CN107592174A - A kind of high-efficiency frequency spectrum cognitive method in intelligent grid communication - Google Patents

Abstract

The invention discloses a high-efficiency spectrum sensing method and system in smart grid communication. The system is composed of a smart meter terminal and a substation control center. Cognitive radio technology is used to dynamically utilize the idle spectrum of the TV band to realize the communication between the terminal and the control center. The two-way communication; the method includes: the control center uses a three-time exponential smoothing model to predict the occupancy of different TV channels according to the stored historical detection data, and selects a group of channels with a lower occupancy according to the prediction results to form A smaller set of channels to be detected. The smart meter terminal uses the known characteristics of the PN sequence code of the digital TV signal and adopts a correlation detection algorithm to detect only this small channel set. Compared with the traditional spectrum sensing method, the method not only has high detection accuracy, but also detects The time is short, and the probability of detecting an idle channel can be improved, and the efficiency of spectrum sensing in smart grid communication is improved.

Description

An Efficient Spectrum Sensing Method for Smart Grid Communications

technical field

The invention relates to the field of smart grid communication, and mainly solves the problem of high-efficiency spectrum detection when the smart grid is wirelessly connected, and specifically relates to a prediction-based high-efficiency spectrum sensing method in smart grid communication.

Background technique

The smart grid is based on the traditional grid framework, adopts advanced sensing, measurement and equipment and efficient control methods, and uses high-speed, two-way, integrated communication network and computer information network as a platform to achieve a high degree of integration of power, information and business , making the power grid more reliable and secure. Compared with the traditional power grid, the smart grid guarantees the high-speed operation of the network structure during the transmission process, can realize the interaction with users, meet the diverse needs of users and provide value-added services to users, thus giving full play to the comprehensive advantages of the smart grid. Function. In the construction of smart grid, the role of communication technology is very important, and the establishment of efficient, real-time and integrated communication system is the basis for realizing smart grid.

In the smart grid system, the user terminal electricity meter is connected to the communication part of the local control center, which is usually called the last mile connection. Compared with other existing wireless communication systems, the last mile wireless access of the smart grid has its own characteristics. First, the transmitter and receiver are fixed, which alleviates many problems in mobile communication, such as fast channel changes, frequent switching, etc.; second, smart grid needs to achieve more end-user interaction, such as real-time meter monitoring. To realize the above requirements, the challenges faced are as follows: (1) The time delay requirement is high. If the control center does not receive the message input from the terminal sensor due to excessive time delay, the control center may send a wrong control message to the smart meter terminal. (2) The amount of transmitted information is large. Due to the huge number of end users, the amount of information transmission must be large, and the communication network should be able to carry the transmission of large data volumes. The above factors must be considered when designing a smart grid last mile access system for reliable and efficient communication.

The present invention considers the situation that the communication link between the electric meter of the user terminal and the local control center adopts a wireless link. At this time, the primary task of the last mile access is to solve the problem of wireless spectrum selection. According to the development of existing technologies, there are currently three alternative spectrum solutions: (1) the power grid system applies to the spectrum management department for a dedicated licensed spectrum; (2) uses public spectrum without authorization; (3) adopts cognitive radio technology, Dynamic utilization of licensed spectrum. Due to the current shortage of spectrum resources, it is not realistic to allocate dedicated spectrum for the power grid system. However, there are many application technologies in the public unlicensed frequency band, and the spectrum interference is relatively large, which cannot meet the needs of the power grid system for reliable and efficient communication. Therefore, the present invention mainly focuses on the communication problem when using the cognitive radio technology to realize the last mile access of the smart grid. Cognitive radio is an intelligent software-defined radio technology that, based on radio environment detection, adjusts the portion of the radio spectrum it configures to facilitate efficient, reliable, and dynamic use of underutilized licensed spectrum. IEEE802.22 is the first standardization project for cognitive radio based on opportunistic utilization of TV spectrum bands. Using IEEE 802.22 to realize the last mile access scheme of the smart grid has the following advantages: (1) Cognitive radio makes full use of unused TV frequency bands, which can greatly alleviate the shortage of spectrum resources and meet the wireless access requirements of the smart grid. Spectrum needs. (2) A transmission rate of up to tens of Mbps can be realized. (3) Due to the long-term support of the transmission characteristic band of TV, the coverage area can reach 100 kilometers. (4) Cognitive radio is an adaptive, programmable and flexible technology that can realize dynamic utilization of idle spectrum in TV band.

According to the application mode of cognitive radio technology, when flexibly using the white space of TV frequency band to realize the last mile access of smart grid communication, each cognitive device (including smart meter terminal and power grid control center) should first use the TV frequency band in its geographical location The situation is detected, and then the communication parties wait for an opportunity to access the unused TV channel after negotiation. In the smart grid communication system based on cognitive radio described in the present invention, considering that the storage and computing capabilities of the wireless module of the smart meter terminal are limited, while the power grid control center has the characteristics of strong storage and computing capabilities, in order to improve the power of the meter terminal Spectrum sensing efficiency, the present invention proposes an efficient spectrum sensing method based on TV channel occupancy prediction.

