KR102179791B1 - Apparatus and method for adjusting duty cycle for coexistence of heterogeneous networks and recording medium for performing the method - Google Patents

Abstract

Disclosed is a duty cycle adjustment method for coexistence of heterogeneous networks in which the duty cycle of the long term evolution unlicensed (LTE-U) and a wireless body area network (WBAN) using overlapping frequency bands in an unlicensed band is adjusted to a duty cycle capable of optimal data processing using machine learning. According to the present invention, a duty cycle adjusting device determines an allocation ratio of a first cycle and a second cycle constituting a duty cycle according to a predetermined queue-value, receives a first throughput which is a part of raw data generated by a sensor device through a first network during a period of the first cycle, transmits a second throughput which is a part of refined data processed with the raw data to a relay server through a second network during a period of the second cycle, computes a fairness index according to the first throughput and the second throughput, and learns a Q value to adjust the allocation ratio.

Description

A duty cycle adjustment device and method for coexistence of heterogeneous networks, and a recording medium for performing the method.

The present invention relates to a duty cycle adjustment apparatus and method for coexistence of heterogeneous networks, and to a recording medium for performing the method, and more particularly, LTE-U and WBAN having overlapping frequency bands in an unlicensed band can coexist. The present invention relates to an apparatus and method for adjusting the duty cycle so as to be possible, and a recording medium for performing the method.

The content described in this section merely provides background information on the present embodiment and does not constitute the prior art.

The remote wireless healthcare system is a system that can reduce the burden on hospitals including doctors by reducing the number of patients visiting the hospital.

With the advancement of wireless communication technology, healthcare systems are able to collect, manage, and transmit medical information, including biometric signals, even in an accurate and non-invasive way.

The healthcare management system is based on an ultra-low-power sensor node that collects biometric data and an edge node that can transmit important signals to medical personnel in a timely manner, and a general mobile healthcare system uses two-hop communication. Consists of

The first hop communication is a wireless body area network (WBAN), which forms a link between a sensor node and an edge node.

At this time, ZigBee is widely used in WBAN because of its low data rate and long battery life.

The second hop communication configures communication between an edge node and an Internet system through a communication method such as a wireless local area network (WLAN) for additional processing and remote monitoring.

The main point of second hop communication is to deliver messages in a timely manner and to ensure quality of service (QoS) required for medical communication.

A collision occurs because the frequency band of WBAN and the frequency band of WLAN use the same 2.4 GHz ISM band (ISM Band, Industrial Science and Medical).

Since the transmission power of Wi-Fi, a WLAN technology, is about 5 to 20 dB stronger than that of ZigBee, a WBAN technology, WBAN transmission does not reach the energy detection (ED) threshold. Accordingly, an error is detected, and the WLAN always searches for a WBAN free channel and interferes with the transmission of WBAN.

In addition, the MAC layer time slot is 9 µs or 20 µs for WLAN and 320 µs for WBAN. Because WLAN has a shorter slot time than WBAN, it creates faster channel access than WBAN transmission.

Therefore, when WLAN and WBAN coexist, it is difficult to ensure that the WBAN stably transmits a signal to an edge node.

In addition, since WBAN and WLAN are independent technologies and cannot exchange control messages with each other, transmission cannot be scheduled in consideration of each other's transmission to avoid collision. Therefore, it can be seen that the physical parameters of the WLAN network greatly affect the performance of WBAN.

WLAN cannot be used as a coexistence technology for data communication in a healthcare system, but may perform second hop communication using "WLAN Like" communication such as LTE-U technology.

LTE (Long-Term Evolution) or simply LTE is a high-speed wireless data communication standard for mobile phones that is more advanced than High Speed Downlink Packet Access (HSDPA). It is called the 3.9th generation wireless communication standard along with HSPA+ (Evolved High Speed Packet Access), an evolved standard of HSDPA.

LTE is largely divided into a licensed spectrum that pays a fee and secures frequency details, and an unlicensed spectrum that can be used without paying a fee.

