KR101187527B1 - Method and system for stable throughput of cognitive radio with relaying capabilities - Google Patents

Abstract

In one embodiment, a cognitive radio system includes a first transmitter in communication with a first receiver over a wireless channel. The first transmitter receives the first plurality of packets and transmits the first packets to the first receiver over the channel. A second transmitter in communication with the second receiver and the first receiver over the channel receives a plurality of second packets, receives the plurality of first packets from the first transmitter, and the second over the channel. Send packets to the second receiver. The second transmitter is configured to detect an idle state of the channel. When an idle state of the channel is detected, the second transmitter optionally transmits at least one of the second packets to the second receiver, or relays at least one of the first packets to the first receiver. configured to relay.

Description

METHOD AND SYSTEM FOR STABLE THROUGHPUT OF COGNITIVE RADIO WITH RELAYING CAPABILITIES

This application is related to US Provisional Application No. 60,874,145, filed December 11, 2006 and entitled “Method and System for a Stable Throughput of Cognitive Radio with Relaying Capabilities”, which is incorporated herein in its entirety It is referenced and is the basis for the priority claim.

Embodiments of the present invention relate to the field of cognitive radio principles for communication systems. Example embodiments relate to a method and system for providing stable throughput of cognitive radio using relay capabilities.

Based on the evidence that fixed (licensed) spectrum allocations result in very inefficient resource use, cognitive radio is licensed (or primary) and unlicensed (secondary or cognitive) over the same bandwidth. Prescribe the coexistence of wireless nodes. Although the first group is allowed to access the spectrum at any time, the second group explores the opportunity for transmission by utilizing the idle periods of the primary nodes, which was published in the February 2005 IEEE Journal on Selected Areas Commun. Vol. 23, no. This is described in more detail in S. Haykin's "Cognitive radio: brain-empowered wireless communications," pages 2, 201-220. The main condition is that the activity of the secondary nodes must be "transparent" to the primary nodes so as not to interfere with the authorized use of the spectrum.

Centralized or decentralized protocols at the media access control (MAC) layer that enforce this restriction are described in Proc. Asilomar Conf. "Joint design and separation principle for opportunistic spectrum access" by Y. Chen, Q. Zhao and A. Swami in on Signals, Systems and Computers, and IEEE Journ in March 2006. Selected Areas Commun., Vol. 24, no. 3, "Dynamic spectrum access in open spectrum wireless networks" by Y. Xing, R. Chandramouli, S. Mangold and S. Shankar N, pages 626-637, which are incorporated by reference in their entirety. Here, the radio channel is modeled as busy (eg, primary user is active) or usable (eg, primary user is idle) according to the Markov chain. An information theoretical study of cognitive radio at the physical layer, taking into account the asymmetry between primary and secondary users: IEEE Trans. Inform. Theory, vol. 52, no. 5, "Achievable rates in cognitive radio" by N. Devroye, P. Mitran and V. Tarokh, pages 1813-1827; "Cognitive radio: an information-theoretic perspective" by A. Jovicic and P. Viswanath, available online at http://lanl.arxiv.org/PS_cache/cs/pdf/0604/0604107.pdf ; "Capacity limits of cognitive radio with distributed and dynamic spectral activity" by SA Jafar and S. Srinivasa (sample [http://arxiv.org/abs/cs.IT/0509077]); And Proc. 2005. International Conference on Wireless Networks, Communications and Mobile Computing, vol. 2, pages 1313-1318, in Kyounghwan Lee and A. Yener, "On the achievable rate of three-node cognitive hybrid wireless networks", which are hereby incorporated by reference in their entirety. Optionally, in 2002, in J. Neel, J. Reed, R. Gilles'"The Role of Game Theory in the Analysis of Software Radio Networks" at the SDR Forum Technical Conference, game theory is competitive in cognitive networks. It is claimed as a suitable structure for studying spectral access. Finally, the concept of cognitive radio was included by the IEEE 802.22 Working Group and is currently working on the definition of the Wireless Regional Area Network (WRAN) standard for the secondary use of spectrum allocated to television services. Proc. IEEE DySPAN, page. 328-337, C. Cordeiro, K. Challapali, D. Birru and Sai Shankar N, described in more detail in "IEEE 802.22: the first worldwide wireless standard based on cognitive radio". It became.

A cognitive network in which two source-destination links, a primary link and a secondary link share the same spectral resource (eg, a cognitive interference channel, as shown in FIG. 1) has recently been developed. From the theoretical perspective of information, it was investigated in breakthrough papers such as Devroye et al. And Jovicic et al. In these references, it is assumed that a cognitive transmitter has complete transfer information about the signal transmitted by the primary transmitter (IEEE Trans. Inform. Theory, vol. 52, no. 4, page 4, 2006). See “On compound channels with side information at the transmitter” by P. Mitran, N. Devroye and V. Tarokh, 1745-1755, which are hereby incorporated by reference in their entirety. Incomplete information on, for example, primary activities, is expected to be an important impediment to the implementation of cognitive principles. A. Sahai, N. Hoven, and R. Tandra's "Some fundamental limits on cognitive radio" at Allerton Conference on Communication, Control, and Computing, are described in more detail, and are hereby incorporated by reference in their entirety. In addition, while the dynamics of primary traffic are of great importance in defining the performance of cognitive radio, random packet arrival can be easily incorporated into pure information theoretical analysis.

Aspects of example embodiments are directed to increasing the average throughput of a secondary link of a cognitive radio system and ensuring the stability of the system.

In one embodiment, a cognitive radio system is a first transmitter in communication with a first receiver over a wireless channel, wherein the first transmitter receives a plurality of first packets and sends the first packets over the channel to the first transmitter. Transmitting to a receiver; and a second transmitter in communication with a second receiver and the first receiver over the channel, wherein the second transmitter receives a plurality of second packets and receives the plurality of second packets from the first transmitter. Receive first packets, and transmit the second packets to the second receiver over the channel. The second transmitter is configured to detect an idle state of the channel. When an idle state of the channel is detected, the second transmitter optionally transmits at least one of the second packets to the second receiver, or relays at least one of the first packets to the first receiver. configured to relay.

The second transmitter selectively transmits at least one of the second packets or based on the detected state of transmission of at least one of the first packets from the first transmitter to the first receiver. And configured to relay at least one of the 1 packets.

The second transmitter is configured to relay at least one of the first packets if it is detected that the transmission of at least one of the first packets was unsuccessful.

The second transmitter is configured to refrain from relaying at least one of the first packets when it is detected that transmission of at least one of the first packets is successful.

The first transmitter is configured to acknowledge acknowledgment of at least one of the first packets by the second transmitter.

After the first transmitter transmits at least one of the first packets, the first transmitter detects that the transmission of at least one of the first packets was unsuccessful, and the first transmitter detects the first transmission. And upon acknowledgment of repair of at least one of the first packets by the two transmitters, refrain from re-transmitting at least one of the first packets.

The second transmitter has a transmit power, and the second transmitter is configured to ensure service stability of the first transmitter by controlling the transmit power.

The second transmitter optionally ensures service stability of the first transmitter in transmitting at least one of the second packets to the second receiver or relaying at least one of the first packets to the first receiver. It is configured to.

The first transmitter corresponds to an authorized user of the channel, and the second transmitter corresponds to an unlicensed user of the channel.

In another embodiment, a method of operating a cognitive radio system includes directing a plurality of first packets to a first transmitter and to a second transmitter for transmission to a first receiver over a wireless channel; Sending a plurality of second packets to the second transmitter for transmission to a second receiver over the channel; Transmitting at least one of the first packets from the first transmitter to the first receiver; Detecting an idle state of the channel at the second transmitter; And after detecting the idle state of the channel, optionally transmitting at least one of the second packets from the second transmitter to the second receiver or from the second transmitter to the first receiver. Relaying at least one of them.

Optionally transmitting at least one of the second packets or relaying at least one of the first packets is based on a detected state of transmission of at least one of the first packets to the first receiver. do.

The first transmitter corresponds to an authorized user of the channel and the second transmitter corresponds to an unauthorized user of the channel.

1 is a diagram of a system configuration in accordance with an exemplary embodiment.

FIG. 2 is a graph showing the arrival rate for the primary user versus the maximum power allocated to the secondary user of the system configuration of FIG.

3 is a graph showing the sensitivity of the error detection probability component of the system configuration of FIG. 1 to the average arrival rate for the primary user.

FIG. 4A is a graph illustrating the upper limit of the arrival rate for the primary user versus the throughput-maximum power allowed for the secondary user of the system configuration of FIG. 1.

FIG. 4B is a graph showing the arrival rate for the primary user versus the maximum throughput of the secondary user of the network configuration of FIG. 1.

Fig. 5 is a diagram of a system configuration according to another exemplary embodiment.

FIG. 6A is a graph showing the arrival rate for the primary user versus the maximum power allowed for the secondary user of the system configuration of FIG. 5.

FIG. 6B is a graph showing the arrival rate for the primary user versus the service probability of the secondary user in the system configuration of FIG. 5.

