A train on-board communication network scheduling method and system

Abstract

This application belongs to the field of train communication network scheduling technology. It provides a train onboard communication network scheduling method and system. The method includes: acquiring communication data streams in the train communication network; dividing them into periodic data streams and aperiodic data streams, further dividing the periodic data streams into first-level periodic data streams and second-level periodic data streams; dividing each basic time slot into first-level, second-level, and third-level time slices, transmitting the three types of data streams respectively; allocating the periodic data streams according to an inter-slot load balancing algorithm and setting the start and end threshold times for each time slice, reserving third-level time slices for aperiodic data streams, thus obtaining a communication scheduling gating list; when aperiodic data streams appear, using a trust mechanism to optimize the scheduling of the aperiodic data streams, obtaining the final scheduling configuration. This invention ensures priority transmission of important and real-time data, achieves real-time response and long-term scheduling balance for aperiodic data streams.

Description

Technical Field

[0001] This invention belongs to the field of train communication network scheduling technology, specifically relating to a train onboard communication network scheduling method and system. Background Technology

[0002] With the development of train communication networks, Time-Sensitive Networking (TSN) is increasingly being applied to train control systems to ensure the real-time and reliable transmission of critical control data and monitoring information. In existing technology, a patent document titled "A Train Communication Network Scheduling Method Based on a TSN-based CQF Mechanism," publication number CN120528806A, proposes a flow scheduling optimization algorithm based on joint flow and routing characteristics by analyzing the physical constraints of train communication network switching nodes and flow transmission constraints. This method can select the flow scheduling strategy with the highest priority according to a sorting formula, optimizing the entire flow scheduling process and improving scheduling efficiency.

[0003] However, this method still has several shortcomings in practical applications. First, it performs online dynamic scheduling of all data streams in the train communication network. To ensure communication reliability, practical applications typically require a fixed scheduling time for each data stream within the TSN gating list, especially for data streams closely related to train operation control. Online dynamic scheduling significantly increases the uncertainty of data stream transmission, potentially affecting the reliable transmission of critical data. Second, when high-priority data streams continuously exist in the network, lower-priority data streams may be unable to transmit for extended periods, leading to uneven resource allocation and potential scheduling risks. Therefore, it still has significant shortcomings in ensuring the reliability of critical data, rationally arranging periodic data streams, and ensuring the fairness of dynamic scheduling. Summary of the Invention

[0004] The technical problem this invention aims to solve is that online dynamic scheduling of all data streams can lead to delays in critical periodic data streams related to train operation control, affecting communication reliability, and low-priority data streams may be unable to be transmitted for extended periods. This invention provides a train onboard communication network scheduling method and system.

[0005] The present invention includes: Firstly, the present invention provides a train onboard communication network scheduling method, comprising:

[0006] Acquire communication data streams from the train communication network;

[0007] Based on the nature of the communication data stream, the communication data stream is divided into periodic data streams and aperiodic data streams. The periodic data stream is further divided into primary periodic data streams and secondary periodic data streams. The nature of the communication data stream includes the source node, target node, and task type of the sampling task. Primary periodic data streams are periodic data streams with high importance, secondary periodic data streams are periodic data streams with moderate importance, and aperiodic data streams are bursty data streams caused by equipment failures or configuration updates in the train communication network.

[0008] Within the macro cycle of communication scheduling, each basic time slot is divided into a first-level time slot, a second-level time slot, and a third-level time slot to transmit the first-level periodic data stream, the second-level periodic data stream, and the aperiodic data stream, respectively.

[0009] According to the time-slot load balancing algorithm, the first-level periodic data stream and the second-level periodic data stream are allocated, and the third-level time slice is reserved for the non-periodic data stream. The start and end gate times of each time slice are set for the first-level periodic data stream and the second-level periodic data stream to obtain the communication scheduling gate list. The gate list is a list that controls the transmission order of various data streams by being offline distributed to each switch device through a dedicated network controller or TSN configuration server. TSN is a time-sensitive network.

[0010] During the scheduling process, when the non-periodic data stream appears, the gating list uses a trust mechanism to optimize the scheduling of the non-periodic data stream, obtaining a final scheduling configuration to achieve optimized bandwidth allocation and real-time response to bursty data streams. The final scheduling configuration is a scheduling mechanism that prioritizes the real-time transmission of the first-level periodic data stream, the second-level periodic data stream, and the non-periodic data stream when the non-periodic data stream appears. The second-level periodic data stream is defined as the data stream whose trust value is higher than that of the non-periodic data stream. The non-periodic data stream is defined as the non-periodic data stream that has reached the set trust value and whose third-level time slice bandwidth is insufficient in the current time slot.

[0011] Furthermore, based on the nature of the communication data stream, the communication data stream is divided into periodic data streams and aperiodic data streams. The periodic data streams are further divided into first-level periodic data streams and second-level periodic data streams, including:

[0012] The source node, target node, and task type of each communication data stream sampling task in the train control system are collected;

[0013] Fixed-period sampling tasks are classified as periodic data streams, and the periodic data streams are divided into first-level periodic data streams and second-level periodic data streams according to the sampling task level; non-fixed-period sampling tasks are classified as non-periodic data streams, wherein the task level includes sampling task priority, real-time requirements, and security level.

[0014] Furthermore, within the macro-cycle of communication scheduling, each basic time slot is divided into a first-level time slice, a second-level time slice, and a third-level time slice to transmit the first-level periodic data stream, the second-level periodic data stream, and the aperiodic data stream, respectively, including:

[0015] Within the macro cycle of communication scheduling, the macro cycle is divided into a predetermined number of basic time slots of fixed length.

[0016] The formula for calculating the basic time slot is as follows:

[0017]

[0018] in, Indicates the basic time slot, Indicates the first The transmission time of a periodic data stream, Indicates the first Is the first basic time slot transmitted? Data stream, Indicates transmission, Indicates no transmission. Indicates the total number of periodic data streams; This represents the transmission time of an aperiodic data stream; the formula for calculating the number of periodic basic time slots is as follows:

[0019]

[0020] in, N Indicates the number of basic time slots in a period. The function represents the least common multiple.

[0021] Each basic time slot is divided into a first-level time slice, a second-level time slice, and a third-level time slice, which are respectively allocated to the first-level periodic data stream, the second-level periodic data stream, and the aperiodic data stream for the transmission of various types of data streams.