LTE-U is an LTE communication technology that uses the 5 GHz frequency band, which is an unlicensed band, and as the number of users of smartphones rapidly increases in recent years, it becomes more and more difficult to handle traffic, which is rapidly increasing mobile traffic. It is a technology to easily solve the difficulties of the communication network by distributing LTE wireless data traffic at low cost by utilizing Wi-Fi communication to accommodate the network.

Since the radio frequency band used by LTE-U and the frequency band of some technologies used in WBAN use the same frequency band, there is a problem that the performance of WBAN is deteriorated when LTE-U is operated.

Accordingly, there is a need for a method in which LTE-U and WBAN can transmit data while using the same frequency band.

Korean Patent Publication No. 10-2015-0077684 US Patent Publication No. 2018/0213379

An aspect of the present invention is a duty cycle adjustment device for coexistence of heterogeneous networks that adjusts the duty cycle of LTE-U and WBAN using overlapping frequency bands in an unlicensed band to a duty cycle capable of optimal data processing using machine learning. And a method and a recording medium for performing the method.

The technical problem of the present invention is not limited to the technical problem mentioned above, and other technical problems that are not mentioned will be clearly understood by those skilled in the art from the following description.

In the duty cycle adjustment method for coexistence of heterogeneous networks performed by at least one sensor device, at least one relay server, and a duty cycle adjusting device according to an embodiment of the present invention, the duty cycle adjusting device includes a predetermined Q value Determining an allocation ratio of the first cycle and the second cycle constituting the duty cycle according to the method; The duty cycle adjusting device receiving raw data generated by the sensor device from the sensor device side during the first cycle; The duty cycle adjustment apparatus comprises the steps of processing raw data to generate refined data; The duty cycle adjustment apparatus includes: transmitting the refined data to the relay server during the second cycle; And the duty cycle adjusting apparatus calculates a fairness index according to a first throughput, which is an amount of raw data received during the first cycle, and a second throughput, which is an amount of refined data transmitted during the second cycle, and calculates a Q value. It may include; learning step of adjusting the allocation ratio by learning.

In one embodiment, the ratio of the first cycle may be any one of 0.2, 0.4, 0.6, and 0.8, and the ratio of the second cycle may have a value obtained by subtracting the ratio of the first cycle from 1.

In one embodiment, determining the allocation ratio comprises: generating a random value; And comparing an arbitrary value with a predetermined probability value to select one of an exploration and an exploration.

In one embodiment, the step of selecting any one of exploration and exploration includes selecting an exploration if a random value is greater than a probability value, and arbitrarily determining the ratio of the first cycle and the second cycle. And, if the arbitrary value is less than the probability value, an expoit may be selected, and a ratio of the first cycle and the second cycle may be determined according to a predetermined ratio.

In one embodiment, the first network may be a Wireless Body Area Network (WBAN).

In an embodiment, the second network may be Long Term Evolution Unlicensed (LTE-U) using an unlicensed band.

In one embodiment, the fairness index may be proportional to a value obtained by a squared value of the sum of the first throughput and the second throughput, and may be inversely proportional to the sum of the squared value of the first throughput and the squared value of the second throughput.

In one embodiment, the step of adjusting the allocation ratio comprises: evaluating in one of the first process chart, the second process chart, the third process chart, and the fourth process chart according to the fairness index according to the fairness index. ; Learning a Q value according to a fairness index and a state; And determining whether to adjust the ratio of the first cycle and the second cycle by comparing a reference value whose Q value is predetermined. And transmitting the adjusted allocation ratio to the relay server using the licensed band.

In one embodiment, determining whether to adjust the ratio comprises adjusting the ratio of the first cycle and the second cycle when the Q value is greater than a predetermined reference value, and the first cycle when the Q value is smaller than the predetermined reference value. The ratio of the cycle and the second cycle can be maintained.

A computer program for performing a duty cycle adjustment method for coexistence of heterogeneous networks is recorded in a computer-readable recording medium according to an embodiment of the present invention.