FIG. 6C is a graph illustrating the arrival rate and maximum throughput of the secondary user for the primary user of the system configuration of FIG. 5.

FIG. 7 is a graph showing the maximum throughput of secondary users of the network of FIG. 5.

The embodiments described herein relate to a scenario in which there are two single-user links, one authorized (primary) to use spectral resources, and one unauthorized (secondary or cognitive). According to the aware radio principle, the operation of the secondary link is required not to interfere with the performance of the primary link. Thus, a recognition link is understood here so that it can only be accessed when the channel is detected to be idle. Further exemplary embodiments include: 1) random packet arrival; 2) detection of errors due to fading on the secondary link; And 3) power allocation at the secondary transmitter based on long-term measurements. According to an exemplary embodiment, the maximum stable throughput of the recognition link (in packets / slot) is derived for the fixed throughput selected by the primary link.

According to another embodiment, the secondary transmitter is configured to operate as a "transparent" relay to the primary link. In particular, packets not correctly received by the intended endpoint can be successfully decoded by the secondary transmitter. The latter queues these packets and forwards them to the intended receiver. Stable throughput of the secondary link using relaying can be derived under the same (or similar) conditions as described above. As will be described in more detail below, the specific features (or advantages) of relaying may depend on the topology of the network (eg, average channel power).

In exemplary embodiments, the cognitive interference channel as shown in FIG. 1 is further investigated by accounting for measurement error associated with the primary transmitter 14 and random packet arrival. In more detail, the primary transmitter 14 and the secondary transmitter 24 have respective queues 12, 22 of a particular size (eg, a defined size), with time slotting. (slot) Here, the secondary transmitter (or node) 24 can infer the timing of the primary link from the signal received during the observation phase. At the beginning of each slot, the acknowledgment node 24 senses the channel 30 and, if detected to be idle, sends a packet (if the node 24 has a packet in the queue 22). . Detection of primary operation causes errors (or encounters errors) due to a failure on the wireless fading channel 30, and thus possible interference from the secondary link 20 to the primary link 10. Causes Since the cognitive principle is based on the idea that the presence of the secondary link 20 should be "transparent" to the primary link 10, appropriate countermeasures (e.g., power control) are the secondary nodes ( Applies in 24). In an exemplary embodiment, the stability of the system (ie, the finiteness of the system's queues 12, 22 at all times) is chosen as the main performance criterion. In another embodiment, given the average throughput independently selected by the primary transmitter 14, the maximum average throughput that the secondary link 20 can sustain while ensuring the stability of the system is determined. .

In another embodiment, the secondary transmitter 54 provides a relaying capability (eg, see FIG. 5). Here, the direct channel on the primary link 40 is weak (or weaker) for the channel from the primary transmitter 44 to the secondary transmitter 54. In this case, having packets relayed by the secondary transmitter 54 may help to empty the queue 42 of the primary transmitter 44, thus creating transmission opportunities for the secondary transmitter. do. According to another embodiment, relaying primary packets by the cognitive transmitter 54 increases the stable throughput of the secondary link 50 (for a fixed and selected throughput of the primary link 40). Let's do it.

I. System Model

Referring again to FIG. 1, the single-link primary communication 10 is activated, and the secondary (cognitive) single link 20 is interested in using the spectrum resources 30 whenever possible. The system includes physical layer parameters and MAC dynamics as follows.

A. MAC  Hierarchy model

Both primary and secondary transmitting nodes 14, 24 include buffers 12, 22 with a specific capacity (eg, limited capacity) for storing incoming packets. The time is slotted, and the transmission of each packet occupies one slot (all packets have the same number of bits). Each node is stationary and independent of the packet arrival process independent of the mean λ _P [packets / slot] for the primary user 14 and the λ _S [packets / slot] for acknowledgment 24 ( See FIG. 1). Due to a failure (eg fading) of the radio channel 30, the packet may be received by mistake by the intended endpoint, which requires retransmission. Here, those skilled in the art will appreciate that there is also overhead for the transmission of ACK (ACKnowledgement) and Not-ACKnowledgement (NACK) messages.

According to the recognition principle, the primary link 10 uses the channel 30 if there are any packets to send to its queue 12. On the other hand, the secondary (cognitive) transmitter 24 detects the channel 30 in each slot and, if the transmitter detects an idle slot, sends a packet from its queue 22 (if present). do. Here, the slot is long enough to allow an appropriate detection time interval for the cognitive node 24. As will be described below, due to fading due to fading, the secondary transmitter 24 causes errors (or encounters errors) while detecting the presence of the primary user 14. In another embodiment, as will be described in more detail below, the MAC layer will be configured to allow the secondary node to act as a relay for the primary node (see, eg, FIG. 5).

B. Physical Layer Model

인 독립된 고정식 Rayleigh 플랫-페이딩 채널들

에 의해 영향을 받는 것으로 추정된다(t는 시간을 나타내며 타임-슬롯들을 통해 흐른다). 여기서 순간 전력

의 누적 분포 함수는

이다. 채널은 각각의 슬롯(블록-페이딩)에 대하여 일정하다. (섀도우잉(shadowing) 및 경로 손실로 인한) 평균 전력 이득은

으로 표시되며, i는 1차적인 연결에 대해 "P"로, 2차적인 것에 대해 "S"로, 2차적인 송신기로부터 1차적인 수신기 사이의 채널에 대해 "SP"로, 1차적인 송신기와 2차적인 송신기 사이의 채널에 대해 "PS"로 판독한다.Referring to FIG. 1, any radio propagation of any nodes

Independent Fixed Rayleigh Flat-Fade Channels

Is assumed to be affected by (t represents time and flows through the time-slots). Instantaneous power here

The cumulative distribution function of

to be. The channel is constant for each slot (block-fading). Average power gain (due to shadowing and path loss)

Where i is "P" for the primary connection, "S" for the secondary, "SP" for the channel between the secondary transmitter and the primary receiver, the primary transmitter Read as "PS" for the channel between and the secondary transmitter.

로 전송하며, 일반성을 잃지 아니하고(WLOG), 모든 수신기에서의 잡음 전력 스펙트럼 밀도는 또한 1로 정규화된다. (활성화된 경우) 2차적인 노드에 의해 전송되는 전력은

이다. 하나의 실시예에서, 주어진 패킷의 전송은 순간 수신 신호-대-잡음 비(SNR)

가 전송 모드의 주어진 선택에 대해 고정된, 주어진 임계값

이상인 경우에 성공적인 것으로 생각된다. 따라서, 1차적 또는 2차적 링크의 중단(outage)(성공적이지 않은 패킷 수신)의 가능성은 다음과 같다(i 는 "P" 또는 "S"와 같다)In one embodiment, the primary node is normalized power.

WLOG, the noise power spectral density at all receivers is also normalized to one. The power transmitted by the secondary node (if enabled)

to be. In one embodiment, the transmission of a given packet results in an instantaneous received signal-to-noise ratio (SNR).

Threshold given, fixed for a given selection of transmission modes

In the above case, it is considered to be successful. Thus, the possibility of outage of the primary or secondary link (unsuccessful packet reception) is as follows (i is equal to "P" or "S"):

)을 사용할 수 있다.In one embodiment, the primary and secondary links have different signal-to-noise ratio requirements for the transmission modes.

) Can be used.

이 임계값 α보다 큰 경우(

임을 상기) 1차적인 사용자의 전송을 정확하게 검출할 수 있다. 검출 프로세스에서의 에러 확률은 다음과 같다.Cognitive Node SNR

Is greater than this threshold α

It is possible to accurately detect the transmission of the primary user. The error probability in the detection process is as follows.

) 이해할 것이다. 또한, 1차적인 사용자가 전송하고 있지 않은 때에는 언제라도, 2차적인 송신기는 0의 에러(거짓 알람) 확률로 유휴 슬롯을 검출할 수 있음을 이해할 것이다. 이러한 이해는 다른 시스템들로부터의 간섭이 무시할 수 있는 것으로 가정되는 시나리오에서 합리적이다. 아래에 더 자세히 설명될 바와 같이, 당업자는 거짓 알람의 0이 아닌 확률의 존재가 여기에 나타난 설명에 따른다는 것을 이해할 수 있을 것이다.Here, whenever the secondary node can decode the primary signal, the secondary node can detect its presence (i.e.,

Will understand). It will also be appreciated that at any time when the primary user is not transmitting, the secondary transmitter can detect idle slots with a zero error (false alarm) probability. This understanding is reasonable in scenarios in which interference from other systems is assumed to be negligible. As will be described in more detail below, those skilled in the art will appreciate that the presence of a nonzero probability of false alarms depends on the description presented herein.

