CN110602073B - Unmanned aerial vehicle flight control protocol field division method based on information theory - Google Patents

Abstract

The invention provides an unmanned aerial vehicle flight control protocol field division method based on an information theory. The method aims to improve the accuracy of continuous change field division by mining the information quantity expressed by data as a feature, and comprises the following steps: the method comprises the steps of preprocessing an unmanned aerial vehicle flight control protocol data set, acquiring a regular field bit number set L and a random field bit number set R based on an information theory, acquiring a separation point bit number set E of the regular field bit number set L based on the information theory, and dividing the random field bit number set R based on the information theory to acquire all field separation points. The statistical characteristics extracted by the method are the information quantity obtained from the data, the potential semantic information of the fields is fully mined, the unmanned plane flight control protocol is divided into completely random fields, flight control related fields and a plurality of regular fields, and the accuracy of dividing the continuous change fields in the unmanned plane flight control protocol is improved.

Description

Unmanned aerial vehicle flight control protocol field division method based on information theory

Technical Field

The invention belongs to the technical field of wireless communication, relates to a method for dividing a flight control protocol field of an unmanned aerial vehicle, and particularly relates to a method for reversely analyzing a syntax of the flight control protocol of the unmanned aerial vehicle based on an information theory.

Background

The unmanned aerial vehicle communication system is a system for transmitting an operation instruction between an operator and an unmanned aerial vehicle through a wireless network according to a set binary protocol. These predetermined binary protocols are the drone communication protocol, and the protocol responsible for the transmission of the flight instructions of the drone is the drone flight control protocol.

Along with the popularization of unmanned aerial vehicles, the lack of a method for controlling the unmanned aerial vehicles becomes a significant problem, the technology of completely controlling the unmanned aerial vehicles by a third party can become an optimal solution for the control of the unmanned aerial vehicles, and the aim of controlling the unmanned aerial vehicles by forging data packets through a reverse unmanned aerial vehicle flight control protocol can be fulfilled.

The unmanned aerial vehicle flight control protocol reverse direction refers to a process of extracting grammar and semantics of the unmanned aerial vehicle flight control protocol by analyzing a large number of unmanned aerial vehicle flight control protocol data packets or monitoring and analyzing the construction and analysis flows of the flight control protocol data packets without depending on protocol description. The unmanned plane flight control protocol reverse analysis comprises unmanned plane flight control protocol field division and unmanned plane flight control protocol field semantic analysis. The unmanned plane flight control protocol field division refers to division of fields with different meanings under the condition of not depending on the semantics, and the unmanned plane flight control protocol field semantic analysis refers to excavation of the meaning to be expressed by each field.

The unmanned plane flight control protocol reverse analysis mainly comprises unmanned plane flight control protocol reverse analysis based on an execution track and unmanned plane flight control protocol reverse analysis based on a datagram. The unmanned aerial vehicle flight control protocol reverse analysis based on the execution track means that the protocol format is extracted by analyzing the generation and analysis processes of the unmanned aerial vehicle flight control protocol data packet in the control equipment, the unmanned aerial vehicle chip and the like. The datagram-based unmanned aerial vehicle flight control protocol reverse analysis means that a large number of unmanned aerial vehicle flight control protocol data packets are intercepted through tools or software, the unmanned aerial vehicle flight control protocol data packets are used as analysis objects, and a protocol format is obtained through inference according to some statistical characteristics of values of protocol fields.

The existing unmanned plane flight control protocol field division method based on the numerical value change rate is a mainstream method, the method firstly analyzes common design ideas of unmanned plane flight control protocol designers, and then summarizes statistical characteristics of the numerical value change rate of common fields, however, many unmanned plane flight control protocol designers already design fields by using complex rules, and the statistical characteristics cannot essentially reflect the relation between two pieces of data in flight control protocol data.

