CN104683987A - Transmission scheduling method for scalable video coding (SVC) video in cognitive radio multiple channels - Google Patents

Abstract

The invention provides a transmission scheduling method for a scalable video coding (SVC) video in cognitive radio multiple channels. The method provides a flexible time slot based dynamic spectrum access scheme taking activity routines of a master user into account, thereby reducing unnecessary channel detection and further improving the use ratio of wireless channel bandwidth; and the invention also provides a grade matching strategy for the SVC video and the cognitive radio multiple channels. According to the grade matching strategy, through reasonable allocation of priority of the SVC video data, and arrangement of transmission order of the SVC video data and dynamic adaptation of the SVC video data and the cognitive radio multiple channels, more important SVC data can be sent by highly reliable wireless channel resources in a long time range, and less important SVC video data can be sent by low-reliability channel resources. Due to optimization of the two aspects, more video data are transmitted successfully, the effectiveness of the received video data is high, and finally, the quality of the received video is optimized.

Description

The transmission dispatching method of SVC video in cognitive radio multichannel

Technical field

The invention belongs to Video coding and transmission field, be specifically related to the transmission dispatching method of a kind of SVC video in cognitive radio multichannel.

Background technology

Video Applications in wireless environment is more and more subject to the favor of user, but this also brings problems simultaneously.First, we know that a video traffic in most cases can take more channel resource than common audio frequency and text services, the wireless terminal emerged in multitude also makes the total amount of the video flow proportion in wireless environment and video flowing increasing to the request of video traffic, and therefore radio spectrum resources also just seems relative scarcity.Secondly, wireless channel is comparatively large by the impact of surrounding environment, and radio link quality has larger time-varying characteristics, so make video flowing preferably adaptive this time the wireless channel that becomes be a very large challenge.Again, wireless terminal is of a great variety, and its resolution, disposal ability, power reservoir capacity all possibility gap are comparatively large, and single video source is often difficult to directly adaptive various different wireless terminal.Problems also promote the development of transmission of video correlative study in wireless environment.

Cognitive radio (Cognitive Radio, CR) is a kind of by frequency spectrum perception and system intelligence study, realizes the dynamic assignment of frequency spectrum and the technology of frequency spectrum share.It is not affecting the idle frequency spectrum these frequency spectrums of Appropriate application that the basis of authorizing frequency range proper communication find wherein.CR improves the low present situation of frequency spectrum resource utilization rate that the fixed allocation policy because of frequency spectrum causes to a certain extent, alleviates the situation of radio spectrum resources relative scarcity.Have in CR network primary user and secondary user's point, secondary user's detects the state of multiple channel, and is only occurring that channel idle could access channel and transmit data.The dynamic spectrum access scheme adopting fixing time slot in existing research more, namely time slot comprises a fixing detection duration and a fixing transmission duration, and only have when judging that channel is the free time according to result of detection, can channel be accessed and transmit data.The not concrete mechanics considering the primary user that channel is corresponding of access scheme of this fixing time slot, therefore likely causes many unnecessary frequent detections, this reduces the utilance of channel width.Some researchs are considered to reduce unnecessary channel detection by self-adaptative adjustment detection duration and transmission duration, thus improve the utilance of channel width.Mostly these researchs are to be optimized for single channel, and do not consider the mutual coordination between multiple channel, and the detection period therefore in each channel will have no rule with the arrangement of the separation of transmission period.When a secondary user's monitors multiple channel simultaneously, secondary user's needs to judge channel status in detection period and transmission period separation place, and whether decision-making accesses, sends which type of data.This irregular decision point arrangement is difficult to arrive optimum by making the data of secondary user's assign decision-making.

" opportunism " access way that this wait idle channel of secondary user's occurs makes the available bandwidth of secondary user's change greatly.The unsteadiness of bandwidth has increased the weight of the problem two in first paragraph to a certain extent, and the multichannel namely in the better adaptive CR network of the video traffic of packet loss sensitivity will be more difficult.Telescopic video more can adapt to the CR network multiple-channel scene of this network condition instability compared with other non-telescopic video.Adopt the Scalable Video Coding(SVC of layered encoding structure) in time, space and quality three dimensions, all there is scalability, a video frequency source coding can be become a SVC video flowing comprising multiple subflow by it, wherein each subflow can reconstruct a source video, just in the resolution of time, space or quality, has difference.The telescopic nature of SVC makes the video anti-packet loss ability of encoding give prominence to, even if network condition poor and in the wireless network of instability still heat energy access video quality relatively preferably.

More existing researchs optimize the transmission of telescopic video in cognitive radio networks further, but these research work just apply the Inherent advantage of telescopic video, and the flexible configuration excavating telescopic video is not further had to improve video to the effectiveness in the adaptability of channel.

Summary of the invention

The object of the invention is to, for overcoming the problems referred to above, improving channel bandwidth utilization ratio by the dynamic spectrum access scheme optimized further in cognitive radio technology, this is to promoting that the development of the wireless video business that bandwidth consumption is larger has larger benefit; Have employed on the other hand and have better adaptive SVC video to channel, go forward side by side one-step optimization SVC video pair time become, the multi channel adaptability of cognitive radio that reliability is different, the quality of final optimization pass receiving terminal video.

For achieving the above object, the invention provides the transmission dispatching method of a kind of SVC video in cognitive radio multichannel environment, described method comprises: be first the flexible factor of channel definition of certain secondary user's detection, this flexible factor is for describing the slot length of channel; Then, the flexible factor of optimum of each channel detected is obtained based on flexible Summing Factor primary user mechanics; Finally, judge that the channel that user can access and user access the length of the transmission time slot of certain channel according to the optimum factor flexibly.

Optionally, above-mentioned transmission dispatching method comprises following steps:

Step 101) secondary user's is at T _sdetect free time or the busy condition of N number of wireless channel in time period, the slot length then defining each channel in N number of channel is θ times of a fixing slot length, and described θ is the flexible factor;

Step 102)

Set up the mechanics model of primary user, obtain primary user according to model within certain period of time period, keep the probability of idle condition or busy condition, and obtain primary user and be converted to the free time by busy or be converted to busy probability by the free time within this period;

Keep the probability of idle condition or busy condition based on the flexible factor, primary user, primary user is converted to the free time by busy and changed into busy probability by the free time, obtain the collision probability of secondary user's and primary user;

Collision probability according to secondary user's and primary user be no more than primary user the principle of patient maximum crash probability, calculate the flexible factor of optimum of each channel in N number of channel

Step 103) flexible for the optimum factor is more than or equal to 1 channel as the available channel of secondary user's, and the number of all available channels is G, and wherein G is less than or equal to N;

Step 104) secondary user's is accessed available channel, adopt G channel to send the SVC video data of secondary user's, return step 101), until all data of secondary user's are sent completely;

Wherein, the duration that the channel n in secondary user's access G idle channel carries out SVC video data transmission is described fixing slot length is " T _s+ T _t", T _tit is the time period of a fixed size.

In such scheme, the computing formula of the flexible factor of optimum of any one channel n is as follows:

Wherein, for the probability that secondary user's and primary user collide at channel n, by the patient maximum collision probability of primary user on channel n, described channel n is certain channel in N number of channel.

Optionally, above-mentioned computing formula be:

$P_c^n\left(\mathrm{\θ}\right)=1-P_I^n\left(S^E\right)\·e^{{-\mathrm{\α}}_n\·(\mathrm{\θ}\·(T_t+T_s)-T_s)}$

Wherein, represent that channel n is at S ^ethe probability of free time in moment, and S ^efor the detection period T of secondary user's sounding channel _sthe ending moment.