In the duty cycle adjustment apparatus for coexistence of heterogeneous networks connected to at least one sensor device and a first network and connected to at least one relay server and a second network according to an embodiment of the present invention, a predetermined Q value A determination unit determining an allocation ratio of the first cycle and the second cycle constituting the duty cycle according to the present invention; A first communication unit for receiving raw data generated by the sensor device from the sensor device side during the first cycle; A processing unit that processes raw data to generate refined data; A second communication unit for transmitting the refined data to the relay server during the second cycle; The fairness index is calculated according to the first throughput, which is the amount of raw data received during the first cycle, and the second throughput, which is the amount of refined data transmitted during the second cycle, and adjusts the allocation ratio by learning the Q value. It may include; learning unit to do.

According to an aspect of the present invention described above, by separating the duty cycle so that LTE-U and WBAN coexist, the optimal duty cycle is set through queue-learning-based machine learning to determine the throughput and LTE-U of WBAN. There is an advantageous effect that can maximize the throughput of.

1 is a diagram illustrating a connection relationship between a sensor device, a duty cycle adjusting device, and a relay server according to an embodiment of the present invention.
2 is a block diagram showing a duty cycle adjustment apparatus according to an embodiment of the present invention.
3 is a diagram illustrating a duty cycle of LTE-U and WBAN according to an embodiment of the present invention.
4 is a diagram illustrating a relationship between a fairness index and a state according to an embodiment of the present invention.
5 is a flowchart illustrating a method of adjusting a duty cycle for coexistence of heterogeneous networks according to an embodiment of the present invention.
6 is a flowchart illustrating a process of determining an allocation ratio of a first cycle and a second cycle according to an embodiment of the present invention.
7 is a flowchart illustrating a process of adjusting an optimal allocation ratio according to a data throughput according to an embodiment of the present invention.

For a detailed description of the present invention to be described later, reference is made to the accompanying drawings, which illustrate specific embodiments in which the present invention may be practiced. These embodiments are described in detail sufficient to enable a person skilled in the art to practice the present invention. It should be understood that the various embodiments of the present invention are different from each other, but need not be mutually exclusive. For example, specific shapes, structures, and characteristics described herein may be implemented in other embodiments without departing from the spirit and scope of the present invention in relation to one embodiment. In addition, it is to be understood that the location or arrangement of individual components in each disclosed embodiment may be changed without departing from the spirit and scope of the present invention. Accordingly, the detailed description to be described below is not intended to be taken in a limiting sense, and the scope of the present invention, if appropriately described, is limited only by the appended claims, along with all scopes equivalent to those claimed by the claims. In the drawings, like reference numerals refer to the same or similar functions over several aspects.

Hereinafter, preferred embodiments of the present invention will be described in more detail with reference to the drawings.

1 is a diagram illustrating a connection relationship between a sensor device, a duty cycle adjusting device, and a relay server according to an embodiment of the present invention.

Referring to FIG. 1, a sensor device 100, a duty cycle adjustment device 300, and a relay server 500 may be included.

The sensor device 100 may be mounted on a user's body and collect information such as blood pressure, glucose level, heart rate, body temperature, and the like to generate raw data.

In addition, the sensor device 100 may transmit data through a wireless body area network (WBAN) such as a ZigBee communication technology that satisfies a low data rate and a long battery life.

The duty cycle adjustment device 300 may be connected to the sensor device 100 through a WBAN communication method such as ZigBee to receive raw data generated by the sensor device 100 and transmit the processed data to the relay server 500. .

In addition, the duty cycle adjustment apparatus 300 may perform a process for adjusting the duty cycle, and a detailed description thereof will be described below with reference to FIG. 2.

The relay server 500 may be connected to the duty cycle adjustment device 300 through an LTE-U communication method, and transmit data received from the duty cycle adjustment device 300 to a user such as a management server for healthcare or a medical staff. have.

In addition, since the relay server 500 is capable of carrier sense adaptive transmission (CSAT), it is possible to adjust a channel access time by receiving a control signal.

2 is a block diagram showing a duty cycle adjustment apparatus according to an embodiment of the present invention.

2, the duty cycle adjustment apparatus 300 according to an embodiment of the present invention includes a determination unit 310, a first communication unit 330, a processing unit 350, a second communication unit 370, and a learning unit. (390) may be included.