[0022] Furthermore, based on the inter-slot load balancing algorithm, the first-level periodic data stream and the second-level periodic data stream are allocated, a third-level time slice is reserved for the non-periodic data stream, and start and end gate times for each time slice are set for the first-level periodic data stream and the second-level periodic data stream, resulting in a communication scheduling gate list, including:

[0023] Calculate the total bandwidth of the periodic data stream within each macro cycle. The formula for calculating the total bandwidth of the periodic data stream is as follows:

[0024]

[0025] Where Bp represents the total bandwidth within the macro period, Indicates the number of transmissions of the periodic data stream within the macro period. Represents data stream Number of bytes, This represents the macro period, and the formula for calculating the macro period is as follows:

[0026] ;

[0027] Calculate the initial load for each time slot and obtain the load variance, where the initial load for each time slot is... The calculation formula is:

[0028]

[0029] The formula for calculating the load variance is as follows:

[0030]

[0031] in, Indicates the average load within the macro period. ;

[0032] Based on the load variance, within the time slot range allocated to various data streams, the starting transmission time slot of the data stream is iteratively adjusted using a random offset method. After the iteration terminates, the adjusted load variance is calculated.

[0033] Based on the adjusted load variance, a target allocation scheme is selected according to the inter-slot load balancing algorithm, wherein the target allocation scheme is the allocation scheme with the smallest adjusted load variance.

[0034] According to the target allocation scheme, the first-level periodic data stream and the second-level periodic data stream are allocated, the third-level time slice is reserved for the non-periodic data stream, and the start and end gate times of each time slice are set for the first-level periodic data stream and the second-level periodic data stream to generate the gate list.

[0035] Furthermore, the criterion for terminating the iteration is when the load variance is 0 or when the preset number of iterations is reached.

[0036] Furthermore, during the scheduling process, the gating list employs a trust mechanism to optimize the scheduling of the aperiodic data stream when it occurs, obtaining a final scheduling configuration to achieve optimized bandwidth allocation and real-time response to bursty data streams, including:

[0037] During the scheduling process, the gating list monitors the arrival of non-periodic data streams in real time.

[0038] When the aperiodic data stream arrives, an initial trust value is set for each aperiodic data stream based on its priority, latency requirements, and current time slot occupancy.

[0039] An improved bubble sort algorithm is used to calculate the trust value and sort the data. The transmission order of the aperiodic data streams within the three-level time slice is adjusted, and the trust value is compensated for the aperiodic data streams that are delayed in transmission.

[0040] Based on trust value compensation, when the set non-periodic data stream occurs, the distribution of the secondary periodic data streams in the gating list is temporarily adjusted to generate a dynamic scheduling table so that the set non-periodic data stream can perform preemptive scheduling.

[0041] After the preemptive scheduling is completed, the dynamic scheduling table is restored to the distribution of the secondary periodic data streams in the original gating list to achieve optimized bandwidth allocation and real-time response to bursty data streams.

[0042] Furthermore, the improved bubble sort method includes:

[0043] Define the trust value threshold for determining the trust value level of non-periodic data streams, and the method for calculating the trust value;

[0044] For any two detected aperiodic data streams, a trust value level is determined. If the difference in trust values ​​between the two aperiodic data streams is greater than the trust value threshold, then the trust value of the former aperiodic data stream is determined to be greater than that of the latter. Conversely, if the positions of the trust values ​​of the two aperiodic data streams are swapped to satisfy the condition that the difference in trust values ​​between the two aperiodic data streams is greater than the trust value threshold, then the trust value of the latter is compensated. The formula for determining the trust value level is as follows:

[0045]

[0046] in, This represents the trust value of the former non-periodic data stream. This represents the trust value of the latter's non-periodic data stream. Indicates the trust threshold;

[0047] The formula for calculating trust value compensation is as follows:

[0048]

[0049] in, Indicates the trust threshold. It is a regulating factor.

[0050] Furthermore, the temporary adjustment of the distribution of the secondary periodic data streams within the gating list includes:

[0051] The non-periodic data streams that are currently detected and cannot be transmitted within the pre-allocated three-level time slice are taken as the remaining data streams, and the trust value of the remaining data streams is calculated.

[0052] Based on the stream trust value of the remaining data, a set of secondary periodic data streams for replacing the remaining data streams is found within the current time slot, and the calculation formula is as follows:

[0053]

[0054] in, This represents the set of secondary periodic data streams that can be used to replace the remaining data streams. This indicates the number of secondary periodic data streams in the current time slot. Indicates the first time slot in the current time slot A secondary periodic data stream, Indicates the first time slot in the current time slot j Trust value of a secondary periodic data stream. This indicates the number of remaining data streams.

[0055] The set of secondary periodic data streams is used to find the secondary periodic data stream that minimizes the trust value. The calculation formula is as follows:

[0056]

[0057] in, This represents the second-level periodic data stream with the lowest trust value.

[0058] Place the data stream with the highest trust value in the remaining data stream at the transmission position of the second-level periodic data stream with the lowest trust value in the second-level periodic data stream;

[0059] Trust value compensation is performed on the secondary periodic data stream with the lowest trust value.

[0060] Repeat all the steps except the first step until all remaining data streams have been allocated, or until all secondary periodic data streams in the set of secondary periodic data streams have been traversed.

[0061] Furthermore, the formula for calculating the trust value compensation for the secondary periodic data stream with the smallest trust value is as follows:

[0062]

[0063] in, This represents the second-level periodic data stream with the lowest trust value.

[0064] Secondly, the present invention provides a train onboard communication network dispatching system, comprising:

[0065] The data acquisition module is used to acquire communication data streams from the train communication network.

[0066] A data segmentation module is used to divide the communication data stream into periodic data streams and aperiodic data streams based on the nature of the communication data stream. The periodic data stream is further divided into primary periodic data streams and secondary periodic data streams. The nature of the communication data stream includes the source node, target node, and task type of the sampling task. Primary periodic data streams are periodic data streams with high importance, secondary periodic data streams are periodic data streams with moderate importance, and aperiodic data streams are bursty data streams caused by equipment failures or configuration updates in the train communication network.

[0067] The time slot allocation module is used to divide each basic time slot into a first-level time slot, a second-level time slot, and a third-level time slot within the macro cycle of communication scheduling, so as to transmit the first-level periodic data stream, the second-level periodic data stream, and the aperiodic data stream, respectively.