Ji Ran et al discloses an unmanned aerial vehicle Flight Control protocol Automatic Reverse analysis method in a published paper of Automatic Reverse Engineering of Private Flight Control Protocols of UAVs (Security and Communication Networks, 2017), the method determines a field semantic as a Flight Control field by using time information of an unmanned aerial vehicle Flight Control protocol data packet and recorded Flight states, and determines a field semantic as a check bit field by using a check bit generation matrix reduction algorithm, so that contribution is provided for unmanned aerial vehicle Flight Control protocol semantic Automatic analysis, but the unmanned aerial vehicle Flight Control protocol field division method still performs division by using numerical value change or not as a numerical value change rate characteristic, so that the method has lower division accuracy for continuous change fields.

Disclosure of Invention

The invention aims to overcome the defects in the prior art, provides an unmanned aerial vehicle flight control protocol field division method based on an information theory, and aims to solve the technical problem that the division accuracy rate of continuously-changing fields in the prior art is low.

In order to achieve the purpose, the invention adopts the technical scheme that:

an unmanned aerial vehicle flight control protocol field division method based on an information theory comprises the following steps:

(1) preprocessing an unmanned aerial vehicle flight control protocol data set:

(1a) collecting M binary flight control protocol data packets in a time range of T, converting the M binary flight control protocol data packets into hexadecimal data, wherein each hexadecimal flight control protocol data packet comprises h hexadecimal data, T is more than 10min, M is more than 10000, and h is more than or equal to 1;

(1b) arranging M hexadecimal flight control protocol data packets into a hexadecimal flight control data set D according to the sequence of binary flight control protocol data packet acquisition, wherein D is { D ═ D₁,D₂,...,D_j,...,D_M}，D_jIndicating the jth data packet, D_j＝{d_1,j,d_2,j,...,d_i,j,...,d_h,j}，d_i,jRepresents the hexadecimal data in the jth data packet at the ith bit, and the value is expressed as | d_i,jI, the i-th hexadecimal data in each hexadecimal flight control protocol data packet form a hexadecimal data set S_iWherein i is more than or equal to 1 and less than or equal to h, and j is more than or equal to 1 and less than or equal to M;

(1c) when j is less than or equal to m, adding S_iMiddle data d_i,jComposing a hexadecimal data set X_1,iAnd will | d_i,jL constitutes the hexadecimal value set X'_1,iWill S_iThe rest of data d_i,jComposing a hexadecimal data set X_2,iAnd will | d_i,jL constitutes the hexadecimal value set X'_2,iWherein, in the step (A),

represents rounding down;

(2) acquiring a regular field bit number set L and a random field bit number set R based on an information theory:

(2a) setting a threshold value V, constructing a set HEX containing 16 hexadecimal values, wherein the u-th element in the HEX is represented as HEX, and the u-th element is represented as HEX {0,1,2,3,4,5,6,7,8,9, a, b, c, d, e and f }_u；

(2b) When d is_i,j∈X_1,iThen all will satisfy | d_i,j-1|＝HEX_uD of_i,j| constitutes a set of hexadecimal values I_1,i,uWhen d is_i,j∈X_2,iThen all will satisfy | d_i,j-1|＝HEX_uD of_i,j| constitutes a set of hexadecimal values I_2,i,u；

(2c) By means of the set X_1,iAnd set I_1,i,uCalculating X'_1,iD in_i,j-1I and | d_i,jConditional entropy between | H_1,iBy means of the set X_2,iAnd set I_2,i,uCalculating X'_2,iD in_i,j-1I and | d_i,jConditional entropy between | H_2,iAnd will satisfy

All i form a regular field bit number set L, and the rest i form a random field bit number set R;

(3) acquiring a separation point bit number set E of a regular field bit number set L based on an information theory:

(3a) when d is_i,j∈X_1,iAnd i, i +1 ∈ L, all will satisfy | d_i-1,j|＝HEX_uD of_i,jL constitutes the hexadecimal value set I'_1,i,uWhen d is_i,j∈X_2,iAnd i, i +1 ∈ L, all will satisfy | d_i-1,j|＝HEX_uD of_i,jL constitutes the hexadecimal value set I'_2,i,u；

(3b) By means of the set X_1,iAnd aggregate I'_1,i,uCalculating X'_1,i-1D in_i-1,jL and X'_1,iD in_i,jConditional entropy between l H'_1,iBy means of the set X_2,iAnd aggregate I'_2,i,uCalculating X'_2,i-1D in_i-1,jL and X'_2,iD in_i,jConditional entropy between l H'_2,iAnd will satisfy