C. Issue Analysis Issues Analyzed )

를 다음의 두 개의 (잠재적인) 충돌(conflict)하는 목표들에 대해 채널들(

)의 통계치 및 시스템 파라미터들(

)에 기반하여 선택할 수 있음을 이해할 것이다: (i) 자신의 동작을 1차적인 링크에 대해 "투명"하게 함(아래에 더 자세히 설명); (ii) 자신의 안정적 스루풋을 최대화. 두 개의 설명(remark)은 순차적이다:Secondary transmitter has its own transmit power

For the following two (potential) conflicting targets:

Statistics and system parameters

Will be able to make a choice based on: (i) making their behavior "transparent" to the primary link (described in more detail below); (ii) Maximize their stable throughput. The two remarks are sequential:

1. The perception of the primary user's "transparency" is defined as the unit of stability of the primary user's queue. That is, due to the operation of the secondary, the primary node can ensure that its queue remains stable. Conversely, in certain embodiments, no constraints on the increase in average delay experience by the primary are imposed. As such, certain embodiments are suitable for delay-insensitive applications.

)의 지식(knowledge)은 인지 송신기에서 가정된다. 가정된 고정식 페이딩 시나리오에서, 여기서의 전제(premise)는 인지 노드가 인지 사이클의 관찰 기간동안 이러한 파라미터들을 추론할 충분한 시간을 가진다는 것이다. 이러한 목표를 이루기 위한 가능한 해법은 유휴 슬롯들에 서의 2차적인 수신기와의 협력으로 이뤄질 수 있다(인지 네트워크에서 협력 검출에 대한 더 자세한 정보에 대해서는, 2005년의 Proc. DySPAN, 페이지. 137- 143의, G. Ganesan 및 Y. Li의 "Cooperative spectrum sensing in cognitive radio networks", 그리고, 2006년의 Proc. IEEE ICC에서 S. M. Mishra, A, Sahai 및 R. W. Brodersen의"Cooperative sensing among cognitive radio"를 참조하기 바라며, 이들의 전체적인 내용은 여기에 참조된다). 하나의 실시예에서, 시스템 파라미터

는 1차적인 통신에 관한 2차적인 링크에서 가용한 이전의 지식의 일부이다. 또한, 다른 실시예에서, 1차적인 링크에 의해 선택된 스루풋

는 유휴 슬롯들의 부분(fraction)을 관찰하고, 2차적인 수신기에 의해 전송된 ACK/NACK 메시지들을 측정함으로서 추정될 수 있다.2. (average) channel parameters (

Knowledge is assumed at the cognitive transmitter. In the hypothesized fixed fading scenario, the premise here is that the cognitive node has enough time to infer these parameters during the observation period of the cognitive cycle. A possible solution to achieving this goal could be to work with a secondary receiver in idle slots (for more information on cooperative detection in cognitive networks, see Proc. DySPAN, 2005, page 137-). See 143, "Cooperative spectrum sensing in cognitive radio networks" by G. Ganesan and Y. Li, and "Cooperative sensing among cognitive radio" by SM Mishra, A, Sahai, and RW Brodersen, in Proc. IEEE ICC, 2006. The full contents of which are incorporated herein). In one embodiment, system parameters

Is part of the previous knowledge available on the secondary link on primary communication. Also, in another embodiment, the throughput selected by the primary link

Can be estimated by observing the fraction of idle slots and measuring ACK / NACK messages sent by the secondary receiver.

III . Stable of Cognitive Nodes Throughput

에 대해 2차적인 노드에 의해 유지될 수 있는 최대 스루풋(즉, 평균 도착 레이트)

에 관련된 것이다. 다시 말해서, 1차적인 사용자는 2차적인 노드의 존재를 무시하면서 자신의 도착 레이트

를 선택한다. 인지 노드의 임무는 시스템의 안정성에 영향을 미치지 아니하면서 1차적인 활동에 의해 사용가능하도록 남겨진 유휴 슬롯들을 가능한 많이 발견하기 위해 자신의 전송 모드(예를 들어, 전력

)를 선택하는 것이다.The embodiments described below provide a given (fixed) throughput for the system to be stable.

Maximum throughput (ie, average arrival rate) that can be maintained by the secondary node for

Related to. In other words, the primary user has his arrival rate ignoring the existence of the secondary node.

. The task of the cognitive node is to transmit its own transmission mode (eg, power) to discover as many idle slots as possible left available for use by primary activity without affecting the stability of the system.

).

Stability is defined as all queues in the system are stable. In one embodiment, the queue is said to be stable only if the probability that it remains empty for a time t that grows to infinity remains nonzero.

는 시간 t에서 i번째 큐의 완료되지 않은 업무(패킷들에서)를 나타낸다. 안정성의 대안적인 정의는 1988년 9월의 IEEE Trans. Inform. Theory, vo. 34, 페이지 918-930의 R. Rao 및 A. Ephremides의 "On the stability of interacting queues in a multi-access system"에 더 자세히 설명되어 있으며, 이들의 전체적인 내용은 여기에 참조된다. 큐잉 시스템의 도착 및 출발 레이트들이 고정적인 경우, 안정상은 Loynes의 정리(theorem)를 사용함으로써 체크될 수 있다(1962년의 Proc. Cambridge Philos. Soc. 58, 페이지 497-520의 R. M. Loynes의 "The stability of a queue with non-independent inter-arrival and service times"를 참조, 이들의 전체적인 내용은 여기에 참조됨). 여기서, 평균 도착 레이트

가 평균 출발 레이트

보다 낮은 경우(

), i번째 큐는 안정적이다; 반면에, 평균 도착 레이트

가 평균 출발 레이트

보다 높은 경우(

), i번째 큐는 불안정한 것이다; 마지막으로,

이면, 큐는 안정적이거나, 불안정하다. Loynes의 정리가 적용가능한 경우는 언제라도, 일 실시예에서, 평균 출발 레이트

는 i번째 큐의 최대 안정 스루풋으로 정의된다. here,

Denotes the incomplete task (in packets) of the i th queue at time t. An alternative definition of stability is described in IEEE Trans. Inform. Theory, vo. It is described in more detail in "On the stability of interacting queues in a multi-access system" by R. Rao and A. Ephremides on page 34, pages 918-930, the entire contents of which are here. If the arrival and departure rates of the queuing system are fixed, the stability phase can be checked by using Loynes' theorem (Proc. Cambridge Philos. Soc. 58, 1962, RM Loynes, pages 497-520). stability of a queue with non-independent inter-arrival and service times ", the full details of which are referenced here). Where average arrival rate

Average departure rate

Lower than

), the i-th cue is stable; On the other hand, average arrival rate

Average departure rate

Higher than

), the i th cue is unstable; Finally,

If it is, the cue is stable or unstable. Whenever Loynes' theorem is applicable, in one embodiment, the average starting rate

Is defined as the maximum stable throughput of the ith cue.

를 가지는 (채널 페이딩 프로세스

의 고정성으로 인한) 고정 출발 레이트에 의해 특정된다. 또한, Loynes의 정리에 의해, 레이트

는 1차적인 사용자에 의해 "감지된(perceived)" 최대 안정적인 스루풋이다. 다시 말해서, 1차적인 사용자는 다음을 만족하는 임의의 레이트

를 선택하도록 허용된다:In more detail, a system from the perspective of the primary transmitter is contemplated. According to the recognition principle, the primary link does not know the existence of a secondary node that wants to use its bandwidth when it is available. Thus, as long as the primary node is involved, the system consists of a single queue of its own, which is averaged

Channel fading process with

Due to a fixed start rate). In addition, by Loynes' theorem, rate

Is the maximum stable throughput that is "perceived" by the primary user. In other words, the primary user is any rate that satisfies

You are allowed to select:

의 고정성으로 인하여 고정되어 있으며,

와 같은 평균을 가진다.

에서의 두 번째 항은 2차적인 노드가 1차적인 노드가 자신의 큐안에 어떠한 패킷도 가지고 있지 않은 경우에만 그 채널에 액세스하는 제약을 가한다. Little의 정의에 의하면,

이고, 이는 수학식 1로부터 다음과 같이 주어진다.In a system without measurement errors, the recognition link does not cause any errors (or encounter errors) while detecting the primary user's motion. Thus, the recognition link can access the channel in an idle slot without causing any interference to the primary link, and the cues at the two transmitters do not interact. The departure rate at the secondary transmitter is the channel process

It is fixed due to the fixing of

Has the same average as

The second term in Eqs. Constrains the secondary node to access the channel only if the primary node does not have any packets in its queue. By Little's definition,

This is given by Equation 1 as follows.

가 자신의 최대(

)와 같은 경우에 대해 이루어질 수 있다. 또한, 2차적인 링크에서의 중단의 확률에 따른 1차적인 링크의 동작에 의해 사용가능하도록 남겨진 "잔여(residual)" 스루풋

의 분수가 존재한다.Thus, in the absence of measurement errors, the maximum throughput of the secondary link is the transmit power

Is his maximum (

Can be done for In addition, "residual" throughput left available by the operation of the primary link according to the probability of interruption in the secondary link.

There is a fraction of.