[0068] The load balancing module is used to allocate the first-level periodic data stream and the second-level periodic data stream according to the time slot load balancing algorithm, reserve the third-level time slot for the non-periodic data stream, and set the start and end gate times for each time slot of the first-level periodic data stream and the second-level periodic data stream to obtain the communication scheduling gate list. The gate list is a list that controls the transmission order of various data streams by being offline distributed to each switch device through a dedicated network controller or TSN configuration server. TSN is the abbreviation for Time Sensitive Networking.

[0069] The scheduling configuration module is used by the gating list to optimize the scheduling of the non-periodic data stream when it appears during the scheduling process, using a trust mechanism to obtain the final scheduling configuration. This configuration aims to achieve optimized bandwidth allocation and real-time response to bursty data streams. The final scheduling configuration is a scheduling mechanism that prioritizes the real-time transmission of the first-level periodic data stream, the second-level periodic data stream, and the non-periodic data stream when it appears. The second-level periodic data stream is defined as a data stream whose trust value is higher than that of the non-periodic data stream. The non-periodic data stream is defined as a non-periodic data stream that has reached a set trust value and whose bandwidth of the third-level time slice is insufficient in the current time slot.

[0070] The beneficial effects of this invention are:

[0071] 1. This invention divides the data stream of the train's onboard communication network according to its periodicity and task importance, and divides each basic time slot into time slices of different levels for transmission within the macro-period. This achieves reasonable priority management and bandwidth allocation for periodic data streams, ensuring that important and real-time data are transmitted first, and avoiding delays in critical data caused by non-periodic data streams or high loads.

[0072] 2. This invention optimizes the allocation of periodic data streams based on a time-slot load balancing algorithm. By traversing random offset combinations and using the minimum variance criterion, an optimized gating schedule is generated and written into the TSN scheduling controller. This provides a static scheduling strategy for periodic data streams, reducing the uncertainty of transmission scheduling for important data streams related to train operation. At the same time, it provides an optimal bandwidth allocation scheme for the transmission of non-periodic data streams, reducing the risk of network congestion and improving the stability and reliability of communication scheduling.

[0073] 3. This invention combines a trust mechanism to temporarily adjust the scheduling strategy of low-trust-value data streams in the gating list when non-periodic data streams arrive, generating a dynamic scheduling table to perform preemptive scheduling on non-periodic data streams; at the same time, it compensates for the trust value of delayed data streams to prevent the problem that data streams with low trust values ​​may not be able to be transmitted for a long time. Attached Figure Description

[0074] Figure 1 This is a flowchart illustrating a train onboard communication network scheduling method according to the present invention.

[0075] Figure 2 This is a schematic diagram of a TSN gated list according to the present invention;

[0076] Figure 3 This is a block diagram of a train-mounted communication network dispatching system according to the present invention. Detailed Implementation

[0077] like Figure 1 , 2 As shown, the present invention provides a train onboard communication network scheduling method, comprising:

[0078] S1. Obtain the communication data stream in the train communication network;

[0079] S2. Based on the nature of the communication data stream, the communication data stream is divided into periodic data streams and aperiodic data streams. The periodic data stream is further divided into first-level periodic data streams and second-level periodic data streams. The nature of the communication data stream includes the source node, target node, and task type of the sampling task. The first-level periodic data stream is a periodic data stream with high importance, the second-level periodic data stream is a periodic data stream with moderate importance, and the aperiodic data stream is a bursty data stream caused by equipment failure or configuration update in the train communication network.

[0080] Burst-type data streams include: equipment failures, configuration updates, etc.

[0081] S21. Collect the source node, target node, and task type of each communication data stream sampling task in the train control system;

[0082] In one embodiment, firstly, information on all communication data stream sampling tasks that need to be transmitted is obtained from the train control system. This includes: source nodes such as sensor units and onboard controllers; target nodes such as the train control center or other onboard terminals; and task types such as temperature sampling, location data, and control commands. Each information type has different transmission requirements and priorities. Then, by collecting information on the source node, target node, and task type of each sampling task, the transmission path and characteristics of each data stream in the communication network are fully obtained.

[0083] S22. The fixed-period sampling task type is determined as a periodic data stream, and the periodic data stream is divided into the first-level periodic data stream and the second-level periodic data stream according to the importance of the sampling task; the non-fixed-period sampling task is determined as a non-periodic data stream, wherein the task level is the importance of the task, including the sampling task priority, real-time requirements and security level.

[0084] In one embodiment, data streams sent at fixed time intervals for fixed-period sampling tasks are classified as periodic data streams; data streams triggered by events or generated temporarily are classified as non-periodic data streams. Then, periodic data streams are further subdivided based on the importance of the tasks along the data stream transmission path: task levels include sampling task priority, real-time requirements, and security level. Combining these characteristics, periodic data streams are divided into: Level 1 periodic data streams: data streams with extremely high importance and strict real-time and security requirements; and Level 2 periodic data streams: data streams with moderate importance and relatively low real-time and security requirements. Through this classification method, the train control system provides clear classification criteria for communication scheduling based on the transmission characteristics and importance of different data streams, ensuring that data transmission for critical tasks is prioritized when resources are limited.

[0085] S3. Within the macro cycle of communication scheduling, each basic time slot is divided into a first-level time slot, a second-level time slot, and a third-level time slot to transmit the first-level periodic data stream, the second-level periodic data stream, and the aperiodic data stream, respectively.

[0086] S31. Within the macro cycle of communication scheduling, the macro cycle is divided into a set number of basic time slots of fixed length;

[0087] in, Indicates the basic time slot, Indicates the first The transmission time of a periodic data stream, Indicates the first Is the first basic time slot transmitted? Data stream, Indicates transmission, Indicates no transmission. Indicates the total number of periodic data streams; This represents the transmission time of an aperiodic data stream; the formula for calculating the number of periodic basic time slots is as follows:

[0088]

[0089] in, N Indicates the number of basic time slots in a period. The function represents the least common multiple.

[0090] In one embodiment, the macro cycle length of communication scheduling is first determined. The macro cycle refers to the time interval for completing a full data stream scheduling in the train communication network. For example, the macro cycle can be set to several seconds or minutes according to the operating characteristics of the train control system. Then, the macro cycle is divided into a number of basic time slots with a fixed length. Each basic time slot has the same length and can accommodate a certain number of data stream transmissions. The above method evenly divides the transmission tasks in the macro cycle, enabling the orderly arrangement of data stream transmissions within each basic time slot, facilitating time slice division and scheduling optimization. Among them, each basic time slot can be regarded as an independent scheduling unit for allocating data stream transmission tasks with different priorities.