All i of (a) form a separation point digit set E;

(4) dividing a random field bit number set R based on an information theory:

(4a) setting a threshold value N;

(4b) when i ∈ R, the number of strips

As a unit, adding S_iThe medium data are divided into f groups, S_i＝{S_i，1,S_i，2,...,S_i,z,...,S_i,fIn which S is_i,zDenotes S_iWhen d is the group z data of_i,j∈S_i,zWhen, will | d_i,j| constitute a set of hexadecimal values W_i,zWherein

Represents rounding down;

(4c) computing a set W_i,zEntropy of aroma concentration C of each element_i,zAnd then S is_iIn each group S_i,zC of (A)_i,zForm a set G of fragrance concentration entropy_i；

(4d) All C's were clustered using KMeans algorithm_i,zClustering into two types of large central value and small central value of the fragrance concentration entropy, taking the type of large central value of the fragrance concentration entropy as a completely random type A, and taking the type of small central value of the fragrance concentration entropy as a flight control related type B;

(4e) g is to be_iIn satisfy C_i,zThe number of elements belonging to the group of A is G_iThe ratio of the total number of elements in the formula is represented by w_iAnd all satisfy w_iI more than or equal to N forms a complete random field bit number set RT, and the rest forms a flight control related field bit number set FC;

(5) acquiring all field separation points:

and adding all i which satisfy that i and i +1 do not belong to the set RT or the set FC or the set L at the same time into the separation point digit set E, and finally, all elements in the set E are all field separation points.

Compared with the prior art, the invention has the following advantages:

the statistical characteristics used when the continuous change fields are divided are the information quantity obtained from the data, the potential semantics of the fields are fully mined, and the unmanned plane flight control protocol is divided into the completely random fields, the flight control related fields and the plurality of regular fields, so that the problem that the relation among the data cannot be essentially reflected by using the numerical value change rate as the statistical characteristics in the prior art is solved, and compared with the prior art, the accuracy of the division of the continuous change fields in the unmanned plane flight control protocol is effectively improved.

Drawings

FIG. 1 is a flow chart of an implementation of the present invention.

Detailed Description

The invention is described in further detail below with reference to the figures and the specific embodiments.

Referring to fig. 1, the present invention includes the steps of:

step 1) preprocessing an unmanned aerial vehicle flight control protocol data set:

step 1a) acquiring M binary flight control protocol data packets within a time range T, wherein the binary system only comprises two values of 0 and 1, and the entropy values are very close to each other due to coincidence when the binary system is directly applied to the calculation of the information entropy, so that the binary system is converted into a hexadecimal system, each hexadecimal flight control protocol data packet comprises h hexadecimal data, T is more than 10min, M is more than 10000, h is more than or equal to 1, in the example, T is 10.25min, M is 24651, and h is 40;

step 1b) arranging M hexadecimal flight control protocol data packets into a hexadecimal flight control data set D according to the sequence of binary flight control protocol data packet acquisition, wherein D is { D ═ D }₁,D₂,...,D_j,...,D_M}，D_jIndicating the jth data packet, D_j＝{d_1,j,d_2,j,...,d_i,j,...,d_h,j}，d_i,jRepresents the hexadecimal data in the jth data packet at the ith bit, and the value is expressed as | d_i,jI, the i-th hexadecimal data in each hexadecimal flight control protocol data packet form a hexadecimal data set S_iWherein i is more than or equal to 1 and less than or equal to h, and j is more than or equal to 1 and less than or equal to M;

step 1c) the hexadecimal flight control data set D is divided into two data sets in a balanced way, namely when j is less than or equal to m, S is divided_iMiddle data d_i,jComposing a hexadecimal data set X_1,iAnd will | d_i,jL constitutes the hexadecimal value set X'_1,iWill S_iThe rest of data d_i,jComposing a hexadecimal data set X_2,iAnd will | d_i,jL constitutes the hexadecimal value set X'_2,iWherein, in the step (A),

represents rounding down, in this example m-12325;

