the length with S nonzero value is N is known as S, and comprise that N subtracts the individual null value of S (" N-S ").

In current embodiment, compression sampling device 431 can use for example equation (1) to carry out compression sampling to input signal (" f ").

make

J &Subset; {1, . . ., N} - - - (1)

Wherein, bracket <> represents inner product, relevance function or other similar functions.

In addition, detector 452 can solve sensing signal (" y "), to use for example equation (2) to find information signal (" x ").

meet

&ForAll; k &Element; J - - - (2)

Wherein l ₁norm be the absolute value sum of the element of its independent variable, and bracket <> represents inner product, relevance function or other similar functions.Can be applied to l ₁a minimized method is simplex accelerated process (simplex method) for example.Solve sensing signal (" y ") and use for example l to find the additive method of information signal (" x ") to comprise ₀the combination of norm algorithm, additive method or method.

Incoherent sampling is the sense waveform (" φ depending on the abundant incoherent sensing matrix of rarefaction representation matrix Ψ (" Φ ") _j") compression sampling form, rarefaction representation matrix Ψ is used for making input signal (" f ") sparse.In order to minimize the sense waveform (" φ of sensing matrix (" Φ ") _j") required number, rarefaction representation the waveform (" ψ of rarefaction representation matrix (" Ψ ") _j") and the sense waveform (" φ of sensing matrix (" Φ ") _j") between coherence (" μ ") should represent that these waveforms are fully incoherent, corresponding to lower coherence (" μ "), wherein, ψ _jrefer to j the waveform of rarefaction representation matrix (" Ψ ").Coherence (" μ ") can be represented by for example equation (3).

μ (Φ, Ψ) = \sqrt{N} \max_{1 \leq k, j \leq N} {| | < φ_{k}, ψ_{j} > | |}_{l_{1}} - - - (3)

Wherein l ₁norm be the absolute value sum of the element of its independent variable, and bracket <> represents inner product, relevance function or other similar functions.

Fig. 5 is that the method can be for for example designing compression sampling system according to the flow chart of the embodiment of the compressive sampling method 500 of the various aspects of setting forth herein.In Fig. 5, method 500 starts from piece 570, in piece 570, method 500 can be to input signal (" f ") modeling, and find rarefaction representation matrix (" Ψ "), in rarefaction representation matrix (" Ψ "), input signal (" f ") is that S is sparse.At piece 571, method 500 can be selected sensing matrix (" Φ "), and (" Ψ ") is fully irrelevant for this sensing matrix and rarefaction representation matrix.At piece 572, method 500 can be randomly, certainty ground or the two are side by side selected M sense waveform (" φ of sensing matrix (" Φ ") _j"), wherein M can be more than or equal to S.At piece 573, method 500 can be used selected M sense waveform (" φ _j") input signal (" f ") is sampled, to produce sensing signal (" y ").At piece 574, method 500 can pass to detector by rarefaction representation matrix (" Ψ "), sensing matrix (" Φ ") and sensing signal (" y "), to recover information signal (" x ").

Fig. 6 is exemplified with according to another embodiment of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.In the present embodiment, the system 600 robust high bandwidth real-time radio communication of high user density that can provide support.System 600 comprises subscriber equipment 606, transducer 610 and base station 602.In Fig. 6, even if transducer 610 for example cannot operate, system 600 also can allow subscriber equipment 606 to communicate by letter with for example bottom cellular system.System 600 can allow transducer 610 to be distributed widely, to meet for example office buildings environment.System 600 can allow base station 602 not to be limited to for example combination of calculated capacity, memory, other resources or resource.System 600 can allow to provide down link signal by for example conventional honeycomb steel tower.System 600 can allow subscriber equipment 606 to carry out minimizing power dissipation by its transmission power level being limited to for example about ten to 100 microwatts.System 600 can allow transducer 610 to use communication link 614 to be coupled to base station 602, and wherein, communication link 614 can be supported for example optical fiber cable connection, coaxial cable connection, other connections or their arbitrary combination.System 600 can allow transducer 610 to be operated by power source, and described power source is such as being the combination of battery, photovoltaic power source, alternating current (" AC ") electric power source, other power sources or power source.

In Fig. 6, system 600 can allow transducer 610 quite a lot ofly more cheap than base station 602.In addition, system 600 can allow the ongoing operation within one period of time period being expanded of the transducer 610 use powers of battery, such as about one to two year.In order to realize this, persons of ordinary skill in the art will recognize that such as the specific function of input, demodulation code must be carried out by for example base station 602.

In Fig. 6, transducer 610 can have receiving element, such as the antenna with the coupling of RF down-conversion chain, for receiving uplink signal (" f ").In the disclosure, uplink signal (" f ") can also be called uplink signal (" g ").Uplink signal (" g ") comprises dissemination channel effect and the environmental effect relevant to uplink signal (" f ").For example, the channel gain of channel 620 (" a ") 621 can representation case as dissemination channel effect, simultaneously the interchannel noise (" v ") 622 of channel 620 can representation case as ambient noise effect.In addition, transducer 610 can support to base station 602 to send for example communication link of sensing signal (" y ").Transducer 610 can not have the computing capability that when sends uplink signal (" f ") for for example identifying subscriber equipment 606.Transducer 610 can be from base station 602 receives and instruction that for example combination of RF down-conversion, compression sampling, other functions or function is associated.

There is the method that much makes subscriber equipment access to wireless communication system.The cut-in method of one type is Aloha accidental access method for example, when Unidentified subscriber equipment is attempted access network, carries out the method.To subscriber equipment, providing the license of use system and distributing after any up link and downlink channel, can occur and base station between two-way communication.

Fig. 7 is exemplified with according to an embodiment of the cut-in method 700 in sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.Various exemplary architecture have been shown in the bottom of Fig. 7, so that the understanding to method 700.In addition, Fig. 7 is exemplified with base station 702 twice, but should be interpreted as same base station 702.Correspondingly, method 700 comprises the communication between base station 702, subscriber equipment 706, transducer 710 or their arbitrary combination.Subscriber equipment 706 can have the event 770 that for example powers on, and starts to observe 702 Overhead Messages 771 that send from base station.Persons of ordinary skill in the art will recognize that base station can be used the combination of for example broadcast communication, point-to-multipoint delivery, point-to-point communication or other communication meanss or communication means, with communications of user equipment.Overhead Message 771 can comprise system parameters, this system parameters comprise message frame for example length, with sense waveform (" φ _j") the value of the M that is associated of number and the degree of rarefication S of the uplink signal (" f ") that is sent out.

In Fig. 7, base station 702 can send for example Overhead Message, with configure user equipment 706, uses degree of rarefication S ₁with rarefaction representation matrix (" Ψ "), as shown in 772.Then, subscriber equipment 706 can send and use degree of rarefication S ₁existence signal, as shown in 780.Existence signal can comprise subscriber equipment 706 to base station 702 send can compressed sampling any signal.In another embodiment, when subscriber equipment 706 determines that it is approaching base station 702, subscriber equipment 706 can send and use S ₁existence signal, as shown in 780.In this case, subscriber equipment 706 can, via base station 702, another base station or Overhead Message 771 that the two sends, determine that it is approaching base station 702.

In Fig. 7, base station 702 can also send the Overhead Message that for example comprises system information, described system message such as be framing, regularly, the combination of system identification, other system information or information, as shown in 773.In addition, for example M of sensing matrix (" Φ ") can be used by indication sensor 710 in base station 702 ₁individual sense waveform (" φ _j"), as shown in 791.Then, transducer 710 can be processed received uplink signal (" f ") continuously, and to base station 702 M that send to use sensing matrixes (" Φ ") ₁individual sense waveform (" φ _j") sensing signal (" y "), as shown in 790.

In Fig. 7, base station 702 can send for example Overhead Message, with configure user equipment 706, uses degree of rarefication S ₂with rarefaction representation matrix (" Ψ "), as shown in 774.Then, subscriber equipment 706 can send and for example use degree of rarefication S ₂existence signal, as shown in 781.In addition, for example M of sensing matrix (" Φ ") can be used by indication sensor 710 in base station 702 ₂individual sense waveform (" φ _j"), as shown in 792.Then, transducer 710 can be processed received uplink signal (" f ") continuously, and to base station 702 M that send to use sensing matrixes (" Φ ") ₂individual sense waveform (" φ _j") sensing signal (" y "), as shown in 793.Subscriber equipment 706 can continue to send use S ₂existence signal, as shown in 781, until for example base station 702 detects and uses S ₂existence signal, as shown in 794.At this some place, base station 702 can send identification message to subscriber equipment 706, and described identification message comprises the request of a part for the Electronic Serial Number (" ESN ") that for example sends subscriber equipment 706, and uses degree of rarefication S ₃with the request of rarefaction representation matrix (" Ψ "), as shown in 775.In addition, base station 702 can send instruction to transducer 710, to use for example M ₃new value and new sensing matrix (" Φ "), as shown in 795.Then, transducer 710 can be processed received uplink signal (" f ") continuously, and to base station 702 M that send to use sensing matrixes (" Φ ") ₃individual sense waveform (" φ _j") sensing signal (" y "), as shown in 796.

In Fig. 7, subscriber equipment 706 can 702 transmissions comprise use S to base station ₃the uplink message of a part of ESN of subscriber equipment, as shown in 782.Once base station 702 has received this uplink message, base station 702 just can send downlinlc message to subscriber equipment 706, to ask subscriber equipment 706 to send the complete ESN of subscriber equipment for example and for the request of resource, as shown in 776.Then, subscriber equipment 706 can send uplink message, comprises and uses S ₃subscriber equipment complete ESN and for the request of resource, as shown in 783.After base station 702 receives these uplink messages, the complete ESN that base station 702 can verifying user equipment 706, qualified in system to determine it, and distribute any resource to it, as shown in 798.Then, base station 702 can send downlinlc message to subscriber equipment 706, with to its Resources allocation, as shown in 777.

Transducer 710 can receive constantly to be had with centre frequency (" f _c") centered by the uplink signal (" f ") of frequency bandwidth (" B ").Transducer 710 can use receiving element to carry out down-conversion to uplink signal (" f "), then carries out compression sampling.By for example received uplink signal (" f ") being sampled, then calculate the product of sensing matrix (" Φ ") and this sampling, to generate sensing signal (" y "), carry out compression sampling.Can locate for example to carry out sampling at the frequency bandwidth corresponding to Nyquist speed (" B "), thereby with preserve received uplink signal (" f ") according to Shannon theorem consistent.Can be for example periodically, aperiodicity ground or the two side by side sample to received uplink signal (" f ").

Described sampling process can obtain N sampling, and the product of calculating sensing matrix (" Φ ") and N sampling can obtain M value of sensing signal (" y ").Sensing matrix (" Φ ") can have dimension N * M.Can, on communication link, M the value of these resulting sensing signals (" y ") be sent to base station 702.Compression sampling can reduce to M sampling from N sampling of conventional scheme by the number of samples that is sent to base station 702, and wherein M can be less than N.If transducer 710 does not have enough timings, can adopt higher sampling rate to carry out sampling, thereby obtain for example 2N sampling.For this situation, transducer 710 can calculate the product of 2N sampling of sensing matrix (" Φ ") and uplink signal (" f "), thereby obtains 2M sampling of sensing signal (" y ").Therefore, described compression sampling device can reduce to 2M sampling from 2N sampling of conventional scheme by the number of samples that is sent to base station 702, and wherein M can be less than N.For this situation, sensing matrix (" Φ ") can have dimension 2N * 2M.