[0091] S32. Each basic time slot is divided into a first-level time slice, a second-level time slice, and a third-level time slice, which are respectively allocated to the first-level periodic data stream, the second-level periodic data stream, and the aperiodic data stream for the transmission of various data streams.

[0092] In one embodiment, the scheduling arrangement of each basic time slot is further refined and divided into three different time slices: First-level time slice: used for transmitting the first-level periodic data stream, that is, the data stream with extremely high importance, strict requirements for real-time performance and security, to ensure the priority completion of key tasks. Second-level time slice: used for transmitting the second-level periodic data stream, that is, the data stream with general importance and relatively low requirements for real-time performance and security. Third-level time slice: used for transmitting the aperiodic data stream. The aperiodic data stream is a bursty data stream that appears when there are equipment failures or configuration updates in the train communication network, belonging to event-triggered or temporarily generated data streams. The arrival of this type of data stream is unpredictable, but it still needs to be reasonably arranged within the basic time slot. Through the above division method, the transmission order and priority of data streams within each basic time slot are clearly defined, ensuring that high-priority data streams can be preferentially scheduled under limited network resources, while providing scheduling space for aperiodic data streams, and realizing the bandwidth optimization and burst message response ability of the overall communication network.

[0093] S4. According to the load balancing algorithm between time slots, the first-level periodic data stream and the second-level periodic data stream are allocated, the third-level time slice is reserved for the aperiodic data stream, and the start and end gating times of each time slice are set for the first-level periodic data stream and the second-level periodic data stream, obtaining a gating list for communication scheduling. The gating list is a list for controlling the transmission order of various data streams that is offline distributed to each switch device by a dedicated network controller or a TSN configuration server. TSN is the Time-Sensitive Network;

[0094] Among them, TSN (Time-Sensitive Networking) is translated as Time-Sensitive Network in Chinese.

[0095] S41. Calculate the total bandwidth of the periodic data stream within each macro cycle, where the formula for calculating the total bandwidth of the periodic data stream is as follows:

[0096]

[0097] Where Bp represents the total bandwidth within the macro period, Indicates the number of transmissions of the periodic data stream within the macro period. Represents data stream Number of bytes, This represents the macro period, and the formula for calculating the macro period is as follows:

[0098]

[0099] In one embodiment, the network bandwidth required by all periodic data streams throughout the entire macrocycle is calculated to provide a basis for time slot allocation. The transmission requirements of each data stream are determined by the number of transmissions within the macrocycle and the amount of data transmitted each time. By calculating the total bandwidth, the communication load throughout the entire macrocycle can be understood.

[0100] S42. Calculate the initial load for each time slot and obtain the load variance, where the initial load for each time slot is... The calculation formula is:

[0101]

[0102] The formula for calculating the load variance is as follows:

[0103]

[0104] in, Indicates the average load within the macro period. ;

[0105] In one embodiment, the initial load for each basic timeslot is calculated to assess the load distribution balance. The load for each basic timeslot is calculated by statistically analyzing the data stream transmission volume allocated within that timeslot. The average load over the macrocycle is the average load of all basic timeslots. Load variance measures the difference between the load of each timeslot and the average load over the macrocycle; by calculating the load variance, the balance of the load distribution is determined.

[0106] For example, the purpose of the inter-slot load balancing algorithm is to balance the load in each basic time slot. In China, how to pre-allocate and The data flow scheduling strategy ensures that the load within each basic time slot is approximately equal. Among these, F A This represents a first-order periodic data stream. FB This indicates a two-level periodic data stream.

[0107] Assume a macrocycle of TSN includes One basic time slot, namely ; Data streams have species, that is , No. The transmission period of this data stream is ; Data streams have species, that is , No. The transmission period of this data stream is The inter-slot load balancing algorithm then becomes about determining the combination of data streams within each basic time slot based on the transmission cycle requirements of each data stream. Specifically:

[0108] set up The transmission period is , The transmission period is , The transmission period is Therefore, one feasible approach is the allocation scheme shown in Table 1. That is, in... Real-time transmission , and Three data streams, in Real-time transmission Data flow, in Real-time transmission Data stream. While this allocation scheme can meet transmission requirements, different basic time slots... The load distribution is uneven, and the order of some data streams within each macro cycle is not fixed. In the allocation scheme shown in Table 1, Three data streams are transmitted at any time. Always free and Only one data stream needs to be transmitted at any given time. Therefore, an optimal allocation scheme needs to be found to make the load within each basic time slot relatively uniform, so as to leave more time for non-periodic data streams. F C .

[0109] Table 1. Data Stream Allocation Diagram Between Time Slots

[0110]

[0111] S43. Based on the load variance, within the time slot range allocated to various data streams, the starting transmission time slot of the data stream is iteratively adjusted using a random offset method. After the iteration terminates, the adjusted load variance is calculated.

[0112] The iteration termination condition is when the load variance is 0 or the preset number of iterations is reached.

[0113] In one embodiment, an iterative optimization method is used to adjust the starting transmission time slots of data streams in each basic time slot to reduce load variance and achieve load balancing: For each data stream, a random offset is generated to adjust its starting transmission time slot position within the macro-cycle; in each iteration, the starting time slot of the data stream is adjusted according to the random offset, and the adjusted load of each time slot and the macro-cycle load variance are calculated; when the iteration termination condition is met, such as when the load variance is 0 or when a preset number of iterations is reached, the adjustment stops, and the adjusted load variance is calculated. While ensuring priority transmission of periodic data streams, the load of each time slot is balanced as much as possible to avoid excessively high loads or conflicts in specific time slots, thereby improving the overall network transmission efficiency.

[0114] For example, a random traversal method is adopted, which, for each periodic data stream, is applied according to a random offset. Among them, satisfying The matrix X is adjusted to obtain a series of possible combinations; then, the value that minimizes the variance is found. Indicates the first The algorithm uses a periodic data stream. The step size is randomly offset by O, and the termination conditions are: 1) the sum of variances is 0; or 2) the number of iterations reaches a certain number, such as 1000, 10000, etc. For complex cases with a large number of data streams, particle swarm optimization, depth-first search, etc. are used.

