Routing algorithm for optimizing routing by utilizing link prediction and routing reconstruction in mobile ad hoc network

Abstract

The invention discloses an algorithm for optimizing routing by utilizing link prediction and route reconstruction in a mobile ad hoc network, aiming at solving the problems of delay, overlarge overhead, excessive number of failed routes and the like caused by continuous node movement, frequent link disconnection and continuous route reconstruction in a mobile Ad Hoc network AODV routing protocol. The invention provides an algorithm for optimizing a routing by utilizing link prediction and routing reconstruction in a mobile Ad Hoc network. According to the algorithm, a link state prediction mechanism is introduced, and the Newton mean difference interpolation polynomial is used for predicting the link disconnection time, so that the local route repair algorithm is started before the route fails, the route reconstruction is realized, and the algorithm has better performance in the aspects of route overhead, end-to-end delay, packet delivery rate and the like.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of mobile ad hoc network routing protocol, and particularly relates to a link prediction and routing reconstruction optimization routing algorithm in a mobile ad hoc network. BACKGROUND

[0002] Mobile Ad Hoc network, also known as MANET (Mobile Ad Hoc Networks), is a kind of dynamic network formed by mobile nodes communicating with each other through shared wireless channels, which does not depend on any fixed infrastructure. The nodes in the network are equal in status and participate in route discovery and route maintenance. Mobile Ad Hoc network has the characteristics of independent networking, self-organization, non-centralization and dynamic, and is more robust and durable than the central wireless network, and has higher invulnerability, and is more suitable for temporary communication occasions such as transportation, battlefield, rescue, rescue, exploration and target monitoring in dangerous environments. Therefore, mobile Ad Hoc network has become a research hotspot today.

[0003] At present, AODV (Ad hoc On-Demand Distance Vector Routing) routing protocol is a kind of on-demand routing protocol which can better adapt to the characteristics of Ad Hoc network. The routing protocol does not need to maintain the route to all nodes, and only obtains the route when needed, and does not maintain the route after the communication ends, thereby saving network resources. However, if the route is interrupted during communication, a local route repair mechanism needs to be used for repair or a new route needs to be found. In the process of finding a new alternative route after the route is broken, the delay is increased, packet loss may occur, network overhead is increased, battery energy is wasted, network performance is rapidly decreased, and better network communication cannot be provided. If the link can be predicted in advance before the link fails, and the route is repaired, the packets will be discarded due to the use of outdated routes and the time for waiting for route reconstruction will be avoided, so the average end-to-end delay of packet transmission will be obviously reduced, which is very important for real-time transmission.

[0004] The existing method for solving link prediction provides position information for each node through GPS, uses position information such as geographic coordinates, current speed and motion direction to predict the position of the target node, realizes link failure prediction, but needs GPS auxiliary equipment, and GPS has the problems of low precision and being easily affected by the surrounding environment. The link failure time is estimated by using Lagrange interpolation, and then local repair is performed before the link fails, but when the number of interpolation points is changed, the entire formula changes, and the operation is too cumbersome.

[0005] This invention addresses the efficient and reliable handling of link failures, proposing an improved scheme based on link prediction and route reconstruction for on-demand routing mechanisms. The AODV routing protocol in mobile Ad Hoc networks should add a new function: link failure time prediction. This function detects link instability before it breaks and employs strategies to stabilize network links, reducing the number of routing failures, packet loss, and latency. This algorithm introduces a link state prediction mechanism, using the power value of received data packets and a Newtonian mean-difference interpolation polynomial to obtain a predicted link failure time. Before a node leaves the effective communication range between nodes, a new route repair process is initiated in advance. To avoid excessive control overhead, this invention employs a proactive local repair strategy, obtaining a new route before the link actually fails. The greatest advantage of this improvement is that it avoids latency jitter and packet loss caused by link failures. Summary of the Invention

[0006] This invention proposes a routing optimization algorithm for mobile ad hoc networks using link prediction and route reconstruction. The aim is to establish a model combining a link prediction mechanism and a proactive local repair algorithm to avoid latency and packet loss caused by link failures, and to minimize control overhead during route repair. This invention introduces a link state prediction mechanism, which uses the power value of received data packets and a Newton-Raphson interpolation polynomial to obtain a predicted link failure time. By comparing this predicted time with the link reconstruction time, a new proactive local repair algorithm is initiated in advance, enabling the link to obtain a new route before actual failure.

