A cross-layer perception based joint congestion control method for internet of vehicles

Abstract

The application discloses a kind of based on cross-layer perception's vehicle networking joint congestion control method, comprising: based on link load condition adjustment link transmission rate to obtain adjustment rate;Obtain the minimum contention window and single data packet average transmission time of current channel;Based on Markov chain from the minimum contention window selects any integer as backoff value to control data packet sending state based on the backoff value;Based on the adjustment rate and single data packet average transmission time, the probability of successful sending of data packet is calculated, when the probability of successful sending of data packet is greater than preset threshold, the data packet is marked as sending success.The application utilizes network resource allocation utility function to control sending rate, reduces the amount of data injected into network, prevents packet loss;According to the control of the size of transmission power of adjustment factor and the priority of data stream, the transmission rate of link is changed, to avoid the occurrence of network congestion.

Description

Technical Field

[0001] This invention belongs to the field of vehicle networking, specifically relating to a joint congestion control method for vehicle networking based on cross-layer perception. Background Technology

[0002] In recent years, with the rapid development of mobile internet, the Internet of Things (IoT), and sensor technologies, the Internet of Vehicles (IoV) has become one of the key driving forces for the development of future intelligent transportation. IoV is an evolution of the traditional Vehicular AdHoc Network (VANET), enabling intelligent interaction between people, vehicles, roads, and the environment. IoV consists of three layers: a data acquisition layer, a transmission layer, and a service layer, and has high requirements for latency and throughput. Based on application scenarios, IoV services can be divided into three main categories: traffic control, information services, and driving safety assurance. These three categories have different requirements in terms of network bandwidth, connectivity performance, latency, and network deployment. In particular, information services such as dynamic maps and AR navigation have high latency requirements and face severe transmission latency problems when processing massive amounts of traffic data. Due to its high-speed mobility, the broadcast characteristics of wireless channels, and extremely high communication demands, IoV is prone to interruptions and high error rates during transmission, increasing the probability of congestion in the vehicle environment and drastically degrading network performance. Therefore, congestion has become one of the most important factors affecting the data transmission performance of vehicle-to-everything (V2X) networks, and how to solve congestion control in V2X networks is a hot research topic worthy of study.

[0003] In existing technologies, congestion control algorithms for vehicular networks (V2X) mainly fall into two categories: one analyzes the causes of packet loss in the V2X and uses TCP for end-to-end congestion control at the transport layer; the other implements hop-by-hop congestion control at the data link layer for data transmission channels. Both methods are congestion control solutions designed for a single layer. Furthermore, neighboring nodes in V2X share channels, and communication links are significantly affected by the external environment. The transmission rate of each link depends on the transmission status of neighboring links, and adjusting the transmission rate involves issues such as sending rate and power. V2X networks belong to multiple different layers, including the physical layer, MAC layer, and transport layer. In traditional layered approaches, information is difficult to share quickly between layers, leading to high latency.

[0004] In summary, the dynamic nature of vehicle-to-everything (V2X) networks leads to rapid changes in network topology and channel environment, as well as scarce network bandwidth. Furthermore, most V2X congestion control strategies are designed for specific layers and do not combine actual V2X communication conditions and service requirements to achieve cross-layer joint congestion control. This results in high latency for network services, making it difficult to meet the diverse service needs of a large number of vehicle users and often leading to unsatisfactory performance in V2X networks. Summary of the Invention

[0005] To address the aforementioned problems in the existing technology, this invention provides a joint congestion control method for vehicle-to-everything (V2X) networks based on cross-layer perception. The technical problem to be solved by this invention is achieved through the following technical solution:

[0006] A joint congestion control method for vehicle-to-everything (V2X) networks based on cross-layer perception includes:

[0007] The adjusted rate is obtained by adjusting the link transmission rate based on the link load.

[0008] Obtain the minimum contention window and average transmission time of a single data packet for the current channel;

[0009] Based on a Markov chain, any integer is selected from the minimum contention window as a backoff value to control the packet transmission status based on the backoff value;

[0010] The probability of a data packet being successfully sent is calculated based on the adjustment rate and the average transmission time of a single data packet. When the probability of a data packet being successfully sent is greater than a preset threshold, the data packet is marked as successfully sent.