의 전력으로 전송을 시작한다. 그러나, 검출 프로세스에서의 에러로 인하여, 2차적 노드는 (다시, 적어도 하나의 패킷이 큐에 존재하는 경우에)

의 확률로 1차적인 전송에 의해 점유된 슬롯들에서도 (동일한 전력

로) 전송을 시작한다(또는 시작할 수 있다). 이는 1차적인 링크의 통신에 간섭을 발생시키며, 차례로 1차적인 링크의 실제 스루풋을 감소시킨다. 여기서, 1차적인 및 2차적인 송신기의 큐잉 시스템이 상호작용하는 것을 알 수 있다. 따라서, 출발 레이트의 고정성이 보증될 수 없으며, loynes의 정리가 적용되지 못할 수 있다(1999년 7월의, IEEE Trans. Inform. Theory, vol. 45, no. 5, 페이지 1579-1587에서, Rao 등, 및 W. Luo 및 A. Ephremides의 "Stability of interacting queues in random-access systems"를 참조, 이들의 전체 내용은 여기에 참조됨).The cognitive node senses the channel in each slot and, if no activity is measured from the primary node (at least one packet is present in the queue)

Start transmission with power. However, due to an error in the detection process, the secondary node may again (if at least one packet is present in the queue).

Even in slots occupied by primary transmission with a probability of equal power

Begin transmission (or may start). This interferes with the communication of the primary link, which in turn reduces the actual throughput of the primary link. It can be seen here that the queuing systems of the primary and secondary transmitters interact. Thus, the fixedness of the starting rate cannot be guaranteed and the loynes theorem may not be applied (see, July 1999, IEEE Trans. Inform. Theory, vol. 45, no. 5, pages 1579-1587, Rao et al., And “Stability of interacting queues in random-access systems” by W. Luo and A. Ephremides, the entire contents of which are incorporated herein).

이다. Described in more detail below, the maximum power allowed for the acknowledgment node to transmit to ensure the stability of the primary queue.

to be.

및 시스템 파라미터들

:Proposition 1: Given channel parameters

And system parameters

:

이면, 2차적인 사용자는 1차적인 노드의 큐의 안정성에 영향을 미치지 아니하고 임의의 전력

를 사용할 수 있으며, 특정한 실시예에서,

는 자신의 최대(

)과 같도록 설정된다;

If the secondary user does not affect the stability of the primary node's queue,

Can be used, and in certain embodiments,

Is your maximum (

Is set equal to;

이면, 인식 노드가 사용할 수 있는 최대 전력은 다음 과 같다.

In this case, the maximum power available to the recognition node is as follows.

이면, 2차적인 노드는 "더미(dummy)" 패킷들을, 유휴 채널을 감지할 때마다 전송하고, 따라서 그 큐가 비어있는지 여부와 관계없이 1차적인 사용자와 가능한 간섭을 지속한다. Luo 등에 따르면, 지배적인 시스템은 원래 시스템이 안정적인 경우에만 안정적이다. 사실, 한 측면에서, 지배적인 시스템의 큐들은 원래 시스템의 큐들보다 항상 더 큰 크기를 갖는다(따라서 지배적인 시스템이 안정적이면, 원래 시스템도 안정적이다). 다른 면에서, 포화(saturation)되는 경우, 지배적인 시스템에서 "더미" 패킷을 전송하는 확률은 0이며, 두 시스템들은 구별가능하지 않다(따라서, 지배적인 시스템이 안정적이지 않으면, 원래 시스템도 안정적이지 않다). 이후에 더 자세하게 설명될 바 와 같이, 지배적인 시스템에서 출발 레이트들은 고정 프로세스이며, Loynes의 정리는 원래 시스템의 안정성에 관한 결론을 끌어내는데 사용가능하다.Due to the interaction between primary and secondary transport nodes, Loynes' theorem cannot be used directly to investigate the stability of the system. To overcome this, a modified system, referred to as dominant, is considered. It has the same stability characteristics as the original system and at the same time represents non-interacting cues. In making the configuration, the dominant system is here established by modifying the original system as described below.

If so, the secondary node sends "dummy" packets each time it detects an idle channel, thus persisting possible interference with the primary user regardless of whether the queue is empty. According to Luo et al, the dominant system is stable only if the original system is stable. In fact, in one aspect, the queues of the dominant system are always larger than the queues of the original system (so if the dominant system is stable, the original system is also stable). On the other hand, when saturated, the probability of sending a "dummy" packet from the dominant system is zero, and the two systems are indistinguishable (so, if the dominant system is not stable, the original system is also stable). not). As will be explained in more detail below, the starting rates in the dominant system are fixed processes, and Loynes' theorem can be used to draw conclusions about the stability of the original system.

에 대하du 자신의 최대 전력(

)을 사용할 수 있다.As in the direct process of Theorem 1, the secondary node is

For your own maximum power (

) Can be used.

(수학식 6 참조) 대 1차적인 사용자에 의해 선택된 스루풋을 도시한 것이며, 여기서,

(이는

임을 의미함),

=10, 15, 20dB 및

이다. 여기서, 예를 들어, 도 2에서 도시된 결과를 더 양호하게 시각화하기 위하여, 확률

에 대하여 큰 값이 선택되었다. 당업자는 다른 값을 가지는 파라미터들에 대한 이 방식의 퍼포먼스들이 표시된 결과들로부터 추론될 수 있음을 이해할 수 있을 것이다. 여기서, 수학식 4에서 1차적인 사용자가 선택할 수 있는 최대 레이트는

이며, 2차적인 노드가 1과 비교하여 자신의 전력을 감소시켜야 하는 수학식 7의 1차적인 레이트

는

가 증가함에 따 라 감소된다. 검출 에러 확률에 대한

의 민감성(sensitivity)은 도 3에 도시되었으며, 여기서,

는

에 대해 도시되었으며, 다시,

=10, 15, 20dB이다:

이 되면, (

가 증가하면), 이 비는

에 가까워지며, 반면에

가 되면, (

가 감소하면), 그 비는 1에 가까워지고, 인식 노드는 도 4의 임의의

에 대해 자신의 최대 전력을 사용하도록 허용된다.2 illustrates a maximum power allowed for a secondary user, in accordance with each embodiment.

(See Equation 6) versus throughput selected by the primary user, where

(this is

Means that

= 10, 15, 20 dB and

to be. Here, for example, to better visualize the result shown in FIG. 2, the probability

A large value was chosen for. Those skilled in the art will appreciate that performances in this manner for parameters with different values can be inferred from the displayed results. Here, in Equation 4, the maximum rate selectable by the primary user is

Where the secondary node has to reduce its power compared to 1

The

Decreases with increasing. For detection error probability

The sensitivity of is shown in Figure 3, where

The

Has been shown for, and again,

= 10, 15, 20 dB:

Then, (

Increases), this rain

Closer to,

Becomes (,

Decreases), the ratio approaches 1, and the recognition node

Are allowed to use their maximum power for.

응 통한 최적으로 감소한다. 실제로,

를 선택하는 데 있어서 고유의 트레이드-오프가 존재한다. 한 면에서, 증가하는

는 1차적인 링크의 간섭을 증가시키고, 이는 인식 노드에 대한 전송 기회의 확률을 제한한다. 다른 면에서, 증가하는

는 2차적인 링크에서의 정확한 수신의 확률을 향상시킨다. The queuing process at the recognition node will be considered in more detail. Here, the original issue that addresses the maximum throughput sustained by the secondary node is addressed under the constraint that the system must be stable. As shown below, the issue is related to the power transmitted (under the constraints described in Theorem 1 above).

Optimally reduces coagulation. in reality,

There is a unique trade-off in selecting. On one side, increasing

Increases the interference of the primary link, which limits the probability of transmission opportunities for the recognition node. In other ways, increased

Improves the probability of accurate reception on the secondary link.

) 및 시스템 파라미터들(

), 1차적인 사용자의 큐들의 안정성이 보존된다(인지 노드의 "투명 성")는 것에 대한 이해하에서, 인지 사용자의 최대 안정적인 스루풋은 다음의 최적화 방정식을 품으로써 획득된다:Proposition 2: Given channel parameters

) And system parameters (

In understanding that the stability of the primary user's cues is preserved (“transparency” of the cognitive node), the maximum stable throughput of the cognitive user is obtained by having the following optimization equation:

Here, the throughput of the secondary link is read as follows,

는 수학식 7에 의해 주어진다. 수학식 8에 설명된 바와 같은 최적화 방정식은 1차원적 탐색을 요구하며, 2003년의 Athena Scientific의 D. P. Bertsekas의 Nonlinear Programming과 같은 표준 방법을 사용함으로서 해결될 수 있으며, 이러한 전체 내용은 여기에 참조된다. 여기서, 2차적인 것에서의 거짓 알람

의 0이 아닌 확률을 가정하면 (Sec. II 참조), 당업자는 결과가 수학식 9의 성취가능한 스루풋

의 스케일링(scaling)임을 이해할 수 있을 것이다. 정리 2와 관련된 더 자세한 내용은 이후에 설명된다.

Is given by equation (7). Optimization equations, as described in Equation 8, require a one-dimensional search and can be solved by using standard methods such as Athena Scientific's DP Bertsekas's Nonlinear Programming in 2003, which is hereby incorporated by reference in its entirety. . Where false alarms in the secondary

Assuming a non-zero probability of (see Sec. II), one skilled in the art will know that the result is

It will be appreciated that the scaling of. More details regarding Theorem 2 are described later.