[0115] S44. Based on the adjusted load variance, select a target allocation scheme according to the inter-slot load balancing algorithm, wherein the target allocation scheme is the allocation scheme with the smallest adjusted load variance;

[0116] In one embodiment, based on the iteratively adjusted load distribution, the optimal allocation scheme, also known as the target allocation scheme, is selected. The target allocation scheme is the scheme with the smallest adjusted load variance, that is, the scheme with the most balanced slot load.

[0117] S45. According to the target allocation scheme, allocate the first-level periodic data stream and the second-level periodic data stream, reserve the third-level time slice for the non-periodic data stream, and set the start and end gate times for each time slice of the first-level periodic data stream and the second-level periodic data stream to generate the gate list.

[0118] In one embodiment, after determining the target allocation scheme, start and end gating times are set for each time slice, clarifying the transmission time period for each type of data stream, and finally generating a gating list. This list records each basic time slot and its time slice allocation, serving as a reference for scheduling and execution. Thus, the transmission order, time slices, and load balancing of periodic data streams are clearly recorded, enabling precise control of train onboard communication network scheduling.

[0119] S5. During the scheduling process, when the non-periodic data stream appears, the gating list uses a trust mechanism to optimize the scheduling of the non-periodic data stream, obtaining a final scheduling configuration to achieve optimized bandwidth allocation and real-time response to bursty data streams. The final scheduling configuration prioritizes the real-time transmission of the first-level periodic data stream, the second-level periodic data stream, and the non-periodic data stream when the non-periodic data stream appears. The second-level periodic data stream is defined as a data stream whose trust value is higher than that of the non-periodic data stream. The non-periodic data stream is defined as a non-periodic data stream that reaches a set trust value and where the bandwidth of the third-level time slice is insufficient in the current time slot. S51. During the scheduling process, the gating list monitors the arrival status of the non-periodic data stream in real time.

[0120] In one embodiment, during the scheduling process, the arrival of non-periodic data streams, their bandwidth usage, latency, and other conditions are monitored in real time, and real-time compensation is provided after the arrival status is detected.

[0121] S52. When the aperiodic data stream arrives, set an initial trust value for each aperiodic data stream according to its priority, latency requirements, and current time slot occupancy.

[0122] In one embodiment, when the gating list encounters an aperiodic data stream during scheduling, it first assigns an initial trust value to each aperiodic data stream. This initial trust value, such as 1.0 or 0.7, is assigned by the train control system at the start of the scheduling task, indicating the system's initial level of trust in the data stream's transmission behavior. It represents the initial priority or reliability of the data stream in the scheduling process. This initial trust value can be set based on the data stream type, historical transmission stability, or predefined parameters of the train control system. For example, frequent bursts of data streams may cause the trust value to decrease, while stable transmissions may increase the trust value.

[0123] S53. The improved bubble sort method is used to calculate the trust value sorting, adjust the transmission order of the aperiodic data stream in the three-level time slice, and compensate the delayed transmission of the aperiodic data stream with trust value.

[0124] In one embodiment, based on the priority, latency requirements, and bandwidth occupancy of the current time slot of the aperiodic data stream, a dynamic trust value is calculated for all aperiodic data streams to be scheduled. All dynamic trust values ​​are then sorted using an improved bubble sort algorithm to generate a priority queue. Messages with higher trust values ​​will receive available bandwidth first, ensuring timely transmission of the aperiodic data stream.

[0125] S54. Based on trust value compensation, when the set non-periodic data stream occurs, the distribution of the secondary periodic data streams in the gating list is temporarily adjusted to generate a dynamic scheduling table so that the set non-periodic data stream can perform preemptive scheduling.

[0126] In one embodiment, if there are aperiodic data streams in the queue that have reached a set trust level, and the available bandwidth in their respective time slots is insufficient to meet their transmission requirements, the gating list is temporarily adjusted to generate a dynamic scheduling table, causing the set aperiodic data streams to undergo preemptive scheduling. Preemptive scheduling temporarily delays the transmission of aperiodic data streams with the set trust level, inserting them into the available bandwidth to achieve real-time response. This mechanism ensures that the train communication network can still transmit critical messages according to priority even in emergency situations.

[0127] S55. After completing the preemptive scheduling, the dynamic scheduling table is restored to the distribution of the secondary periodic data streams in the original gating list to achieve optimized bandwidth allocation and real-time response to bursty data streams.

[0128] In one embodiment, after preemptive scheduling is completed, an updated gating schedule is generated and reused as the base scheduling template for the next round of periodic transmission control. This forms an adaptive regression scheduling mechanism after the response to aperiodic data streams, ensuring both the real-time transmission requirements of aperiodic data streams and the long-term transmission stability and overall bandwidth optimization of periodic tasks.

[0129] Improved bubble sort methods include:

[0130] a1. Define the trust value threshold for determining the trust value level of non-periodic data streams, and the method for calculating the trust value;

[0131] For example, the trust value calculation method divides the priority of non-periodic data streams into three different levels: high, medium, and low, with corresponding trust values ​​defined as 5, 3, and 1; divides the real-time requirements of non-periodic data streams into three different levels: high, medium, and low, with corresponding trust values ​​defined as 7, 2, and 1; and then adds the trust values ​​corresponding to the three priority levels and their corresponding real-time requirement levels to obtain the trust value of each non-periodic data stream.

[0132] a2. For any two data streams in the detected aperiodic data streams, determine their trust value levels. If the difference in trust values ​​between the two aperiodic data streams is greater than the trust value threshold, then the trust value of the former aperiodic data stream is determined to be greater than that of the latter. Conversely, if swapping the positions of the trust values ​​of the two aperiodic data streams is necessary to satisfy the condition that the difference in trust values ​​between them is greater than the trust value threshold, then trust value compensation is applied to the latter's trust value. The formula for calculating the trust value level is as follows:

[0133]

[0134] in, This represents the trust value of the former non-periodic data stream. This represents the trust value of the latter's non-periodic data stream. Indicates the trust threshold;

[0135] The formula for calculating trust value compensation is as follows:

[0136]

[0137] in, As a regulating factor, Indicates the first time slot in the current time slot j Trust value of an aperiodic data stream. Its function is to limit the trust value of the j-th aperiodic data stream in the current time slot, so as to ensure that, after compensation, the trust value of the j-th aperiodic data stream in the current time slot does not exceed the trust value of the j-th aperiodic data stream in the current time slot. i Trust value of an aperiodic data stream.