Described compression sampling device can be by independent the sense waveform (" φ selecting with for example sensing matrix (" Φ ") by the uplink signal receiving of having sampled (" f ") _j") relevant, calculate sensing signal (" y ").Can be in the situation that without the knowledge of information signal (" x "), select the sense waveform (" φ of sensing matrix (" Φ ") _j").Yet the selection of M can be according to for example estimation to the degree of rarefication S of received uplink signal (" f ").Therefore, selected M sense waveform (" φ of sensing matrix (" Φ ") _j") can be independent of rarefaction representation matrix (" Ψ "), but M can depend on the estimation to the attribute of received uplink signal (" f ").In addition, the degree of rarefication S of the uplink signal receiving (" f ") can be by for example send base station 702 controls of downlinlc message to subscriber equipment 706, this downlinlc message is used for identifying subscriber equipment 706, and configure user equipment 706 is used degree of rarefication S ₃with new rarefaction representation matrix (" Ψ ") 775.

The successful detection of 702 pairs of base stations information signal (" x ") can require M to be more than or equal to degree of rarefication S.Can estimate degree of rarefication S and adjust afterwards by for example base station 702, overcome the knowledge that lacks degree of rarefication S.For example, base station 702 can be initialized as M for example value of N, and N can be corresponding to not compressing benefit.When the activity grade of the frequency bandwidth B that receives in transducer 710 places is estimated in base station 702, the value of M can for example be adjusted in base station 702.By like this, base station 702 can be by adjusting M sense waveform (" φ _j") number, affect the power consumption of transducer 710; Therefore, adjusted the bandwidth of the sensing signal (" y ") that is sent to base station 702 on communication link.

In addition, base station 702 can send instruction to transducer 710, for example periodically to increase the value of M, thereby allows base station 702 thoroughly to assess the degree of rarefication S in frequency bandwidth B.In addition, base station 702 can send and select sense waveform (" φ to transducer 710 _j") the relevant instruction of method, such as the combination of random selection, other systems of selection or the system of selection of selecting, carrying out according to scheduling.In some instances, transducer 710 may need it sense waveform (" φ _j") selection send base station 702 to.

Subscriber equipment 706 can send existence signal, 702 to notify its existence to base station.Each existence signal can be by for example selecting and combine rarefaction representation the waveform (" ψ of rarefaction representation matrix (" Ψ ") _j") information signal that generates.Rarefaction representation waveform (" ψ to rarefaction representation matrix (" Ψ ") _j") the selection Overhead Message that can be sent by base station for example 702 configure.For example, base station 702 can broadcast overhead messages, and this Overhead Message has been specified rarefaction representation the waveform (" ψ of rarefaction representation matrix (" Ψ ") _j") subset.

Base station 702 can also broadcast downlink Overhead Message, is used for using specific rarefaction representation the waveform (" ψ of rarefaction representation matrix (" Ψ ") for Unidentified subscriber equipment 706 _j"), it can be called pilot signal (" ψ ₀").Transducer 710 is receiving uplink signal (" f ") continuously, and uplink signal (" f ") is carried out to compression sampling to generate sensing signal (" y "), and 702 sends sensing signals (" y ") to base station.Then, pilot signal (" ψ can be detected in base station 702 in sensing signal (" y ") ₀").Once the pilot signal of detecting (" ψ ₀"), base station 702 can estimating user equipment 706 and transducer 710 between channel gain and can indicate pilot signal transmitted (" ψ ₀") any subscriber equipment 706 send a for example part of its ESN.If clashed between the up link from different user devices 706 sends, can come the subsequent uplink that separated different user devices 706 carries out to send trial with the contention resolution such as Aloha algorithm.

Transducer 710 can also in the situation that the communication between base station 702 and subscriber equipment 706 of not considering operate.For example M rarefaction representation waveform (" ψ of rarefaction representation matrix (" Ψ ") can be used by indication sensor 710 in base station 702 _j").In addition, base station 702 can be based on expecting that for example the uplink signal of subscriber equipment 706 (" f ") is measured the value that changes M.For example, if the degree of rarefication S of base station 702 expection uplink signals (" f ") changes, base station 702 can indication sensor 710 changes the value of M.For specific certainty sensing matrix (" Φ "), when M equals the value of N, the sensing matrix in transducer 710 (" Φ ") can become discrete Fourier transform (" DFT ") effectively.

Fig. 8 is exemplified with according to another embodiment of sensor-based wireless communication system 800 of the use compression sampling of the various aspects of setting forth herein.In the present embodiment, the system 800 robust high bandwidth real-time radio communication of high user density that can provide support.In Fig. 8, system 800 comprises subscriber equipment 806, transducer 810 and base station 802.Base station 802 can receive sensing signals (" y ") from transducer 810, as the input of the detector 851 of base station 802, and with the estimation of information generated signal (" x "), also referred to as then, base station 802 can quantize this estimation, to generate for example quantitative estimation of information signal (" x "), also referred to as can use for example simplex accelerating algorithm, l ₁norm algorithm, l ₀the estimation of information signal (" x ") is determined in the combination of norm algorithm, other algorithms or algorithm.In the present embodiment, all elements of the estimation of information signal (" x ") may have nonzero value.Therefore can carry out the hard decision to the estimation of information signal (" x "), to determine the information signal (" x ") that is subtracted the individual null value formation of S (" N-S ") by for example S nonzero value and N.

Fig. 9 is exemplified with according to an embodiment of the quantization method 900 of the detector in sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.Fig. 9 relates to the step in the quantizer 953 of step in base station 902 and base station 902.Method 900 can start at 902 transducer 910 places that send sensing signal (" y ") to base station.At piece 952, method 900 can solve sensing signal (" y "), to determine the estimation of information signal (" x "), also referred to as at piece 970, method 900 can be by the element sequence in the estimation of information signal (" x ") for for example from peak to peak.

In Fig. 9, information signal (" x ") is applied to quantizer 953.At piece 971 places, method 900 can be used the estimation of for example sensing signal (" y "), information signal (" x ") or the two to determine degree of rarefication S.In addition, base station 902 can, by sending downlinlc message to subscriber equipment, fix the S value for subscriber equipment.The desired value of S also can for example,, by sending different S values to transducer and determining the degree of rarefication S of uplink signal (" f ") of during section sometime (to two second), periodically be scanned in base station 902.Because subscriber equipment can carry out repeatedly access attempts, so the bad luck estimation of S is had an opportunity to identify in base station 902, and indication sensor is adjusted its M value.In the situation that relevant to scanning for S enough low duty ratios (duty cycle) can retain the power consumption advantage of using sensor-based cordless communication network.In this manner, the compression sampling activity of transducer 910 can be followed the tracks of the degree of rarefication of the signal that may affect it adaptively.Therefore,, even when carrying out compression sampling continuously, transducer 910 also can minimize its power consumption.

At piece 972, method 900 can be used the degree of rarefication S definite at step 971 place, to keep the index of maximum S element of the estimation of information signal (" x ").At piece 973, method 900 can be used at S definite index of piece 972, with maximum S element of the estimation of information signal (" x "), is set to the first value 974.At piece 975, residue N-S the element of method 900 in subsequently can the estimation of information signal (" x ") is set to the second value 976.The output of quantizer 953 can be the quantitative estimation of information signal (" x "), is called the first value 974 can be logical one for example.In addition, the second value 976 can be logical zero for example.

Figure 10 is exemplified with according to the chart 1000 of the example of the type of the rarefaction representation matrix using in sensor-based wireless communication system 100,200,300,400,600 and 800 of the use compression sampling of the various aspects of setting forth herein.In one embodiment, use sensor-based wireless communication system of compression sampling can use random matrix as rarefaction representation matrix (" Ψ ") and sensing matrix (" Φ ").Random matrix consists of for example independent same distribution (" iid ") Gauss's value.

In another embodiment, use sensor-based wireless communication system of compression sampling can use certainty matrix as rarefaction representation matrix (" Ψ ") and sensing matrix (" Φ ").Certainty matrix consists of the unit matrix of for example rarefaction representation matrix (" Ψ ") and the cosine matrix of sensing matrix (" Φ ").Persons of ordinary skill in the art will recognize that a lot of dissimilar matrixes and combination thereof can be for being used sensor-based wireless communication system of compression sampling.

Figure 11 is exemplified with can be according to an embodiment of the subscriber equipment 1100 using in sensor-based wireless communication system 100,200,300,400,600 and 800 of the use compression sampling of the various aspects of setting forth herein.In Figure 11, subscriber equipment 1100 can comprise for modulates uplink message to form the modulator 1140 of information signal (" x ").Maker 1141 can receiving information signal (" x "), and can be to information signal (" x ") application rarefaction representation matrix (" Ψ ") 1143, with generating uplink signal (" f "), this signal is used for example antenna 1364 to send by link transmitter 1142.Subscriber equipment 1100 can also comprise that the down link signal for antenna 1164 is received carries out the descending link receiver 1148 of down-conversion.The down link signal receiving can be processed by demodulator 1149 subsequently, to generate downlinlc message.

In the present embodiment, subscriber equipment 1100 can comprise oscillator 1162, for to subscriber equipment 1100 timing and maintain timing; Power supply 1163, such as battery 1361, for powering to subscriber equipment 1100; Input-output apparatus 1367, such as keypad and display; Memory 1360, is coupled to for controlling the controller 1147 of the operation of subscriber equipment 1100; The combination of other unit or unit.The typical unit that those of ordinary skills find understanding in subscriber equipment.

Figure 12 shows can be according to an embodiment of the transducer 1200 using in sensor-based wireless communication system 100,200,300,400,600 and 800 of the use compression sampling of the various aspects of setting forth herein.In Figure 12, transducer 1200 can comprise the receiving element 1230 that carries out down-conversion for the uplink signal to for example antenna 1264 receives (" f ").Compression sampling device 1231 can be applied to uplink signal (" f ") by sensing matrix (" Φ ") 1233, and to generate sensing signal (" y "), this sensing signal (" y ") can send with sensor emission machine 1232.

In the present embodiment, transducer 1200 can comprise: oscillator 1262, for to transducer 1200 timing and maintain timing; Power supply 1263, such as battery 1261, for powering to subscriber equipment 1100; Memory 1260, is coupled to for controlling controller or the state machine 1237 of the operation of transducer 1200; The combination of other unit or unit.Can utilize hardware, software, firmware or their arbitrary combination to realize controller 1237.In addition, controller 1237 can comprise microprocessor, digital signal processor, memory, state machine or their arbitrary combination.

Figure 13 is exemplified with can be according to an embodiment of the base station 1300 of using in sensor-based wireless communication system 100,200,300,400,600 and 800 of the use compression sampling of the various aspects of setting forth herein.In Figure 13, at uplink direction, base station 1300 can comprise for collecting the gatherer 1350 of sensing signal (" y ").Detector 1351 can receive collected sensing signal (" y "), and can use sensing matrix (" Φ ") 1233 and rarefaction representation matrix (" Ψ ") 1143, in collected sensing signal (" y "), estimate and detection information signal (" x ").Controller 1357 can be assessed detected information signal to determine uplink message.At downlink side, make progress, base station 1300 can comprise: modulator 1359, for modulating downlinlc message; And downlink transmitter interface 1358, for sending modulated down link signal.

In the present embodiment, base station 1300 can comprise: oscillator 1362, for to base station 1300 timing and maintain timing; Power supply 1363, for 1300 power supplies to base station; Memory 1360, is coupled to for controlling the controller 1337 of the operation of base station 1300; Sensor controller 1355, for controlling transducer; Downlink transmitter controller, for controlling downlink transmitter; The combination of other unit or unit.