및 도 2 및 3과 동일한 조건들(

,

및

)에서의 인식 노드에 대한 대응하는 최대 스루풋

대 1차적인 사용자의 선택된 스루풋

을 도시한 것이다. 특히 도 4a는 상한 경계(수학식 6 참조)(점선) 및 스루풋-최대화 전력

를 모두 도시하며, 도 4b는 대응하는 최대 스루풋

을 도시한다. 수학식 9로부터 예상되는 바와 같이, 최대 스루풋

는 최적

가 1과 같은 한,

에 대하여 선형적으로(또는 실질적으로 선형적으로) 감소한다. 참조를 위해, 수학식 5의 최대 안정적 스루풋

이 도시되었으며, 이는 인지 링크에서 어떠한 감지 에러들도 발생하지 않는 경우와 관련된다.4A and 4B show optimal transmit power

And the same conditions as in FIGS. 2 and 3 (

,

And

Corresponding throughput for the listening node at

Primary user's selected throughput

It is shown. In particular, FIG. 4A shows an upper bound (see Equation 6) (dashed line) and throughput-maximum power.

4b shows the corresponding maximum throughput

To show. As expected from Equation 9, maximum throughput

Is optimal

As long as 1 is equal to 1,

Decrease linearly (or substantially linearly) relative to. For reference, the maximum stable throughput of equation (5)

Is shown, which relates to the case where no sensing errors occur in the cognitive link.

IV . Relaying  Stability of Cognitive Nodes Throughput

)이 직접 채널

과 비교하여 유리한 경우, 제 2 채널에 의해 패킷들이 릴레이되도록 하는 것은 1차적인 송신기(44)의 큐를(42) 비우는 것을 도울 수 있다. 이는 2차적인 것에 대한 전송 기회를 생성한다. 당업자는 인식 노드(54)에 대한 사용가능한 슬롯들의 증가한 수가 자신의 패킷들 및 릴레이된 패킷들의 전송 사이에서 공유되어야만 한다는 것을 이해할 수 있을 것이다. 이러한 수정된 구조의 특징들(또는 이점들)을 평가하는 것은 사소한 것이 아니며, 이후에 더 자세히 설명될 것이다.Here, the Sec. Protocol allows the acknowledgment node to act as a (“transparent”) relay to the primary link. The modified version of the system model shown in II is explained in more detail. More precisely, in one embodiment, the secondary node is allowed to forward packets of the primary user that were not successfully received by the intended endpoint. By doing so, the system is designed so as not to violate the radio principle whether the secondary node is " invisible " for the primary node (see, for example, FIG. 5 showing the system). As described above, the characteristics obtained by adding the relaying capability to the secondary transmitter 54 are as follows. A propagation channel from the first transmitter 44 to the second transmitter 54 (

) This direct channel

If advantageous compared to, causing the packets to be relayed by the second channel may help to empty the queue 42 of the primary transmitter 44. This creates a transmission opportunity for the secondary. Those skilled in the art will appreciate that the increased number of available slots for the acknowledgment node 54 must be shared between the transmission of its packets and the relayed packets. Evaluating the features (or advantages) of this modified structure is not trivial and will be described in more detail later.

A. Relaying  Used MAC  Hierarchical System Model

를 가지는 2차적 노드에 의한 채널 감지)은 변하지 않는다. 하나의 실시예에 따르면, 차이점들은 오직 인지 노드의 전송 전략 및 ACK/NACK 메시지들의 교환과 관련된 세부 내용과 관련된다. 도 5를 참조하면, 인지 노드(54)는 두 개의 큐들을 가지고있으며, 큐(52)는 자신의 패킷들(

)을 수집하고, 큐(51)는 1차적인 끝점으로 릴레이될 1차적인 송신기에 의해 수신된 패킷들(

)을 수집한다. 1차적인 노드(44)에 의해 전송된 패킷은, 실제로 의도된 끝점(예를 들어, 수신기 46)에 의해 잘못 수신될 수 있으나(NACK 메시지와 함께 이벤트를 시그널링함), (ACK 메시지를 전송하는) 2차적 송신기(54)에 의해 정확하게 수신될 수 있다. 이러한 경우에, 1차적인 소스(44)는, (끝단(46)에 의해 정확하게 수신된 것 처럼) 자신의 큐(42)로부터 패킷을 떨어뜨리고, 2차적인 소스(54)는 그것을 자신의 큐(51)

에 넣는다 . (이는 1차적인 사용자에 대한 2차적인 사용자의 투명성의 인지 무선 원칙으로부터의 작은 일탈이다: 실제로, 2차적인 동작으로 인하여, 1차는 동일한 패킷에 대한 두 개의 승인 확인을 수신할 수 있다. 여기서, 일 실시예에서, 적어도 하나의 승인 확인이 긍정적인 경우에는 패킷이 정확하게 수신된 것으로 고려할 것이다.) 여기서, 1차적인 끝단(46) 및 2차적인 송신기(54)가 정확하게 신호를 디코딩한 경우, 2차적인 것(54)은 (끝단으로부터 ACK 메시지가 수신되면) 자신의 큐에서 후자를 포함하지 않는다.Here, according to one embodiment, Sec. The MAC layer model described in II is modified to account for the relaying capability added to the secondary node 54. Key assumptions (e.g. limited buffers, slotted transmissions, fixed process arrivals, detection error probability in Equation 2)

Channel sensing by the secondary node with < RTI ID = 0.0 > According to one embodiment, the differences relate only to details relating to the transmission strategy of the acknowledgment node and the exchange of ACK / NACK messages. 5, acknowledgment node 54 has two queues, and queue 52 has its own packets (

), The queue 51 receives packets received by the primary transmitter to be relayed to the primary endpoint.

Collect). Packets sent by the primary node 44 may actually be incorrectly received by the intended endpoint (e.g., receiver 46) (signaling an event with a NACK message), but sending an ACK message. ) Can be correctly received by the secondary transmitter 54. In this case, the primary source 44 drops the packet from its queue 42 (as if it has been correctly received by the end 46) and the secondary source 54 sends it to its own queue. (51)

Put in. (This is a small deviation from the principle of wireless awareness of the secondary user's transparency to the primary user: in fact, due to the secondary operation, the primary can receive two acknowledgments for the same packet. In one embodiment, the packet will be considered correctly received if at least one acknowledgment is positive. In this case, the primary end 46 and the secondary transmitter 54 correctly decode the signal. The secondary 54 does not include the latter in its queue (if an ACK message is received from the end).

로부터(1차적인 것의 패킷들) 1-ε의 확률로 제 2 큐(52)로부터 (자신의 패킷들) 패킷들을 전송한다. 따라서, 릴레잉을 이용하지 않는 경우와 유사하 게, 여기서, 2차적 노드(54)는 다음의 목표들에 대한 자신의 전송 전력(

)및 채널들(

) 및 시스템 파라미터들(

)의 통계에 기반한 확률 ε를 선택할 수 있다는 것을 이해할 수 있을 것이다:(i) 1차적인 노드의 큐의 안정성을 유지(인식 노드의 "투명성"); (ii) 자신의 안정적인 스루풋을 최대화.In one embodiment, whenever the secondary node 54 detects an idle slot (and it does so with the error probability P _e in Equation 2), the secondary slot has a probability 51 of probability ε.

Send packets (its packets) from the second queue 52 with a probability of 1-ε from (packets of the primary). Thus, similar to the case of not using relaying, here, the secondary node 54 has its own transmit power for the following targets:

) And channels (

) And system parameters (

It will be appreciated that it is possible to select a probability ε based on the statistic of (i): (i) maintain the stability of the queue of the primary node (“transparency” of the recognition node); (ii) Maximize their stable throughput.

B. System Analysis

는 아래에 도출된다.The primary user does not keep in mind the behavior of the secondary user. Thus, in the absence of a relay, the primary user selects an average rate in the range described in equation (4). Maximum power allowed for the cognitive node to transmit to guarantee the stability of the queue of primary

Is derived below.

) 및 시스템 파라미터들(

):Theorem 3: Given Channel Parameters

) And system parameters (

):

이고 if

ego

이면,

If so,

을 사용할 수 있으며, 특히

는 자신의 최대치(

=1)로 설정될 수 있다;Secondary users can use arbitrary power without affecting the stability of the primary node's queue.

Can be used, especially

Is your maximum (

= 1);

이면, 인지 노드가 사용할 수 있는 최대 전력은 다음과 같다:if,

, Then the maximum power available to the cognitive node is:

(또는

)인 경우, 2차적인 사용자는 자신이 유휴 채널을 감지하고 제 1(또는 제 2 큐)가 선택될 때마다 "더미" 패킷들을 전송하는 것을 계속할 수 있으며, 따라서, 자신의 큐들이 비워져 있는지 여부에 관계없이 1차적인 사용자와 잠재적으로 간섭하는 것을 계속한다. 이는 아래에서 더 자세히 설명될 것이다.As in Sec III, due to the interaction of the queues of the system, the concept of the dominant system is used. In the embodiment shown in Fig. 5, the dominant system can be defined by the following modifications to the original system.