Contents of the invention

Based on the above analysis, in order to improve the spectrum sensing efficiency of the meter terminal in the smart grid communication system, the present invention proposes the following efficient spectrum sensing method and system:

The present invention only considers the use of cognitive radio to solve the problem of accessing the last mile of smart grid communication. The system is composed of smart meter terminals, terminal substation control centers and regional substation control centers. Cognitive radio technology is used to dynamically utilize the idle time of the TV band. The spectrum, the smart meter terminal and the terminal substation control center are cognitive devices. The cognitive device first detects the TV channel before communicating and obtains a list of available channels. After negotiation, the communication parties select a common available channel for communication.

The method includes a correlation detection method based on digital TV signal PN sequence codes and a channel occupancy prediction and selection method based on a cubic exponential smoothing model.

The basic idea of a correlation detection method based on the digital TV signal PN sequence code is: according to the known digital TV signal frame header PN sequence mode, the smart meter terminal generates the PN sequence locally, and compares the received signal with the local PN sequence Correlation operation to obtain the relevant signal; compare the peak value of the relevant signal with the judgment threshold to obtain the judgment result of the TV signal being idle or occupied. The essence of the method is to use the known characteristics of the PN sequence code of the digital TV signal and adopt the correlation detection method. Compared with the traditional detection method, the detection reliability is high and the anti-noise performance is good.

The basic idea of a channel occupancy prediction and selection method based on the cubic exponential smoothing model is: the terminal substation control center uses the cubic exponential smoothing model to predict the TV channel occupancy of different meter terminals according to the stored historical detection data, and According to the prediction result, a group of channels with lower occupancy is selected to form a smaller set of channels to be detected. The essence of this method is to use the characteristics of strong storage and computing capabilities of the smart grid control center. The smart grid control center predicts the occupancy of the TV channel based on the historical detection results of the smart meter terminal. Spectrum detection is performed on lower channels, which saves spectrum sensing time and increases the probability of detecting idle spectrum.

Specifically, an efficient spectrum sensing method based on correlation detection and prediction includes the following steps:

Step 101) The terminal substation control center divides all smart meter terminals within its management scope into several clusters according to geographical locations;

Step 102) After the system is started, the smart meter terminal of the i-th cluster uses the correlation detection method to detect the occupancy of the TV channels at the location, and sets the TV channel set as A, which includes N channels. The smart meter terminal selects a channel j, and uses the correlation detection method to detect the occupancy of the channel j. The specific implementation of the correlation detection method is as follows: suppose the received signal is r(n), and the length of the PN sequence of channel j is L, then the correlation calculation The result is recorded as:

Compare the peak value R _j-max (n) of the related signal R _j (n) with the decision threshold λ, if R _j-max (n)>λ, it is determined that the TV channel is occupied, otherwise it is determined that the TV channel is idle.

Step 103) The smart meter terminals in the i-th cluster send the detection results of the N channels to the control center;

Step 104) The control center predicts and sorts the occupancy of the N channels in the set A according to the historical detection data of the stored i-th cluster, and the specific prediction and sorting methods are as follows:

Step 104-1) The control center conducts statistical analysis on the historical detection data for a certain channel in the television channel set to be detected, and obtains the time series of the channel occupancy, and sets the channel occupancy sequence as {s _t };

Step 104-2) Use the cubic exponential smoothing model to predict the occupancy time series {s _t }, and obtain the prediction result y _t of the channel occupancy. Triple exponential smoothing prediction is re-smoothing on the basis of quadratic smoothing. Its basic idea is: the predicted value is the weighted sum of previous observations, and different data are given different weights. New data is given greater weight, and old data is given weight. lesser right. Its prediction formula is:

The prediction model of triple exponential smoothing method is:

in,

Step 104-3) The control center sorts the channels from low to high according to the predicted spectrum occupancy, and forms the top channels into a smaller subset of channels to be detected. According to the order of occupancy from low to high, the first M (M<N) channels are taken out to form a new subset of channels to be detected A1, A1∈A;

Step 105) The control center broadcasts the channel subset A1 to be detected to the terminal in the ith cluster;

Step 106) The terminal in the i-th cluster only performs spectrum detection on the channels in the channel subset A1 to be detected, obtains the detection results, and reports the detection results of the M channels to the control center. The next step is to transfer to step 104 to continue execution.

Description of drawings

Accompanying drawing 1 is the structural diagram of the smart grid communication system based on cognitive radio proposed by the present invention.

Figure 2 is an efficient spectrum sensing method based on correlation detection and prediction.

Detailed ways

The scheme proposed by the present invention will be described in detail below in conjunction with accompanying drawings 1 and 2 .

Figure 1 is a structural diagram of a smart grid communication system based on cognitive radio. The system consists of a smart meter terminal, a terminal substation control center and a regional substation control center. The present invention only considers the last mile wireless access of smart grid communication , using cognitive radio technology to dynamically utilize the idle spectrum in the TV band. The smart meter terminal and the terminal substation control center are cognitive devices. Before communication, the cognitive device first detects the TV channel and obtains a list of available channels. Select a common available channel for communication. The present invention only considers the efficient spectrum detection problem of the system shown in FIG. 1 , and uses the method shown in FIG. 2 to perform spectrum detection.