step 2) acquiring a regular field bit number set L and a random field bit number set R based on the information theory:

step 2a) sets a threshold V, constructs a set HEX containing 16 hexadecimal values {0,1,2,3,4,5,6,7,8,9, a, b, c, d, e, f }, where the u-th element in HEX is denoted HEX_uIn this example, V ═ 0.1;

step 2b) when d_i,j∈X_1,iThen all will satisfy | d_i,j-1|＝HEX_uD of_i,j| constitutes a set of hexadecimal values I_1,i,uWhen d is_i,j∈X_2,iThen all will satisfy | d_i,j-1|＝HEX_uD of_i,j| constitutes a set of hexadecimal values I_2,i,u；

Step 2c) even if the median values of both regular and random fields are constantly changing, for the regular field | d_i,jI is to d according to a certain rule_i,j-1The result of the calculation is made, | so although the law is unknown, it can be passed | d depending on the acquired data_i,j-1I predict | d_i,jFor random fields, | d_i,jI and | d_i,j-1I has no operational relationship between I, therefore, can not pass through | d_i,j-1I predict | d_i,jL. Conditional entropy in information theory reflects uncertainty between data and thus can be used to reflect whether | d can be passed_i,j-1I predict | d_i,jFor the two data sets X obtained in step 1c) |, the data sets X are stored in a memory_1,iAnd X_2,iIn the rule field, no matter X_1,iOr X_2,iCan all pass through d_i,j-1I predict | d_i,jI, so that the conditional entropy does not vary much, but no matter X in the random field_1,iOr X_2,iAll can not pass through d_i,j-1I predict | d_i,jAnd therefore the conditional entropy fluctuation is large. Since the values in each data packet are likely to change in the regular field and the random field, and the change rates of the values are likely to be similar, the change rate of the values as the statistical characteristic cannot essentially reflect the essential difference between the regular field and the random field, and the conditional entropy as the statistical characteristic is mined from | d_i,j-1Derived in | regarding | d_i,jThe information amount of | can reflect the regular field and the random in natureThe field difference can effectively improve the accuracy of dividing the continuous random field and the regular field, so that the method is based on a formula

By means of the set X_1,iAnd set I_1,i,uCalculating X_1,iD in_i,j-1I and | d_i,jCondition H between |_1,iAccording to the formula

By means of the set X_2,iAnd set I_2,i,uCalculating X_2,iD in_i,j-1I and | d_i,jCondition H between |_2,iAnd will satisfy

All i form a regular field bit number set L, and the rest i form a random field bit number set R, wherein X is X'_t,iP (X) is X'_t,iProbability of occurrence of x in (2)₁Is I_1,i,uAll first occurring elements in (A), p (y)₁| x) is HEX_uWhen x is equal to I_1,i,uY in the set₁Probability of occurrence, y₂Is I_2,i,uAll first occurring elements in (A), p (y)₂| x) is HEX_uWhen x is equal to I_2,i,uY in the set₂The probability of occurrence;

step 3) acquiring a separation point bit number set E of the regular field bit number set L based on the information theory:

step 3a) when d_i,j∈X_1,iAnd i, i +1 ∈ L, all will satisfy | d_i-1,j|＝HEX_uD of_i,jL constitutes set I'_1,i,uWhen d is_i,j∈X_2,iAnd i, i +1 ∈ L, all will satisfy | d_i-1,j|＝HEX_uD of_i,jL constitutes set I'_2,i,u；