Optionally, obtain step be:

Step 201) obtain the step of primary user mechanics model, be specially:

The mechanics of primary user obeys Markov model continuous time, namely Markov characteristic is obeyed in the transfer between free time of primary user and busy condition, the duration of " busy and idle " two states of primary user obeys following quantum condition entropy respectively simultaneously, obtains following quantum condition entropy and namely obtains primary user's mechanics model:

$f_I\left(t\right)={\mathrm{\α}}_n{e}^{-{\mathrm{\α}}_{n}t}$

$f_B\left(t\right)={\mathrm{\β}}_n{e}^{-{\mathrm{\β}}_{n}t}$

Above-mentioned f _it () represents the value of probability density function in t of the primary user's free time in channel n, α _nvalue obtain according to the historical data of primary user's activity in channel n, this value equals primary user in channel n and is in the inverse of the desired value of the duration of idle period;

F _bt () represents the value of the busy probability density function of the primary user in channel n in t, β _nvalue equal to obtain according to the historical data of primary user's activity in channel n, this value equals primary user in channel n and is in the inverse of the desired value of the duration of busy period;

Step 202) according to the mechanics model of primary user obtained, according to following formula predictions at (0, ε) in the time period, the free time of channel n shared by primary user or busy probability:

$F_I\left(\mathrm{\ϵ}\right)=1-{\∫}_{t=0}^{\mathrm{\ϵ}}{f}_I\left(t\right)\mathrm{dt}$

$F_B\left(\mathrm{\ϵ}\right)=1-{\∫}_{t=0}^{\mathrm{\ϵ}}{f}_B\left(t\right)\mathrm{dt}$

Wherein, F _i(ε) and F _b(ε) primary user is respectively at the probability keeping idle condition and busy condition from the t=0 moment to the time period in t=ε moment; " 1-F _i(ε) " for primary user is transformed to busy probability by the free time from the t=0 moment to the time period in t=ε moment; 1-F _b(ε) for primary user is transformed to idle probability by busy from the t=0 moment to the time period in t=ε moment;

Step 203) definition the actual free time of channel n when the detection finish time SE of secondary user's probability be:

$\begin{array}{c}{P}_I^n\left(S^E\right)=\frac{P(\mathrm{RS}=X,\mathrm{AS}=0)}{P(\mathrm{RS}=X)},X=0\mathrm{or}1\\ =\frac{P(\mathrm{AS}=0)\·\left({X+(-1)}^X(1-{\mathrm{\γ}}_n)\right)}{P(\mathrm{AS}=0)\·\left({X+(-1)}^X(1-{\mathrm{\γ}}_n)\right)+P(\mathrm{AS}=1)\·\left({X+(-1)}^X{\mathrm{\δ}}_n\right)}\end{array}$

Wherein, secondary user's to the detection period of channel n from S ^bmoment, to S ^emoment terminates; RS is that secondary user's completes the moment in each detection to channel n, i.e. S ^ein the moment, obtain a result of detection, and the result of RS=0 representative detection is idle condition, the result of RS=1 representative detection is busy condition; γ _nfor the probability that false-alarm occurs, be namely idle condition, i.e. AS=0 when channel is actual, but result of detection is busy, i.e. RS=1; δ _nfor the probability that false dismissal occurs, be namely busy condition, i.e. AS=1 when channel is actual, but result of detection is idle condition, i.e. RS=0;

Step 204) according to F _i(ε), F _b(ε), 1-F _i(ε) and 1-F _b(ε) P(AS=0 is obtained) and value P(AS=1), be specially:

$P(\mathrm{AS}=0)=P_I^n\left(S^B\right)\·F_{I\→I}^{T_s}+(1-P_I^n\left(S^B\right))\·F_{B\→I}^{T_s}$

$P(\mathrm{AS}=1)=P_I^n\left(S^B\right)\·F_{I\→B}^{T_s}+(1-P_I^n\left(S^B\right))\·F_{B\→B}^{T_s}$

Wherein, for channel n is at S ^bmoment actual idle probability, this probable value relies on S ^bthe state of the upper non-slotted channel n before the moment; When at S ^bsecondary user's success before transmits data in channel n, then think that the probability of channel n free time is when at S ^bsecondary user's and primary user think after colliding that the probability of channel n free time is before when secondary user's does not transmit data in the channel, then can calculate according to the channel idle probability after the upper time slot detection period , and prediction equation as follows:

$P_I^n\left(S^B\right)=P_I^n\left(S_p^E\right)\·F_{I\→I}^{T_t}+(1-P_I^n\left(S_p^E\right))\·F_{B\→I}^{T_t}$

represented the upper detection time slot finish time of channel n in secondary user's time the free time probability;

represent the state of channel n respectively at transmission duration T _tinteriorly keep idle probability and transform to idle probable value from busy.

Optionally, step 104) described in the step of SVC video data of transmission secondary user's comprise further:

Step 104-1) SVC video data arranged priority and starting point and the cut-off point of each frame are set, the NALU being positioned at starting point and cut-off point scope is effective, and is placed on according to priority by effective NALU and sends in buffering area;

Step 104-2) each NALU fractionation that will send in buffer area encapsulates, and the transmission of packet fractionation encapsulated carries out dynamic adaptation with G the interchannel of free time, thus the more reliable channel of packet employing channel performance getting over high priority is sent.

Optionally, above-mentioned steps 104-1) comprise further:

Step 104-1-1) will the encode SVC video data that obtains of GOP be adopted to carry out following principle and carry out priority division:

The priority setting the frame that time stratum level is lower in a GOP is higher, and in the identical frame of multiple time stratum level, playing sequence frame priority is more early higher;

And meet:

The NALU priority that in a frame in a GOP, mass segregation lower grade is higher;

And meet:

The priority of all NALU in a certain mass segregation grade is in, the priority of overall all NALU higher than being positioned on more high-quality stratum level in a GOP, and lower than the priority of all NALU be positioned on more low quality stratum level;

Step 104-1-2) according to the sending order of all frames in priority level initializing GOP of definition, and be that in this GOP, each frame distributes a starting point and a cut-off point, when being between the starting point of certain frame and cut-off point sometime, all NALU then in this frame are effective, and are sorted in transmission buffering area according to the priority of definition by these effective NALU;

Wherein,

When cut-off point arrives, be arranged in buffering area but the NALU data be not yet sent out will be lost efficacy and abandon, and start the effective NALU in next GOP to put into buffering area;

The setting principle of above-mentioned starting point and cut-off point is: the time difference between adjacent two frame starting points equals time interval during two frame normal play; And all frames share same cut-off point in a GOP, described shared cut-off point is the starting point of the first frame comprised in next GOP.

Optionally, above-mentioned steps 104-2) comprise further:

Step 104-2-1) be C={c when detecting by this secondary user's the set that the available channel that obtains forms when certain decision-making ₁, c ₂..., c _g..., c _g, and be arranged in the decision-making moment that to send the set expression of buffering area NALU be Φ={ φ ₁, φ ₂..., φ _k..., φ _k;

Wherein, φ _k< φ _k+1, namely index is φ _kthe priority of NALU to be greater than index be φ _k+1the priority of NALU;

Step 104-2-2) descending of NALU priority in buffering area is sent according to above-mentioned detection moment, maximize each element φ in the set of NALU successively _ksuccessfully accepted and the probability of decoding, the transmission finally obtaining the optimum sending video source data arranges vector;

Wherein, mate being arranged in the NALU sending buffering area with the channel of available channel set C according to optimal vector.