The efficient spectrum sensing method based on correlation detection and prediction shown in Figure 2 includes the following steps:

Step 101) In the system shown in Figure 1, the terminal substation control center divides all smart meter terminals within its management scope into several clusters according to their geographic locations;

Step 102) After the system is started, the smart meter terminal of the i-th cluster uses the correlation detection method to detect the occupancy of the TV channels at the location, and sets the TV channel set as A, which includes N channels. The smart meter terminal selects a channel j, and uses the correlation detection method to detect the occupancy of the channel j. The specific implementation of the correlation detection method is as follows: suppose the received signal is r(n), and the length of the PN sequence of channel j is L, then the correlation calculation The result is recorded as:

Perform a peak search on the calculation result of the above formula, and compare the highest peak R _j-max (n) with the decision threshold γ. If R _j-max (n)>γ, it is determined that the TV channel is occupied, otherwise it is determined that the TV channel is idle. Among them, the value of the decision threshold γ is determined by the false alarm probability _Pf , and the relationship between the two is:

in, is the average noise power, is the average power of the PN sequence.

Step 103) The smart meter terminals in the i-th cluster send the detection results of the N channels to the control center;

Step 104) The control center predicts and sorts the occupancy of the N channels in the set A according to the historical detection data of the stored i-th cluster, and the specific prediction and sorting methods are as follows:

Step 104-1) The control center conducts statistical analysis on the historical detection data for a certain channel in the television channel set to be detected, and obtains the time series of the channel occupancy, and sets the channel occupancy sequence as {s _t };

Step 104-2) Use the cubic exponential smoothing model to predict the occupancy time series {s _t }, and obtain the prediction result y _t of the channel occupancy. Triple exponential smoothing prediction is re-smoothing on the basis of quadratic smoothing. Its basic idea is: the predicted value is the weighted sum of previous observations, and different data are given different weights. New data is given greater weight, and old data is given weight. lesser right. Its prediction formula is:

The prediction model of triple exponential smoothing method is:

in,

Step 104-3) The control center sorts the channels from low to high according to the predicted spectrum occupancy, and forms the top channels into a smaller subset of channels to be detected. According to the order of occupancy from low to high, take out the first M (M<N) channels to form a new subset of channels to be detected A1, A1 _∈ A;

Step 105) The control center broadcasts the channel subset A1 to be detected to the terminal in the ith cluster;

Step 106) The terminal in the i-th cluster only performs spectrum detection on the channels in the channel subset A1 to be detected, obtains the detection results, and reports the detection results of the M channels to the control center. The next step is to transfer to step 104 to continue execution.

Claims (3)

1. An effective spectrum sensing method in smart grid communication is characterized by comprising the following operation steps:

step 101) dividing all intelligent electric meter terminals in a terminal substation management range into a plurality of clusters according to geographical positions;

step 102), after the system is started, detecting the television channel occupation condition of the position of the intelligent electric meter terminal of the ith cluster by adopting a correlation detection method, and setting a television channel set as A, wherein the television channel set comprises N channels;

step 103) the intelligent electric meter terminal in the ith cluster sends the detection results of the N channels to the control center;

step 104) the control center performs data mining and prediction on the occupancy rates of the N channels in the set A according to the stored historical detection data of the ith cluster, the occupancy rates are sorted from low to high, the front M (M < N) channels are taken out to form a new channel subset A1 to be detected, and A1 belongs to A;

step 105) the control center broadcasts the channel subset A1 to be detected to the ith cluster terminal;

and 106) the ith intra-cluster terminal only carries out spectrum detection on the channels in the channel subset A1 to be detected to obtain detection results, and reports the detection results of the M channels to the control center. The next step is to step 104 to continue execution.

2. The method for spectrum sensing for television signal correlation detection as claimed in claim 1, wherein said step 102) further comprises:

step 102-1), according to a known PN sequence mode of a digital television signal frame header, the intelligent electric meter terminal locally generates a PN sequence, and performs correlation operation on the received signal and the local PN sequence to obtain a correlation signal. Assuming that the received signal is r (n) and the PN sequence length of the channel j is L, the formula of the correlation calculation is:

step 102-2) correlating the signal R _j (n) peak value R _j-max (n) comparing with a judgment threshold lambda if R _j-max (n)&And gt, lambda, determining that the television channel occupies, otherwise, determining that the television channel is idle.

3. The occupancy rate data mining and prediction method according to claim 1, wherein the step 104) comprises the following specific steps:

step 104-1) the control center carries out statistical analysis on historical detection data aiming at a certain channel in the television channel set to be detected to obtain a time sequence { s ] of the occupancy rate of the channel _t }；

Step 104-2) using a cubic exponential smoothing model for the occupancy time series s _t Predicting to obtain a prediction result y of the occupancy rate of the channel _t . The prediction formula is as follows:

the prediction model of the cubic exponential smoothing method is as follows:

wherein,

and step 104-3) the control center sorts the channels from low to high according to the predicted spectrum occupancy rate, and the channels arranged in the front are combined into a smaller channel subset to be detected.