In one embodiment, sensor-based wireless communication system 100,200,300,400,600 and 800 can be processed uplink signal (" f ") with a plurality of transducers 110 to 113,210 to 213,310,610,710,810,1200 and 1310, to allow by using the combination of antenna array signals treatment technology, MIMO signal processing technology, beam forming technique, other technologies or technology, 102,202,302,602,702,802 and 1302 place's joint-detection existence signals in base station.A plurality of transducers 110 to 113,210 to 213,310,610,710,810,1200 and 1310 use can allow in the M value at each transducer 110 to 113,210 to 213,310,610,710,810,1200 and 1310 places lower.Therefore, can, by place a plurality of transducers 110 to 113,210 to 213,310,610,710,810,1200 and 1310 with for example more intensive deployment, reduce the power consumption of each transducer 110 to 113,210 to 213,310,610,710,810,1200 and 1310.

In another embodiment, sensor-based wireless communication system 100,200,300,400,600 and 800 can be disposed transducer 110 to 113,210 to 213,310,610,710,810,1200 and 1310, typically to allow two transducers 110 to 113,210 to 213,310,610,710,810,1200 and 1310 to receive the uplink signal (" f ") being sent by subscriber equipment 706.This deployment can be in indoor environment, and in this indoor environment, transducer 110 to 113,210 to 213,310,610,710,810,1200 and 1310 can be deployed as 30 meters, interval distance, and the path loss index having is between two and three.Each in transducer 110 to 113,210 to 213,310,610,710,810,1200 and 1310 can be deployed as the region that covering is larger; Yet path loss index may be less.For successful detection, the probability that detects single existence signal can be higher than 10.

In another embodiment, sensor-based wireless communication system 100,200,300,400,600 and 800 can be disposed transducer 110 to 113,210 to 213,310,610,710,810,1200 and 1310 in Microcell, to support for example combination of vehicle-carrying communication, other communications or communication.In addition, transducer 110 to 113,210 to 213,310,610,710,810,1200 and 1310 can be deployed in Microcell supporting that for example pedestrian communicates by letter, the combination of indoor communications, office communication, other communications or communication.

In system 100,200,300,400,600 and 800, channel 620 and 820 can be static, there is channel gain (" a ") 621 and 821, and interchannel noise (" v ") 622 and 821 can be additive white Gaussian noise (" AWGN ").Interchannel noise (" v ") 622 and 821 can comprise additive signal, and this additive signal can be so that the information of interest distortion that receiver is watched.The source of interchannel noise (" v ") can be the combination of thermal noise, cochannel interference, adjacent-channel interference, other noise sources or the noise source at for example reception antenna place.In addition, transducer 110 to 113,210 to 213,310,610,710,810,1200 and 1310; Subscriber equipment 106,206,306,606,706,806 and 1100; Base station 102,202,302,602,702,802 and 1302; Or their arbitrary combination can be fully synchronous in the combination of timing, frequency, phase place, their condition of other conditioned disjunctions.In addition, can only there is a transducer 110 to 113,210 to 213,310,610,710,810,1200 and 1310; A subscriber equipment 106,206,306,606,706,806 and 1100; A base station 102,202,302,602,702,802 and 1302; Or their arbitrary combination.

Compression sampling scheme can be used rarefaction representation matrix (" Ψ ") and sensing matrix (" Φ "), they be for example at random to, certainty to or above-mentioned arbitrary combination.For these matrixes, can be to base station 102,202,302,602,702,802 and 1302, transducer 110 to 113,210 to 213,310,610,710,810,1200 and 1310, subscriber equipment 106,206,306,606,706,806 and 1100 or their arbitrary combination, rarefaction representation matrix (" Ψ "), sensing matrix (" Φ ") or the two, information (such as for generating rarefaction representation matrix (" Ψ "), sensing matrix (" Φ ") or the seed of the two) are for example provided.Base station 102,202,302,602,702,802 and 1302 can be known and using which rarefaction representation matrix (" Ψ ") and sensing matrix (" Φ ").Base station 102,202,302,602,702,802 and 1302 can indication sensor 110 to 113,210 to 213,310,610,710,810,1200 and 1310 M sense waveform (" φ that use sensing matrixes (" Φ ") _j") specific collection.In addition, base station 102,202,302,602,702,802 and 1302 can to subscriber equipment 106,206,306,606,706,806 and 1100 and transducer 110 to 113,210 to 213,310,610,710,810,1200 and 1310 indication uplink signals by for example N interval or chip, formed.

Aforementioned random matrix, certainty matrix or the two can only generate once, if or again generate, do not change.In addition, can a period of time (for example several seconds) afterwards, regenerate these matrixes.Equally, can, when each these matrixes of use, regenerate these matrixes.In arbitrary situation, the detector that comprises solver in base station 102,202,302,602,702,802 and 1302 can be known the rarefaction representation matrix (" Ψ ") of subscriber equipment 706 uses and the sensing matrix (" Φ ") that sampler uses.This area this spectral technology personnel will recognize, this does not mean that base station must provide described matrix.On the other hand, for example, subscriber equipment 106,206,306,606,706,806 and 1100 and base station 102,202,302,602,702,802 and 1302 can change rarefaction representation matrix (" Ψ ") according to the pseudo noise of for example system time (" pn ") function.Similarly, for example, transducer 110 to 113,210 to 213,310,610,710,810,1200 and 1310 and base station 102,202,302,602,702,802 and 1302 can change sensing matrix (" Φ ") according to the pseudo noise of for example system time (" pn ") function.

Figure 14 is exemplified with according to detecting the simulation result of an embodiment of subscriber equipment herein in sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth, wherein, use N=10, M=5, S=1 or 2 and random matrix carry out the performance of measuring system 800.The integral body of graphical illustration is known as 1400.The logarithm amplitude of signal to noise ratio (" SNR ") illustrates on abscissa 1401, and has been plotted the scope from 0 decibel (" dB ") to 25dB.Detection probability (" Pr (detection) ") is shown in ordinate 1402, and has been plotted from zero (corresponding to zero probability) to one the scope of (corresponding to probability absolutely).The simulation result of Figure 140 3,1404 and 1405 expression systems 800, wherein, N is that 10, M is that 5, S is 1 or 2, and fills rarefaction representation matrix (" Ψ ") and sensing matrix (" Φ ") by random iid Gauss value.Figure 140 3 shows the probability that a nonzero term detected in the quantitative estimation of information signal (" x "), and wherein S is 1.Figure 140 4 shows the probability that a nonzero term detected in the quantitative estimation of information signal (" x "), and wherein S is 2.Figure 140 5 shows the probability that two nonzero terms detected in the quantitative estimation of information signal (" x "), and wherein S is 2.

Figure 15 is exemplified with according to the simulation result of the performance of an embodiment of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein, wherein, use N=20, M=10, S=1 or 2 and random matrix carry out the performance of measuring system 800.The integral body of graphical illustration is known as 1500.The logarithm amplitude of SNR ratio illustrates on abscissa 1501, and has been plotted the scope from 0dB to 25dB.Detection probability (" Pr (detection) ") is shown in ordinate 1502, and has been plotted from zero (corresponding to zero probability) to one the scope of (corresponding to probability absolutely).The simulation result of Figure 150 3,1504,1505,1506 and 1507 expression systems 800, wherein, N is that 20, M is that 10, S is 1 or 2, and fills rarefaction representation matrix (" Ψ ") and sensing matrix (" Φ ") by random iid Gauss value.Figure 150 3 shows the probability that a nonzero term detected in the quantitative estimation of information signal (" x "), and wherein S is 1.Figure 150 4 shows the probability that two nonzero terms correctly detected in the quantitative estimation of information signal (" x "), and wherein S is 2.Figure 150 5 shows in the quantitative estimation of information signal (" x ") and the probability that there is no nonzero term correctly detected, and wherein S is 1.Figure 150 6 shows in the quantitative estimation of information signal (" x ") and the probability that there is no nonzero term correctly detected, and wherein S is 2.Figure 150 7 shows the probability that a nonzero term correctly detected in the quantitative estimation of information signal (" x "), and wherein S is 2.

Figure 16 is exemplified with according to the simulation result of the performance of an embodiment of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein, wherein, use N=10, M=3, S=1 and certainty or random matrix carry out the performance of measuring system 800.The integral body of graphical illustration is known as 1600.The logarithm amplitude of SNR ratio illustrates on abscissa 1601, and has been plotted the scope from 0dB to 25dB.Detection probability (" Pr (detection) ") is shown in ordinate 1602, and has been plotted from zero (corresponding to zero probability) to one the scope of (corresponding to probability absolutely).The simulation result of Figure 160 3,1604,1605,1606 and 1607 expression systems 800, wherein, N is that 20, M is that 10, S is 1 or 2, and deterministic value is used for to rarefaction representation matrix (" Ψ ") and sensing matrix (" Φ ").Figure 160 3 shows the probability that a nonzero term correctly detected in the quantitative estimation of information signal (" x "), and wherein S is 1.Figure 160 4 shows the probability that two nonzero terms correctly detected in the quantitative estimation of information signal (" x "), and wherein S is 2.Figure 160 5 shows in the quantitative estimation of information signal (" x ") and the probability that there is no nonzero term correctly detected, and wherein S is 1.Figure 160 6 shows in the quantitative estimation of information signal (" x ") and the probability that there is no nonzero term correctly detected, and wherein S is 2.Figure 160 7 shows the probability that a nonzero term correctly detected in the quantitative estimation of information signal (" x "), and wherein S is 2.

Figure 17 is exemplified with according to the simulation result of the performance of an embodiment of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein, wherein, use N=10, M=3, S=1 and random or certainty matrix carry out the performance of measuring system 800.The integral body of graphical illustration is known as 1700.The logarithm amplitude of SNR ratio illustrates on abscissa 1701, and has been plotted the scope from 0dB to 45dB.Detection probability (" Pr (detection) ") is shown in ordinate 1702, and has been plotted from zero (corresponding to zero probability) to one the scope of (corresponding to probability absolutely).The simulation result of Figure 170 3,1704,1705 and 1706 expression systems 800, wherein, N is that 10, M is that 3, S is 1.Figure 170 3 shows the probability that a nonzero term correctly detected in the quantitative estimation of information signal (" x "), wherein uses certainty matrix.Figure 170 4 shows the probability that a nonzero term correctly detected in the quantitative estimation of information signal (" x "), wherein uses iid gaussian random matrix.Figure 170 5 shows in the quantitative estimation of information signal (" x ") and the probability that there is no nonzero term correctly detected, wherein uses iid gaussian random matrix.Figure 170 6 shows in the quantitative estimation of information signal (" x ") and the probability that there is no nonzero term correctly detected, wherein uses certainty matrix.

Figure 18 shows according to the simulation result of the performance of an embodiment of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein, wherein, uses N=10, M=5, and S=2 and random matrix carry out the performance of measuring system 800.In addition, each, send information signal (" x ") before, change rarefaction representation matrix (" Ψ ") and sensing matrix (" Φ ").The integral body of graphical illustration is known as 1800.The logarithm amplitude of SNR ratio illustrates on abscissa 1801, and has been plotted the scope from 0dB to 50dB.Detection probability (" Pr (detection) ") is shown in ordinate 1802, and has been plotted from zero (corresponding to zero probability) to one the scope of (corresponding to probability absolutely).The simulation result of Figure 180 3,1804,1805 and 1806 expression systems 800, wherein, N is 10, M is 5, S is 2, and random iid Gauss matrix is used for rarefaction representation matrix (" Ψ ") and sensing matrix (" Φ "), and is regenerating described random matrix before each transmission.Figure 180 3 shows the probability that two nonzero terms detected in the quantitative estimation of information signal (" x ").Figure 180 4 shows the probability that two nonzero terms detected in the quantitative estimation of information signal (" x "), wherein any two sense waveform (" φ of sensing matrix (" Φ ") _j") substantially irrelevant.Figure 180 5 shows the probability that a nonzero term detected in the quantitative estimation of information signal (" x "), wherein any two sense waveform (" φ of sensing matrix (" Φ ") _j") substantially irrelevant.Especially, Figure 180 4 and 1805 also represents that any two in refusal sensing matrix (" Φ ") have the sense waveform (" φ that correlation amplitude is greater than 0.1 _j") effect.Figure 180 6 shows the probability that a nonzero term detected in the quantitative estimation of information signal (" x ").