[0138] The temporary adjustment of the distribution of the secondary periodic data streams within the gating list includes:

[0139] b1. Take the currently detected non-periodic data streams that cannot be transmitted within the pre-allocated three-level time slices as the remaining data streams, and calculate the trust value of the remaining data streams;

[0140] In one embodiment, within the current macro cycle, the train control system monitors the arrival status of aperiodic data streams in real time. For aperiodic data streams that cannot be transmitted within the pre-allocated three-level time slices, they are marked as remaining data streams. For each remaining data stream, its corresponding trust value is calculated based on its transmission success rate, latency requirements, and other factors during historical scheduling.

[0141] b2. Based on the stream trust value of the remaining data, find the set of secondary periodic data streams to replace the remaining data streams within the current time slot, calculated as follows:

[0142]

[0143] in, This represents the set of secondary periodic data streams that can be used to replace the remaining data streams. This indicates the number of secondary periodic data streams in the current time slot. Indicates the first time slot in the current time slot A secondary periodic data stream, Indicates the first time slot in the current time slot j Trust value of a secondary periodic data stream. This indicates the number of remaining data streams.

[0144] In one embodiment, a set of secondary periodic data streams is searched within the current time slot, and candidate data streams that meet the replacement criteria are selected from this set. The candidate data streams have a trust value lower than the trust value of the remaining data streams to be scheduled under the current load. The formation of this set ensures that the replacement operation does not compromise the overall scheduling reliability, while providing scheduling space for non-periodic data streams.

[0145] b3. Find the second-level periodic data stream in the set of second-level periodic data streams that minimizes the trust value of the second-level periodic data streams. The calculation formula is as follows:

[0146]

[0147] in, This represents the second-level periodic data stream with the lowest trust value.

[0148] In one embodiment, the trust value of each secondary periodic data stream is calculated in the candidate set, and the secondary periodic data stream with the smallest trust value is selected. This data stream has low transmission stability or importance in the current time slot, and its transmission time slice is dynamically adjusted to improve the overall real-time performance and scheduling fairness.

[0149] b4. Place the data stream with the highest trust value in the remaining data stream at the transmission position of the secondary periodic data stream with the lowest trust value in the secondary periodic data stream;

[0150] In one embodiment, after the replacement is completed, a trust value compensation mechanism is performed on the replaced secondary periodic data stream. Based on the number of times it has been replaced, the latency, and the queue waiting time, its trust value is dynamically compensated, allowing the data stream's priority to gradually recover in subsequent time slots, preventing it from being kept at a low priority for an extended period and thus preventing data stream transmission failure.

[0151] b5. For the secondary periodic data stream with the lowest trust value, perform trust value compensation, using the following formula:

[0152]

[0153] in, This represents the second-level periodic data stream with the lowest trust value.

[0154] b6. Repeat steps b2-b5 above until all remaining data streams are allocated, or all secondary periodic data streams in the set of secondary periodic data streams are traversed.

[0155] In one embodiment, steps b2-b5 are repeated to process all remaining data streams sequentially. The scheduling process terminates when all remaining data streams have been reallocated or when the entire set of secondary periodic data streams has been traversed. At this point, the trust values ​​of the secondary periodic data streams are higher than those of the remaining data streams. Throughout the entire scheduling process, priority is always given to ensuring the real-time transmission needs of the primary periodic data streams and the secondary periodic data streams, ensuring that the train communication network maintains high reliability and latency determinism even under sudden loads.

[0156] This invention divides the data stream of the train's onboard communication network according to its periodicity and task importance, and further divides each basic time slot into different levels of time slices for transmission within a macro-period. This achieves reasonable priority management and bandwidth allocation for periodic data streams, ensuring priority transmission of important and real-time data and avoiding delays in critical data caused by non-periodic data streams or high loads. It optimizes the allocation of periodic data streams based on a time slot load balancing algorithm, generating an optimized gating schedule by traversing random offset combinations and using the minimum variance criterion, and writing it into the TSN scheduling controller. This provides a static scheduling strategy for periodic data streams, reducing the uncertainty of transmission scheduling for important data streams related to train operation. Simultaneously, it provides an optimal bandwidth allocation scheme for the transmission of non-periodic data streams, reducing network congestion risks and improving the stability and reliability of communication scheduling. By incorporating a trust mechanism, when non-periodic data streams arrive, the scheduling strategy for low-trust-value data streams in the gating list is temporarily adjusted, generating a dynamic scheduling table for preemptive scheduling of non-periodic data streams. At the same time, trust value compensation is applied to delayed data streams to prevent the problem of low-trust-value data streams potentially being unable to transmit for extended periods.

[0157] Figure 3 This is a block diagram of a train-mounted communication network dispatching system according to the present invention.

[0158] Based on the same concept, the present invention also provides a train onboard communication network dispatching system, comprising:

[0159] Data acquisition module 31 is used to acquire communication data streams in the train communication network;

[0160] The data segmentation module 32 is used to divide the communication data stream into periodic data streams and aperiodic data streams according to the nature of the communication data stream. The periodic data stream is divided into first-level periodic data streams and second-level periodic data streams. The nature of the communication data stream includes the source node, target node, and task type of the sampling task. The first-level periodic data stream is a periodic data stream with high importance, the second-level periodic data stream is a periodic data stream with moderate importance, and the aperiodic data stream is a bursty data stream caused by equipment failure or configuration update in the train communication network.

[0161] The time slot allocation module 33 is used to divide each basic time slot into a first-level time slot, a second-level time slot, and a third-level time slot within the macro cycle of communication scheduling, so as to transmit the first-level periodic data stream, the second-level periodic data stream, and the aperiodic data stream, respectively.

[0162] The load balancing module 34 is used to allocate the first-level periodic data stream and the second-level periodic data stream according to the inter-slot load balancing algorithm, reserve the third-level time slice for the non-periodic data stream, and set the start and end gate times of each time slice for the first-level periodic data stream and the second-level periodic data stream to obtain the communication scheduling gate list. The gate list is a list that controls the transmission order of various data streams by being offline distributed to each switch device through a dedicated network controller or TSN configuration server. TSN is a time-sensitive network.