Optionally, above-mentioned steps 104-2-2) optimum transmission arranges vector to adopt following formula to determine

${\stackrel{\&RightArrow;}{A}}_{\mathrm{opt}}=\{A_{{\mathrm{\&phi;}}_1-\mathrm{opt}},A_{{\mathrm{\&phi;}}_2-\mathrm{opt}},...,A_{{\mathrm{\&phi;}}_{K\&prime;}-\mathrm{opt}}\},K^{\&prime;}\&le;K$

Wherein,

Above-mentioned expression index is φ _knALU be successfully received and the probability that can decode, represent bag according to packet loss of link probability when arranging, representative bag according to collision drop probabilities when arranging, represent index to be φ _knALU carry out splitting the index of each bag encapsulating and obtain, the span of k is 1≤k≤K '.

Optionally, greedy algorithm is adopted to obtain optimal vector value, be specially:

When set of available channels is C={c ₁, c ₂..., c _g..., c _g, send the set Φ={ φ of NALU in buffering area ₁, φ ₂..., φ _k..., φ _ktime:

Step 301) by optimal vector be initialized as: wherein, represent empty set;

Step 302) each NALU sent in the set Φ of effective NALU in buffering area is split and package;

Step 303) judge to send in buffering area whether have the bag be not sent out, if enter step 304); Otherwise enter step 306);

Step 304) judge whether there is available channel resource in set C, if had, enter step 305); Otherwise enter step 306);

Step 305) send buffering area to take out from data and come the most front bag (priority is the highest), be assumed to be bag calculate it respectively and remain Successful transmissions probability when available channel transmits in set C;

Selection can provide the channel of maximum Successful transmissions probability and concrete package location as the transmission arrangement of this bag, as position (c _g, q);

To arrange join set in, be in an element;

Upgrade in set C and remain concrete remaining available volume of resources information in available channel information and channel; Upgrade the residue packet information to be sent sent in the Φ of buffering area; Return step 303);

Step 306) obtain the best fit strategy of SVC data and Wireless Multi-Channel

Compared with prior art, technical advantage of the present invention is:

First, the present invention with the mechanics of primary user for foundation, a kind of dynamic spectrum access scheme of Flexible timeslot is proposed, this scheme is not still in order to improve the utilance of channel width, and the coordination considered between CR multichannel, improve the validity of secondary user's distribute data in CR multichannel.Secondly, the present invention, based on the dynamic spectrum access scheme of Flexible timeslot, devises a SVC video and the multi channel graduation matching strategy of CR.The scheme of carrying out data transmission arrangement with priority orders that one is considered data age is comprised in this strategy, the available channel that multiple reliability differs larger may be had when considering each decision-making, the program makes the importance of the data sent in buffering area during decision-making also have larger gap, this just makes more important data in for a long time can carry out transmission by more reliable channel becomes possibility, and the quality receiving video also can be optimized further.

Accompanying drawing explanation

Fig. 1 is the dynamic spectrum access scheme exemplary plot of the fixing time slot of prior art; In this figure, Sense represents the detection period, and IDLE represents idle condition, and BUSY represents busy condition;

Fig. 2 is the rough schematic that the transmitting and scheduling of SVC video provided by the invention in cognitive radio multichannel environment performs flow process;

Fig. 3 is that SVC video provided by the invention mates schematic flow sheet with available cognitive radio is multi channel;

Fig. 4 is the dynamic spectrum access scheme exemplary plot of Flexible timeslot of the present invention; In this figure, Sense represents the detection period, and IDLE represents idle condition, and BUSY represents busy condition;

Fig. 5 is SVC data priority allocation example figure of the present invention;

Fig. 6 (a) and 6(b) be coded sequence and the sending order example of SVC frame of video of the present invention;

Fig. 7 is that the priority orders data of the consideration data age of the embodiment of the present invention send arrangement scheme exemplary plot.

Embodiment

Below in conjunction with drawings and Examples, the method for the invention is described in detail.

One, the dynamic spectrum access scheme based on Flexible timeslot provided by the invention is as described below:

The transmission dispatching method of a kind of SVC video provided by the invention in cognitive radio multichannel environment, described method comprises:

Step 101) secondary user's is at T _sdetect free time or the busy condition of N number of wireless channel in time period, the slot length then defining each channel in N number of channel is θ times of a fixing slot length, and described θ is the flexible factor;

Step 102)

Set up the mechanics model of primary user, obtain primary user according to model within certain time period, keep the probability of idle condition or busy condition, and obtain primary user and be converted to the free time by busy or changed into busy probability by the free time within this period;

Keep the probability of idle condition or busy condition based on the flexible factor, primary user, primary user is converted to the free time by busy and changed into busy probability by the free time, obtain the collision probability of secondary user's and primary user;

Collision probability according to secondary user's and primary user be no more than primary user the principle of patient maximum crash probability, calculate the flexible factor of optimum of each channel in N number of channel;

Step 103) flexible for the optimum factor is more than or equal to 1 channel as the available channel of secondary user's, and the number of all available channels is G, and wherein G is less than or equal to N;

Step 104) secondary user's is accessed available channel, adopt G channel to send the SVC video data of secondary user's, return step 101), until all data of secondary user's are sent completely;

Wherein, the duration that the channel n in secondary user's access G idle channel carries out SVC video data transmission is described fixing slot length is " T _s+ T _t".

Logical relation for above-mentioned 4 steps can with reference to shown in Fig. 2.

The principle of above steps is as described below with all formula related to:

May there is multiple channel in cognitive radio networks, each channel has the primary user of its correspondence, and primary user has access to priority to channel, and namely only have when primary user does not need transmission data, secondary user's could access channel and transmit data.And from secondary user's one side, only have by primary user can be understood to the detection of channel whether in transmission data, and then could judge channel whether the free time, can access and transmit data.Therefore secondary user's just relies on the mechanics of primary user to the detecting strategy of channel, secondly also needs to design the secondary user's access scheme coordinating multiple channel.This part will divide three these thoughts of explanation of itemizing.

1.1, primary user's mechanics modeling---continuous time Markov model

Such as a secondary user's monitors 8 channels, channel n(n=1,2, ..., 8) corresponding primary user has two kinds of discrete state BUSY/IDLE, corresponding primary user's busy condition (the busy BUSY of channel) of BUSY, corresponding primary user's idle condition (channel idle IDLE) of IDLE.The mechanics of primary user obeys Markov model continuous time, namely Markov characteristic is obeyed in the transfer between primary user's two states, simultaneously the duration of two states BUSY/IDLE all obeys quantum condition entropy, but its respective parameter is not necessarily identical.Primary user P _n(n=1,2 ..., 8) mean down time be the average busy time is suppose in this embodiment all equal 2 seconds, the probability density function of corresponding free time length is parameter is α _nexponential function, its expression formula is as shown in Equation (1); The probability density function of busy duration is parameter is β _nexponential function, its expression formula as shown in Equation (2),

$f_I\left(t\right)={\mathrm{\&alpha;}}_n{e}^{-{\mathrm{\&alpha;}}_{n}t},---\left(1\right)$

$f_B\left(t\right)={\mathrm{\&beta;}}_n{e}^{-{\mathrm{\&beta;}}_{n}t}.---\left(2\right)$

According to the memoryless characteristic of exponential function, primary user P _nwithin at least ε time from some moment (t=0), keep the probability of its state ID LE or BUSY to be respectively F _i(ε) and F _b(ε), it calculates as shown in formula (3) and (4).According to F _i(ε) and F _b(ε) also and then can calculating primary user Pn state, to be transformed to busy probability by the free time be 1-F _i(ε) and by busy to be transformed to idle probability be 1-F _b(ε).According to above primary user's mechanics model, the activity situation of primary user's a period of time just can be predicted, and then free time of channel n or busy situation also just can predict.