Figure 19 is exemplified with according to the simulation result of the performance of an embodiment of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein, wherein, use N=10, M=3, S=1, the test of random matrix and various numbers carrys out the performance of measuring system 800.In addition, each, send information signal (" x ") before, change rarefaction representation matrix (" Ψ ") and sensing matrix (" Φ ").The integral body of graphical illustration is known as 1900.The logarithm amplitude of SNR ratio illustrates on abscissa 1901, and has been plotted the scope from 0dB to 50dB.Detection probability (" Pr (detection) ") is shown in ordinate 1902, and has been plotted from zero (corresponding to zero probability) to one the scope of (corresponding to probability absolutely).The simulation result of Figure 190 3,1904,1905,1906 and 1907 expression systems 800, wherein, N is 10, M is 3, S is 1, and random iid Gauss matrix is used for rarefaction representation matrix (" Ψ ") and sensing matrix (" Φ "), and is regenerating described random matrix before each transmission.Figure 190 3 shows the probability that a nonzero term detected in the quantitative estimation of information signal (" x "), wherein any two sense waveform (" φ of sensing matrix (" Φ ") _j") substantially irrelevant, and carried out 200 tests.Especially, Figure 190 3 also represents that any two in refusal sensing matrix (" Φ ") have the sense waveform (" φ that correlation amplitude is greater than 0.1 _j") effect.Figure 190 4 shows the probability that a nonzero term correctly detected in the quantitative estimation of information signal (" x "), has wherein carried out 200 tests.Figure 190 5 shows the probability that a nonzero term correctly detected in the quantitative estimation of information signal (" x "), wherein carries out 4,000 tests.Figure 190 6 shows the probability that a nonzero term correctly detected in the quantitative estimation of information signal (" x "), wherein carries out 1,000 tests.Figure 190 7 shows the probability that a nonzero term correctly detected in the quantitative estimation of information signal (" x "), wherein carries out 2,000 tests.

Figure 20 is according to the example of the certainty matrix using in an embodiment of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.The example of certainty matrix is referred to as does 2000.Matrix 2001 and 2002 represents the certainty matrix can using in system 100,200,300,400,600 and 800, wherein N be 10 and M be 5.Matrix 2001 can represent the conversion of sensing matrix (" Φ ").Matrix 2002 can represent rarefaction representation matrix (" Ψ ").

Figure 21 is according to the example of the random matrix using in an embodiment of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein.The example of random matrix is referred to as does 2100.Matrix 2101 and 2102 represents the random matrix can using in system 100,200,300,400,600 and 800, wherein N be 10 and M be 5.Matrix 2101 can represent the conversion of sensing matrix (" Φ ").Matrix 2102 can represent rarefaction representation matrix (" Ψ ").

Figure 22 shows a kind of different sample mode.This figure is based on [CW08].Sampler in Figure 22 is the set Φ of sense waveform.If f is sparse, can be in the situation that there is no error restoring signal x.If in representing f=Ψ x, x only has S nonzero term (referring to [CW08, the 23rd page]), N dimensional signal is that S is sparse.Represent that parameter is for characterizing the parameter of feature of the variable of expression formula f=Ψ x.These parameters comprise the number (being S) of nonzero term in the value of line number (being N) in Ψ, the element in Ψ and x.By new operation, " sensing " and " solving " substituted to " sampling " and " recovery " step in Figure 22.

Step 1, sensing

make

J &Subset; {1, . . ., N} - - - (4)

Step 2, solve

meet

&ForAll; k &Element; J - - - (5)

Equation (1) and (2) are from [CW08, equation 4 and 5].In equation (1), bracket <> represents inner product, also referred to as correlation.By the l of indication ₁norm is the absolute value sum of the element of its independent variable.

In order to use the least possible sense waveform, what the vector sum of basic matrix Ψ took out from Φ must low [CW08, equation 3 and 6] for the coherence between the vector of sensing.Coherence μ is provided by following equation:

μ (Φ, Ψ) = \sqrt{N} \max_{1 \leq k, j \leq N} {| | < φ_{k}, ψ_{j} > | |}_{l_{1}} - - - (6)

For designing the incoherent method of sampling (than [CW08]) of sampling system, be:

1, to f modeling, and find Ψ, in Ψ, f is that S is sparse.

2, the incoherent Φ of selection and Ψ.

3, select randomly the M row of Φ, wherein M > S.

4, use selected vector is sampled to f, to produce y.

5, Ψ, Φ and y are passed to l ₁minimize device, and recover x.

Can be applied to l ₁a kind of minimized method is simplex accelerated process [LY08].

Embodiments of the invention shown in Figure 23 comprise low-power receiver.Can, as what instruct in [ESY05, KJR+06], realize the RF part of described low-power receiver.This figure represents multiple access system 2300.The multiple access access scheme that can use in native system comprises the TD/DMA[Cas04 of FDMA, TDMA, DS-CDMA, use FDD and tdd mode, 23-45 page, 109] and OFDM access scheme [AAN08].Described system comprises subscriber equipment or UE 2206 and foundation structure 2210.UE 2206 comprises mobile radio station, is equipped with laptop computer and the smart phone of cellular radio.Foundation structure 2210 comprises in cellular system not being the part of UE, such as long-range sampler 2212, base station 2216, central brain and DL steel tower 2222.Long-range sampler 2212 comprises the equipment consisting of following part: antenna, downconverting RF parts, associated components, in backhaul, receive the controller of instruction or state machine, for the memory of stored configuration and for optical fiber (backhaul) upper to base station 2216 optical transmitter that send correlated results or value.Each base station 2216 generally will be fed to by more than one long-range sampler 2212.Can in use entreat in the system of brain concept, or in not using the system of central brain concept, dispose long-range sampler 2212.

Certain other forms that conversion comprises being applicable to sending or calculating represent to input waveform.Example be shift signal frequency (down-conversion), from analog form, change to digital form (A to D changes).

Central authorities' brain is high power infrastructure component, and it can, with acceptable cost, be carried out and calculate with very high speed.Central authorities brain comprise can with the infrastructure component of base station high-speed traffic, thereby make to carry out at central brain place a lot of physical layer computational activity.The controlled in wireless of carrying out via base station and DL steel tower is not to making to overcome the communication failure being associated with channel fading speed, to be infeasible slowly.Central authorities' brain and base station can be identical computer or infrastructure component physically.Base station transmitter is positioned at DL (down link) steel tower 2222 places, and DL steel tower comprises conventional honeycomb steel tower, is arranged on the low-power unit in cellular transmitter, auroral poles or the office on building.

Down link DL 2220 is the information carrying RF energy flows of from foundation structure to subscriber equipment (being UE).This comprises the wireless signal being received by the wireless signal of DL steel tower 2222 transmissions and UE 2206.

Decline comprises can how from a lot of reflectors, flicking wireless signal, and resulting reflection and the description of attribute.More information for relevant to decline, refers to [BB99, the 13rd chapter].

Environmental parameter comprises: the distance from UE to long-range sampler, the distance from UE to DL steel tower, in SNR and any co-channel signal presenting and any decline of any interested long-range sampler.

In cellular system, there is several access.When UE wishes to arrive foundation structure, but when foundation structure does not know that this UE exists, there is Aloha and access at random.After using the license of system to UE and having distributed UL and DL channel, there is bidirectional data exchange.More discussion for access, refer to [Cas04, the 119th page].

" channel " comprise by time, frequency, code and/or spatial limitation parameterized licensed waveform.Example will be specific T DMA time slot in particular cell sectors in gsm system.On channel, send for maintaining honeycomb and connect required user data and/or signaling information.

Term " base station " is generally used for the description comprising entity, this entity receives the signal of bearing optical fiber from long-range sampler, 11 solvers and quantizer are provided, and operation intelligently (, operation computer software) identify the detected message of quantizer, thus with DL carry out and UE between agreement exchange.Base station is created on the upper Overhead Message sending of DL.Base station is a part for the central brain concept of RIM (2012244 Ontario Inc.) establishment in function." solver " comprises the equipment that uses 11 distance metrics.This distance is measured as the absolute value sum of the difference in each dimension.For example, the distance between (1.0,1.5,0.75) and (0,2.0,0.5) is | 1-0|+|1.5-2.0|+|0.75-0.5|=1.75." quantizer " comprises the equipment for following operation, and this equipment is accepted the estimation as input, and one of produces in the finite aggregate of information symbol or word as output

Base station receiver, solver, quantizer and controller are known as the some place of " base station " 2216 in figure.Base station 2216 and DL steel tower 2222 place that can coexist, and in any event, they are connected completely for the object of signaling.Up link 2224 is the information carrying RF energy flows from UE 2206 to foundation structure 2210.This comprises UE 2206 wireless signal that send and that received by one or more long-range samplers 2212.

Cellular system provides multiple access access to a lot of mobile subscribers, for real time bidirectional communication.The example of these systems is GSM, IS-95, UMTS and UMTS-Wi-Fi[Cas04, the 559th page].

Mixed type is grand/and microcellulor network comprises for the Big Residential District of the vehicles with for pedestrian's small cell [Cas04, the 45th page].For the general points of view of cellular system design, GSM or WCDMA system are suitable frames of reference.That is, they have shown the layout of mobile radio station (UE), base station, base station controller etc.In these systems, the phase place that depends on the communication between UE and foundation structure, use various signaling systems, such as random access, paging, resource are distributed (channel allocation), expense signaling (regularly, navigation system id, be permitted for the channel of access), switched training or pilot signal and stable state communication (voice or data, grouping or circuit) on message, up link or down link.

In [CG91], presented and via optical fiber, to base station, be fed to unsampled analog signal.In Chu, the transducer of a type is additional to antenna, and to optical fiber feed.Transducer in [CG91] is not to RF signal sampling, and it only uses simulated laser transmitter that signal is converted to optical energy.A part of novelty of this invention be from remote antenna be sent to base station value number and attribute and how to control number and attribute.

Conventionally in the context of harmless sampling, consider Figure 24.If for | f| > fmax, the power spectrum of signal A (f) is zero, can the discrete sampling based on speed 2fmax sampling represent time-domain signal a (t) [Pro83, the 71st page].In this general scene, the unique thing about A (f) that sampler is known is that A (f) is zero on fmax.

For sampler being locked onto in the wireless system of spreading rate, in general, harmless sampling will once be sampled and form by each chip.For N Chip Waveform (it is included in the frame of N discrete sequential time point place definition), this will mean N sampling of each chip-level code word.This frame may be the frame that is ready to be transformed into the passband for sending, or it may be only the frame with Boolean, real number value, complex values in computing equipment or memory.In one embodiment of the invention, utilize M value to carry out a sensing N Chip Waveform, wherein, M < N." frame " comprises the set of the time sampling that order is caught.The set of boolean's (or real number or plural number) value that frame also can description sequentially produces.

" noise " comprises additive signal, its information distortion of searching that this additive signal makes that receiver watches.Source can be that the thermal noise at receiver antenna place or it can be from cochannel wireless signals less desirable or other expectation sources, or it can be from other sources.In [BB99, the 2.6th chapter], discussed the basic theories of detection signal in noise.