[0163] The scheduling configuration module 35 is used to optimize the scheduling of the non-periodic data stream using a trust mechanism when the non-periodic data stream appears during the scheduling process of the gating list, so as to obtain the final scheduling configuration, so as to realize bandwidth optimization allocation and real-time response to bursty data streams. The final scheduling configuration is a scheduling mechanism that prioritizes the real-time transmission of the first-level periodic data stream, the second-level periodic data stream, and the non-periodic data stream when the non-periodic data stream appears. The second-level periodic data stream is defined as the data stream whose trust value is higher than that of the non-periodic data stream. The non-periodic data stream is defined as the non-periodic data stream that has reached the set level of trust value and whose third-level time slice bandwidth is insufficient in the current time slot.

[0164] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of protection of this application is limited to these examples; within the framework of this application, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of different aspects of one or more embodiments of this application as described above, which are not provided in detail for the sake of brevity.

[0165] One or more embodiments in this application are intended to cover all such substitutions, modifications, and variations that fall within the broad scope of this application. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of one or more embodiments in this application should be included within the protection scope of this application.

Claims

1. A train onboard communication network scheduling method, characterized in that, include: Acquire communication data streams from the train communication network; Based on the nature of the communication data stream, the communication data stream is divided into periodic data streams and aperiodic data streams. The periodic data stream is further divided into primary periodic data streams and secondary periodic data streams. The nature of the communication data stream includes the source node, target node, and task type of the sampling task. Primary periodic data streams are periodic data streams with high importance, secondary periodic data streams are periodic data streams with moderate importance, and aperiodic data streams are bursty data streams caused by equipment failures or configuration updates in the train communication network. Within the macro cycle of communication scheduling, each basic time slot is divided into a first-level time slot, a second-level time slot, and a third-level time slot to transmit the first-level periodic data stream, the second-level periodic data stream, and the aperiodic data stream, respectively. According to the time-slot load balancing algorithm, the first-level periodic data stream and the second-level periodic data stream are allocated, and the third-level time slice is reserved for the non-periodic data stream. The start and end gate times of each time slice are set for the first-level periodic data stream and the second-level periodic data stream to obtain the communication scheduling gate list. The gate list is a list that controls the transmission order of various data streams by being offline distributed to each switch device through a dedicated network controller or TSN configuration server. TSN is a time-sensitive network. During the scheduling process, when the aperiodic data stream occurs, the gating list uses a trust mechanism to optimize the scheduling of the aperiodic data stream, obtaining a final scheduling configuration to achieve optimized bandwidth allocation and real-time response to bursty data streams, including: During the scheduling process, the gating list monitors the arrival of non-periodic data streams in real time. When the aperiodic data stream arrives, an initial trust value is set for each aperiodic data stream based on its priority, latency requirements, and current time slot occupancy. An improved bubble sort algorithm is used to calculate the trust value and sort the data. The transmission order of the aperiodic data streams within the three-level time slice is adjusted, and the trust value is compensated for the aperiodic data streams that are delayed in transmission. Based on trust value compensation, when a set non-periodic data stream occurs, the distribution of the secondary periodic data streams in the gating list is temporarily adjusted to generate a dynamic scheduling table so that the set non-periodic data stream can perform preemptive scheduling. After completing the preemptive scheduling, the dynamic scheduling table is restored to the distribution of the secondary periodic data streams in the original gating list to achieve optimized bandwidth allocation and real-time response to bursty data streams. The final scheduling configuration is as follows: when the non-periodic data stream occurs, a scheduling mechanism that prioritizes real-time transmission is adopted for the first-level periodic data stream, the set second-level periodic data stream, and the set non-periodic data stream. The set second-level periodic data stream is the data stream whose trust value is higher than that of the non-periodic data stream. The set non-periodic data stream is the data stream that reaches the set level of trust value and is the non-periodic data stream when the bandwidth of the third-level time slice is insufficient in the current time slot.

2. The train onboard communication network scheduling method as described in claim 1, characterized in that, Based on the nature of the communication data stream, the communication data stream is divided into periodic data streams and aperiodic data streams. The periodic data streams are further divided into first-level periodic data streams and second-level periodic data streams, including: The source node, target node, and task type of each communication data stream sampling task in the train control system are collected; Fixed-period sampling tasks are classified as periodic data streams, and the periodic data streams are divided into first-level periodic data streams and second-level periodic data streams according to the sampling task level; non-fixed-period sampling tasks are classified as non-periodic data streams, wherein the task level includes sampling task priority, real-time requirements, and security level.

3. The train onboard communication network scheduling method as described in claim 1, characterized in that, Within the macro-cycle of communication scheduling, each basic time slot is divided into a first-level time slice, a second-level time slice, and a third-level time slice to transmit the first-level periodic data stream, the second-level periodic data stream, and the aperiodic data stream, respectively, including: Within the macro cycle of communication scheduling, the macro cycle is divided into a predetermined number of basic time slots of fixed length. The formula for calculating the basic time slot is as follows: in, Indicates the basic time slot, Indicates the first The transmission time of a periodic data stream, Indicates the first Is the first basic time slot transmitted? Data stream, Indicates transmission, Indicates no transmission. Indicates the total number of periodic data streams; This represents the transmission time of an aperiodic data stream; the formula for calculating the number of periodic basic time slots is as follows: in, N Indicates the number of basic time slots in a period. The function representing the least common multiple; Each basic time slot is divided into a first-level time slice, a second-level time slice, and a third-level time slice, which are respectively allocated to the first-level periodic data stream, the second-level periodic data stream, and the aperiodic data stream for the transmission of various types of data streams.