[0011] In one specific implementation, adjusting the link transmission rate based on link load conditions to obtain the adjusted rate includes:

[0012] When a data packet is detected arriving on the link, the link queue length is calculated;

[0013] When it is determined that the length of the link queue does not exceed a preset threshold, an adjustment factor is calculated, and the link transmission rate is adjusted according to the adjustment factor to obtain the adjusted rate;

[0014] Update the queue length to iterate and obtain the adjustment rate based on the updated queue length.

[0015] In one specific implementation, adjusting the link transmission rate according to the adjustment factor to obtain the adjusted rate includes:

[0016] When the adjustment factor is determined to be non-zero, the transmission power of each link is calculated;

[0017] The data stream transmission rate from the source node to the destination node is calculated based on the transmission power of the link to obtain the adjusted rate.

[0018] In one specific implementation, the adjustment factor is:

[0019] Where γ represents the expected utilization rate of the queue, s i (t) represents the maximum transmission rate of link i. L represents the sum of the transmission rates of the upstream one-hop link. n express.

[0020] In one specific implementation, the transmission power of each link is:

[0021] Among them, w a For data stream weights, This is the sum of the lengths of each link.

[0022] In one specific implementation, the adjustment rate is:

[0023] q i (t) represents the queue length of link i, s i (t) represents the maximum transmission rate of link i.

[0024] In one specific implementation, the queue length is updated as follows:

[0025] in, It represents the sum of the rates of all data streams using the link, s i (t) is the maximum transmission rate of the link, LE i (t) represents the congestion factor.

[0026] In one specific implementation, the probability of successfully sending a data packet is: P a =2Rρτ·e -2Rρ Where R represents the transmission radius of the node, ρ represents the vehicle density, and τ represents the probability of data transmission at any given time.

[0027] In one specific implementation, the probability of data transmission at any given time is:

[0028] Where β represents the probability of a busy channel, W represents the minimum contention window, V represents the adjustment rate, and T... V The time to transmit a data packet; the probability of a busy channel is: β = 1 - e -2Rρτ .

[0029] The beneficial effects of this invention are:

[0030] The cross-layer perception-based joint congestion control method for vehicle networks of the present invention avoids channel congestion by allocating and controlling channel resources and using a channel backoff mechanism based on Markov chains to achieve channel sharing; it controls the transmission rate using a network resource allocation utility function to reduce the amount of data injected into the network and prevent packet loss; and it changes the transmission rate of the link by controlling the transmission power according to the adjustment factor and the priority of the data stream to avoid network congestion.

[0031] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0032] Figure 1 This is a schematic flowchart of a vehicle-to-everything (V2X) joint congestion control method based on cross-layer perception provided in an embodiment of the present invention.

[0033] Figure 2 This is a schematic diagram of a simplified backoff model based on Markov chains provided in an embodiment of the present invention. Detailed Implementation

[0034] The present invention will be further described in detail below with reference to specific embodiments, but the implementation of the present invention is not limited thereto.

[0035] Example 1

[0036] Figure 1 This is a schematic flowchart of a vehicle-to-everything (V2X) joint congestion control method based on cross-layer perception provided by an embodiment of the present invention, including:

[0037] S1. Adjust the link transmission rate based on the link load to obtain the adjusted rate;

[0038] In one specific implementation, adjusting the link transmission rate based on link load conditions to obtain the adjusted rate includes:

[0039] When a data packet is detected arriving on the link, the link queue length is calculated;

[0040] When it is determined that the length of the link queue does not exceed a preset threshold, an adjustment factor is calculated, and the link transmission rate is adjusted according to the adjustment factor to obtain the adjusted rate;

[0041] Adjusting the link transmission rate according to the adjustment factor to obtain the adjusted rate includes:

[0042] When the adjustment factor is determined to be non-zero, the transmission power of each link is calculated;

[0043] The data stream transmission rate from the source node to the destination node is calculated based on the transmission power of the link to obtain the adjusted rate.

[0044] Update the queue length to iterate and obtain the adjustment rate based on the updated queue length.