Step 3b) even if the values in the plurality of rule fields obeying different rules are constantly changed, the fields obeying the same rule are long due to the field lengthThe value change in the field must be cyclic, so that if there are a large number of packets to analyze (the value in the field should complete many cycles), the value can be analyzed via | d |_i-1,jI predict | d_i,jDistribution rule of | for | d between fields subject to different rules_i-1,jI and | d_i,jI, the following rules of the front and back fields are different, so that the I can not pass through the I d_i-1,jI predict | d_i,jThe distribution rule of | and the conditional entropy in the information theory reflect the uncertainty between data, so that the method can be used for reflecting whether | d can be passed_i-1,jI predict | d_i,jI distribution rule of the two data sets X obtained in step 1c)_1,iAnd X_2,iIn the same rule field, | d_i-1,jI and | d_i,jI obey a uniform law no matter X_1,iOr X_2,iCan all pass through d_i-1,jI predict | d_i,jThe distribution rule of | is that the conditional entropy will not change greatly, but for | d obeying different rules_i-1,jI and | d_i,jL, no matter X_1,iOr X_2,iAll can not pass through d_i-1,jI predict | d_i,jThe distribution rule of | is regular, so the conditional entropy fluctuation is large. Because the values in the regular fields all change according to a certain rule, the change rate of the values can be correctly divided as statistical characteristics only when the change rates of the lowest bit of the previous regular field and the highest bit of the next regular field are clearly changed and distinguished, and the fields which obey the same rule can be divided into two fields due to improper characteristic selection, and the conditional entropy is used as statistical characteristics for mining_i-1,jDerived in | regarding | d_i,jThe information content of | can essentially reflect whether the connection exists between the front and back two-bit data, and can effectively improve the accuracy of dividing two continuous fields which obey different rules, so that the method is based on a formula

By means of the set X_1,iAnd aggregate I'_1,i,uCalculating X_1,i-1D in_i-1,jI and X_1,iD in_i,jCondition between |Entropy H'_1,iAccording to the formula

By means of the set X_2,iAnd aggregate I'_2,i,uCalculating X_2,i-1D in_i-1,jI and X_2,iD in_i,jConditional entropy between l H'_2,iAnd will satisfy

All i of (a) constitute a set of bin bit numbers E, wherein X is X'_t,iAll of the first occurring elements of (a), (b), (c), (d'₁Is l'_1,i,uAll of the first occurring elements of (a), p (y'₁| x) is HEX_uX is'_1,i,uY 'in the set'₁Probability of occurrence, y'₂Is l'_2,i,uAll of the first occurring elements of (a), p (y'₂| x) is HEX_uX is'_2,i,uY 'in the set'₂The probability of occurrence;

step 4) dividing the random field bit number set R based on the information theory:

step 4a) setting a threshold value N;

step 4b) when i ∈ R, counting by number

As a unit, adding S_iThe medium data are divided into f groups, i.e. S_i＝{S_i，1,S_i，2,...,S_i,z,...,S_i,fIn which S is_i,zDenotes S_iWhen d is the group z data of_i,j∈S_i,zWhen, will | d_i,j| constitute a set of hexadecimal values W_i,zWherein

Denotes rounding down, i.e. grouping by the average number of packets collected per second, which is calculated in this example from T ═ 10.25min and M ═ 24651, obtained in step 1a)

f＝615；

Step 4c) the flight control field and its related fields in the unmanned plane flight control protocol are changed according to the flight instructions sent by the operator, the flight control field and the related field thereof are difficult to distinguish from the completely random field in the step 2), but because the time interval of the data packets of the unmanned plane flight control protocol is generally short, tens of packets are often sent every second, the frequency of the operator changing the flight command is generally low, and the operator often sends the same flight command within one second or even several seconds, therefore, the uncertainty of the flight control field and the related field in the local data is obviously reduced, the fragrance concentration entropy can reflect the uncertainty in a group of data, therefore, the fragrance concentration entropy of local data in the flight control field and the related field is obviously reduced, and the local fragrance concentration entropy is still larger for a completely random field. Because different flight instructions are likely to be controlled by different bits, according to the number of different flight instructions sent by a manipulator, the numerical value change rate characteristics of the different bits have larger difference, the numerical randomness in a completely random field is extremely large, and the numerical value change rate characteristic of each bit is also random, so that the numerical value change rate is taken as a statistical characteristic, the continuous flight control field and the completely random field are difficult to be correctly divided, the information entropy of local data is taken as the statistical characteristic, the differences of the flight control field, related fields and the completely random field can be reflected essentially, and the accuracy of the division of the continuous flight control field and the completely random field is further effectively improved. Then according to the formula