(or

), The secondary user may continue to transmit "dummy" packets whenever he detects an idle channel and the first (or second queue) is selected, thus whether his queues are empty Continue to potentially interfere with the primary user regardless. This will be explained in more detail below.

에 대하여 자신의 최대 전력(

)을 사용할 수 있으며 여기서, According to the direct result of Theorem 3, the secondary

Against their maximum power (

), Where:

이다.

to be.

에 의하여) 1차적인 것의 평균 출발 레이트를 향상시키며, 따라서, 인지 노드가 전체 전력으로 전송하도록 허용된 경우에 1차적인 사용자 스루풋의 범위를 증가시킨다.Therefore, relaying is (

Improve the average starting rate of the primary, thereby increasing the range of primary user throughput when the cognitive node is allowed to transmit at full power.

을 찾는 이슈는 (정리 3에 의한 제약 세트 하에서) 전송 전력

를 통한 최적화로 정리한다. 그러나, 여기의 분석은, 일 실시예에서, 2차적인 노드는 큐(51)

및 큐(52)

사이에서 어떠한 큐가 서빙될지를 구분하는 확률 ε를 선택하는 추가된 자유도를 지닌다는 사실에 의해 복잡해진다. 주요한 결과는 릴레잉이 없는 경우에 대한 정리3을 반영하는 정리 4에 요약되어있다.Sec. Secondary maximum stable throughput, similar to the absence of the relaying described above in III (D)

The issue of finding the transmit power (under the constraint set by theorem 3)

Clean up with optimization. However, the analysis here is, in one embodiment, that the secondary node is queue 51.

And queues (52)

It is complicated by the fact that it has an added degree of freedom to choose the probability ε that distinguishes which queues will be served between. The main results are summarized in Theorem 4 which reflects Theorem 3 for the absence of relaying.

) 및 주어진 시스템 파라미터들(

), 1차적인 사용자의 안정성이 보존된다(인지 노드의 "투명성")는 가정하에서, 인지 사용자의 최대 안정적인 스루풋은 다음의 최적화 방정식에 의해 정의된다Theorem 4: Given channel parameters

) And given system parameters (

Assuming that the primary user's stability is preserved ("transparency" of the cognitive node), the maximum stable throughput of the cognitive user is defined by the following optimization equation:

Where the throughput of the secondary link is

이며,

Is,

This depends on the following primary throughput:

Similar to Equation 8, this equation can be solved using the standard methods described in Bertsekas. Similar to Theorem 2, it is assumed by 1-P _fa to assume the nonzero probability P _fa of a false alarm in the secondary. This results in scaling of the achievable throughput μ _S (P _S ) of Equation 14.

의 안정성을 보장하기 위하여, 정리 3의 결과에 따라, 전송된 전력을 제한한다. 다른 면에서, 두 번째 제약은 큐

의 안정성을 보증하는데 필요한 확률 ε는 실제로 하나의 확률이라는 것을 (

으로 만들기 때문에, 동등한 경우는 제외된다)강제한다. 여기서, 릴레잉의 경우와는 반대로(예를 들어, 정리 2 참조), 최적화 방정식(수학식 13)은

의 안정성에 대한 (예를 들어, ε에 대한) 제약으로 인하여 일부

에 대하여 실행가능한 지점을 가지지 않을 수 있다. 예를 들어, 1차적인 송신기 및 2차적인 송신기 사이의 중단 가능성

은

보다 훨씬 작으며, 동시에 2차적인 송신기 및 1차적인 수신기 사이의 중단 확률

은 크다; 이러한 경우에, 대부분의 트래픽이 2차적인 수신기에 대한 낮은 도착 레이트로 인하여 오버플로우하는 큐

를 통해 통과하는 것이 명백하다.In equation (13), the first constraint is the primary queue

In order to ensure the stability of, according to the result of theorem 3, the transmitted power is limited. On the other hand, the second constraint is queue

The probability ε necessary to guarantee the stability of is actually one probability (

As it is made, it is excluded) Here, in contrast to the case of relaying (see, for example, theorem 2), the optimization equation (Equation 13)

Some due to constraints on the stability (eg, for ε)

It may not have a viable point for. For example, the possibility of interruption between the primary transmitter and the secondary transmitter

silver

Much smaller and at the same time the probability of interruption between the secondary transmitter and the primary receiver

Is large; In this case, the queue that most traffic overflows due to the low arrival rate for the secondary receiver

It is obvious to pass through.

)을 수락하도록 하는 것일 수 있다. 수학식 13과 유사한 최적화 방정식은 이러한 경우에 셋업될 수 있으며, 2차적인 것은 전력

및 확률들 ε및

를 선택하는 자유도를 가진다.A possible solution to this problem is that only a fraction of the packets that the secondary transmitter has successfully received (and are incorrectly decoded at the intended end) by the primary (

) May be accepted. An optimization equation similar to Eq. 13 can be set up in this case, the second being power

And probabilities ε and

Has the freedom to choose.

C. Numerical Results

대 1차적인 노드에 의해 선택된 스루풋

를 보여준다. 일 실시예에서, 파라미터들은,

이고,

인 것으로 선택된다. 여기서, "릴레이"에 대한 또는 "릴레이"로부터의 평균 채널 이득은 직접적인 1차적인 링크

보다 6dB 더 양호하다. 1차적인 최대 레이트는

이다(예를 들어, 수학식 4 참조). 도 6a는 비-릴레잉 모드에서 인지 노드가 최대

부근의 최대 전력만을 전송할 수 있으며(예를 들어, 수학식 7 참조), 릴레잉의 경우에서, 인지 노드가 수학식 4에 설명된 전체 범위의 최대 전력으로 전송할 수 있음을 보여준다. 또한, 도 6b에 도시된 바와 같이, 이러한 경우의 큐(51)

는 언제나 안정화가능하다(즉, 수학식 13에 의한 최적 확률 ε는 관심 영역에서 일(1)보다 작다). 마지막으로, 도 6c는 비-릴레잉 경우(정리 2)에 대한 최대 스루풋과 릴레잉 경우(정리 1)를 비교하며, 충분히 큰

에 대한 릴레잉의 관련된 이점들을 보여준다.The results of Theorem 4 are confirmed in FIGS. 6A, 6B, 6C and 7. 6A, 6B and 6C show the optimal power ε and maximum stable throughput obtained from the optimization power P _S , theorem 4

Throughput selected by primary node

Lt; / RTI > In one embodiment, the parameters are:

ego,

It is chosen to be. Here, the average channel gain for or from "relay" is a direct primary link.

6 dB better than. The primary maximum rate is

(See, for example, Equation 4). 6A shows that the cognitive node is maximum in non-relaying mode.

Only a nearby maximum power can be transmitted (see, for example, Equation 7), and in the case of relaying, it is shown that the cognitive node can transmit at the full range of maximum power described in Equation 4. Further, as shown in Fig. 6B, the queue 51 in this case

Is always stabilizeable (ie, the optimal probability ε by Equation 13 is less than one (1) in the region of interest). Finally, FIG. 6C compares the maximum throughput for the non-relaying case (theorem 2) and the relaying case (theorem 1), which is large enough

It shows the related advantages of relaying for.

은 릴레잉 케이스에서 고정된

에 대해 도시되어 있다(수학식 4 참조). 더 자세하게, 도착 레벨이 수학식 4에 대한 모델에 의해 제한되기 때문에, 임의의 작은

에 대하여,

이다. 여기서,

에 대해, 릴레잉을 이용하지 않는 인지 노드(54)의 스루풋은 0이며, 따라서, 여기의 도면은 릴레잉에 의해 획득되는 이득을 측정한다(도 6c참조). 다르게 설명되지 않는 한, 파라미터들은 위에서 예로서 선택된 것이다. 도 7은

와 관련하여 인지 노드(54)로 또는 인지 노드(54)로부터의 채널(

및

)의 품질을 (동일한 레이트로) 증가시키는 것은 릴레잉의 이득을 증가시키며, 그 이득은 직접 채널 이득이 더 작은 경우에 더 의미 있다(

=4dB 및 7dB인 두 곡선을 비교)는 것을 보여준다.The benefit of using relaying is further shown in FIG. 7, where the maximum throughput of the secondary user 54 is shown.

Fixed in the relay case

Is shown (see Equation 4). More specifically, since the arrival level is limited by the model for equation (4), any small

about,

to be. here,

For the perception node 54 that does not use relaying, the throughput is zero, so the figure here measures the gain obtained by relaying (see FIG. 6C). Unless otherwise stated, the parameters have been selected by way of example above. Figure 7

In connection with or from the cognitive node 54

And

Increasing the quality (at the same rate) increases the gain of relaying, which is more meaningful when the direct channel gain is smaller (

Comparing the two curves = 4dB and 7dB).

