A multipath information packet coding and scheduling method with minimal delay

Abstract

The present invention discloses a multipath information packet encoding and scheduling method for minimizing latency. The method includes: updating lost information packets at the transmitting end through feedback information; planning the transmission order of each packet within a packet scheduling window based on the different transmission rates, feedback delays, and path erasure probabilities of multiple paths; estimating the total number of lost information packets based on the feedback information at the beginning of the scheduling window, and arranging the transmission of repair coding packets at the beginning of the scheduling window and forward error correction coding packets at the end of the scheduling window based on the feedback information delay; and overcoming data packet loss caused by channel errors. The present invention estimates packet loss within the feedback delay, transmits information packets and coding packets in the order they are received, and introduces repair coding packets and forward error correction coding packets to achieve, with virtually no information redundancy, reduced end-to-end latency for information packet transmission in multipath networks with feedback lag.

Description

Technical Field

[0001] The present invention belongs to the field of wireless communication standardization process, and in particular relates to a multi-path information packet encoding and scheduling method for minimizing time delay. Background Art

[0002] With the continuous development of modern communication technology, communication devices are often equipped with various network access methods. For example, mobile phones can communicate simultaneously via Wi-Fi, LTE, and Bluetooth. Therefore, multipath communication has become a promising technology for improving the reliability and capacity of wireless networks, while also effectively reducing communication latency. Multipath communication can aggregate the spatial / frequency diversity of various paths and provide great flexibility between diversity and multiplexing. However, the correlation, dynamics, and asymmetry of each path in the multipath make packet scheduling in multipath communication quite challenging. In addition, the design of packet scheduling also needs to adapt to different Quality of Service (QoS) objectives. Therefore, in order to improve the latency and throughput performance of packet transmission, the design of packet scheduling algorithms under the multipath system model has become an important part of today's communication protocols.

[0003] The main challenges of multipath packet scheduling lie in the following four aspects: First, multipath packet scheduling should account for the asymmetry of network paths—that is, different paths may have different transmission rates, path elimination probabilities, and round-trip times (RTTs)—to effectively utilize multipath scenarios. Second, packet scheduling algorithms must balance multiple QoS objectives, such as maximizing average throughput, reducing overall data transmission time, minimizing end-to-end latency or minimizing the size of the receiver's out-of-order buffer, and maintaining jitter smoothing. Packet erasure events across paths may be correlated. Finally, multipath packet scheduling algorithms must adapt to the dynamic nature of the network environment. In real-world environments, network conditions fluctuate over time and vary significantly across different environments. Summary of the Invention

[0004] The present invention aims to provide a multipath packet coding and scheduling method that minimizes latency. In the case where each path has asymmetric packet transmission rates, erasure probabilities, and feedback delays, feedback information is used to estimate the number of lost packets. This method, combined with existing random linear network coding techniques, utilizes both a priori and a posteriori error correction techniques to minimize the end-to-end delay of packet transmission while maximizing the capacity of the multipath channel.

[0005] In order to solve the above technical problems, the specific technical solutions of the present invention are as follows:

[0006] A method for multipath information packet encoding, scheduling and real-time transmission under delayed feedback conditions, comprising the following steps:

[0007] Step 1: Before transmission begins, the length of the scheduling window is determined based on the different transmission rates of each path in the multipath channel. That is, the number of paths is set to Z, and the time it takes for each path to transmit a packet is T1, T2, ...T Z , then the length of the scheduling window is the least common multiple of the transmission time of each path, that is, w = lcm(T1, T2, ... T Z );

[0008] Step 2: When the transmission starts, the sender has N packets waiting to be transmitted, and the channel erasure probability of each path is e1, e2, ...e Z , the transmission time is T1, T2, ...T Z , the total number of packets transmitted in a scheduling window is n w ,Only when a single path completes the transmission of an information packet or coding packet can the next packet be transmitted on the path, and the transmission of each path is carried out in parallel;

[0009] Step 3: When the receiving end successfully receives the information packet, it immediately uploads the information packet to the application layer and feeds back the feedback information of the information packet to the sending end. The feedback information of each path is fed back after the feedback delay D1, D2, ..., D Z Arrives at the sending end;

[0010] Step 4: When the sender starts transmitting, if it is the beginning of the scheduling window, the sender first confirms the packets that the receiver has not successfully received through the received feedback information, and then uses the feedback delay D and the erasure probability of each path e1, e2, ...e Z , estimate the total number of unsuccessfully received packets, encode these estimated unsuccessfully received packets into repair coded packets, and decide the number of packets, repair coded packets, and forward error correction coded packets to be sent, the total number is n w , arrange these n according to the receiving time sequence of each sending position in the scheduling window w The sending position of each packet is determined to eliminate the problem of out-of-order reception caused by different transmission rates of multiple paths. The repair code packet is received first, followed by the information packet, and finally the forward error correction code packet. After the scheduling is completed, the sender sends the information packet on the currently available path according to the current scheduling.