[0045] Specifically, in the Internet of Vehicles (IoV), adjusting the link transmission rate essentially means adjusting the link transmission power. When the link output rate is less than the data arrival rate, the link may become congested, leading to delayed data processing and forwarding, and ultimately, packet loss. If a link becomes congested, incoming data packets will be dropped. Upstream links will retransmit the lost packets. However, because the congestion is not under control, the congested link will continue to drop newly arriving data packets. If congestion on one link is not controlled in time, it will spread outwards, eventually causing congestion throughout the entire network.

[0046] Assume the transmission power of link i is G. i The transmission power of all links in the network is represented by a vector. In the Internet of Vehicles (IoV), the transmission rate of each link is no longer fixed, but rather a function of many global variables, such as modulation scheme and power vector. and channel conditions. Maximum transmission rate of the link (s) i As shown in the following formula:

[0047] s i (G i ) = log2(1+G i |h i | 2 );

[0048] Among them, h i This represents the channel gain of link i, where noise in the channel is normalized to the channel gain.

[0049] In the Internet of Vehicles (IoV), assume that the data stream emitted by source node a has a transmission rate of x. a The satisfaction of source node a with the network resource allocation function is defined as the utility function. As shown in the following formula:

[0050]

[0051] When the transmission rate of a link is adjusted, the transmission power of the link also changes. In vehicular networks, the transmission power of a link affects its transmission rate. This study aims to investigate the relationship between link transmission power and data traffic volume in the network, with the goal of maximizing user utility, and to use link queue length to predict network congestion.

[0052] Where ε represents the sensitivity factor for transmitting different service data in the network, as shown in the following formula:

[0053] ε=ΥRE i (t);

[0054] Where Υ represents the priority weight for transmitting different service data, RE i (t) represents the delay weight of the link, as shown in the following formula:

[0055] RE i (t)=R i (t) / R imax ;

[0056] Among them, R i (t) is the delayed response time of link i at time t; R imax This is the maximum delay of link i. When R i (t)≤R imax At that time, no message was lost, RE i (t)∈(0,1]; when R i (t)>R imax When a message is lost, the RE is set. i (t) = 1. Therefore, RE i The range of values ​​for (t) is (0,1).

[0057] Each link maintains a queue. When a data packet arrives, the link queue is updated, and the queue length changes accordingly. (Using q...) i Indicates the queue length, and has When the queue length exceeds a certain threshold, network congestion is suspected, and congestion information is sent to all source nodes using the link and all data sources. Upon receiving this information, the source nodes adjust their flow rates V. a Adjustments will be made.

[0058] Assume the link transmission rate of link i at time t is d. i The arrival rate of the link is the sum of the transmission rates of the upstream one-hop link, denoted as . The departure rate of link i is the link rate r(i) determined by the transmission power of that link. When congestion occurs at link i, the link transmission rate within two upstream hops is as follows:

[0059]

[0060] in, Let θ be the utility function, defined as an adjustment factor used to adjust the transmission rate of the node; here, θ is a constant. k represents the hop count range of the node whose rate needs to be adjusted.

[0061] Based on the link queue status, link transmission rate update information is transmitted to links within a two-hop range upstream. The calculation process for the adjustment factor θ is shown in the following formula:

[0062]

[0063] Establish virtual power queue Q i (t) Track the power consumption of the link and solve the power allocation problem, as shown in the following formula. The virtual queue of link power does not exist in actual application. In the simulation process, its role is equivalent to a mathematical counter.

[0064] Q i (t+1)=[Q i (t)+(G i (t)-G avg )] + ;

[0065] Given that the transmission power of link i at time t is G i G avg This represents the average power.

[0066] The transmission power of each link is calculated based on the link's transmission rate, and the specific calculation process is shown in the following formula:

[0067]

[0068] Data stream transmission rate from source node to destination node Use functions Maximize, as shown in the following formula:

[0069]

[0070] Where, q i This can be represented as the queue length of the link, or it can be understood as the implicit cost of link i. Clearly, q i The smaller the value of w, the lower the cost of using link i, and the larger the value of the objective function for using link i. a It represents the weight of the data stream.