At this time, the duty cycle adjustment device 300 according to the present invention may have mobility or may be fixed. For example, the duty cycle adjustment apparatus 300 may be in the form of a server or an engine, and may be a smartphone, a device, an apparatus, a terminal, or a user equipment (UE). ), MS (mobile station), wireless device (wireless device), handheld device (handheld device) may be referred to as other terms.

In addition, the duty cycle adjustment apparatus 300 may be implemented by more components than the components shown in FIG. 2, and may be implemented by fewer components. Alternatively, the duty cycle adjustment device 300 includes at least two components of the determination unit 310, the first communication unit 330, the processing unit 350, the second communication unit 370, and the learning unit 390. It is integrated as a component of, so that one component can perform a complex function. Hereinafter, the above-described components will be described in detail.

The determiner 310 may determine an allocation ratio of the first cycle and the second cycle constituting the duty cycle according to a predetermined Q value.

3 is a diagram illustrating a duty cycle of LTE-U and WBAN according to an embodiment of the present invention.

The duty cycle is one of the machine control methods and is expressed as a percentage.

Referring to FIG. 3, it is the time T1 taken until the signal completes the on & off cycle, and is divided into an ON cycle T2 and an OFF cycle T3.

In the ON cycle (T2), LTE-U may transmit data while occupying the channel, and in the OFF cycle (T3), WBAN rather than LTE-U may occupy the channel and transmit data.

LTE-U can adjust cycle parameters for coexistence with a communication network using the same frequency band.

Since the first cycle means the OFF cycle T3 and the second cycle means the ON cycle T2, when the first cycle and the second cycle are combined, one duty cycle is obtained.

The determiner 310 may determine an allocation ratio of the first cycle and the second cycle according to the Q value.

In this case, the period of the duty cycle is 40 ms, and the ON cycle (T2) time of the LTE-U may be changed for each period of the duty cycle.

The ratio of the first cycle, which is the ON cycle (T2) of LTE-U, may have any one value of 0.2, 0.4, 0.6, and 0.8, and the ratio of the second cycle is 0.8 minus the ratio of the first cycle from 1 , 0.6, 0.4, and 0.2 may have a value.

For example, if the ratio of the first cycle is 0.2, LTE-U occupies the channel for 8 ms (40 ms × 0.2) in the first cycle, and WBAN is occupied for 32 ms (40 ms × 0.8) in the second cycle. It means to occupy the channel.

Q-Learning is primarily used to explore the optimal strategy for taking appropriate actions for the finite-Markov Decision Process, such as when the environment is unknown.

The Q value is initially designated as a specific value, but after data transmission/reception occurs, it may be designated by a value calculated according to the data throughput.

The determination unit 310 may generate a random value, compare the random value with a predetermined probability value, and divide it into an exploration and an exploration.

If the random value is greater than the probability value, the allocation ratio of the first cycle and the second cycle can be arbitrarily determined by selecting explore. If the random value is less than the probability value, the first cycle and the first cycle are selected by selecting expoit. The allocation ratio of two cycles can be determined according to the predetermined ratio described above.

The first communication unit 330 may receive a first throughput, which is a portion of raw data generated by the sensor device 100 through the first network, during the first cycle.

The sensor device 100 is at least one intelligent sensor implanted in the patient's body, and can generate raw data by detecting blood pressure, glucose level, heart rate, body temperature, and the like.

The first communication unit 330 is wirelessly connected to the sensor device 100 through a first network, and the first network means WBAN.

The first communication unit 330 may receive raw data from the sensor device 100 during the first cycle in which the WBAN occupies the channel, and the first throughput is the size of the raw data received during the first cycle. Means.

The processing unit 350 may process raw data to generate refined data.

Since raw data is generated by a plurality of sensors, the data is large, and since unnecessary information for data transmission is included, only necessary information must be transmitted, and processed refined data must be generated to increase transmission efficiency.

Therefore, refined data always has a feature of smaller data size than raw data.

The second communication unit 370 may transmit a second throughput, which is a part of the refined data, to the relay server through the second network during the second cycle.

The second communication unit 370 is wirelessly connected to the relay server 500 through a second network, and the second network refers to LTE-U.