" performance " comprises the expection operation according to designer, and how wireless system work obtains.For example, designer may wish when UE powers on and identify overhead signal, and UE is by the message sending to base station alarm.The performance that this signal is detected in base station comprises the probability that base station sends the single of this message of identification.Described performance changes according to system parameters and environmental factor." system parameters " comprises the length of message frame, the degree of rarefication of the number of sense waveform and the message that is sent out.

Up link is the information carrying RF energy flow from UE to foundation structure.This comprises the wireless signal that UE sends and received by one or more long-range samplers.Irrelevant sampling comprises a kind of compression sampling of type, and it depends on and the incoherent sense waveform of basic matrix Ψ (row of Φ), and in basic matrix Ψ, input signal is sparse.This report discloses simple sampling and low-rate data sends, and to save the power of battery at long-range sampler place, refers to Figure 25.Compression sampling comprises a kind of technology, in this technology, utilizes the specific properties " degree of rarefication " of input signal to reduce the number that does not lose the desirable value of information needed for reliable expression signal (in statistical meaning) simultaneously.Below some general points relevant to this creativeness framework.

Even if 1 sampler quits work, integral honeycomb system also continues with complete capability operation.

2,, in building/urban environment, long-range sampler extensively distributes with interval 30 to 300m.

3, base station is not limited to its computing capability.

4, by conventional community steel tower, provide cellular system down link, there is no uncommon RF Power Limitation simultaneously.

5, will save UE battery, target effective load data transmission power level is 10 to 100 microwatts.

6, by optical fiber, any given long-range sampler is connected with base station.An alternative of disposing for selected sampler will be coaxial cable.

7, if possible, long-range sampler should operate with the power of battery.Using line power (being 110V, 60Hz in the U.S.) is another kind of possibility.

According to total system characteristic, we infer the following characteristic of long-range sampler.

1, long-range sampler is dirt cheap, and is almost droppable.

2, long-range sampler battery must continue 1-2.

3, long-range sampler power budget will not allow to carry out receiver detection/demodulation/decoding algorithm.

4, long-range sampler will have RF down-conversion chain, and for send some schemes of digital sample to base station.

5, long-range sampler identifies by not having the computer intelligence when UE sends signaling.

6, long-range sampler can be from base station reception and down-conversion and the relevant instruction of sampling.

The example of modulation scheme is QAM and PSK and differential variation [Pro83, the 164th, 188 pages], coded modulation [BB99, the 12nd chapter].

According to these characteristics, produce these design rules:

Rule A: all optional calculation tasks are shifted onto base station from sampler.

Rule B: the transmission rate by sampler on optical fiber is reduced to the minimum rank of no-float systematic function.

Rule C: strive for the balance between total system effect and sampler battery save.

Rule D: allow sampler change robust for the physical layer of evolution, do not download and rely on CPU.

According to above-mentioned design rule, we obtain the design of drawing roughly in Figure 23 and 25.In this report, we are absorbed in following problem: when Unidentified UE (subscriber equipment or mobile radio station) exists before, to base station alarm.Except we are untreated power is controlled or disturbed herein, this situation is similar to an access scene of describing in [LKL+08, " situation 1 "].There is the scheme of well-known these problems of control.Sampler operates with the system clock that locks onto base station and provided.

Illustration for the message sending in the cellular system access events of paying close attention at this report, refers to Figure 26.Figure 26 sends an illustrative case of existence signal 2314 exemplified with UE 2206.In the figure, UE 2206 powers on, and observes overhead signal 2312, and starts to send existence alarm signal 2314.Term " existence signal " comprise by UE 2206 be sent to base station can the incoherent sampling of sensed waveform any signal." sense waveform " comprises the row in sensing matrix Φ, and these row are associated to obtain correlation with the frame of input.Described correlation is known as y _i, wherein i is the row of the Φ of use when relevant.Conventionally, whenever UE 2206 determines that by Overhead 2312 it is approaching the community of the current UE of not knowing 2206, UE2206 always can be used existence alarm signal 2314.Long-range sampler 2212 sends continuously sensing and measures y, unless M=0.

Sensor parameter is the parameter for the feature of expression variable.Send continuously expense 2312.In expection existence alarm signal 2314, by be identified in the situation that, send existence alarm signal 2314.UE and base station exchange messages in the following manner: UL is that UE 2206 is to long-range sampler 2212.Long-range sampler 2212 is sensing continuously, and does not detect, and sends sensing measurement y by base station on optical fiber 2216.DL is base station iron tower 2222 to UE2206, and for example, message 2318 indications are when sending existence signal 2314, and UE is used degree of rarefication S ₂.Sparse signal comprises can be by the N Chip Waveform that the S from NxN matrix row summation created.The key property of this signal is the value of S, i.e. degree of rarefication.For non-trivial signal, the scope of S is from 1 to N.Show the instruction 2316 of the value that changes the M that long-range sampler 2212 uses.Indication is a kind ofly to UE, to send message or the mode of the particular value of the particular variables that will use to long-range sampler indication.This figure is intended to accurately illustrate how many message of transmission.

UE is also upper via the overhead transmission from base station at down link (DL), access system clock.The wireless energy bandwidth B of long-range sampler observation centered by certain frequency f c.In general, long-range sampler is not treated B as its only information, thus its really by optical fiber, to base station, to provide speed be the sampling of 2B.On the contrary, sampler obtains N sampling of described N Chip Waveform, and calculates M correlation.By optical fiber, a resulting M value is sent to base station.If sampler does not have chip timings locking, it can obtain 2N sampling in half chip timings, and calculates 2M correlation.The sampling that is sent to base station reduces to 2M from the 2N of conventional scheme.

Sampler can, by relevant to the independent Φ matrix column of selecting, calculate sensing and measure y.Sensor parameter is for being illustrated in the parameter of received signal g and the feature of the relevant variable of Φ matrix column.These parameters comprise: the value of element in the number of the element in y (being M), Φ and the number of the chip samples that g represents (being N).To the selection of used Φ matrix column be except to the selective dependency of the value of M itself in to the estimation of S, in the situation that carry out without the knowledge of any information of x.Therefore, use which row of Φ irrelevant with Ψ, but the number of the row of the Ψ using depend on the estimation to the attribute of f.Or, can as in Figure 26 shown in the time t17, by DL, sent the degree of rarefication of controlling f.

The necessary condition of locating successfully to detect x in base station is that the value of the M that long-range sampler must be used is chosen as and is greater than S.Can, by guessing and adjust afterwards at place, base station, overcome the knowledge that lacks S.For example, M can be from the maximum of N, and along with the activity grade of frequency range B is understood in base station, and M can move to less but abundant large value still.In this manner, can allow the power consumption (power consumption when calculating correlation y and while sending to base station by optical fiber all) at long-range sampler place keep low.Base station can periodically increase M (via the instruction of going to long-range sampler), to assess up hill and dale the degree of rarefication of the signal in frequency range B.Sampler can be instructed in base station, and which row it should use, or sampler can select row according to scheduling, or sampler can be selected randomly row and to base station, notify its selection.

Detection comprises estimated value is operated to obtain the closest approach in the planisphere of limited size.Planisphere comprises set a little.For example, if each point in this planisphere and unique being associated of vector that comprises N, and each to the greatest extent can value 0 or 1 (conventionally, vector entries can be Boolean, real number or plural number), in this planisphere, has the individual or point still less of 2N.

After powering on, UE 2206 wishes to allow system know its existence.For this reason, UE sends existence alarm signal 2314.This existence alarm signal is by selecting the row in Ψ matrix the information signal that its summation is constructed.On the selection of row, can be subject to the impact of base station overhead signal.For example, the subset of the Ψ row that will therefrom select can be specified in base station.

Base station can require the UE being not yet identified to send a particular column via DL Overhead Message 2312, is called Ψ ⁰.This will serve as pilot tone.Long-range sampler 2212 will operate according to incoherent sampling, and 2216 send sampling y to base station.This signal will be processed subsequently in base station 2216, and detect Ψ ⁰existence, estimate the multiple fading channel gain between previous Unidentified UE and long-range sampler then for example indication has sent Ψ ⁰any UE start to send latter two bit in their ESN (Electronic Serial Number, globally unique moving station mark symbol)." sampling " comprises that the signal that signal is constantly had to a value from each changes into discrete series corresponding to the input of discrete time point (periodically or aperiodicity).

If clashed between the transmission from two different mobile radio stations, can use up link (UL) standard A loha random back (back-off) technology to come separated follow-up UL to attempt.

Long-range sampler 2212 is not known this agreement progress, and utilizes simply the row in Φ to keep sensing, and 2216 sends sampling y to base station.Base station 2216 can indicate long-range sampler 2212 to use the individual sensing row of specific quantity (M).This quantity can change from more or less information flow of UE along with base station expects.If base station expects has the S of maximum N, increase, base station will indicate long-range sampler to increase M (maximum that M can get is N).For example, in Figure 26, identification message can comprise the new value S of the S that UE will be used ₃, and base station can configure the larger value that long-range sampler uses M simultaneously, is known as in the drawings M ₃.In the figure, these events occur in time t ₁₃, t ₁₅and t ₁₆.At t ₁₇, base station expectation has degree of rarefication S ₃message, and expect to have may with suitable M value, (particularly, this value be called M herein ₃) sense this message.Exemplified with sequence of events, but this sequential does not mean that it is accurate.In the situation that M is increased to the limit, if Φ be certainty (for example, sine) and be plural, when M gets limiting value N, the Φ in long-range sampler becomes DFT operation (being likely implemented as the discrete Fourier transform of FFT).Continue the description under this scene, once base station has had a part of the ESN of all UE that attempt connecting system, base station can utilize specific part ESN, tells the resource that particular UE sends its complete ESN and wishes request.Or base station can be qualified in this system time at definite UE, Resources allocation.

Long-range sampler/centrencephalic system carries out signalling information in noisy environment, and almost there is no intelligency activity at long-range sampler place.This system has the benefit of feeding back via conventional DL.Link budget be included in consider RF energy source and before receiver is attempted restoring signal, cause wireless system design in lossy situation.For more details, refer to [Cas04,39-45 page, the 381st page].Our initial link circuit budget is calculated and is illustrated: if can realize reuse factor 3, and realize 0 to 10dB reception SNR, UE can the transmitted power with 10 to 100 microwatts operate in 20 to 30m scope.These numerals are the quantity of " magnitude " type, and do not have significant digit.For the detection of existence signal, common more than one sampler can receive the f of noisy different editions, and can carry out joint-detection.This will allow M lower only at a long-range sampler visible f in place in the situation that at each sampler place ratio.Therefore,, by disposing sampler with compact mode, can reduce the battery consumption of each sampler.In brief, sometimes the noisy version of f is called to g.

" reuse " and be included in before model identical occurs geographically again, wireless bandwidth resource has been carried out to the non-overlapped deployment of how many times.

For worst case design, we suppose that the signal from UE is only mapped on (imping on) long-range sampler.Conventionally, for indoor transmission, we expect that two long-range samplers are within the scope of 30m, and path loss index is between 2 and 3.This design is not limited to indoor transmission.Outdoor, scope is larger, but path loss indication will be tended to less.For successful detection, the probability that signal transmission detected should be higher than 10% (assumption error mechanism be that noise causes, and therefore detection trial will be independently).Can in macrocell, dispose long-range sampler to support vehicle-mounted business, and in Microcell, dispose long-range sampler to support pedestrian or office's indoor communications business.

Come concrete example, we have designed following scene so.

1, channel is static (without decline).

2, noise is AWGN.

3, UE, long-range sampler and base station are all locked into the clock of timing, frequency or the phase error without any type.Can process the fault [BB99, the 5.8th chapter, the 9th chapter] such as these with standard mode.

4, there is a UE.