Computing a set W_i,zEntropy of aroma C_i,zWill S_iIn each group S_i,zC of (A)_i,zForm a set G of fragrance concentration entropy_iWherein t is W_i,zP (x) is the set W_i,zThe probability of x occurrence in (2);

step 4d) clustering all C using KMeans algorithm_i,zClustering to obtain a sum entropy center with large central valueThe two types with small values are used, the type with a large central value of the aroma concentration entropy is used as a completely random type A, the type with a small central value of the aroma concentration entropy is used as a flight control related type B, in the example, the cluster number K is 2, the Euclidean distance is used as a distance measurement standard, and clustering is repeated until the central values of the two types of aroma concentration entropies do not change any more;

step 4e) reaction of G_iIn satisfy C_i,zThe number of elements belonging to the group of A is G_iThe ratio of the total number of elements in the formula is represented by w_iAnd all satisfy w_iI more than or equal to N forms a complete random field bit number set RT, and the rest forms a flight control related field bit number set FC;

step 5) acquiring all field separation points:

and adding all i which satisfy that i and i +1 do not belong to the set RT or the set FC or the set L at the same time into the separation point digit set E, and finally, all elements in the set E are all field separation points.

The final field division result of the embodiment is completely consistent with the result obtained by analyzing the function for constructing the data packet in the unmanned aerial vehicle APP.

The above description is only one specific example of the present invention and should not be construed as limiting the invention in any way. It will be apparent to persons skilled in the relevant art(s) that various modifications and changes in form and detail can be made therein without departing from the principles and arrangements of the invention, but these modifications and changes based on the inventive concept are also within the scope of the invention as defined in the appended claims.

Claims (3)

1. An unmanned aerial vehicle flight control protocol field division method based on an information theory is characterized by comprising the following steps:

(1) preprocessing an unmanned aerial vehicle flight control protocol data set:

(1a) collecting M binary flight control protocol data packets in a time range of T, converting the M binary flight control protocol data packets into hexadecimal data, wherein each hexadecimal flight control protocol data packet comprises h hexadecimal data, T is more than 10min, M is more than 10000, and h is more than or equal to 1;

(1b) m pieces of data are collected according to the sequence of binary flight control protocol data packetsThe hexadecimal flight control protocol data packets are arranged into a hexadecimal flight control data set D, D ═ D₁,D₂,...,D_j,...,D_M}，D_jIndicating the jth data packet, D_j＝{d_1,j,d_2,j,...,d_i,j,...,d_h,j}，d_i,jRepresents the hexadecimal data in the jth data packet at the ith bit, and the value is expressed as | d_i,jI, the i-th hexadecimal data in each hexadecimal flight control protocol data packet form a hexadecimal data set S_iWherein i is more than or equal to 1 and less than or equal to h, and j is more than or equal to 1 and less than or equal to M;

(1c) when j is less than or equal to m, adding S_iMiddle data d_i,jComposing a hexadecimal data set X_1,iAnd will | d_i,jL constitutes the hexadecimal value set X'_1,iWill S_iThe rest of data d_i,jComposing a hexadecimal data set X_2,iAnd will | d_i,jL constitutes the hexadecimal value set X'_2,iWherein, in the step (A),

represents rounding down;

(2) acquiring a regular field bit number set L and a random field bit number set R based on an information theory:

(2a) setting a threshold value V, constructing a set HEX containing 16 hexadecimal values, wherein the u-th element in the HEX is represented as HEX, and the u-th element is represented as HEX {0,1,2,3,4,5,6,7,8,9, a, b, c, d, e and f }_u；

(2b) When d is_i,j∈X_1,iThen all will satisfy | d_i,j-1|＝HEX_uD of_i,j| constitutes a set of hexadecimal values I_1,i,uWhen d is_i,j∈X_2,iThen all will satisfy | d_i,j-1|＝HEX_uD of_i,j| constitutes a set of hexadecimal values I_2,i,u；

(2c) By means of the set X_1,iAnd set I_1,i,uCalculating X'_1,iD in_i,j-1I and | d_i,jConditional entropy between | H_1,iBy means of the set X_2,iAnd set I_2,i,uCalculating X'_2,iD in_i,j-1I and | d_i,jConditional entropy between | H_2,iAnd will satisfy

All i form a regular field bit number set L, and the rest i form a random field bit number set R;