$F_I\left(\mathrm{\&epsiv;}\right)=1-{\&Integral;}_{t=0}^{\mathrm{\&epsiv;}}{f}_I\left(t\right)\mathrm{dt},---\left(3\right)$

$F_B\left(\mathrm{\&epsiv;}\right)=1-{\&Integral;}_{t=0}^{\mathrm{\&epsiv;}}{f}_B\left(t\right)\mathrm{dt}.---\left(4\right)$

2.1, the access scheme of secondary user's

The secondary user's access scheme of prior art is the dynamic spectrum access scheme of fixing time slot, and namely each channel is divided into continuous print time slot in time-domain, and the length of each time slot is fixed, and each time slot comprises the detection period T of fixing duration _swith the transmission period T of fixing duration _t, in this embodiment, set T _s=10ms, T _t=50ms.Secondary user's is from the starting position sounding channel of detection period, if judge that channel is idle condition in ending place of detection period, then secondary user's can access channel and transmit data, otherwise secondary user's can not access channel.In cognitive radio networks, the detection period of multiple channel is also thought synchronous with section boundary during transmission.Fig. 1 describes the implication of the dynamic spectrum access scheme of fixing time slot with an example.

In order to improve the utilance of channel width further, according to the mechanics of primary user in this invention, devise the dynamic spectrum access scheme of Flexible timeslot.In the program, each channel is still divided into continuous print time slot in time-domain, but the length of each time slot may be different: detection duration still remains T _s, but transmission duration can the activity situation current according to primary user adjust, and the essence of this design is the reduction of unnecessary frequent channel detection.Simultaneously detect period and section boundary during transmission in each channel to coordinate, specify in this design the length of a Flexible timeslot should be the θ of a fixing slot length doubly (θ=1,2 ...), wherein θ is called as the flexible factor.

Fig. 4 describes the implication of the dynamic spectrum access scheme of Flexible timeslot with an example, wherein in channel 1,2,3, the flexible factor of the time slot that dashed rectangle is surrounded is respectively 2,2,3.Although the detection period in Flexible timeslot structure in not all channel and section boundary all Complete Synchronizations during transmission can be seen from example, but section boundary or Partial synchronization when detection period and transmission, this is that the effective distribution of data in multichannel provides condition.

3.1, the decision-making of optimum factor θ flexibly

Channel n(n=1,2 ..., 8) the detection period from S ^bmoment, to S ^emoment terminates.Secondary user's after each detection to channel n completes, i.e. S ^ein the moment, all can obtain a result of detection RS, the fruit of RS=0 representative detection is IDLE, channel idle; The result of RS=1 representative detection is BUSY, and channel is busy.But the factors such as the instability of signal strength signal intensity, other subscriber signal interference can cause the inaccurate of result of detection, be therefore IDLE state when channel is actual, i.e. AS=0, but result of detection is BUSY, i.e. RS=1, we are referred to as false-alarm, the probability γ that false-alarm occurs _n; Be BUSY state, i.e. AS=1 when channel is actual, but result of detection is IDLE, i.e. RS=0, we are referred to as false dismissal, and the probability that false dismissal occurs is δ _n.

Considering the possibility of false-alarm and false dismissal, and when channel history information, channel n is at S ^emoment, actual idle probability was set to , it calculates as shown in Equation (5), the computing reference formula (6) of (5) and (7),

$\begin{array}{c}{P}_I^n\left(S^E\right)=\frac{P(\mathrm{RS}=X,\mathrm{AS}=0)}{P(\mathrm{RS}=X)},X=0\mathrm{or}1\\ =\frac{P(\mathrm{AS}=0)\&CenterDot;\left({X+(-1)}^X(1-{\mathrm{\&gamma;}}_n)\right)}{P(\mathrm{AS}=0)\&CenterDot;\left({X+(-1)}^X(1-{\mathrm{\&gamma;}}_n)\right)+P(\mathrm{AS}=1)\&CenterDot;\left({X+(-1)}^X{\mathrm{\&delta;}}_n\right)}\end{array},---\left(5\right)$

$P(\mathrm{AS}=0)=P_I^n\left(S^B\right)\&CenterDot;F_{I\&RightArrow;I}^{T_s}+(1-P_I^n\left(S^B\right))\&CenterDot;F_{B\&RightArrow;I}^{T_s}---\left(6\right),$

$P(\mathrm{AS}=1)=P_I^n\left(S^B\right)\&CenterDot;F_{I\&RightArrow;B}^{T_s}+(1-P_I^n\left(S^B\right))\&CenterDot;F_{B\&RightArrow;B}^{T_s}---\left(7\right).$

In above formula the state representing channel n respectively keeps the probability of IDLE, transforms to the probability of IDLE from BUSY, transforms to the probability of BUSY from IDLE and keeps the probability of BUSY within the detection period.Each detection period T _sin or transmission period T _tthe conversion of interior channel status occurs once at most, this is because T _sand T _tin practice much smaller than the duration of a certain state of primary user, therefore according to formula (3) and (4), can obtain being calculated as follows of each value,

$F_{I\&RightArrow;I}^{T_s}=e^{-{\mathrm{\&alpha;}}_n{T}_s},---\left(8\right)$

$F_{B\&RightArrow;I}^{T_s}=1-e^{-{\mathrm{\&beta;}}_n{T}_s},---\left(9\right)$

$F_{I\&RightArrow;B}^{T_s}=1-e^{-{\mathrm{\&alpha;}}_n{T}_s},---\left(10\right)$

$F_{B\&RightArrow;B}^{T_s}=e^{-{\mathrm{\&beta;}}_n{T}_s}.---\left(11\right)$

In formula (6) and (7) for channel n is at S ^bmoment actual idle probability, this probable value relies on S ^bthe state of the upper non-slotted channel n before the moment.When at S ^bsecondary user's success before transmits data in channel n, then think that the probability of channel n free time is ; When at S ^bsecondary user's and primary user think after colliding that the probability of channel n free time is before ; When secondary user's does not transmit data in the channel, then can calculate according to the channel idle probability after the upper time slot detection period , it calculates as shown in Equation (12),

$P_I^n\left(S^B\right)=P_I^n\left(S_p^E\right)\&CenterDot;F_{I\&RightArrow;I}^{T_t}+(1-P_I^n\left(S_p^E\right))\&CenterDot;F_{B\&RightArrow;I}^{T_t}.---\left(12\right)$

In above formula represented channel n in upper time slot detection finish time period time the free time probability; with represent the state of channel n respectively at transmission duration T _tthe probability of interior maintenance IDLE and transform to the probability of IDLE from BUSY, therefore according to formula (3) and (4), can obtain being calculated as follows of two values,

$F_{I\&RightArrow;I}^{T_t}=e^{-{\mathrm{\&alpha;}}_n{T}_t},---\left(13\right)$

$F_{B\&RightArrow;I}^{T_t}=1-e^{-{\mathrm{\&beta;}}_n{T}_t}.---\left(14\right)$

According to above derivation, S can be obtained ^ethe probability of channel n free time in moment .On this basis, the flexible factor arranging this time slot is θ, and namely this Flexible timeslot is θ times of fixing slot length, so from S ^emoment start to the time at this Flexible timeslot end be θ (T _s+ T _t)-T _s.If within this period, channel status is converted to BUSY by IDLE, then there occurs the collision of primary user and secondary user's, the probability stamps of collision is , can calculate according to formula (3) and (4) it is as follows,

$P_c^n\left(\mathrm{\&theta;}\right)=1-P_I^n\left(S^E\right)\&CenterDot;e^{{-\mathrm{\&alpha;}}_n\&CenterDot;(\mathrm{\&theta;}\&CenterDot;(T_t+T_s)-T_s)}.---\left(15\right)$

The primary user that known channel n is corresponding patient maximum collision probability be then should make when decision-making θ be no more than so the flexible factor of optimum that channel n adopted in the decision-making moment calculating as shown in Equation (16), wherein round symbol downwards.