The second communication unit 370 may transmit the refined data to the relay server 500 during the second cycle in which LTE-U occupies the channel, and the second throughput is the amount of refined data transmitted during the second cycle. Means size.

The learning unit 390 may calculate the fairness index according to the first processing amount and the second processing amount and learn the Q value to adjust the allocation ratio.

Q-Learning is primarily used to explore the optimal strategy for taking appropriate action for a finite-MDP decision process (finite-MDP), such as when the environment is unknown.

In reinforcement learning (RL) such as Q-Learning, the tradeoff between exploration and exploration is a characteristic of reinforcement learning (RL) that other types of learning do not have. to be.

Reinforcement learning (RL) is aimed at earning rewards, and in order to obtain significant rewards, it is necessary to obtain rewards by exploiting actions that have been attempted in the past, but exploration is required to select actions that give better rewards. need.

Therefore, learners must try a variety of actions, and in order to obtain the expected reward from a probabilistic point of view, they must try each action several times and continue to support in order to discover the action judged to be the best.

The framework of reinforcement learning (RL) essentially includes Policy, Reward, and Value Function, and may sometimes include a Q-learning model of Environment. .

Policy characterizes the learner's way in which the learner behaves at any given time. In general, a policy is a mapping from an explicit state of the environment to the action to be taken in that state.

This is compared to a set of behavior-response relationships in psychology. The policy could be a simple function or an index table, or it could contain significant computations, such as a pursuit procedure. Policy is the learner's essence of reinforcement learning (RL) because policy alone is sufficient to determine the learner's behavior.

The reward flag characterizes the purpose of the learning problem. At each time step, the environment determines a unique number of rewards.

Since the learner's primary purpose is to maximize the overall rewards they collect over the long term, rewards represent good and bad events for learners.

Reward is an important reason to change policy in a certain state. If the action chosen by the policy results in low rewards, the policy can be changed in the future to choose some other action for that condition.

Reward represents better in the immediate sense, but Value Function represents what is ultimately best.

As such, the value function of a specific state is the sum of rewards that a learner can collect in the long run starting from the initial state.

For example, one state can pay low and fast rewards, but at the same time it has a high value function because another state that pays high rewards can often follow.

The value function is taken into account when making decisions and making evaluations. Action choices are based on value judgment. Since these behaviors provide the learners with the best rewards in the long term, we pursue behaviors with the highest values, not the best rewards. In fact, the most important factor in most reinforcement learning (RL) algorithms is the technique of seamless value estimation.

Model environment is an optional component of the Q-Learning framework.

The environmental model can reflect the behavior of the environment, and is generally a suggestion of how the environment will behave.

For example, given a state and an action, the model can predict the next state and the next reward that will result. The model is used for planning and refers to a method of establishing a set of actions taking into account the foreseeable future situations before actual experiences.

Reinforcement learning (RL) relies strongly on the concept of state as input to policy and value functions.

Informally, we can think of a state as a flag passed to the learner with a little sense of how it exists at a particular time. Much of the reinforcement learning (RL) technique revolves around value function evaluation.

However, it is not entirely important to do this to deal with reinforcement learning (RL) problems.

In the duty cycle adjustment apparatus 300, the learning unit 390 learns using a Q-learning algorithm, and compensation is provided by a Jane fairness index (μ ^β , Jain fairness index).

At this time, the Jane fairness index (μ ^β , Jain fairness index) is calculated according to Equation 1 below.

[Equation 1]

Here, β denotes a duty cycle period during which LTE-U occupies a channel, that is, a period of the second cycle, and thus (1-β) denotes a period of the first cycle.

은 제 1 사이클의 기간 동안 WBAN을 통해 수신되는 원시 데이터의 처리량을 의미하며, 아래 수학식 2에 따라 계산된다.

Denotes the throughput of raw data received through WBAN during the first cycle, and is calculated according to Equation 2 below.

[Equation 2]

= 수신된 원시 데이터의 처리량 / ((1-β) * 40㎳)

= Throughput of received raw data / ((1-β) * 40 ms)

는 제 2 사이클의 기간 동안 LTE-U를 통해 송신되는 정제 데이터의 처리량을 의미하며, 아래 수학식 3에 따라 계산된다.