[0071] Data stream weights w a This is quantified as the urgency of the data to be sent, i.e., the priority of the data stream. As shown in the following formula:

[0072] w a =(ACI+1) / (N) ACI +1);

[0073] Where, N ACI =4 represents the number of ACI categories. In VANET, different data types have different priorities. Based on the importance and urgency of the messages, IEEE 802.11p specifies four access categories (AC), denoted by N. ACIThis indicates that each Access Categories Index (ACI) has an ACI index, where ACI ∈ [0, 1, 2, 3]. The larger the ACI, the higher the priority of the information, and the greater its importance and urgency. Adding 1 to the numerator and denominator avoids values ​​of 0. Clearly, ACI < N. ACI The data priority value range is w a ∈(0,1).

[0074] After rate adjustment, the queue length is updated, and the link queue length is iteratively updated according to the following formula.

[0075]

[0076] Here, γ represents the expected utilization rate of the queue, and γ < 1. When γ is close to 1, it indicates that the current link utilization rate is high; conversely, when γ is close to 0, it indicates that the link utilization rate is low. The value of γ can be set according to the actual situation of the link.

[0077] Within each time slot, the link queue length is calculated by adding the newly added queue length to the link queue length of the previous time slot. (The parentheses on the right side of the equation are...) This represents the length of the newly added link queue, which can be either positive or negative. It represents the sum of the rates of all data streams using the link, s i (t) is the maximum transmission rate of the link, LE i (t) represents the congestion factor. If there is If the sum of the data rates arriving on the link is greater than the link's actual maximum rate, then data forwarding will be delayed and the data will queue up, increasing the queue length. Conversely, if the data is not forwarded in time, the link can continue to transmit data after the arrival data has been forwarded, so some of the data that was waiting in the queue will be transmitted, thus shortening the link's queue length.

[0078] Among them, the congestion factor LE i (t) represents the congestion level of the link load, as shown in the following formula:

[0079] LE i (t)=L i (t) / L imax ;

[0080] Among them, L imax L is the capacity of the transmit buffer of link i. i (t) is the size of the occupied portion at time t. LE i The range of values ​​for (t) is (0,1).

[0081] S2. Obtain the minimum contention window and average transmission time of a single data packet for the current channel;

[0082] S3. Select any integer from the minimum contention window as a backoff value based on the Markov chain to control the data packet sending state based on the backoff value;

[0083] S4. Calculate the probability of successful data packet transmission based on the adjustment rate and the average transmission time of a single data packet. When the probability of successful data packet transmission is greater than a preset threshold, mark the data packet as successfully transmitted.

[0084] Please see Figure 2 , Figure 2 This is a schematic diagram of a simplified backoff model based on Markov chains provided in an embodiment of the present invention. In the diagram, W is the minimum contention window size, the probability that at least one node has a data packet waiting to be sent at any given time is α, and the probability that the channel is busy is β.

[0085] When a node has data to send, it enters a waiting state, randomly selecting an integer from (0, 1, ..., W-1) as the backoff value. The node then listens to the channel to check if it is busy. If the channel is idle, the counter is decremented by 1; otherwise, it remains unchanged. When the counter reaches 0, data is sent. If there is still data to send, the node enters the next waiting state. Solving the Markov chain yields the probability b0 of the counter backing up to 0, as shown in the following formula:

[0086] b0 = 2(1-β) / 2-3β+βW;

[0087] The data transmission rate V follows a Poisson distribution. If the average time to transmit one data packet is T... V The probability α that at any given time there is at least one data packet waiting to be sent can be obtained by the following formula:

[0088]

[0089] The probability τ of data transmission at any given time is calculated using the following formula:

[0090]

[0091] The probability β of a busy channel is shown in the following formula:

[0092]

[0093] Here, R represents the transmission radius of the node. The two probability formulas above form a system of nonlinear equations, which can be proven to have a unique solution in (0,1). By solving the system of nonlinear equations, the values ​​of β and τ can be obtained.

[0094] While backoff mechanisms effectively prevent multiple nodes from occupying the channel in an unordered manner, they cannot completely prevent data collisions. Data packets can only be successfully received by other nodes if only one node is transmitting within the sending node's range; otherwise, a collision will occur. The probability P of successful data packet transmission can be derived. a As shown in the formula below:

[0095]

[0096] Simplifying, we get P a =2Rρτ·e -2Rρ .