To channel n, if calculated then mean channel n probably at ensuing transmission period T _tin be transformed to busy condition by idle condition, then secondary user's should not access channel; If then mean that the state of channel n is at ensuing transmission period T _tin to be transformed to the probability of busy condition by the free time less, secondary user's can access channel and transmit data, especially when time, the advantage of the dynamic spectrum access scheme of Flexible timeslot is embodied, and the utilance of channel width is promoted.

Two, the present invention is directed to when sending SVC data and additionally provide based on SVC video and the multi channel graduation matching strategy of cognitive radio, namely optimize above-mentioned steps 104 further) method of SVC video data of transmission secondary user's, the method specifically comprises:

Step 104-1) SVC video data arranged priority and starting point and the cut-off point of each frame are set, the NALU being positioned at starting point and cut-off point scope is effective, and is placed on according to priority by effective NALU and sends in buffering area;

Step 104-2) each NALU fractionation that will send in buffer area encapsulates, and the transmission of packet fractionation encapsulated carries out dynamic adaptation with G the interchannel of free time, thus the more reliable channel of packet employing channel performance getting over high priority is sent.

Logical relation for above-mentioned two each and every one steps can with reference to shown in Fig. 3, and above-mentioned steps 104-1) and being described in detail as follows of concrete principle 104-2) and enforcement:

2, SVC video and the multi channel graduation matching strategy of CR

Relevant portion of the present invention gives multi channel model in CR network, and to be also this part to classify the basis of mating about SVC video and CR multichannel this model.The classify strategy that mates of SVC video and CR multichannel specifically includes three and itemizes: the distribution of SVC video data priority, consider that the priority orders data of data age send the matching algorithm of arrangement scheme, data and channel, and these three subitems interknit and constitute the graduation matching strategy of this part.

2.1, the distribution of SVC video data priority

Prior art adopts the SVC of layered encoding structure to have scalability in three dimensions: quality dimensions, time dimension and Spatial Dimension.SVC can go out video containing multiple quality layers, multiple time horizon and multiple space layer according to configuration codes, each specific layer corresponding specific video definition, frame per second and spatial resolution.When practical application user can according to oneself need extract some specific quality layers, time horizon and space layer in SVC video, thus obtain the compressed video of a certain specific definition, frame per second and spatial resolution.

In the present invention, SVC adopts GOP(Group of Pictures) structured coding video, quality scalable dimension adopts MGS(Medium Grain Scalability) coding structure, time, scalable dimension adopted Hierarchical-B coding structure, and spatial scalability is not used.4 mass segregation (F are had in the SVC video data be encoded in this embodiment ₀, F ₁, F ₂, F ₃) and 4 time layering (T ₀, T ₁, T ₂, T ₃), then each GOP will have 2^ (4-1)=8 frame, 8 scales being called as GOP, the data of each frame are encoded into 4 NALU(Network Abstraction Layer Unit by mass segregation again) in, so total 8*4=32 NALU, Fig. 5 are the data formations of this embodiment in each GOP.

In time scalable dimension, the time layering that the video data that user obtains comprises is more, then mean that the frame that video packets contains is more intensive, and video is decoded and video that is that obtain when playing will be more smooth.But the predictive coding adopted in SVC coding and layered encoding structure make the frame that in video, time stratum level is higher directly or indirectly rely on the lower frame of some time stratum level.If the frame that certain time stratum level that is that the frame that namely time stratum level is higher directly relies on or that indirectly rely on is lower does not exist, so this frame just can not be decoded smoothly.Therefore in this invention, the priority of the frame that setting-up time stratum level is lower is higher.Can there is the identical frame of multiple time stratum level in a GOP, in the frame that in this invention, setting-up time stratum level is identical, those playing sequences frame priority is more early higher simultaneously.

In quality scalable dimension, the effective mass segregation data comprised in a frame are more, then mean that decodable NALU is more, and the image of the decoded rear gained of this frame is more clear.According to SVC coding principle, the decoding dependency being in the NALU of certain mass segregation grade in a frame is in those NALU lower than this NALU of all mass segregation grades.Only have in time all being existed by those NALU relied on, a NALU just can be decoded smoothly, and is considered to effective.Therefore the NALU priority that in a frame, mass segregation lower grade is set in this invention higher.

When resource-constrained, user is more prone on the basis ensureing video fluency, just goes to obtain the definition that more quality enhancement layer data removes to increase video further.Therefore quality scalable dimension and time scalable dimension is considered, this invention sets the priority being in the NALU on equal in quality stratum level in those different frames, the priority of overall NALU higher than those mass segregation higher grades, the priority of overall NALU lower than those mass segregation lower grades.

Comprehensive above analysis, 32 NALU comprised in the GOP in this embodiment will be assigned with unique corresponding priority, and the index of priority is 1,2 ..., 32, wherein index is less means that the priority of NALU is higher.Fig. 5 is this embodiment video data priority distribution diagram.

2, consider that the priority orders data of data age send arrangement scheme

Frame of video is different by the order of encoding from the playing sequence of frame of video, and in Fig. 6, (a) is the coded sequence of frame of video in this embodiment; The order that frame of video is sent out in the invention simultaneously had both been different from the playing sequence of frame, was different from again frame by the order of encoding.The sending order of the frame rank set in this invention is consistent with the priority of frame rank, and namely the priority of frame is higher, and this frame is more early considered to be sent out, and in Fig. 6, (b) is the sending order of frame of video in this embodiment.A frame is made up of multiple NALU, and this invention, based on the sending order of frame rank, a kind ofly at NALU level design considers that the ageing priority orders data of NALU send arrangement scheme.

According to the sending order of frame rank in a GOP, each frame is assigned with a starting point and a cut-off point, only have when the current time is between the starting point of this frame and cut-off point time, just think that this frame be multiple NALU in effective, this frame is effective, these NALU should be placed into and send in buffering area with to be sent.When cut-off point arrives, inefficacy is all dropped by the data be not yet sent out in buffering area, and starts the effective NALU in next GOP to be placed in buffering area.The setting rule of starting point and cut-off point is: the time difference between adjacent two frame starting points equals time interval during two frame normal play; In a GOP, all frames share same cut-off point, and this cut-off point is also the starting point of next GOP first frame.Therefore at a time, send in buffering area may have the NALU being still in the term of validity in multiple frame wait to be sent, these NALU arrange according to priority in transmission buffering area, the NALU that priority is higher more early considers to be sent out, and namely this be consider that the priority orders data of data age send the implication of arrangement scheme.

Fig. 7 is with this embodiment for background explanation such scheme, based on the sending order of this figure frame rank in Fig. 6 (b), and the NALU in (1,9,17,25) representative frame 1, the NALU in (2,10,18,26) representative frame 2, the like; The implication of starting point and cut-off point also can be found out by Fig. 7 example.At the time point of red arrow logo, frame 1, frame 2 and frame 3 are all in the term of validity, their NALU should be placed to and sends in buffering area and send according to the priority orders of NALU, data in figure mean index be 1,9,2 and 10 NALU be successfully sent, and the NALU that index is 3,11,17,18,19,25,26 and 27 still stay successively send buffering area medium to be sent.