및 약한 직접 채널

이 존재하는 경우, 대부분의 트래픽들은 2차적인 송신기로 다시 보내진다. 한편으로, 이는 2차적인 것에 의한 전송에 대한 사용가능한 슬롯들을 증가시키는 것을 돕는다. 한편으로, 2차적인 송신기(54)로부터 1차적인 끝단(46)으로의 충분히 양호한 채널

이 지원되지 않는 경우, 2차적인 것은 1차적인 것으로부터 오는 추가적인 트래픽을 전달할 수 없다. 따라서, 최적화 방정식 13은 실행가능한 해법을 가지지 않으며, 2차적인 노드의 스루풋은 0이다. 이 방정식에 대한 해법은 위에 설명한 기술들의 구현일 수 있으며, 여기서 2차적인 것은 1차적인 것으로부터의 패킷들의 부분만을 받아들이는 것이다.Also the issues of convenience and stability described above are illustrated by FIG. 7. Good channel to secondary transmitter 54

And weak direct channel

If present, most traffic is sent back to the secondary transmitter. On the one hand, this helps to increase the available slots for transmission by the secondary. On the one hand, a sufficiently good channel from the secondary transmitter 54 to the primary end 46

If this is not supported, the secondary cannot carry additional traffic from the primary. Thus, optimization equation 13 has no viable solution and the throughput of the secondary node is zero. The solution to this equation may be an implementation of the techniques described above, where the secondary is to accept only portions of packets from the primary.

As described herein, cognitive interference channels, with one authorized (primary) link and one unlicensed (secondary or cognitive) link, were studied in a fixed fading environment. The operation of the secondary link is considered "transparent" for the primary link unless it affects the stability of the queues of the primary link. Considering this understanding and power allocation for the secondary transmitter, it has been shown that inevitable errors in detecting the operation of the primary link limit the maximum stable throughput that can be achieved by the secondary link. . To mitigate this problem, in an exemplary embodiment, a modification of the original cognitive interference channel has been proposed, where the secondary transmitter acts as a "transparent" relay for the primary's traffic. The numerical results show that the benefits of this solution depend (can depend on) the topology of the network.

While the exemplary embodiments described in the figures and the foregoing descriptions are now more preferred, it should be understood that these embodiments are provided only in exemplary meanings. The invention is not limited to the specific embodiments, but may be extended to various modifications, combinations and substitutions consistent with the scope and content of the appended claims.

VI . attachment

V. Additional Details Related to Theorem 1

로 전개되며, 여기 서

는 슬롯 t에서 도착하는 숫자를 나타내는 고정식 프로세스이며(

),

는 (고정식으로 증명된) 출발 프로세스이다. 함수 ()⁺는 (x)⁺ = max(x,0)으로서 정의된다. 지배적인 시스템의 정의를 개발하고, 재전송을 요구하는 중단을 리콜함으로써, 출발 프로세스는 다음과 같이 설명될 수 있다.The size of the primary queue in packets

Will be deployed here,

Is a fixed process representing the number arriving at slot t (

),

Is the starting process (fixed). The function () ⁺ is defined as (x) ⁺ = max (x, 0). By developing the definition of the dominant system and recalling the interruption requiring retransmission, the departure process can be described as follows.

는 인지 노드가 1차적인 사용자의 진행중인 동작을 정확하게 식별하는(그리고 간섭하지 않는) 이벤트를 나타낸다, 이는 1-P_e의 확률로 발생한다(수학식 2 참조);

는 1차적인 사용자에 의한 성공적인 전송(2차적인 사용자가 간섭하지 않는다고 가정함)의 이벤트를 나타내며, 이는

의 확률로 발생한다(수학식 1 참조);

는

의 컴플리먼트(compliment)이다;

는 2차적인 사용자가

의 확률로 간섭한다고 가정한 경우에 1차적인 사용자에 의한 성공적인 재전송의 이벤트를 나타낸다(아래의 수학식 19 참조). 수학식 16의 모든 이벤트들이 고정식 채널 프로세스에만 좌우되기 때문에, 출발 프로세스

는

에 의해 주어진 평균으로 고정적이다(유사한 분석에 대한 Luo 등을 참조). 수학식 1, 2 및 19를 치환하면, 평균 출발 레이트는 다음과 같다.Where 1 {?} Is an indicator function of the events enclosed in batches;

Denotes an event in which the cognitive node correctly identifies (and does not interfere with) the primary user's ongoing operation, which occurs with a probability of 1-P _e (see Equation 2);

Represents an event of a successful transmission by the primary user (assuming the secondary user does not interfere), which

Occurs with a probability of (see Equation 1);

The

Is a complement of;

Means that the secondary user

Assuming that the interference with the probability of is represented by the event of a successful retransmission by the primary user (see equation 19 below). Since all events of equation (16) depend only on the fixed channel process, the departure process

The

Fixed to the mean given by (see Luo et al. For similar analysis). Substituting Equations 1, 2 and 19, the average starting rate is as follows.

인 경우에 안정적이라는 결론을 내릴 수 있다(Loynes의 정리). 주어진 선택된 값

에 대하여, 이는 정의 1에 언급된 바와 같은 2차적인 사용자가 사용할 수 있는 전력에 대한 제한을 강요한다.If the process is stationary, the primary queue

It can be concluded that is stable (Lynes' Theorem). Given selected value

For, this imposes a restriction on the power available to the secondary user as mentioned in definition 1.

의 유도 B.

Induction of

In the case where the primary user's transmission is interfered by the cognitive transmission, the signal-to-interference-plus-noise ratio (SINR) at the primary receiver is:

See, IEEE Trans. Inform. Theory's vol. 51, no. 2, pages 506-522, using the results of M.Sharif and B. Hassibi's "On the capacity of MIMO broadcast channels with partial side information" (ie, equation 15), the cumulative distribution function is evaluated as do.

C. Additional Details Regarding Theorem 2

로서 전개되며, 여기서,

는 슬롯 t에서 도착하는 수를 나타내며(

),

는 (고정식인 것으로 증명될) 출발 프로세스이다. 후자는

으로써 설명될 수 있으며, 여기서,

는 2차적인 사용자에 의한(자신의 수신기로의) 성공적인 전송 이벤트이며, 이 확률은

이다(수학식 1참조);

는 슬롯 t가 인지 노드에 의한 전송에 대해 사용가능한 이벤트를 표시한다. 1차적인 사용자의 큐가 설계에 의해 고정적이기 때문에, (

에 의해 정의되는) 인지 노드에 대한 슬롯 사용가능성 프로세스는 고정식이다(Rao 등을 참조). 또한, 고려되는 MAC 모델로 인하여, 사용가능성이 확률은 1차적인 사용자의 큐에 0 패킷들을 가지는 확률에 대응한다(Little의 정리, 1987년 Prentice-Hall의 D. Bertsekas 및 R. Gallager의 Data networks참조, 이 전체 내용은 여기에 참조됨):The size of the queue (in packets) on the secondary node is

As, where

Represents the number arriving from slot t (

),

Is the starting process (which will prove to be fixed). The latter

It can be described as, wherein

Is a successful transmission event by the secondary user (to his receiver), and this probability is

(See Equation 1);

Indicates an event in which slot t is available for transmission by the acknowledgment node. Because the primary user's queue is fixed by design, (

The slot availability process for the cognitive node (as defined by) is fixed (see Rao et al.). Also, due to the MAC model considered, the availability probability corresponds to the probability of having zero packets in the primary user's queue (Little's Theorem, D. Bertsekas of Prentice-Hall, 1987 and Data networks of R. Gallager. See, this full text is referenced here):

는 다음 평균을 가지고 고정식이다in this case,

Is fixed to have the next average

에 의해 제한된다. 마지막 식은 전술한 바와 같은 전송 전력

를 선택하는데 있어서의 트레이드 오프를 명확하게 보여준다. 실제로, 수학식 1 및 17에 기반하여, 수학식 21의 두 개의 항이 전송 전력

에 대해 서로 반대 방향으로 좌우되며,

가 증가함에 따라 첫 째 항은 감소하고, 두 번째 항은 증가한다. 수학식 20, 1 및 17을 수학식 21에 대입하면, 대수적인 연산에 의해 수학식 9가 나타나며, 이는

의 오목함을 보여준다. 관련된 프로세스의 고정성을 보이면, 정리 2는 정리 1 및 Loynes의 정리의 직접적인 결과이다.And, using Loynes' theorem, stable throughput for secondary nodes

Limited by The last equation is the transmit power as described above.

Clearly shows the trade off in choosing. Indeed, based on Equations 1 and 17, the two terms of Equation 21

Are mutually opposite in respect of,

As the first term decreases, the second term increases. Substituting Equations 20, 1, and 17 into Equation 21, Equation 9 is represented by an algebraic operation,

Shows the concaveness. Theorem 2 is a direct result of theorem 1 and Loynes' theorem, showing the fixedness of the processes involved.

D. Additional Details Related to Theorem 3

는,

을 만족하고, 여기서,

, 2차적인 사용자가 간섭하지 않는다고 가정한 경우, 1차적인 사용자에 의한 성공적인 전송의 이벤트를 나타내며, 이는

의 확률로 발생하며, 여기서,Primary queue start rate

Quot;

And where,

, Assuming that the secondary user does not interfere, this indicates an event of a successful transmission by the primary user.