Denotes the throughput of refined data transmitted through LTE-U during the second cycle, and is calculated according to Equation 3 below.

[Equation 3]

= 송신된 정제 데이터의 처리량 / (β * 40㎳)

= Throughput of transmitted refinement data / (β * 40 ms)

4 is a diagram illustrating a relationship between a fairness index and a state according to an embodiment of the present invention.

In the Q-learning algorithm, the state is a finite set, meaning the current situation.

Referring to FIG. 4, a correspondence relationship between states, which is the next threshold value defined based on the fairness index, can be seen.

The state is divided into a 1st process chart, a 2nd process chart, a 3rd process chart, and a 4th process chart.The first process chart is an extremely unfair state, and the second process chart is unfair. ) One state, a third process chart means a fair state, and a fourth process chart means a highly fair state, respectively.

If the value of the Jane's fairness index (μ ^β ) is 0 or more and less than 0.4, it corresponds to the first process chart, and when it is 0.4 or more and less than 0.6, it corresponds to the second process chart.

If the value of the Jane fairness index (μ ^β ) is 0.6 or more and less than 0.8, it corresponds to the third process chart, and when it is 0.8 or more and less than 1, the state may be determined as corresponding to the fourth process chart.

)과 단위 시간당 송신된 정제 데이터의 처리량(

)이 적절하게 분배되는 것을 의미한다.In other words. The larger the Jane Fairness Index (μ ^β ), the more fair it is. Fairness means the throughput of raw data received per unit time (

) And throughput of refined data transmitted per unit time (

) Is properly distributed.

The Q value means the expected value of the utility that will be brought by performing a given action in a given state. In this case, an operation having the highest Q value may be determined as an optimal operation.

The Q value depends on the compensation according to the fairness index above, and is calculated according to Equation 4 below.

[Equation 4]

Q (state, action) ← (1 α) Q (state, action) + α [r + γ max _action′ Q (state′, action′)]

Here, α means a learning rate, and γ means a discount factor. State` represents the next state, and action` means the action with the highest Q value in the state`.

The learning unit 390 compares the Q value with a reference value for the number of repetitions, and if the Q value is less than the reference value, the process of maintaining the duty cycle and adjusting the duty cycle may be terminated.

The learning unit 390 compares the Q value with a reference value for the number of repetitions, and if the Q value is greater than the reference value, the duty cycle needs to be adjusted, and thus a new allocation ratio is transmitted to the relay server 500. Duty cycle can be adjusted.

5 is a flowchart illustrating a method of adjusting a duty cycle for coexistence of heterogeneous networks according to an embodiment of the present invention.

The duty cycle adjustment method for coexistence of heterogeneous networks according to the present invention may be performed by the duty cycle adjustment apparatus 300 for coexistence of heterogeneous networks according to the present invention. To this end, the duty cycle adjustment apparatus 300 for coexistence of heterogeneous networks according to the present invention may be pre-installed with an application (software) for performing each step of configuring a method for estimating a data classification rule to be described later. For example, in the user's computer, a platform for the duty cycle adjustment method for coexistence of heterogeneous networks according to the present invention may be pre-installed in the form of software, and the user executes software installed on the computer to perform data Various services provided by the classification rule estimation method can be provided.

The duty cycle adjustment method for coexistence of heterogeneous networks according to an embodiment of the present invention is performed on substantially the same configuration as the duty cycle adjustment apparatus 300 for coexistence of heterogeneous networks shown in FIGS. 1 and 2. The same reference numerals are assigned to the same components as those of the duty cycle adjusting apparatus 300 for coexistence of heterogeneous networks of 2, and repeated descriptions will be omitted.

Referring to FIG. 5, a heterogeneous network performed by a duty cycle adjustment device 300 connected to at least one sensor device 100 and a first network, and connected to at least one relay server 500 and a second network You can see how to adjust the duty cycle for coexistence.

The duty cycle adjustment apparatus 300 may determine an allocation ratio of the first cycle and the second cycle constituting the duty cycle according to a predetermined Q value (S100).