Consider that the priority orders data of data age send the Advantages found of arrangement scheme in two.1) in CR network multiple-channel environment, secondary user's may have multiple idle channel to use in some decision-making moment simultaneously, and multiple available channel may have larger gap in the reliability of transmission data.Adopt the priority orders data of consideration data age as above to send arrangement scheme, the priority limit of the medium NALU to be sent in transmission buffering area can be made comparatively large, and namely the NALU of higher priority, high medium priority and lower priority exists.Under such arrangement scheme, the data that will be able to make referrals to next subitem and the matching algorithm of channel, make the higher NALU of importance in the long term time frames channel resource that often application reliability is higher transmit; The NALU channel resource that often application reliability is lower that importance is lower transmits.2) this delivery plan is based on the priority orders of frame rank, therefore the frame that importance is higher often can apply the transmission that channel resource arranges its NALU earlier, NALU in the frame that importance is lower then likely when cut-off point closes on because CR channel resource deficiency has little time to send, thus to be simply discarded.When resource-constrained, the program makes limited channel resource for the transmission of relatively important video data, simultaneously make again data more important in these data be transmitted be transmitted by more reliable channel resource, the relatively poor data of importance are transmitted by the channel resource that reliability is relatively poor.Which ultimately increases the validity of the video data received, the video data that can not decode received reduces, and the quality of video is improved.

3, SVC data and the multi channel matching algorithm of CR

In above-described embodiment, a secondary user's with multiple wave point detects 8 channels simultaneously, when certain decision-making can channel form set be C={c ₁, c ₂..., c _g..., c _g, wherein G is no more than the subset that 8, C is 8 channels; The set expression sending NALU in buffering area in the decision-making moment is Φ={ φ ₁, φ ₂..., φ _k..., φ _k, wherein K is no more than 32, and φ _k< φ _k+1, index is φ _kthe priority of NALU to be greater than index be φ _k+1the priority of NALU.

SVC video data is all finally send with data packet form.Index is φ _knALU length be carry out splitting the package bag number that can split into it and be wherein L _spufor tearing the length of packetized units open, L in this embodiment _spube set to 1400, so each bag that will dispatch has an index by index be bag being operating as when being arranged in CR channel by index be bag be arranged in channel c after the current decision moment _gin q package location on.Index is φ _knALU in the set expression of transmission arrangement of all bags be and then in channel width limited field, the transmission arrangement set of the individual NALU of K ' finally arranged constitutes a vector sVC data in this invention and the multi channel matching algorithm of CR are exactly to find an optimum vector

Packet may because radio link quality difference or there occurs with primary user and collide and lose in the process of transmission; And only have in a NALU after all subpackages all have received, just to think this NALU decodable code, index is φ _knALU to be successfully received and the probability that can decode can be expressed as sVC data in this invention and CR multichannel matching scheme maximize them successively according to the descending sending NALU priority in buffering area successfully to be accepted and the probability of decoding, and obtains with this and reach effect of optimization, as shown in formula (16) and (17).

${\stackrel{\&RightArrow;}{A}}_{\mathrm{opt}}=\{A_{{\mathrm{\&phi;}}_1-\mathrm{opt}},A_{{\mathrm{\&phi;}}_2-\mathrm{opt}},...,A_{{\mathrm{\&phi;}}_{K\&prime;}-\mathrm{opt}}\},K^{\&prime;}\&le;K---\left(16\right)$

$s.t.{\mathrm{Data}}_{c_g}\left(\stackrel{\&RightArrow;}{A}\right)\&le;R_{c_g},g=1,2,...,G$

if represent and adopt feasible solution time be assigned to channel c _gin data volume; represent channel c _gthe stock number that can provide, can pass through formula (18) and obtain, wherein represent channel c _gbandwidth, in this embodiment, the bandwidth range of each channel is 400Kbps-1Mbps, channel c _gthe flexible factor of optimum.

$R_{c_g}=B_{c_g}\&CenterDot;({\mathrm{\&theta;}}_{\mathrm{opt}}^{c_g}\&CenterDot;(T_s+T_t)-T_s)---\left(18\right)$

In formula (17) represent bag respectively in arrangement packet loss of link probability in situation and collision drop probabilities.Any bit in a packet makes a mistake because of channel link qualities, all thinks that this bag there occurs packet loss of link, the bit error rate of a channel and the SNR(signal-to-noise ratio of this channel, signal to noise ratio) and modulating mode be closely related, channel c _gbit error rate be identified as it calculates as shown in Equation (19), and wherein R represents this channel SNR, a _mand b _mbe empirical coefficient when adopting index to be the modulating mode of m, they can by 1 acquisition of tabling look-up.And then can according to channel c _gbit error rate derive bag packet loss of link probability as shown in Equation (20).The scope of the modulating mode adopted in this embodiment to be code efficiency that index m=2 is corresponding the be QPSK of 1/2, SNR is 6-50dB.

$P_{\mathrm{ber}}^{c_g,m}=\frac{a_m}{\exp (R*b_m)}---\left(19\right)$

Parameter list under form 1 different modulating pattern

AMC（m）

m=1

m=2

m=3

m=4

m=5

m=6

Modulation

BPSK

QPSK

QPSK

16-QAM

16-QAM

64-QAM

Coding Rate

1/2

1/2

3/4

9/16

3/4

3/4

a m

1.1369

0.3351

0.2197

0.2081

0.1936

0.1887

b m

7.5556

3.2543

1.5244

0.6250

0.3484

0.0871

To bag collision drop probabilities calculating be actually estimate before this bag is successfully received primary user resume channel c _gprobability, if namely resume channel and secondary user's collides some moment primary users, so schedule ahead in the channel, the bag that is not yet successfully received all will lose.Channel c _gbandwidth be bag be arranged at channel c _gin q package location, its transmission need time be and then can calculate to transmit data to from after decision-making and comprise bag in the time that interior q bag expection is successfully received needs be if at this t _qin time, channel status is converted to busy by the free time, and namely primary user has resumed channel, so wraps just there occurs collision packet loss, its probability calculation formula is as shown in (21).

The above is SVC data and the multi channel matching idea of CR, and the present invention also follows this thought and devises a greedy algorithm to obtain best fit strategy namely the present invention adopts greedy algorithm to realize SVC data to mate with CR is multi channel.The bag that when performing this algorithm in this embodiment, current priority is the highest is always arranged prior to other bag, and is be arranged to utilize the channel resource of the Successful transmissions probability that can provide the highest to send.By the solution that this greedy algorithm obtains consistent with the expection solution in SVC data above and the multi channel matching idea of CR; And adopt this greedy algorithm to avoid traversal to numerous feasible solutions, greatly reduce amount of calculation.