5, incoherent sampling plan is used at random to (Ψ _r, Φ _r) or certainty to (Ψ _d, Φ _d), in arbitrary situation, solver is known all information except signal x, f and noise.

6, base station has indicated sampler to use the specific collection of M the row of Φ.

7, base station consists of N interval or chip to UE and sampler indication transmission waveform.

Figure 27 is according to the diagram of an embodiment of the long-range sampler/central brain cellular architecture of the various aspects of setting forth herein.Described information is x 3240.F 3242 is that S is sparse, and is discussed as other places of this paper, and S has been estimated in base station.The input of long-range sampler 3212 is noisy versions of f, and it is known as g 3244 here sometimes.Long-range sampler 3212 calculates M correlation between the row of selecting in advance in g 3244 and Φ, produces Mx1 vector y 3215 (equation 1).By y 3215 via optical fiber going down to base station 3216.

" estimation " is statistics term, and it comprises attempts selection numerical value from show the finite aggregate (such as real number) of minimum range in some sense, from the true value of x, select.The method of measurement of the minimum range of often using is mean square deviation (MSE).A lot of estimators are designed to minimize MSE, i.e. expectation in [Pro83, chapter 1], contained statistics computing, such as expectation.In reality, often with fixed-point value, represent the numerical value of estimator output.

For real number, g with between correlation or inner product be calculated as wherein g (k) represents k the element of g.

For plural number, correlation will be wherein, g ^*represent complex conjugate.

12 norms of signal g are expression formula for real number is identical, and complex conjugate is inoperative in this case.

First base station 3216 produces the estimation of x, is known as 3246, then produce and be known as 3248 hard decision.By forming linear programming (linear program), then use simplex accelerating algorithm to solve, produce and estimate 3246.Algorithm is explored and is produced vectorial y ^*nx1 vector x ^*the border of area of feasible solutions of realization.Search does not rely on degree of rarefication.Because signal is sparse, therefore 11 minimize and work, but minimize device, acting on without attempt utilizing degree of rarefication in the situation that.

Therefore, x ^*n be all non-zero conventionally.Generation is met to y ^*the y of=y ^*and have least absolute value and x ^*be elected to be (equation 5). conventionally be not equal to x, therefore carry out hard decision with find by S 1 and N-S 0 that form with immediate vector

Linear programming comprises the set of equation and possible inequality.Variable only occurs with linear forms.For example, if x1 and x2 are variablees, form variable do not occur.

The probability that this quantification identifies the one or more correct nonzero term in x is the probability that emulation is designed to determine.There are a variety of definition of " the most approaching ".We determine as follows quantizer 3230 is first right element carry out arithmetic sequence, and retain to start the index (for example+1.5 are greater than-2.1) of S element.Secondly, quantizer will all items be set to logical zero.The 3rd, quantizer will in those elements of the index with the index that equals retained be set to logical one.This result is the output of quantizer.

Quantizer 3230 obtains S according to variety of way.Example will be omniscient smart (all-knowing genie) (determining for limiting performance), or the value of the DL S that fixedly mobile radio station will be used has been used in base station, or base station is for example, by attempting different values (via the instruction of going to long-range sampler) and determining the degree of rarefication of f during certain grand time period (1-2 second), " scanning " S periodically.Because UE will repeatedly attempt, so base station has an opportunity to identify, the estimation of S is missed to (miss-estimate), and indicate long-range sampler to reduce or increase the value of the S of its use.In the situation that the abundant low duty ratio relevant to scanning S will retain the power save aspect of detection technology.In this manner, the movable degree of rarefication that is mapped to the signal on this long-range sampler of following the tracks of of the sensing of long-range sampler.Therefore, long-range sampler, conventionally always in sampling, still only has the battery consumption enough operating for system, and does not produce the battery consumption that enough supplies the battery consumption of system operation considerably beyond this.Particularly, for not there is long-time section of UE, long-range sampler is not sampled with full Nyquist speed.

Y ^*it is the symbol from [CW08, the 24th page]. not the symbol from [CW08], because this quotes the signal of not processing by noise corrupted.General use in industrial quarters for estimating and detecting exported with symbol, and can for example in [Pro83, the 364th page, Fig. 6 .4.4 " Adaptive zero-forcing equalizer "], see these symbols.

Figure 27 shows functor (functional piece) and the signal in Computer Simulation.In table 1, specified the attribute of the matrix using.Described row are normalized to unit length.Refer to the example of these matrixes in Figure 20 and 21.

Table 1: the attribute of matrix

Certainty matrix only generates once, and if again generate and will not change.Random matrix can only generate once, or can after a period of time (such as several seconds), regenerate random matrix.In addition, can in the time that random matrix will be used, regenerate random matrix at every turn.In arbitrary situation, solver 3228 must know UE 3206 is used what Ψ matrix and sampler 3212 to use what Φ matrix at any time.This does not mean that solver 3228 must stipulate what matrix (dictate) is used.If UE changes Ψ according to pseudorandom (" the pn ") function of system time (time obtaining via DL expense), solver 3228 can find what Ψ is with identical pn function maker.Unless Stated otherwise, the probability providing in this report is for all SNR and the test at these SNR places, to be fixing situation for once generating random matrix and this random matrix.

Emulation is restricted to real number, so that the restriction that it is real number that exploitation easily, does not exist their application restric-tion in the scheme still presenting herein.Identical structure piece technology (such as correlation and linear programming) can be applied to conventionally utilize in the system of plural modeling.This is genuine, because any plural a+jb can be written as to full real number 2x2 matrix, wherein the first behavior [a-b], the second behavior [b a].

Can in scalar or matrix level, complete above-mentioned.Therefore, any set of complex numbers of equation can be rewritten as full set of real numbers.

?SNR(dB)	S	Pr{ always misses }	Pr{j=1 hits }	Pr{j=2 hits }
					?0	1	0.67	0.32	n/a
?10	1	0.29	0.71	n/a
					?20	1	0.12	0.87	n/a
?0	2	0.44	0.46	0.09
					?10	2	0.22	0.47	0.30
?20	2	0.16	0.28	0.55

Table 2: at M=5, N=10, AWGN, has the detector performance in the Ψ of iid Gauss item and the situation of Φ.Referring to Figure 27.

In these emulation, the performance that we find be surpass about 10% those.Because only chance event is that noise, signal and matrix generate, therefore do not need the test of large quantity.In most of the cases, with 100 or 200 tests of each point, collect data point.In about 0.5% test, our 11 solvers are realized while exiting in the situation that it should be in existing solution, attempt continuing to optimize reject these few tests.Even using its as success or unsuccessfully comprise, on the impact of result, will be also ND, because we are finding any performance that is greater than 10%.

In Figure 14, draw the data in table 3.S is the number of the nonzero term in x, and in Figure 14, is known as " pulse ".Event " j=1 hits " means that detector just correctly detects a nonzero term in x.The in the situation that of S=1, this is the optimal cases that detector can be accomplished.Event " j=2 hits " means that detector just correctly detects two nonzero terms in x.

I have also done at M=3, the emulation in the situation of N=10 and S=1 (referring to Figure 17 discussed below).

?SNR(dB)	S	Pr{ always misses }	Pr{j=1 hits }	Pr{j=2 hits }
					?0	1	0.64	0.36	n/a
?10	1	0.13	0.87	n/a

20	1	0.03	0.97	n/a
					0	2	0.42	0.49	0.09
10	2	0.13	0.40	0.47
					20	2	0.07	0.19	0.74

Table 3: at M=5, N=10, AWGN, has the detector performance in the Ψ of certainty item and the situation of Φ.

Figure 15,16 and 17 has provided the detection performance for the various combinations of M, N, S, SNR and matrix attribute.In each of these figure, j be by 11 minimize device and quantizer combination the correct number (Figure 25) of the nonzero term in definite x.

For system, important probability is that detector obtains the probability of message in once observing entirely truely.Suppose that system use repeatedly sends, with regard to uncontrolled effect (as noise), sending each time will be independently.In this case, the probability that existence signal detected in sending for C time or still less time sending is 1-Pr (missing) C.Can be defined as event j=0 or event j < S by missing.When S=1 and use random matrix, event j=S sentences the probability generation that is greater than 10% at the SNR lower than 0dB, and sentences at SNR and the S=2 of about 3dB the probability generation that is greater than 10%.As shown in figure 15,90% point is respectively about 12 and 17dB place.As shown in figure 16, for certainty matrix and S=1, performance is better.

In order to see how detector will work when degree of rarefication condition (M > > S is non-true) is weak, and we use S=1 and M=3, generate data as shown in figure 17.Random and certainty configuration can both detect at low SNR place, but random arrangement is saturated, approaches 70%, rather than arrives 90% point.With M=5, the situation of N=10 is compared, the performance of random arrangement will be almost (for example, at SNR=10dB place, Pr{ detects }=0.55, simultaneously for M=5, this probability is 0.71).In Figure 17, at high SNR place, for determinacy case, probability approaches 1.

Therefore, we see that we have approached the Candes noiseless result that reaches 100% reliable Exact recovery in the situation that of M and SNR increase.Yet for low M and noisy signal, solver sometimes produces and is not equal to x's important quantization characteristic is: for certainty configuration, deterioration is gradually.Unless M > > S, otherwise for random arrangement, there is the threshold effect in noise.In Figure 17, M=3S, and in every other figure, when S=1, M >=5S.

The uncommon characteristic of the incoherent method of sampling is incoherence.Most of detectors seek to attempt a lot of candidate's waveforms, to check which waveform and the Waveform Matching being received, then use " selecting maximum " function of certain type to determine identity or the index of sent waveform.The local replica waveform identical with the identity of sent waveform that be had identity.In incoherent sampling, unique requirement is that Ψ and Φ are weak at most relevant.This means various sensing matrix (Φ _s) can be for any Ψ.For random case, we explore the effect that changes two matrixes when each transmission.The result of this situation has been shown in Figure 18 and 19.Accordingly, we notice some variations in performance, even at high SNR place.We confirm following guess: this has poor auto-correlation attribute " bad " matrix owing to having generated.High correlation in any matrix will weaken estimated capacity, because for Ψ, it is by the support of distinguishing between the value reducing for the x in any two related column, and for Φ, it will reduce the ability of distinguishing between the contribution of the candidate for two related column at Φ of solver.For the mechanism localization in high SNR these variations of place, we refuse following Φ matrix: in this Φ matrix, any two row have the correlation magnitude of the threshold value of being greater than.In the drawings, threshold value is 0.1.Utilize other threshold values to be studied.Threshold value 0.4 does not almost have effect.We obtain accordingly: yes, actual Φ matrix is extensively to change to the influential effect of performance.The another kind of mode of rationally inferring is: exist us not wish for carrying out the Φ matrix of sensing " bad ".Described performance is the stochastic variable with the distribution correlation of matrix.This means and can define outage probability (probability of outage).Particularly, outage probability is that detection probability is by the probability lower than probability threshold value.For example, system can be designed as and makes not only average detected probability be greater than 40%, and detection probability is less than 1% by the probability that is less than 10%.We can reduce the number of " bad " matrix, to reduce outage probability.A kind of mode of so doing is the correlation in constraint Φ matrix.Constraint Ψ matrix also will be beneficial to, particularly when S increases.

For robust high bandwidth real time service and high user density are provided by radio, we have created based on distributing antenna and the framework to the centralized processing of wireless signal.We claim that this system is remote conversion or long-range sampling.Mobile radio station is simple low-power equipment, and foundation structure core is supercomputer class, and base station is connected with the cheap wireless senser of mobile radio station by redundancy magnanimity.Figure 28 is the diagram of our cellular network that herein proposes.Figure 28 shows the simple sensor 2712 of a series of a large amount of deployment, thereby makes common more than one transducer 2712 in the scope of mobile subscriber (MS) equipment 2206.In this programme, can also by these scales, be long-range sampler or remote conversion equipment.Can be by the distance of ten meters, these transducer intervals to hundreds of rice.Between the required capacity of transducer power demand, the easness of disposing transducer and system, exist and dispose balance.UE can be used than the general higher frequency range of frequency range of using in cell phone.