[0007] The specific method used in this invention is as follows:

[0008] (1) Introduce a link state prediction mechanism. When the power value of the received data packet is less than the danger threshold, use Newton's mean difference interpolation polynomial to obtain the predicted value of the link failure time.

[0009] A link in a dangerous state refers to a link availability time that is less than or equal to the time required for route reconstruction. Since calculating link availability time is complex, frequent calculations can lead to system performance degradation. Therefore, the protocol introduces a power threshold P. warning When the received power signal power is less than P warning The link failure time calculation begins. This invention uses the Newton's mean-difference interpolation polynomial to predict the link failure time. A significant advantage of the Newton's mean-difference interpolation polynomial is its relatively low computational cost, overcoming the problem of excessive computation caused by changes in the entire formula when the interpolation point changes in the Lagrange interpolation polynomial. The Newton's mean-difference interpolation polynomial is as follows:

[0010] N n(x) = f(x0) + f[x0, x1](x - x0) + f[x0, x1, x2](x - x0)(x - x1) +... + f[x0, x1,..., x n ](x - x0)(x - x1)... (x - xn-1) n -1)

[0011] If the signal power received at a certain node in the network is less than the danger threshold P warning , the node needs to start the link break time prediction mechanism to predict the break time of the link between the node and the predecessor node. In the link break time prediction mechanism, by periodically evaluating the signal power received by the node, the node routing table saves three sets of predecessor values, i.e. the power values P1, P2, P3 of the three predecessor node data received at the latest three times t1, t2, t3, and the above data needs to be updated when a new packet is received.

[0012] The signal strengths of the data packets received in the last three times are saved in the routing table as P1, P2, P3, respectively corresponding to the times t1, t2, t3. And P min is taken as the threshold signal power acting at t p , which can predict t p . When the received signal power decreases to the threshold signal power, the link will be disconnected. The signal power of the received data packet can be predicted using the Newton divided difference interpolation polynomial, where Δ and Δ 2 are the first-order divided difference and the second-order divided difference, respectively.

[0013] P min = P1 + (t p - t1) Δ + (t p - t1)(t p - t2) Δ 2 (1)

[0014]

[0015] Let

[0016]

[0017] Substituting (2) can obtain

[0018] P min = P1 + (t p - t1) A + (t p - t1)(t p - t2) B (3)

[0019] Rearranging (3) can obtain

[0020] Bt p 2+ (A - Bt1 - Bt2)t P + (P1 - P min - At1 + t1t2B) = 0

[0021] This equation is a monomial quadratic polynomial, wherein a = B, b = A - Bt1 - Bt2, c = P1 - P min - At1 + t1t2B can be obtained

[0022] at p 2 + bt P + c = 0

[0023] Therefore, the link disconnection time t can be solved P

[0024]

[0025] (2) The advance type local repair is used to start a new route repair process in advance before the node leaves the effective communication range of the node, and in order to avoid excessive control overhead, the advance type local repair strategy is adopted in the application to obtain a new route before the link is really invalid.

[0026] The route protocol needs a time to establish a new or alternative path, and a time parameter t is introduced, which is the time required for route reconstruction. r The time variable is just less than the link duration t P , and is enough to send error information to the upper node and the packet source node. r After that, the node enters the repair state, and the node needs to find an alternative route.