The greedy algorithm that the invention described above adopts is specially:

When set of available channels is C={c ₁, c ₂..., c _g..., c _g, send the set Φ={ φ of NALU in buffering area ₁, φ ₂..., φ _k..., φ _ktime:

Step 301) by optimal vector be initialized as: wherein, represent empty set;

Step 302) each NALU sent in the set Φ of effective NALU in buffering area is split and package;

Step 303) judge to send in buffering area whether have the bag be not sent out, if enter step 304); Otherwise enter step 306);

Step 304) judge whether there is available channel resource in set C, if had, enter step 305); Otherwise enter step 306);

Step 305) send buffering area to take out from data and come the most front bag (priority is the highest), be assumed to be bag calculate it respectively and remain Successful transmissions probability when available channel transmits in set C;

Selection can provide the channel of maximum Successful transmissions probability and concrete package location as the transmission arrangement of this bag, as position (c _g, q);

To arrange join set in, in an element;

Upgrade in set C and remain concrete remaining available volume of resources information in available channel information and channel; Upgrade the residue packet information to be sent sent in the Φ of buffering area; Return step 303);

Step 306) obtain the best fit strategy of SVC data and Wireless Multi-Channel

In sum, the present invention relates to a kind of SVC(Scalable Video Coding) transmission dispatching method of video in cognitive radio multichannel environment, it utilizes cognitive radio technology can improve the feature of wireless signal-path band width utilance, more radio channel resources can be excavated out, and then the development of the Video Applications of promotion high bandwidth; Utilize again the adaptability that SVC video scalability characteristic is good to time varying channel, improve the validity of transmission of video, suitably can improve the quality of receiver, video under equivalent channel condition.This invention is based on the inherent advantage of cognitive radio and SVC video, optimize further: propose a dynamic spectrum access scheme based on Flexible timeslot, the program considers the mechanics of primary user, reduce unnecessary channel detection, further increase the utilance of wireless signal-path band width; Devise a SVC video and the multi channel graduation matching strategy of cognitive radio, this strategy is by the priority of reasonable distribution SVC video data, the sending order arranging SVC video data and SVC video data and the multi channel dynamic adaptation of cognitive radio, the SVC data that importance is higher in a long term time frames can be made to be sent by the radio channel resource that reliability is higher, the SVC video data that importance is lower is sent by the channel resource that reliability is lower, improves video transmission efficiency.By the optimization of two aspects, more video data Successful transmissions can be had, and the validity of video data received is higher, the quality of the video that this final optimization pass receives.

The dynamic spectrum access scheme based on Flexible timeslot in the present invention decreases the number of detection period in certain limit, is conducive to the utilance improving channel resource.SVC video and CR multi channel graduation matching strategy make the data that importance is higher in for a long time can be transmitted by the channel that reliability is higher, and the data that importance is lower can be transmitted by the channel that reliability is relatively low.These 2 advantages of this invention can make the video quality received be promoted further.

It should be noted last that, above embodiment is only in order to illustrate technical scheme of the present invention and unrestricted.Although with reference to embodiment to invention has been detailed description, those of ordinary skill in the art is to be understood that, modify to technical scheme of the present invention or equivalent replacement, do not depart from the spirit and scope of technical solution of the present invention, it all should be encompassed in the middle of right of the present invention.

Claims (10)

1. the transmission dispatching method of SVC video in cognitive radio multichannel, described method comprises: be first the flexible factor of channel definition of certain secondary user's detection, this flexible factor is for describing the slot length of channel; Then, the flexible factor of optimum of each channel detected is obtained based on flexible Summing Factor primary user mechanics; Finally, judge that the channel that user can access and user access the length of the transmission period of certain channel according to the optimum factor flexibly.

2. the transmission dispatching method of SVC video according to claim 1 in cognitive radio multichannel, it is characterized in that, described transmission dispatching method specifically comprises:

Step 101) secondary user's is at T _sdetect free time or the busy condition of N number of wireless channel in time period, the slot length then defining each channel in N number of channel is θ times of a fixing slot length, and described θ is the flexible factor;

Step 102)

Set up the mechanics model of primary user, obtain primary user according to model within certain period of time period, keep the probability of idle condition or busy condition, and obtain primary user and be converted to the free time by busy or be converted to busy probability by the free time within this period;

Keep the probability of idle condition or busy condition based on the flexible factor, primary user, primary user is converted to the free time by busy and changed into busy probability by the free time, obtain the collision probability of secondary user's and primary user;

Collision probability according to secondary user's and primary user be no more than primary user the principle of patient maximum crash probability, calculate the flexible factor of optimum of each channel in N number of channel

Step 103) flexible for the optimum factor is more than or equal to 1 channel as the available channel of secondary user's, and the number of all available channels is G, and wherein G is less than or equal to N;

Step 104) secondary user's is accessed available channel, adopt G channel to send the SVC video data of secondary user's, return step 101), until all data of secondary user's are sent completely;

Wherein, the duration that the n-th channel in secondary user's access G idle channel carries out SVC video data transmission is described fixing slot length is " T _s+ T _t", T _tit is the time period of a fixed size.

3. the transmission dispatching method of SVC video according to claim 2 in cognitive radio multichannel, is characterized in that, the computing formula of the flexible factor of optimum of any one channel n is as follows:

Wherein, for the probability that secondary user's and primary user collide at channel n, by the patient maximum collision probability of primary user on channel n, described channel n is certain channel in N number of channel.

4. the transmission dispatching method of SVC video according to claim 3 in cognitive radio multichannel, is characterized in that, described in computing formula be:

$P_c^n\left(\mathrm{\&theta;}\right)=1-P_I^n\left(S^E\right)\&CenterDot;e^{{-\mathrm{\&alpha;}}_n\&CenterDot;(\mathrm{\&theta;}\&CenterDot;(T_t+T_s)-T_s)}$

Wherein, represent that channel n is at S ^ethe probability of free time in moment, and S ^efor the detection period T of secondary user's sounding channel _sthe ending moment.

5. the transmission dispatching method of SVC video according to claim 4 in cognitive radio multichannel, is characterized in that, obtains step be:

Step 201) obtain the step of primary user mechanics model, be specially:

The mechanics of primary user obeys Markov model continuous time, namely Markov characteristic is obeyed in the transfer between free time of primary user and busy condition, the duration of " busy and idle " two states of primary user obeys following quantum condition entropy respectively simultaneously, obtains following quantum condition entropy and namely obtains primary user's mechanics model:

$f_I\left(t\right)={\mathrm{\&alpha;}}_n{e}^{-{\mathrm{\&alpha;}}_{n}t}$

$f_B\left(t\right)={\mathrm{\&beta;}}_n{e}^{-{\mathrm{\&beta;}}_{n}t}$

Above-mentioned f _it () represents the value of probability density function in t of the primary user's free time in channel n, α _nvalue obtain according to the historical data of primary user's activity in channel n, this value equals primary user in channel n and is in the inverse of the desired value of the duration of idle period;

F _bt () represents the value of the busy probability density function of the primary user in channel n in t, β _nvalue equal to obtain according to the historical data of primary user's activity in channel n, this value equals primary user in channel n and is in the inverse of the desired value of the duration of busy period;

Step 202) according to the mechanics model of primary user obtained, according to following formula predictions at (0, ε) in the time period, the free time of channel n shared by primary user or busy probability:

$F_I\left(\mathrm{\&epsiv;}\right)=1-{\&Integral;}_{t=0}^{\mathrm{\&epsiv;}}{f}_I\left(t\right)\mathrm{dt}$

$F_B\left(\mathrm{\&epsiv;}\right)=1-{\&Integral;}_{t=0}^{\mathrm{\&epsiv;}}{f}_B\left(t\right)\mathrm{dt}$

Wherein, F _i(ε) and F _b(ε) primary user is respectively at the probability keeping idle condition and busy condition from the t=0 moment to the time period in t=ε moment; " 1-F _i(ε) " for primary user is transformed to busy probability by the free time from the t=0 moment to the time period in t=ε moment; 1-F _b(ε) for primary user is transformed to idle probability by busy from the t=0 moment to the time period in t=ε moment;

Step 203) definition the actual free time of channel n when the detection finish time SE of secondary user's probability be:

$\begin{array}{c}{P}_I^n\left(S^E\right)=\frac{P(\mathrm{RS}=X,\mathrm{AS}=0)}{P(\mathrm{RS}=X)},X=0\mathrm{or}1\\ =\frac{P(\mathrm{AS}=0)\&CenterDot;\left({X+(-1)}^X(1-{\mathrm{\&gamma;}}_n)\right)}{P(\mathrm{AS}=0)\&CenterDot;\left({X+(-1)}^X(1-{\mathrm{\&gamma;}}_n)\right)+P(\mathrm{AS}=1)\&CenterDot;\left({X+(-1)}^X{\mathrm{\&delta;}}_n\right)}\end{array}$