[0097] When the probability P of successfully sending the data packet a If the value is greater than 0.95, the data packet is marked as successfully sent.

[0098] The simulation analysis based on the scheme in this embodiment is as follows: The simulation environment is built on a Linux system, with SUMO and NS-3 simulation software installed. SUMO performs traffic simulation, while NS3 is responsible for network simulation. The Linux system is Ubuntu 16.04, the NS-3 version is NS-33.27, and the SUMO version is SUMO 0.32.0.

[0099] First, a traffic simulation scenario is built using SUMO. The parameters of the urban traffic scenario are shown in Table 1. Considering a real urban traffic scenario, this scenario contains 0-500 vehicle nodes. The generated scenario is random; the faster the average speed of the nodes, the more frequent the topology changes.

[0100] Table 1 Parameters of Traffic Simulation Scenarios

[0101]

[0102] Secondly, the traffic flow information from SUMO was imported into the NS3 simulation software as parameters to verify the performance of the strategy in this embodiment. The network simulation scenario parameters are shown in Table 2.

[0103] Table 2 Network Simulation Parameters

[0104]

[0105]

[0106] At different bit error rates, the method of this embodiment was compared with NewReno, TOSCL, VC-TCP / NC, and CLCC to complete data transmission of information services such as dynamic maps and AR navigation, and the average latency was compared. As the number of vehicles increased from 50 to 500, the average latency of this embodiment increased from 9.5ms to 105ms at a low bit error rate, and increased to a maximum of 150ms at a high bit error rate. In this embodiment, the average latency throughput of transmitted data remained largely within 150ms under both low and high bit error rates, with only slight fluctuations at high bit error rates. In contrast, the average latency of the NewReno, TOSCL, VC-TCP / NC, and CLCC algorithms was higher than that of this embodiment, regardless of whether the bit error rate was low or high.

[0107] The method in this embodiment consistently delivers lower average latency than TOSCL, VC-TCP / NC, and CLCC, primarily because the cross-layer congestion algorithm combines queue expectation values ​​with a utility function. And adjust the transmission rate by factor θ, combined with the data stream priority weight w a Adjusting the link rate reduces the computational latency at the MAC layer, thereby reducing the average network latency.

[0108] To more effectively analyze algorithm performance, assuming the vehicles are in a high-density state, under the same simulation environment, files ranging from 50MB to 500MB are transferred using an FTP application. The average throughput of this embodiment is compared with that of NewReno, TOSCL, VC-TCP / NC, and CLCC under different packet loss rates.

[0109] When the packet loss rate is low, all five algorithms exhibit high throughput, with the algorithm in this embodiment outperforming CLCC, VC-TCP / NC, TOSCL, and NewReno. As the packet loss rate increases, the throughput of all five algorithms decreases. However, the algorithm in this embodiment maintains a consistently high and relatively stable average throughput during file transmission. In contrast, NewReno's throughput significantly decreases as the packet loss rate increases because it cannot quickly identify the cause of packet loss.

[0110] Then, the average throughput of different strategies under different vehicle densities was compared when the packet loss rate was relatively high (0.007-0.009). When the number of vehicles increased from 50 to 500, the average throughput of the algorithm in this embodiment increased by 28%, which was always higher than that of CLCC, VC-TCP / NC, TOSCL and NewReno methods, while the throughput of NewReno only increased by 7.7%.

[0111] When there are 500 vehicles, the packet loss rate of the algorithm in this embodiment is 17.5%, while the packet loss rate of the NewReno strategy is 57.4%. The packet loss rate of the algorithm in this embodiment is always lower than that of CLCC, VC-TCP / NC, TOSCL, and NewReno strategies. When there are 500 vehicles, the collision probability of the algorithm in this embodiment is 0.19, while the collision probability of the NewReno strategy is 0.95.