(3) acquiring a separation point bit number set E of a regular field bit number set L based on an information theory:

(3a) when d is_i,j∈X_1,iAnd i, i +1 ∈ L, all will satisfy | d_i-1,j|＝HEX_uD of_i,jL constitutes the hexadecimal value set I'_1,i,uWhen d is_i,j∈X_2,iAnd i, i +1 ∈ L, all will satisfy | d_i-1,j|＝HEX_uD of_i,jL constitutes the hexadecimal value set I'_2,i,u；

(3b) By means of the set X_1,iAnd aggregate I'_1,i,uCalculating X'_1,i-1D in_i-1,jL and X'_1,iD in_i,jConditional entropy between l H'_1,iBy means of the set X_2,iAnd aggregate I'_2,i,uCalculating X'_2,i-1D in_i-1,jL and X'_2,iD in_i,jConditional entropy between l H'_2,iAnd will satisfy

All i of (a) form a separation point digit set E;

(4) dividing a random field bit number set R based on an information theory:

(4a) setting a threshold value N;

(4b) when i ∈ R, the number of strips

As a unit, adding S_iThe medium data are divided into f groups, S_i＝{S_i，1,S_i，2,...,S_i,z,...,S_i,f}，Wherein S_i,zDenotes S_iWhen d is the group z data of_i,j∈S_i,zWhen, will | d_i,j| constitute a set of hexadecimal values W_i,zWherein

Represents rounding down;

(4c) computing a set W_i,zEntropy of aroma concentration C of each element_i,zAnd then S is_iIn each group S_i,zC of (A)_i,zForm a set G of fragrance concentration entropy_i；

(4d) All C's were clustered using KMeans algorithm_i,zClustering into two types of large central value and small central value of the fragrance concentration entropy, taking the type of large central value of the fragrance concentration entropy as a completely random type A, and taking the type of small central value of the fragrance concentration entropy as a flight control related type B;

(4e) g is to be_iIn satisfy C_i,zThe number of elements belonging to the group of A is G_iThe ratio of the total number of elements in the formula is represented by w_iAnd will all satisfy _iw ≥NI of (2) forms a complete random field bit number set RT, and the rest forms a flight control related field bit number set FC;

(5) acquiring all field separation points:

and adding all i which satisfy that i and i +1 do not belong to the set RT or the set FC or the set L at the same time into the separation point digit set E, and finally, all elements in the set E are all field separation points.

2. The method for partitioning fields of the flight control protocol of unmanned aerial vehicle based on information theory as claimed in claim 1, wherein the step (2c) is to calculate X_1,iD in_i,j-1I and | d_i,jConditional entropy between | H_1,iCalculating X_2,iD in_i,j-1I and | d_i,jConditional entropy between | H_2,iThe calculation of X as described in step (3b)_1,i-1D in_i-1,jI and X_1,iD in_i,jConditional entropy between l H'_1,iCalculating X_2,i-1D in_i-1,jI and X_2,iD in_i,jConditional entropy between l H'_2,iThe calculation formulas are respectively as follows:

wherein X is X'_t,iP (X) is X'_t,iProbability of occurrence of x in (2)₁Is I_1,i,uAll first occurring elements in (A), p (y)₁| x) is HEX_uWhen x is equal to I_1,i,uY in the set₁Probability of occurrence, y₂Is I_2,i,uAll first occurring elements in (A), p (y)₂| x) is HEX_uWhen x is equal to I_2,i,uY in the set₂Probability of occurrence, y'₁Is l'_1,i,uAll of the first occurring elements of (a), p (y'₁| x) is HEX_uX is'_1,i,uY 'in the set'₁Probability of occurrence, y'₂Is l'_2,i,uAll of the first occurring elements of (a), p (y'₂| x) is HEX_uX is'_2,i,uY 'in the set'₂The probability of occurrence.

3. The UAV flight control protocol field according to claim 1The method of partitioning, wherein the set of computations W described in step (4c) is_i,zEntropy of aroma concentration C of each element_i,zThe expression is as follows:

wherein t is W_i,zP (x) is the set W_i,zProbability of x occurrence in (c).

Images