[0027] The link state analysis judges whether the link enters the repair state by calculating the link disconnection time. When the signal power of the received data packet is less than the danger threshold P warning , and the sample value saved in the route table reaches 3 groups, the link disconnection time is calculated by using the Newton difference interpolation polynomial. If the estimated link disconnection time t P is less than the time required for route reconstruction t r , it indicates that the link enters the repair state, and at this time, the local route discovery module needs to be called to find an alternative route; if the effective time of the link is greater than the reaction time, the normal mode is performed.

[0028] The purpose of the advance type local route repair is to ensure the reliability of the route while controlling the overhead of the network. The process is as follows.

[0029] (1) After the intermediate node discovers that the link state between the intermediate node and the predecessor node enters the repair state, the intermediate node sends an advance repair notification packet to the predecessor node.

[0030] (2) The predecessor node receives the notification, calls the local route discovery process, and starts a timer

[0031] ① The predecessor node broadcasts a local route request packet with 2 hops.

[0032] ② The neighbor node receives the packet, and if the node is not the local target node, the node first saves the reverse route of the arriving node, and then continues to broadcast the packet. The packet that does not meet the condition is discarded.

[0033] ③ The local target node obtains the request message, and responds to the first local route request received.

[0034] (3) If the node receives a local response packet before the end of the timer, the original local route is updated with the received response; if the timer ends but no local response is received, the node clears the route to the target node, and reports the information of the invalid route to the source node to avoid the use of the invalid route.

[0035] The application proposes a new algorithm, a link prediction mechanism and an early local repair algorithm combined model, for link breakage time prediction and route reconstruction in a mobile Ad Hoc network. The algorithm introduces a link state prediction mechanism, obtains a predicted value of the link failure time by using a Newton divided difference interpolation polynomial through the power value of the received data packet, and compares the predicted value with the link reconstruction time to pre-start a new route repair process of the early local repair algorithm, so that the link obtains a new route before the real failure, and has better performance in terms of route overhead, end-to-end delay and packet delivery rate. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 Dangerous threshold model;

[0037] Figure 2 Local route switching process;

[0038] Figure 3 Algorithm flowchart. DETAILED DESCRIPTION

[0039] The technical solutions of the application are further described below with reference to the drawings.

[0040] Figure 1 The dangerous threshold model is a dangerous state of a link, which means that the available time of the link is less than or equal to the time required for route reconstruction. Since the calculation of the available time of the link is relatively complex, frequent calculation of the available time of the link will lead to a decrease in system performance. Therefore, the protocol introduces a power threshold P warning When the received power signal power is less than P warningThe link break time calculation is started. The Newton divided difference interpolation polynomial is used to obtain the prediction of the link failure time. The Newton divided difference interpolation polynomial has the advantage of small amount of calculation, and overcomes the problem of large amount of calculation caused by change of the interpolation point of the Lagrange interpolation polynomial. The Newton divided difference interpolation polynomial is:

[0041] N n (x)=f(x0)+f[x0,x1](x-x0)+f[x0,x1,x2](x-x0)(x-x1)+…+f[x0,x1,...,x n ](x-x0)(x-x1)…(x-x n-1 )

[0042] Figure 1 R is the maximum transmission distance of the node, and r is the stable transmission distance. When the neighbor node of node A is located in the stable area, the transmission path is robust. When the neighbor node moves to the maximum distance area, the transmission path starts to become unstable. Therefore, when the node moves from the stable transmission distance to the maximum distance, it is set as P warning , which is represented by the dashed circle in the figure. The area between the dashed circle and the solid circle represents a dangerous state of the link. When node C reaches the boundary of the stable transmission distance, node C broadcasts a route warning message to node A.

[0043] Under the dual-path ground reflection model, for a typical WaveLAN environment, the maximum communication distance of the node is 250m, the minimum signal power P min identified is 3.65×10 10 w. The maximum speed of the node is set as V max , the longest time required for maintaining the route is T max , the dangerous threshold P warning is the received power when the two nodes are apart by (R-2V max T max )m, and the dangerous threshold P warning can be obtained as follows:

[0044]

[0045] When the signal power received by a node in the network is less than the dangerous threshold P warning , the node needs to start the link break time prediction mechanism to predict the break time of the link between the node and the predecessor node. Therefore, the dangerous threshold is directly related to whether the prediction mechanism is started. In the link break time prediction mechanism, the received signal power of the node is periodically evaluated, and three sets of predecessor values are saved in the route table of the node, i.e., the power values P1, P2, P3 of the three predecessor node data received at the latest three time points t1, t2, t3. When a new packet is received, the above data needs to be updated.