The optical fiber backhaul 2714 of going to central base station 2716 is provided to transducer.Described backhaul also can be provided by another kind of medium, such as coaxial cable.In deployment, can there are some base stations, their communication transmission of informations.Transducer has one or more antenna that appends to RF front end, and has and be designed to cheap Base-Band Processing.The transducer with an antenna can be used as to array, and can be made MIMO air interface.

The air interface that wave beam forms allows MS to send with low-power.The upper-layer protocol using between MS and base station can be for example, an agreement from standardized cellular network (LTE).The upper-layer protocol of becoming privileged in low-power and short distance (for example Bluetooth) is the substitution model for the communication between MS and base station.The storehouse at transducer place will only comprise the part of the 1st layer (physical layer).This has reduced cost and battery power consumption.Transducer is likely powered by AC (being 110V power line in the U.S.).Can use the low hybrid ARQ repeating transmission two-way time feed technique for the treatment of application in real time; The 2nd layer of unit processing ARQ will be in transducer, and in BS or central brain.Provide innovation region herein, complete new topological structure, wherein, transducer compresses the high bandwidth movable signal receiving in short distance, and foundation structure is at full speed carried out physical layer calculating.

1, the instruction between base station and transducer, communication protocol and hardware interface

A, remote conversion instruction

B, oscillator return instruction

C, wave beam are controlled (phase sample) instruction

2, communication protocol and the hardware interface between MS and BS or central brain

A, can support the MS of real time service and the high bandwidth MAC hybrid ARQ link between BS.

3, communication protocol and the treatment technology between MS and central processing unit/central brain

A, in the situation that without the existence signaling code of working from the cooperation of transducer

B, the empty time-code mixing for this new topological structure and channel knowledge

C, the fountain codes of registering and sending in real time for mobile radio station

D, large-scale array signal process technique

E, utilize higher frequency to send the signal processing technology of frequency range

4, following activity is supported to comprise in base station:

A, transmitting system Overhead

The existence of b, the mobile radio station of detection in the scope of one or more transducers

Bidirectional real-time between c, base station and mobile radio station.

This memorandum has proposed transducer or the sampler that will in cell phone framework, use.These transducers are more cheap than base station, and the RF signal of high bandwidth (for example bandwidth B) is sampled.By optical fiber, to base station, send compressed signal.Transducer is not carried out Nyquist sampling conventionally.So do For several reasons.A reason is that two-forty sampling consumes too much energy.Our object is to provide low power sensor technology.Expectation by Redundancy Design in system, thereby make easily to overcome the loss of single-sensor.For a lot of important signals, can to the signal existing, carry out low error reconstruct at place, base station.Transducer can be equipped with direct sequence despreader or FFT equipment.Transducer does not carry out demodulation judgement.The frequency spectrum bands that the direct sequence code using in despreader or the timing of the sub-chip of despreader or the frequency separation number using in FFT or transducer will be applied FFT is that the content of transducer is told in base station by instruction.In one embodiment, these instructions send with 1ms interval, and transducer is being adjusted its sampling or conversion for receiving being less than of this instruction in 0.1ms.For the object of structure, we suppose that the mobile radio station sending and receiving duration is 1 to 5ms information block or frame.Form can be circuit class NOR circuit class.Overhead information can comprise guiding and synchronizing information, and this guiding and synchronizing information make mobile radio station according to game theory, carry out and copy good cooperation behavior.Suppose that MS need to know possible wireless WAN, can also exist important information.By keeping all communications in this sub-communication network, and external network needn't be monitored, the power of battery can be saved.Mobile radio station sends their message with low-power.Transducer is sampled to wireless channel.Transducer in this proposal compresses sampling.By fiber channel, to base station, send the compression sampling in this proposal.A lot of the 1st layer of activity (demodulation, decoding), the 2nd layer of activity (packet numbering, ARQ) and signaling activity (registration, channel allocation, switching) are responsible in base station.The computing capability of base station is high.Real-time solution equation system can be carried out by this computing capability in base station, and in existing system, this equation system is by only emulation under off-line mode.The information (number of mobile radio station antenna related matrix, the transducer when considering mobile radio station) of channel can be used in base station, carrys out to determine link adjustment strategy on 1ms interval.These strategies operate being included in optimum spatial multiplexing gain/diversity gain balance point place.Equally, a plurality of base stations can be carried out almost communication each other simultaneously, and optimally design sends waveform, and described transmission waveform will add at simple mobile radio station place and to obtain undistorted waveform (dirty paper code).Other base stations that receive extra uplink energy from mobile radio station provide the 1ms wiping in other cases frame period to anchor base stations once in a while.Figure 29 shows another schematic diagram of proposed system.It is movable that transducer 2712 in this proposal is only responsible for the 1st sublayer, i.e. compression in sample stage.Base station 2716 in this proposal can send instruction to transducer, such as using multiple access access code 16 (these may be DS code, OFDM code) to compress.Base station can send instruction, and the 2x such as execution with phase place theta samples.In other words, transducer is the long-range path of pulling out, A/D path from conventional base station, just looks like to pull one jiao of taffy, and creates thin connecting line.Described taffy line is the metaphor of the fiber channel from transducer to base station.Very high available computing capability is used in base station, detects the existence of MS signal in packed data.Then, base station in this proposal is used the sampling of good prize MS signal and compress technique (even if transducer is not known the effectiveness of instruction by indication sensor, also the timing, frequency, the coding details that present the packed data that is full of MS signal), detected MS is responded.MS in this proposal can utilize fountain codes to send, at least for call setup.For very high bandwidth, low-power link, mobile radio station can send real-time voice with fountain codes.Grouping transmission rate should have the cycle of 1 to 5ms magnitude.Transducer master is if it were not for decision device; Transducer neither be local adaptive; Transducer is controlled from base station.Spatially dispose thick and fast transducer, every 100mx100m at least one, and may every 10mx10m mono-.Transducer can support or can not support DL to send.Can transmit DL from thering is the traditional base station steel tower of sectorization.The density of this steel tower will be every 1000mx1000m at least one (building deployment), and may mono-of every 300mx300m (street lamp deployment).

Figure 30 shows according to the simulation result of the performance of an embodiment of sensor-based wireless communication system of the use compression sampling of the various aspects of setting forth herein, wherein, uses N=8, and the M value of S=1 and variation is carried out the performance of measuring system 3000.Illustrate at the compression sampling of each value of M and the mutual information between transmitted signal.Based on simulation result, identify the object run region of the compression sampling framework of proposition.The importance of these observations is the following fact: preserving battery life is the determinant attribute of the compression sampling framework of proposition.When M value increases, sampler requires the more power of battery.Yet if M value is too small, the mutual information between sending and receiving signal may be lower than acceptable level.Therefore, for acceptable systematic function, need to identify the value of M, so that stable system to be provided.For this emulation, rarefaction representation matrix (" Ψ ") is Walsh matrix on attribute, and sensing matrix (" Φ ") is random on attribute.Depend on SNR, the mutual information on the selection impact of the expression of using and sensing matrix between transmitted signal and compression sampling.For specific set of circumstances, representing matrix orthogonalization is good.Use certainty matrix to contribute to increase mutual information, yet, more signaling will be required.Thereby, between signaling and the power of battery, there is balance, and accordingly, coordinating also there is balance between matrix and M value.In the situation that signaling is more limited, should use higher M value.Yet, if battery life is more crucial, should use more signaling.In addition, the mutual information between transmitted signal and compression sampling is relevant to additive noise.Therefore,, when feasible, should use certainty matrix.Yet this will increase the signaling request of system again again.In addition it is favourable, selecting to have expression and the sensing matrix that certain form length retains.

This picture specification integral body is called to 3000.The logarithm amplitude of SNR ratio is shown on abscissa 3009, and draws the scope from 0dB to 35dB.On ordinate 3008, mutual information is shown, and draws from-1.0 to 3 scope.The simulation result of curve 3003,3004,3005 and 3006 expression systems 3000, wherein, N is that 8, S is 1, and random iid Gauss matrix is used for to rarefaction representation matrix (" Ψ ") for sensing matrix (" Φ ") and by Walsh matrix.Curve 3003 shows the lower bound (" LB ") of the mutual information when M=1.Curve 3004 shows the LB of the mutual information when M=2.Curve 3005 shows the LB of the mutual information when M=3.Curve 3006 shows the LB of the mutual information when M=4.3001 and 3007 represent respectively the upper bound and lower bound gathering.The example in object run region is depicted as to region 3002.Executed the max computing best Monte Carlo of probability that retains the Φ of each M realize.Optimal boundary when as shown in the figure, the poorest border (Φ, Ψ) during M=3 is better than M=1.Select target operating area is as the region of being indicated by region 3002, to obtain the reasonable limits to signaling delay.The behavior of this analogue system is all set up for any linear modulation system.

When design system, can change or adjust various attributes, to increase systematic function or maximum efficiency.For example, can distribute to all UE of system identical S value, can distribute to all long-range samplers identical M value simultaneously.This not necessarily because for all UE and long-range sampler, the value of S and M can be different.In addition, for low SNR value, the value of S can be reduced, simultaneously for high SNR, the value of S can be increased.The change of these values is logical, because increase S under low SNR leads, only has considerably less benefit.Yet under high SNR leads, it is significant increasing S in order to transmit more users information.If solver is known the S value of distributing to UE, this system also will benefit.Those skilled in the art it is to be further appreciated that the in the situation that of asynchronous-sampling, and the maximum of M will be 2N, because the synchro system for chip locking (chip lock), needs each code word N sampling, otherwise, for the system without chip locking, must minimumly get 2N sampling.Another aspect of the present invention is that controller can be distinguished various types of signals in compression sampling framework, such as distinguishing at WCDMA and GSM.Therefore, controller can send instruction, carrys out the efficiency of maximum signal transmission with the signal type perceiving based on it.System can also be designed to not require the adjustment to the flight time of UE.For example, in gsm system, the fact of time migration has occurred based on relative other signals of signal, system can require UE to adjust its transmission.Yet, in the system proposing, the chip time period that can grow by use, these are adjusted and are considered in system, make not require the adjustment to UE part.

Figure 31 is the sketch of one embodiment of the present of invention, and wherein, some UE are used compression sampling to communicate.Figure 31 shows the UE 3101,3102 and 3103 communicating by letter with 3106 with long-range sampler 3104,3105.Long-range sampler 3104,3105 and 3106 is connected to solver 3108 via fiber optic cables 3107.Except sending for the instruction of solver 3108 own, controller 3109 also sends instructions via fiber optic cables 3107 to long-range sampler 3104,3105 and 3106.Controller 3109 sends instruction by base station towers 3110 to UE 3101,3102 and 3103.One aspect of the present invention is: UE 3101,3102 and 3103 is not restricted to any specific long-range sampler.Each UE sends simply, and a plurality of long-range sampler is reported the sampling that they catch simply.Down link between UE and controller is realized via base station towers 3110.Up link realizes by long-range sampler 3104,3105 and 3106.

In any given system, if increase the number of long-range sampler, can reduce M value, can not endanger significantly systematic function simultaneously.In addition,, although the present invention seeks to retain the battery life of long-range sampler, if there is the long-range sampler with the remarkable utilisable energy more than other long-range samplers in system, it is favourable the M value at these long-range sampler places being increased.Like this, can reduce the M value of other long-range samplers of energy constraint, and not affect systematic function.