The duty cycle adjustment apparatus 300 may receive a first throughput for a period of a first cycle from the sensor apparatus 100 side through a first network of raw data generated by the sensor apparatus (S200 ).

The duty cycle adjustment device 300 may process raw data to generate refined data (S300).

The duty cycle adjustment apparatus 300 may transmit the refined data generated from the raw data to the relay server 500 through the second network for the second throughput during the second cycle (S400).

The duty cycle adjustment apparatus 300 may calculate a fairness index according to the first processing amount and the second processing amount, and adjust the allocation ratio by learning the Q value (S500).

6 is a flowchart illustrating a process of determining an allocation ratio of a first cycle and a second cycle according to an embodiment of the present invention.

Referring to FIG. 6, the duty cycle adjustment apparatus 300 is a step of generating a random value (S110), comparing the random value with a predetermined probability value to select any one of exploration and exploration. It can be seen that step S130 and steps S150 and S160 of determining the allocation ratio according to the selected value are included.

The duty cycle adjustment device 300 may generate an arbitrary value (S110).

The duty cycle adjustment apparatus 300 may select any one of exploration and exploration by comparing an arbitrary value with a predetermined probability value (S130).

If the random value is greater than the probability value, the duty cycle adjustment device 300 may arbitrarily determine the allocation ratio by selecting an exploration (S150).

When the random value is less than the probability value, the duty cycle adjustment apparatus 300 may select an expoit and determine the allocation ratio according to a predetermined ratio (S160).

7 is a flowchart illustrating a process of adjusting an optimal allocation ratio according to a data throughput according to an embodiment of the present invention.

Referring to FIG. 7, the duty cycle adjustment apparatus 300 may know a process of calculating a fairness index according to a first throughput and a second throughput, learning a Q value, and adjusting the allocation ratio.

The duty cycle adjustment apparatus 300 may calculate a fairness index according to the first processing amount and the second processing amount (S510).

In this case, the fairness index may be proportional to a value obtained by squaring the sum of the first processing amount and the second processing amount, and may be inversely proportional to the sum of the square value of the first processing amount and the square value of the second processing amount.

The duty cycle adjustment device 300 may evaluate in any one of extremely unfair, unfair, fair, and highly fair according to the fairness index (S520), and the fairness index And it is possible to learn the Q value according to the state (S530).

The duty cycle adjustment apparatus 300 may determine whether to adjust the allocation ratio by comparing a reference value in which the Q value is predetermined (S540).

When the Q value is greater than a predetermined reference value, the allocation ratio may be adjusted (S550).

When the Q value is less than a predetermined reference value, the allocation ratio may be maintained (S560).

The duty cycle adjustment apparatus 300 may transmit the allocation ratio to the relay server side by using the licensed band for control of the relay server 500 (S570).

Such a method of adjusting a duty cycle for coexistence of heterogeneous networks may be implemented as an application or implemented in the form of program instructions that may be executed through various computer components and recorded in a computer-readable recording medium. The computer-readable recording medium may include program instructions, data files, data structures, etc. alone or in combination.

The program instructions recorded in the computer-readable recording medium may be specially designed and constructed for the present invention, and may be known and usable to those skilled in the computer software field.

Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical recording media such as CD-ROMs and DVDs, and magnetic-optical media such as floptical disks. media), and a hardware device specially configured to store and execute program instructions such as ROM, RAM, flash memory, and the like.

Examples of the program instructions include not only machine language codes such as those produced by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like. The hardware device may be configured to operate as one or more software modules to perform processing according to the present invention, and vice versa.

According to an embodiment of the present invention described above, by separating the duty cycle so that LTE-U and WBAN coexist, the optimal duty cycle is set through queue-learning-based machine learning, so that the throughput and LTE- There is an advantage in that the throughput of U can be maximized.

Although the above has been described with reference to embodiments, those skilled in the art will understand that various modifications and changes can be made to the present invention without departing from the spirit and scope of the present invention described in the following claims I will be able to.