[0046] The routing table stores the signal strengths of the three most recently received data packets as P1, P2, and P3, corresponding to times t1, t2, and t3, respectively. And P... min As in t p The threshold signal power that takes effect at any given time can predict t. p The link will disconnect when the received signal power drops to a threshold signal power. The signal power of the received data packets can be predicted using a Newtonian mean-difference interpolation polynomial, where Δ and Δ 2 These are the first-order mean difference and the second-order mean difference, respectively.

[0047] P min =P1+(t) p -t1)Δ+(t p -t1)(t p -t2)Δ 2 (1)

[0048]

[0049] set up

[0050]

[0051] Substituting into (2), we get

[0052] P min =P1+(t) p -t1)A+(t p -t1)(t p -t2)B (3)

[0053] Simplifying (3) yields

[0054] Bt p 2 +(A-Bt1-Bt2)t P +(P1-P min -At1+t1t2B)=0

[0055] This expression is a quadratic polynomial in one variable, where let a = B, b = A - Bt1 - Bt2, and c = P1 - Pt2. min -At1+t1t2B yields

[0056] at p 2 +bt P +c=0

[0057] Therefore, the link disconnection time t can be calculated. P

[0058]

[0059] Figure 2 Local route switching procedure. The route protocol needs a time to establish a new or alternative path, a time parameter is introduced, the time needed for route reconstruction t r . This time variable is just smaller than the link duration t P , and enough to send error information to the upper node, the source node of the data packet. t r After that, the node enters the repair state, and the node needs to find an alternative route.

[0060] Link state analysis judges whether the link enters the repair state by calculating the link breakage time. When the signal power of the received data packet is less than the danger threshold P waming , and the number of the sample values saved in the route table reaches 3, the Newton's divided difference interpolation polynomial is used to calculate the link breakage time. If the estimated link breakage time t P is less than the time needed for route reconstruction t r , it indicates that the link enters the repair state, and the local route discovery module needs to be called to find an alternative route; if the effective time of the link is greater than the reaction time, it is processed in the normal way.

[0061] The purpose of the early local route repair is to ensure the reliability of the route while controlling the overhead of the network. Its procedure is shown in Figure 2 , assuming that the original local route A--C--F is part of a certain active path.

[0062] (1) After node C finds that the link state between node C and node A enters the repair state (process 1), C sends an early repair notification packet to A (process 2).

[0063] (2) After A receives the notification, the local route discovery process is called, and a timer t

[0064] ① A broadcasts a local route request packet with a hop of 2 (process 3).

[0065] ② After the neighbor node B or D receives the packet, if the node is not the local target node, the node first saves the reverse route to A node, and then continues to broadcast the packet (process 3). Packets that do not meet the conditions are discarded.

[0066] ③ The local target node F obtains the request packet, and responds to the first local route request received (process 4).

[0067] (3) If node A receives a local response packet before the timer ends, the original local route is updated with the received response; if the timer ends but no local response is received, A node clears the route to the target node, and reports the information of the invalid route to the source node to avoid the use of the invalid route.

[0068] Figure 3 Algorithm flow chart. The node receives signal power by periodic evaluation. Three sets of predecessor values are saved in the node routing table, i.e. the power values P1, P2, P3 of the three predecessor node data received at the last three time instants t1, t2, t3. The above data need to be updated when a new packet is received. When the signal power of the data packet received by the node is less than the danger threshold P warning , and the sample values saved in the routing table reach 3 sets, the link breakage time is calculated using Newton's divided difference interpolation polynomial. If the estimated link breakage time t P is less than the time t r required for routing reconstruction, it indicates that the link enters the repair state, at which time the local routing discovery module needs to be called to find a replacement route; if the effective time of the link is greater than the reaction time, it is processed in the normal way.