The framework proposing be that channel coefficients based on known reduces signal complexity on the other hand.If there are a plurality of UE that communicate by letter with a plurality of long-range samplers, channel coefficients can be indicated: because some hinder, particular UE is almost only communicated by letter with single long-range sampler.In this case, the channel coefficient matrix being associated with a plurality of UE can show: in specific region, the vector being associated with specific sense waveform is inappreciable.For example, if UE only communicates by letter with a long-range sampler, the channel coefficients being associated with this UE for other long-range samplers can be 0.Thereby, can in the situation that do not consider the measurement result at any other the long-range sampler place a long-range sampler except communicating by letter with UE, rebuild the signal being associated with this UE.By this signal specific is separated, reduced the complexity that represents the matrix of residual signal.This so will reduce the required rated output of solver.Based on this, controller can send instruction to solver, matrix is split into less matrix, to reduce computation complexity.

Figure 32 represents to use the method for the frequency domain sample of frequency displacement and bank of filters.These are the simulation of the system that proposes or the form of continuous temporal correlation.It should be noted that and can be correlated with discrete time or continuous time.The 3212nd, the figure of the sparse signal sampler of use bank of filters.3212 show by the group of M narrow band filter 3202 and recover 3211.To received signal y 3201 carry out multiplexing, and by the signal group of M narrow band filter 3202 of its feed-in.Bank of filters is carried out matrix operation Φ to analog signal.Output is signal y 3203, is passed to optimizer 3204, and optimizer 3204 recovers estimation 3205.Use the frequency domain sample of bank of filters be characterised in that following some:

1, the number M of sampling is limited to the number of the narrow band filter in equipment.

2, when needs M narrow band filter, hardware requirement increases along with M.

The memory stores that 3, can not need y.

4, non-static state or time-varying signal processing are possible.

The 3213rd, the figure of the sparse signal sampler of use frequency displacement.3213 have shown for time stationary singnal from the direct restoring signal of time-domain signal y method.Voltage-controlled oscillator 3207 and narrow band filter 3208 are carried out the computing of Φ in analog domain.By VCO 3202 by signal y 3206 frequency displacements the passband to narrow band filter 3208.It should be noted that and can use low pass filter, rather than narrow band filter, and there is different results.In memory 3209, store output amplitude and phase place, until all M frequency is sampled.Then y3209 is passed to optimizer 3210, optimizer 3210 produces estimation 3211.Use the frequency domain sample of frequency displacement be characterised in that following some:

1, can dynamically change by controlling VCO the number M of sampling.

2, the memory stores that needs the y value of initial discovery, to recover whole vectorial y.

3, signal must be static or slow time variation.

Figure 33 is the block diagram that utilizes the long-range sampler of the concept of sampling continuous time as herein described.Antenna 3301 receives sparse signal, and signal is passed to down-converter 3305.Due to antenna features, noise 3302 will be a part that receives signal 3304, and indicates it to add (although this is not actual structure, the interpolation of noise 3302 is indicated by adder, so that the attribute that receives signal 3304 to be shown) by adder 3303.At 3305 places, signal is carried out to down-conversion.At 3306 places, the configuration of using long-range sampler to receive from remote central processor (not shown), is correlated with to signal.Then sampling 3307 is sent to analog to digital converter 3308.Then the signal after conversion is sent to solver (not shown) along optical fiber 3309.

At Kaukovuori[KHR+06] in provided the example of low-cost wireless electricity, at Enz[ESY05] in provided another example.

By Chu[CG91] propose and tested use optical fiber remote antenna is connected with base station.

Current Intel processor (similar QX9775) is to surpass 1GHz clock speed, surpass 1GHz bus speed and utilize over 1MB high-speed cache and carry out.According to Moore's Law, by 2015, transistor density will reach 8 times of currency.Based on the reasoning of typical clock rate multiple (clock-rate-times) door number (gate-count), we can expect can obtain the disposal ability of about 10 times of described disposal abilities in single processor in 2015.Therefore,, in 1ms, can carry out cisc instruction 1,000 ten thousand times.A microprocessor is by the physical layer adjustment of 10 transducers of real-time instruction.http://compare.intel.com/pcc/

/ restriction that diversity balance relevant multiplexing to MIMO is by Zheng and Tse, 2L.Zheng and D.Tse, at " Diversity and Multiplexing:A Fundamental Tradeoff in Multiple-Antenna Channels; IEEE Transactions on Info.Theory; May2003, pp.1073-1096 ", derive.

At for example Ng, " C.Ng and A.Goldsmith; Transmitter Cooperation in Ad-Hoc Wireless Networks:Does Dirty-Paper Coding Best Relaying?, IEEE ITW 2004, pp.277-282 " in the current concept of dirty paper code has been discussed.

At for example Hales, in " D.Hales; From Selfish Nodes to Cooperative Networks Emergent Link-based incentives in Peer-to-Peer Networks; IEEE Peer-to-Peer Computing, 2004 ", discussed and instructed selfish user to cooperate.

At Kokalj-Filipovic, the concept of a plurality of nodes that receive clever redundancy transmission has been discussed in " A.Kokalj-Filipovic; P.Spasojevic; R.Yates and E.Soljanin; Decentralized Fountain Codes for Minimum-Delay Data Collection; CISS 2008, pp.545-550 ".

According to these tables and figure, we infer: yes, likely design existence signal, and detect at long-range sampler place, meet design rule qualitatively simultaneously.Particularly, having shown two combinations Ψ and Φ makes to utilize considerably less signal process and do not carry out decision-making and detect existence signal and become possibility at long-range sampler.Please recall, existence signal is the row sum in Ψ matrix.In the situation that the S=1 in an x or S=2 nonzero term detects probability enough height for the SNR in 0 to 10dB scope of existence signal.This is in the situation that long-range sampler sends and realizes than the sampling still less of following situation to base station: carrying out routine hypothesis, when signal has been used N dimension base completely, observed signal is being carried out to the required sampling of conventional conversion.Sparse by consciously signal transmitted being designed to, long-range sampler is designed to simply, and by base station design for intelligence and be equipped with the down link of going to mobile radio station of separate design (with the non-coexistence of long-range sampler one place) to be connected, produce this gain.

Appendix A appended and that be incorporated herein by reference, B, C, D, E, F and G have described the technology consideration relevant to design compression sampling system herein.Particularly, discussed the mutual information in long-range sampler in more detail.In addition, solved the noise problem in sparse signal sampling.Appendix C and D have introduced the computer program that is designed to address these problems.Appendix G is the provisional application of submitting on April 15th, 2009.

List of references:

[AAN08]K.Adachi，F.Adachi，and?M.Nakagawa.Cellular?mimo?channel?capacities?of?mc-cdma?and?ofdm.IEEE，2008.

[BB99]S.Benedetto?and?E.Biglieri.Principles?of?Digital?Trans-mission?With?Wireless?Applications.Kluwer，New?York，1999.

[Cas04]J.P.Castro.All?IP?in?3G?CDMA?Networks.John?Wiley?&?Sons，Ltd.，Chichester，England，2004.

[CG91]T.S.Chu?and?M.J.Gans.Fiber?optic?microcellular?radio.IEEE，pages?339-344，1991.

[CR02]S.Cotter?and?B.Rao.Sparse?channel?estimation?via?matching?pursuit?with?application?to?equalization.IEEE?Trans.On?Communications，pages?374-377，March?2002.

[CW08]E.Candes?and?M.Wakin.An?introduction?to?compressive?sampling?IEEE?Signal?Proc.Mag.，pages?21-30，March?2008.

[ESY05]C.Enz，N.Scolari，and?U.Yodprasit.Ultra?low-power?radio?design?for?wireless?sensor?networks.IEEE?Intl.Workshop?on?RF?Integration?Tech.，pages?1-17，Dec.2005.

[JR08]Y.Jin?and?B.Rao.Performance?limits?of?matching?pursuit?algorithms.IEEE?Intl.Sym.Info.Theory，pages?2444-2448，July?2008.

[KJR+06]J.Kaukovuori，J.A.M.Jarvinen，J.Ryynanen，J.Jussila，K.Kivekas，and?K.A.I.Halonen.Direct-conversion?re-ceiver?for?ubiquitous?communications.IEEE，pages?103-106，2006.

[LKL+08]M.Lee，G.Ko，S.Lim，M.Song，and?C.Kim.Dyanmic?spectrum?access?techniques：Tpc-resilient?initial?access?in?open?spectrum?bands.Intl.Conf.on?Cognitive?Radio?Ori-ented?Wireless?Networks?and?Comm.，pages?1-6，May?2008.

[LY08]D.Luenberger?and?Y.Ye.Linear?and?Nonlinear?Programming.Springer，third?edition，2008

[Pro83]John?G.Proakis.Digital?Communications.McGraw-Hill，New?York，New?York，first?edition，1983.

[TGS05]J.A.Tropp，A.C.Gilbert，and?M.J.Strauss.Simultaneous?sparse?approximation?via?greedy?pursuit.IEEE?ICASSP，pages?V721-V824，2005.

Shown and described example embodiment, those of ordinary skills can complete the further adjustment to method, apparatus and system as herein described by appropriate modification without departing from the scope of the disclosure.Mentioned some this potential modifications, and other modification will be apparent for those skilled in the art.For example, the example of above-mentioned discussion, embodiment etc. are exemplary but not inevitable requirement.Correspondingly, should consider the scope of the present disclosure for following claim, and be to be understood that the disclosure is not limited to shown in specification and accompanying drawing and the details of structure, operation and the function described.

As mentioned above, the described aspect that openly comprises following elaboration.

Claims

1. the method to signal sampling, described method comprises:

By wireless channel, receive the subscriber equipment transmitted signal of the sparse combination of S based on vector set;

Received transmitted signal is carried out to down-conversion and discretization, to create discrete signal;

Described discrete signal is relevant to sense waveform collection, to create sampling set, wherein, the sum of the sampling in described sampling set equals the sum of the concentrated sense waveform of described sense waveform, wherein, described sense waveform collection does not mate with described vector set, and the sum of the concentrated sense waveform of wherein said sense waveform is less than the vectorial sum in described vector set; And

To remote central processor, send at least one sampling in described sampling set.

2. method according to claim 1, wherein, described vector set comprises at least one set of selecting in the list from consisting of the following:

The row of basic matrix and the row of basic matrix.

3. method according to claim 1, also comprises:

In response to receiving instruction from described remote central processor, select new sense waveform collection.

4. method according to claim 1, also comprises:

In response to receiving instruction from described remote central processor, adjust timing reference.

5. method according to claim 1, also comprises:

In response to receiving instruction from described remote central processor, change the number of sense waveform.

6. method according to claim 1, wherein, described vector set and described sense waveform collection have and are less than or equal to the 0.45 subduplicate coherent value that is multiplied by the vectorial dimension in described vector set.

7. a method for communication, described method comprises:

At central processing unit place, receive sampling set;

Use described sampling set, basic matrix and sense waveform collection, form and there is the minimized linear programming of L1;

Solve described linear programming, to produce data estimator collection; And

Quantize described data estimator collection, to produce information symbol collection.

8. method according to claim 7, also comprises:

Select the first subset of described data estimator collection, wherein, the sum of the element in described the first subset means parameter, and all elements of described the first subset is greater than the surplus element of described data estimator collection arithmetically; And

Identify the index of the element of described the first subset.

9. method according to claim 8, wherein, described information symbol collection forms binary vector, the length of described binary vector equals the length of described data estimator collection, and described binary vector consists of logical one and the logical zero at rest position place of each corresponding position of one of index with identified.

10. method according to claim 8, also comprises:

Equipment for indicating user is used described expression parameter.