Occurs at the probability of, where

이다.

to be.

의 확률로) 의도된 끝단에 의해 또는 (

의 확률로) 인지 노드에 의해 정확히 수신되는 경우에 성공적인 것으로 여겨지는 경우에, 수학식 22의 중단 확률은 릴레잉이 없는 경우와 다르다. 따라서,

는 다음의 평균을 가지는 고정식 프로세스이다.Where the transmission by the primary node is a packet (

By the intended end or (

In case it is considered successful if it is correctly received by the cognitive node), the stopping probability of Equation 22 is different from the case without relaying. therefore,

Is a fixed process with the following mean:

를 정의하고, 수학식 15에 의해 판독된다. 수학식 15를 수학식 17의

와 비교하면, 우리는 협력(cooperation)이 P_S와는 독립적인 1차적인 사용자의 스루풋의 추가적인 증가를 초래한다는 결론을 내릴 수 있다. 여기서, 후자의 조건은 전달(delivery) 레이트

가 끝단으로의 실제 릴레잉된 트래픽이 아닌 1차적인 것으로부터 출발한 패킷들을 측정한다는 것을 반영한다. 위의 설명들로부터, 정리 3이 계속된다.By using Equation 22 in Equation 23, the average starting rate of the primary node considered in the relaying scenario is

Is defined and read by the equation (15). Equation 15 in Equation 17

In comparison, we can conclude that cooperation results in an additional increase in the throughput of primary users independent of P _S. Where the latter condition is a delivery rate

Reflects packets originating from the primary rather than the actual relayed traffic to the end. From the above descriptions, theorem 3 continues.

E. Additional Details Related to Theorem 4

는

로 전개되며, 여기서, 도착 레이트

는

과 같이 기록될 수 있다. Sec. III(D)의 표기를 따르면,

는 2차적인 송신기에서 1차적인 송신기에 의해 전송된 패킷의 성공적인 수신 이벤트를 표시하며, 이는

의 확률을 가진다. 페이딩 프로세스들의 고정성 및 1차적인 사용자의 큐들의 안정성으로부터, 도착 프로세스

는 아래의 평균을 지니며 고정식이다.First cue size

The

, Where, arrival rate

The

Can be written as: Sec. According to the notation of III (D),

Denotes a successful reception event of a packet sent by the primary transmitter at the secondary transmitter, which

Has a probability of From the fixedness of fading processes and the stability of the primary user's queues, the arrival process

Has a mean below and is fixed.

임이 이해된다(Sec. II 참조). 다른 면에서, 출발 프로세스는

이며,

는 i번째 타임 슬롯이 2차적인 것의 제 1 큐에 의한 전송에 대해 사용가능하며, 이는

의 확률로 발생한다는 이벤트를 나타낸다;

는 2차적인 노드에 의해 전송된 패킷의 1차적인 끝단으로부터의 성공적인 수신 이벤트이다. 따라서, 출발 프로세스

는 고정식이며, 이 평균은 다음과 같다.Here, Little's theorem (see Equation 20) is applied. Here, in deriving the equation (24),

Is understood (see Sec. II). On the other hand, the departure process

Is,

Is available for transmission by the first queue of the i th time slot being secondary, which is

Indicates an event that occurs with a probability of;

Is a successful receive event from the primary end of the packet sent by the secondary node. Thus, the departure process

Is fixed, and the mean is

의 안정성은

인 조건이 유지되는 경우에 보장되며(Loynes의 정리), 여기서, 차례로 수학식 24 및 25로부터 ε및 P_S에 대한 다음의 조건들을 이끌어낸다:cue

The stability of

Where the condition is maintained (Loynes' theorem), where, in turn, derives the following conditions for ε and P _S from equations (24) and (25):

는 다음의 평균을 지니며 고정식이다.Departure process

Is fixed with the mean:

을 최대화하는 것에 이르며, 이는 Loynes의 정리로부터

이기 때문이다. 최대 성취가능한 스루풋

은 ε의 감소함수이다. 따라서,

을 최대화하기 위하여, ε는 자신의 최소값과 동일하도록 설정된다(수학식 26참조), 따라서,

을 얻으며, 여기서,

는 수학식 14에 있다. 전술한 설명으로부터 정리 4가 계속된다.Optimizing the stable throughput of the cognitive node is related to ε and P _S

To maximize, from Loynes' theorem.

. Maximum Achievable Throughput

Is the decreasing function of ε. therefore,

In order to maximize, ε is set equal to its minimum value (see Equation 26).

, Where

Is in equation (14). Theorem 4 continues from the above description.

Claims (20)

In a cognitive radio system, A first transmitter configured to receive a plurality of first packets and to transmit the plurality of first packets to a first receiver over a wireless channel; And A second transmitter configured to receive a plurality of second packets, receive the plurality of first packets from the first transmitter, and detect an idle state of the wireless channel, If the idle state of the wireless channel is detected, the second transmitter, Sending at least one of the plurality of second packets to a second receiver based on the detected state of transmission of at least one of the plurality of first packets from the first transmitter to the first receiver; And select between relaying the at least one of the plurality of first packets to the first receiver, Based on the selection, configured to transmit the at least one of the plurality of second packets to the second receiver or to relay the at least one of the plurality of first packets to the first receiver. Cognitive wireless system. delete The method of claim 1, wherein the second transmitter, Relay the at least one of the plurality of first packets when it is detected that the transmission of the at least one of the plurality of first packets from the first transmitter to the first receiver is unsuccessful. Cognitive radio system. The method of claim 1, wherein the second transmitter, Refrain from relaying the at least one of the plurality of first packets when the transmission of the at least one of the plurality of first packets from the first transmitter to the first receiver is successful. Cognitive radio system. The method of claim 1, wherein the first transmitter, And acknowledge by the second transmitter an acknowledgment of the at least one of the plurality of first packets. The method of claim 5, wherein after the first transmitter transmits the at least one of the plurality of first packets, the first transmitter, The transmission of the at least one of the plurality of first packets is detected as unsuccessful and the first transmitter acknowledges acceptance of the at least one of the plurality of first packets by the second transmitter. And to resend the at least one of the plurality of first packets. The method of claim 1, The second transmitter has a transmit power, And the second transmitter is configured to ensure service stability of the first transmitter by controlling the transmit power. The method of claim 1, wherein the second transmitter, Optionally, transmitting the at least one of the plurality of second packets to the second receiver or relaying the at least one of the plurality of first packets to the first receiver, the service of the first transmitter. A cognitive radio system, configured to ensure stability. The method of claim 1, Wherein the first transmitter corresponds to an authorized user of the channel and the second transmitter corresponds to an unlicensed user of the channel. In a method for operating a cognitive radio system, Directing the plurality of first packets to a first transmitter to transmit a plurality of first packets to a first receiver over a wireless channel and to transmit the plurality of first packets to a second transmitter; Sending a plurality of second packets to the second transmitter for transmission to a second receiver over the wireless channel; Transmitting at least one of the plurality of first packets from the first transmitter to the first receiver; Detecting an idle state of the wireless channel at the second transmitter; And Upon detecting an idle state of the channel, in the second transmitter, Transmitting at least one of the plurality of second packets from the second transmitter to the second receiver based on the detected state of the transmission of the at least one of the plurality of first packets to the first receiver; Select between relaying the at least one of the plurality of first packets from the second transmitter to the first receiver, and based on the selection, the at least one of the plurality of second packets 2 transmitting from the transmitter to the second receiver or relaying the at least one of the plurality of first packets from the second transmitter to the first receiver A method of operating a cognitive radio system comprising a. delete 11. The method of claim 10, Wherein the first transmitter corresponds to an authorized user of the channel and the second transmitter corresponds to an unauthorized user of the channel. 10. A wireless system comprising at least two single-user communication links, wherein the first link is licensed to use spectral resources and the second link is not authorized. A first receiver configured to communicate with a first transmitter over a communication channel and receive first packets; and A second receiver coupled to a second transmitter by the communication channel, Upon detecting an idle state of the communication channel, the second transmitter, Based on the detected state of transmission of at least one of the first packets to the first receiver, sending second packets to the second receiver and the at least one of the first packets to the first receiver Choose between relaying, Based on the selection, sending the second packets to the second receiver or relaying the at least one of the first packets to the first receiver. The method of claim 13, wherein the second transmitter, Relaying the at least one of the first packets to the first receiver when it is detected that the transmission of the at least one of the first packets by the first transmitter to the first receiver is unsuccessful. The wireless system. The method of claim 13, Wherein the second transmitter controls the transmit power. The method of claim 13, Wherein the first transmitter is an authorized user of the communication channel and the second transmitter is an unlicensed user of the communication channel. The method of claim 13, And the second transmitter derives a maximum stable throughput for the communication channel. The method of claim 13, Wherein the first receiver detects random packet arrivals. The method of claim 13, And the second transmitter detects errors due to fading in the communication channel. 20. The method of claim 19, And if fading in the communication channel is detected, the second transmitter relays the at least one of the first packets to the first receiver.

Images