Claims

1. A routing algorithm for optimizing routing in mobile ad hoc networks using link prediction and route reconstruction, characterized in that, The link state prediction mechanism is introduced, and the Newton difference interpolation polynomial is used to obtain the predicted value of link failure time when the power value of received data packet is less than the dangerous threshold value; The link dangerous state refers to the link available time being less than or equal to the time required for route reconstruction, and the link available time is relatively complex to calculate, so frequent calculation of the link available time will result in the decline of system performance; To this end, the protocol introduces a power threshold P warning When the received power signal power is less than P warning The link break time calculation is started, and the Newton forward difference interpolation polynomial is used to obtain the prediction of the link failure time, and the Newton forward difference interpolation polynomial is: The signal power received at a certain node in the network is less than a dangerous threshold P warning If so, the node needs to start the link breakage time prediction mechanism to predict the breakage time of the link between the node and the predecessor node. In the link break time prediction mechanism, by periodic evaluation of node received signal power, three sets of predecessor values are saved in the node routing table, namely the power values P1, P2, P3 of the three predecessor node data received at the last three times t1, t2, t3, when a new packet is received, the above data needs to be updated, and P min As the threshold signal power acting at t p , the prediction of t p , the prediction of the signal power of the received data packet can use the Newton's divided difference interpolation polynomial, where Δ and Δ 2 are the first-order and second-order divided differences respectively: P min = P1+ (t p -t1)Δ+ (t p -t1)(t p -t2)Δ 2 (1) Substitute (2) into (1) to obtain Adjust (3) to obtain P min = P1+ (t p -t1)A+ (t p -t1)(t p -t2)B (3) In order to avoid excessive control overhead, the strategy of advance local repair is adopted to obtain a new route before the link is really failed; Bt p 2 +(A-Bt1-Bt2)t P +(P1-P min -At1+t1t2B)=0 This equation is a monic quadratic polynomial, where we set a = B, b = A - Bt1 - Bt2, and c = P1 - P min - At1 + t1t2B at p 2 +bt P +c=0 t P is the link disconnection time.

2. The method according to claim 1, wherein, The purpose of the advance local route repair is to ensure the reliability of the route while controlling the network overhead, and the process is as follows: The routing protocol needs a time to establish new or alternative paths, introducing a time parameter, the time needed for the rerouting t r ; This time variable is just less than the link duration t P And enough to send error information to the upper node, the data packet source node; t r After that, the node enters the repair state, and the node needs to find an alternative route; The link state analysis judges whether the link enters into the repair state by calculating the link breakage time, when the signal power of the data packet received by the node is less than the dangerous threshold P warning , and the sample value saved in the routing table reaches 3 groups, the Newton difference interpolation polynomial is used to calculate the link breakage time; if the estimated link breakage time t P is less than the time t r required for the routing reconstruction, the link enters into the repair state, at this time, the local routing discovery module needs to be called to find the replacement routing; if the effective time of the link is greater than the reaction time, the normal mode is performed; (1) After the intermediate node discovers that the link state between the intermediate node and the predecessor node enters the repair state, the intermediate node sends an advance repair notification packet to the predecessor node; (2) After the predecessor node receives the notification, the local route discovery process is called, and the timer is started ① The predecessor node broadcasts a local route request packet with 2 hops; ② After the neighbor node receives the packet, if the node is not a local target node, the reverse route to the arrival node is first saved, and then the packet is continuously broadcasted, and the packet that does not meet the condition is discarded; ③ The local target node obtains the request message, and responds to the first local route request received; (3) If the node receives a local response packet before the timer ends, the original local route is updated with the received response; if the timer ends but no local response is received, the node clears the route to the target node and reports the information of the failed route to the source node, so as to avoid the use of the failed route. ​

Images