Wherein, secondary user's to the detection period of channel n from S ^bmoment, to S ^emoment terminates; RS is that secondary user's completes the moment in each detection to channel n, i.e. S ^ein the moment, obtain a result of detection, and the result of RS=0 representative detection is idle condition, the result of RS=1 representative detection is busy condition; γ _nfor the probability that false-alarm occurs, be namely idle condition, i.e. AS=0 when channel is actual, but result of detection is busy, i.e. RS=1; δ _nfor the probability that false dismissal occurs, be namely busy condition, i.e. AS=1 when channel is actual, but result of detection is idle condition, i.e. RS=0;

Step 204) according to F _i(ε), F _b(ε), 1-F _i(ε) and 1-F _b(ε) P(AS=0 is obtained) and value P(AS=1), be specially:

$P(\mathrm{AS}=0)=P_I^n\left(S^B\right)\&CenterDot;F_{I\&RightArrow;I}^{T_s}+(1-P_I^n\left(S^B\right))\&CenterDot;F_{B\&RightArrow;I}^{T_s}$

$P(\mathrm{AS}=1)=P_I^n\left(S^B\right)\&CenterDot;F_{I\&RightArrow;B}^{T_s}+(1-P_I^n\left(S^B\right))\&CenterDot;F_{B\&RightArrow;B}^{T_s}$

Wherein, for channel n is at S ^bmoment actual idle probability, this probable value relies on S ^bthe state of the upper non-slotted channel n before the moment; When at S ^bsecondary user's success before transmits data in channel n, then think that the probability of channel n free time is when at S ^bsecondary user's and primary user think after colliding that the probability of channel n free time is before when secondary user's does not transmit data in the channel, then can calculate according to the channel idle probability after the upper time slot detection period and prediction equation as follows:

$P_I^n\left(S^B\right)=P_I^n\left(S_p^E\right)\&CenterDot;F_{I\&RightArrow;I}^{T_t}+(1-P_I^n\left(S_p^E\right))\&CenterDot;F_{B\&RightArrow;I}^{T_t}$

represented the upper detection time slot finish time of channel n in secondary user's time the free time probability;

represent the state of channel n respectively at transmission duration T _tinteriorly keep idle probability and transform to idle probable value from busy.

6. the transmission dispatching method of SVC video according to claim 2 in cognitive radio multichannel, is characterized in that, step 104) described in the step of SVC video data of transmission secondary user's comprise further:

Step 104-1) SVC video data arranged priority and starting point and the cut-off point of each frame are set, the NALU being positioned at starting point and cut-off point scope is effective, and is placed on according to priority by effective NALU and sends in buffering area;

Step 104-2) each NALU fractionation that will send in buffer area encapsulates, and the transmission of packet fractionation encapsulated carries out dynamic adaptation with G the interchannel of free time, thus the more reliable channel of packet employing channel performance getting over high priority is sent.

7. the transmission dispatching method of SVC video according to claim 6 in cognitive radio multichannel, is characterized in that, described step 104-1) comprise further:

Step 104-1-1) the employing GOP SVC video data obtained of encoding is carried out priority division according to following principle:

The priority setting the frame that time stratum level is lower in a GOP is higher, and in the identical frame of multiple time stratum level, playing sequence frame priority is more early higher;

And meet:

The NALU priority that in a frame in a GOP, mass segregation lower grade is higher;

And meet:

The priority of all NALU in a certain mass segregation grade is in, the priority of overall all NALU higher than being positioned on more high-quality stratum level in a GOP, and lower than the priority of all NALU be positioned on more low quality stratum level;

Step 104-1-2) according to the sending order of all frames in priority level initializing GOP of definition, and be that in this GOP, each frame distributes a starting point and a cut-off point, when being between the starting point of certain frame and cut-off point sometime, all NALU then in this frame are effective, and are sorted in transmission buffering area according to the priority of definition by these effective NALU;

Wherein,

When cut-off point arrives, be arranged in buffering area but inefficacy is dropped by the NALU data be not yet sent out, and start the effective NALU in next GOP to put into buffering area;

The setting principle of above-mentioned starting point and cut-off point is: the time difference between adjacent two frame starting points equals time interval during two frame normal play; And all frames share same cut-off point in a GOP, described shared cut-off point is the starting point of the first frame comprised in next GOP.

8. the transmission dispatching method of SVC video according to claim 6 in cognitive radio multichannel, is characterized in that, described step 104-2) comprise further:

Step 104-2-1) be C={c when detecting by this secondary user's the set that the available channel that obtains forms when certain decision-making ₁, c ₂..., c _g..., c _g, and be arranged in the decision-making moment that to send the set expression of buffering area NALU be Φ={ φ ₁, φ ₂..., φ _k..., φ _k;

Wherein, φ _k< φ _k+1, namely index is φ _kthe priority of NALU to be greater than index be φ _k+1the priority of NALU;

Step 104-2-2) descending of NALU priority in buffering area is sent according to above-mentioned detection moment, maximize each element φ in the set of NALU successively _ksuccessfully accepted and the probability of decoding, the transmission finally obtaining the optimum sending video source data arranges vector;

Wherein, mate being arranged in the NALU sending buffering area with the channel of available channel set C according to optimal vector.

9. the transmission dispatching method of SVC video according to claim 8 in cognitive radio multichannel, is characterized in that, described step 104-2-2) optimum transmission arranges vector to adopt following formula to determine:

${\stackrel{\&RightArrow;}{A}}_{\mathrm{opt}}=\{A_{{\mathrm{\&phi;}}_1-\mathrm{opt}},A_{{\mathrm{\&phi;}}_2-\mathrm{opt}},...,A_{{\mathrm{\&phi;}}_{K\&prime;}-\mathrm{opt}}\},K^{\&prime;}\&le;K$

Wherein,

Above-mentioned expression index is φ _knALU be successfully received and the probability that can decode, represent bag according to packet loss of link probability when arranging, representative bag according to collision drop probabilities when arranging, represent index to be φ _knALU carry out splitting the index of each bag encapsulating and obtain, the span of k is 1≤k≤K '.

10. the transmission dispatching method of SVC video according to claim 9 in cognitive radio multichannel, is characterized in that, adopts greedy algorithm to obtain optimal vector value, be specially:

When set of available channels is C={c ₁, c ₂..., c _g..., c _g, send the set Φ={ φ of NALU in buffering area ₁, φ ₂..., φ _k..., φ _ktime:

Step 301) by optimal vector be initialized as: wherein, represent empty set;

Step 302) each NALU sent in the set Φ of effective NALU in buffering area is split and package;

Step 303) judge to send in buffering area whether have the bag be not sent out, if enter step 304); Otherwise enter step 306);

Step 304) judge whether there is available channel resource in set C, if had, enter step 305); Otherwise enter step 306);

Step 305) send buffering area to take out from data and come the most front bag (priority is the highest), be assumed to be bag calculate it respectively and remain Successful transmissions probability when available channel transmits in set C;

Selection can provide the channel of maximum Successful transmissions probability and concrete package location as the transmission arrangement of this bag, as position (c _g, q);

To arrange join set in, be in an element;

Upgrade in set C and remain concrete remaining available volume of resources information in available channel information and channel; Upgrade the residue packet information to be sent sent in the Φ of buffering area; Return step 303);

Step 306) obtain the best fit strategy of SVC data and Wireless Multi-Channel