100: sensor device
300: duty cycle adjustment device
310: decision
330: first communication unit
350: processing part
370: second communication unit
390: Learning Department
500: relay server

Claims (11)

In the duty cycle adjustment method for coexistence of heterogeneous networks performed by a duty cycle adjusting device connected to at least one sensor device and a first network, and connected to at least one relay server and a second network,
The duty cycle adjusting device includes determining an allocation ratio of a first cycle and a second cycle constituting a duty cycle according to a predetermined Q value;
The duty cycle adjusting device receiving raw data generated by the sensor device from the sensor device side during the first cycle;
The duty cycle adjusting device processing the raw data to generate refined data;
The duty cycle adjusting device transmitting the refinement data to the relay server during the second cycle; And
The duty cycle adjustment apparatus calculates a fairness index according to a first throughput, which is an amount of raw data received during the first cycle and a second throughput, which is an amount of refined data transmitted during the second cycle, and the A learning step of adjusting the allocation ratio by learning the Q value; and
The fairness index is,
The duty cycle for coexistence of heterogeneous networks is proportional to the squared value of the sum of the first throughput and the second throughput, and inversely proportional to the sum of the squared value of the first throughput and the squared value of the second throughput. How to adjust.
The method of claim 1,
The ratio of the first cycle is any one of 0.2, 0.4, 0.6, and 0.8, and the ratio of the second cycle has a value obtained by subtracting the ratio of the first cycle from 1. How to adjust the duty cycle for.
The method of claim 2,
The step of determining the allocation ratio,
Generating an arbitrary value; And
Comparing the random value with a predetermined probability value to select one of an exploration and an exploration; and a duty cycle adjustment method for coexistence of heterogeneous networks.
The method of claim 3,
The step of selecting any one of the exploration and utilization,
If the random value is greater than the predetermined probability value, the allocation ratio is arbitrarily determined by selecting an explorer,
When the random value is less than the predetermined probability value, an expoit is selected and the allocation ratio is determined according to a predetermined ratio.
The method of claim 1,
The first network,
A method of adjusting a duty cycle for coexistence of heterogeneous networks, characterized in that it is a WBAN (Wireless Body Area Network).
The method of claim 1,
The second network,
Duty cycle adjustment method for coexistence of heterogeneous networks, characterized in that the LTE-U (Long Term Evolution Unlicensed) using an unlicensed band.
delete The method of claim 1,
Adjusting the allocation ratio,
Evaluating in one of a first process chart, a second process chart, a third process chart, and a fourth process chart according to the fairness index;
Learning the Q value according to the fairness index and the state;
Determining and adjusting whether or not the allocation ratio is adjusted by comparing the Q value with a predetermined reference value; And
Transmitting the allocation ratio to the relay server using the licensed band; duty cycle adjustment method for coexistence of heterogeneous networks comprising a.
The method of claim 8,
The step of determining whether to adjust the ratio,
If the Q value is greater than a predetermined reference value, the allocation ratio is adjusted,
When the Q value is less than a predetermined reference value, the allocation ratio is maintained.
A computer-readable recording medium having a computer program recorded thereon for performing the duty cycle adjustment method for coexistence of heterogeneous networks according to any one of claims 1 to 6, 8, and 9.
In the duty cycle adjustment apparatus for coexistence of heterogeneous networks connected to at least one sensor device and a first network, and connected to at least one relay server and a second network,
A determining unit determining an allocation ratio of the first cycle and the second cycle constituting the duty cycle according to a predetermined Q value;
A first communication unit for receiving raw data generated by the sensor device from the sensor device side while the first cycle is in progress;
A processing unit processing the raw data to generate refined data;
A second communication unit for transmitting the refined data to the relay server during the second cycle;
The fairness index is calculated according to the first throughput, which is the amount of raw data received during the first cycle and the second throughput, which is the amount of refined data transmitted during the second cycle, and learns the Q value, Includes; a learning unit for adjusting the allocation ratio,
The fairness index is,
The duty cycle for coexistence of heterogeneous networks is proportional to the squared value of the sum of the first throughput and the second throughput, and inversely proportional to the sum of the squared value of the first throughput and the squared value of the second throughput. Adjustment device.

Images