4. The train onboard communication network scheduling method as described in claim 3, characterized in that, Based on the inter-slot load balancing algorithm, the first-level periodic data stream and the second-level periodic data stream are allocated, and a third-level time slice is reserved for the non-periodic data stream. Start and end gate times for each time slice are set for the first-level periodic data stream and the second-level periodic data stream, resulting in a communication scheduling gate list, including: Calculate the total bandwidth of the periodic data stream within each macro cycle. The formula for calculating the total bandwidth of the periodic data stream is as follows: Where Bp represents the total bandwidth within the macro period, Indicates the number of transmissions of the periodic data stream within the macro period. Represents data stream Number of bytes, This represents the macro period, and the formula for calculating the macro period is as follows: ； Calculate the initial load for each time slot and obtain the load variance, where the initial load for each time slot is... The calculation formula is: The formula for calculating the load variance is as follows: in, Indicates the average load within the macro period. ; Based on the load variance, within the time slot range allocated to various data streams, the starting transmission time slot of the data stream is iteratively adjusted using a random offset method. After the iteration terminates, the adjusted load variance is calculated. Based on the adjusted load variance, a target allocation scheme is selected according to the inter-slot load balancing algorithm, wherein the target allocation scheme is the allocation scheme with the smallest adjusted load variance. According to the target allocation scheme, the first-level periodic data stream and the second-level periodic data stream are allocated, the third-level time slice is reserved for the non-periodic data stream, and the start and end gate times of each time slice are set for the first-level periodic data stream and the second-level periodic data stream to generate the gate list.

5. The train onboard communication network scheduling method as described in claim 4, characterized in that, The iteration termination condition is when the load variance is 0 or the preset number of iterations is reached.

6. The train onboard communication network scheduling method as described in claim 1, characterized in that, The improved bubble sort method includes: Define the trust value threshold for determining the trust value level of non-periodic data streams, and the method for calculating the trust value; For any two detected aperiodic data streams, a trust value level is determined. If the difference in trust values ​​between the two aperiodic data streams is greater than the trust value threshold, then the trust value of the former aperiodic data stream is determined to be greater than that of the latter. Conversely, if the positions of the trust values ​​of the two aperiodic data streams are swapped to satisfy the condition that the difference in trust values ​​between the two aperiodic data streams is greater than the trust value threshold, then trust value compensation is applied to the latter's trust value. The formula for determining the trust value level is as follows: in, This represents the trust value of the former non-periodic data stream. This represents the trust value of the latter's non-periodic data stream. Indicates the trust threshold; The formula for calculating trust value compensation is as follows: in, Indicates the trust threshold. It is a regulating factor.

7. The train onboard communication network scheduling method as described in claim 6, characterized in that, The temporary adjustment of the distribution of the secondary periodic data streams within the gating list includes: The non-periodic data streams that are currently detected and cannot be transmitted within the pre-allocated three-level time slice are taken as the remaining data streams, and the trust value of the remaining data streams is calculated. Based on the trust value of the remaining data stream, a set of secondary periodic data streams for replacing the remaining data streams is found within the current time slot, and the calculation formula is as follows: in, This represents the set of secondary periodic data streams that can be used to replace the remaining data streams. This indicates the number of secondary periodic data streams in the current time slot. Indicates the first time slot in the current time slot A secondary periodic data stream, Indicates the first time slot in the current time slot j Trust value of a secondary periodic data stream. Indicates the number of remaining data streams; The set of secondary periodic data streams is used to find the secondary periodic data stream that minimizes the trust value. The calculation formula is as follows: in, This represents the second-level periodic data stream with the lowest trust value. Place the data stream with the highest trust value in the remaining data stream at the transmission position of the second-level periodic data stream with the lowest trust value in the second-level periodic data stream; Trust value compensation is performed on the secondary periodic data stream with the lowest trust value. This continues until all remaining data streams have been allocated, or until all secondary periodic data streams in the set of secondary periodic data streams have been traversed.

8. The train onboard communication network scheduling method as described in claim 7, characterized in that, The formula for calculating the trust value compensation for the secondary periodic data stream with the lowest trust value is as follows: in, This represents the second-level periodic data stream with the lowest trust value.

9. A train-mounted communication network dispatching system, characterized in that, include: The data acquisition module (31) is used to acquire the communication data stream in the train communication network; The data segmentation module (32) is used to divide the communication data stream into periodic data streams and aperiodic data streams according to the nature of the communication data stream. The periodic data stream is divided into first-level periodic data streams and second-level periodic data streams. The nature of the communication data stream includes the source node, target node and task type of the sampling task. The first-level periodic data stream is a periodic data stream with high importance. The second-level periodic data stream is a periodic data stream with moderate importance. The aperiodic data stream is a bursty data stream caused by equipment failure or configuration update in the train communication network. The time slot allocation module (33) is used to divide each basic time slot into a first-level time slot, a second-level time slot, and a third-level time slot within the macro cycle of communication scheduling, and to transmit the first-level periodic data stream, the second-level periodic data stream, and the non-periodic data stream, respectively. The load balancing module (34) is used to allocate the first-level periodic data stream and the second-level periodic data stream according to the inter-slot load balancing algorithm, reserve the third-level time slice for the non-periodic data stream, and set the start and end gate times of each time slice for the first-level periodic data stream and the second-level periodic data stream to obtain the communication scheduling gate list. The gate list is a list that controls the transmission order of various data streams by being offline distributed to each switch device through a dedicated network controller or TSN configuration server. TSN is a time-sensitive network. The scheduling configuration module (35) is used by the gated list to optimize the scheduling of the aperiodic data stream using a trust mechanism when the aperiodic data stream occurs during the scheduling process, so as to obtain the final scheduling configuration, thereby realizing bandwidth optimization allocation and real-time response to bursty data streams, including: During the scheduling process, the gating list monitors the arrival of non-periodic data streams in real time. When the aperiodic data stream arrives, an initial trust value is set for each aperiodic data stream based on its priority, latency requirements, and current time slot occupancy. An improved bubble sort algorithm is used to calculate the trust value and sort the data. The transmission order of the aperiodic data streams within the three-level time slice is adjusted, and the trust value is compensated for the aperiodic data streams that are delayed in transmission. Based on trust value compensation, when a set non-periodic data stream occurs, the distribution of the secondary periodic data streams in the gating list is temporarily adjusted to generate a dynamic scheduling table so that the set non-periodic data stream can perform preemptive scheduling. After completing the preemptive scheduling, the dynamic scheduling table is restored to the distribution of the secondary periodic data streams in the original gating list to achieve optimized bandwidth allocation and real-time response to bursty data streams. The final scheduling configuration is as follows: when the non-periodic data stream occurs, a scheduling mechanism that prioritizes real-time transmission is adopted for the first-level periodic data stream, the set second-level periodic data stream, and the set non-periodic data stream. The set second-level periodic data stream is the data stream whose trust value is higher than that of the non-periodic data stream. The set non-periodic data stream is the data stream that reaches the set level of trust value and is the non-periodic data stream when the bandwidth of the third-level time slice is insufficient in the current time slot.

Images