[0112] Because the algorithm in this embodiment introduces data flow weight w a By comparing the priorities of different data types using this parameter, the link rate is adjusted reasonably according to the importance and urgency of messages. This effectively avoids frequent and rapid drops in the window, which can lead to a decrease in throughput. Compared to NewReno, TOSCL, VC-TCP / NC, and CLCC, the algorithm in this embodiment has a more stable throughput, which is reflected in a significant improvement in the overall throughput performance of the data transmission process.

[0113] The vehicle-to-everything (V2X) joint congestion control method based on cross-layer perception in this embodiment avoids channel congestion by allocating and controlling channel resources and using a Markov chain-based channel backoff mechanism to achieve channel sharing; it controls the transmission rate using a network resource allocation utility function to reduce the amount of data injected into the network and prevent packet loss; and it changes the transmission rate of the link by controlling the transmission power based on adjustment factors and data stream priorities to avoid network congestion.

[0114] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. In addition, those skilled in the art can combine and integrate the different embodiments or examples described in this specification.

[0115] Although this application has been described herein in conjunction with various embodiments, those skilled in the art, by reviewing the accompanying drawings, disclosure, and appended claims, will understand and implement other variations of the disclosed embodiments in carrying out the claimed application. In the claims, the word "comprising" does not exclude other components or steps, and "a" or "an" does not exclude a plurality. A single processor or other unit can implement several functions listed in the claims. While different dependent claims may recite certain measures, this does not mean that these measures cannot be combined to produce good results.

[0116] The above description, in conjunction with specific preferred embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various simple deductions or substitutions can be made without departing from the concept of the present invention, and all such modifications and substitutions should be considered within the scope of protection of the present invention.

Claims

1. A joint congestion control method for vehicle-to-everything (V2X) networks based on cross-layer perception, characterized in that, include: The adjusted rate is obtained by adjusting the link transmission rate based on the link load. Obtain the minimum contention window and average transmission time of a single data packet for the current channel; Based on a Markov chain, any integer is selected from the minimum contention window as a backoff value to control the packet transmission status based on the backoff value; The probability of a data packet being successfully sent is calculated based on the adjustment rate and the average transmission time of a single data packet. When the probability of a data packet being successfully sent is greater than a preset threshold, the data packet is marked as successfully sent. Adjusting the link transmission rate based on link load conditions to obtain the adjusted rate includes: When a data packet is detected arriving on the link, the link queue length is calculated; When it is determined that the length of the link queue does not exceed a preset threshold, an adjustment factor is calculated, and the link transmission rate is adjusted according to the adjustment factor to obtain the adjusted rate; Update the queue length to iterate and obtain the adjustment rate based on the updated queue length; Adjusting the link transmission rate according to the adjustment factor to obtain the adjusted rate includes: When the adjustment factor is determined to be non-zero, the transmission power of each link is calculated; The data stream transmission rate from the source node to the destination node is calculated based on the transmission power of the link to obtain the adjustment rate; The probability of successfully sending a data packet is: ,in, Indicates the transmission radius of the node. Indicates vehicle density. This represents the probability of data transmission at any given time. The probability of data transmission at any given time is: ,in, Let W represent the probability that the channel is busy, W represent the minimum contention window, and V represent the adjustment rate. The time to transmit a data packet; the probability of a busy channel is: .

2. The joint congestion control method for vehicle-to-everything (V2X) networks based on cross-layer perception as described in claim 1, characterized in that, The adjustment factor is: ,in, The expected utilization rate of the queue is represented. This represents the maximum transmission rate of link i. This represents the sum of the transmission rates of the upstream one-hop link. express.

3. The joint congestion control method for vehicle-to-everything (V2X) networks based on cross-layer perception as described in claim 1, characterized in that, The transmission power of each link is: ,in, For data stream weights, This is the sum of the lengths of each link.

4. The vehicle-to-everything (V2X) joint congestion control method based on cross-layer perception according to claim 3, characterized in that, The adjustment rate is: , This represents the queue length of link i. This represents the maximum transmission rate of link i.

5. The joint congestion control method for vehicle-to-everything (V2X) networks based on cross-layer perception according to claim 1, characterized in that, Update the queue length as follows: ,in, This represents the sum of the rates of all data streams using the link. It is the maximum transmission rate of the link. This represents the congestion factor.

Images