[0011] Step 5: When the sender starts transmission, if it is not the start of the scheduling window, the sender first confirms the packets that the receiver has not successfully received through the received feedback information, and inserts these lost packets into the head of the current sending queue. The lost packets and the remaining unsent packets and coded packets in the scheduling window are rescheduled for transmission according to the order of reception time, so as to give priority to the retransmission of the lost packets. After completing the scheduling arrangement, the sender sends the packets on the currently available path according to the current scheduling arrangement.

[0012] Step 6: When the receiving end successfully receives the data packet, repeat step 3;

[0013] Step 7: After completing steps 1-6, repeat steps 4-6 until N information packets are transmitted.

[0014] Furthermore, the total number of packets that can be transmitted within a scheduling window in step 2 is n w is the length of the scheduling window w and the transmission time of each path T1, T2, ... T Z OK, n w for:

[0015]

[0016] Furthermore, in step 4, the known number of unacknowledged packets u and the erasure probability of each path e1, e2, ... e are obtained through feedback information in each scheduling window. Z and feedback delays D1, D2, ..., D Z The estimated number of unacknowledged packets is recorded as For the transmitter, each path has a feedback delay of D1, D2, ..., D Z The number of unacknowledged packets sent within a period is known and is denoted as k1, k2, ..., k Z , from which we can derive the estimated number of unacknowledged packets, namely for

[0017]

[0018] Furthermore, step 4 uses the erasure probability e1, e2, ... e of each path Z , the length of the scheduling window w and the transmission time of each path T1, T2, ...T Z Determine the number of information packets, repair coding packets, and forward error correction coding packets sent, which are:

[0019] Repair code package:

[0020] Information Packet:

[0021] Forward error correction coding packet:

[0022] The multipath information packet encoding and scheduling method for minimizing delay of the present invention has the following advantages:

[0023] 1. This method fully considers the different transmission rates, channel erasure probabilities, and feedback delays of each path in the multipath scenario, and uses the asymmetric characteristics of each path to make scheduling decisions, thereby eliminating the transmission performance degradation caused by the asymmetric characteristics of each path to a certain extent.

[0024] 2. This method can complete the a priori and a posteriori error correction mechanisms by sending repair coding packets and forward error correction coding packets, and reasonably estimate the number of lost information packets in the presence of feedback delay, thereby efficiently utilizing the channel capacity to complete the information packet transmission in each scheduling window.

[0025] 3. This method reorders the packet sending order within the scheduling window so that the packets in the sending queue are sent in the order of the receiving time of each sending position within the scheduling window, so that the sending queue can be received in order at the receiving end, thereby eliminating the problem of out-of-order reception caused by different transmission rates of multiple paths. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 This is a schematic diagram illustrating an example of the transmission and encoding process at the transmitting end according to an embodiment of the present invention;

[0027] Figure 2 A schematic diagram of generating a coding packet according to an embodiment of the present invention;

[0028] Figure 3 is a schematic flow chart of an information packet encoding and transmission solution according to an embodiment of the present invention;

[0029] Figure 4 FIG1 is a schematic diagram of simulation performance of multipath erasure probability according to an embodiment of the present invention;

[0030] Figure 5 It is a schematic diagram of simulation performance of multipath feedback delay according to an embodiment of the present invention. DETAILED DESCRIPTION

[0031] In order to better understand the purpose, structure and function of the present invention, the following further describes in detail a multi-path information packet encoding and scheduling method for minimizing delay of the present invention in conjunction with the accompanying drawings.

[0032] The present invention is applicable to multipath communication networks with feedback delay. The transmitter formulates a corresponding scheduling strategy based on the characteristics of the asymmetric paths in the multipath channel (transmission rate, channel erasure probability, and feedback delay). To ensure the orderly arrival of information packets, the transmitter divides the information packet scheduling into multiple scheduling windows for scheduling. Within the scheduling windows, each information packet is sent in the order in which it is received. At the beginning of each scheduling window, the transmitter combines known feedback information and channel path characteristics to estimate the actual number of lost information packets. It then uses random linear network coding to encode information packets with unknown reception status into repair coded packets and sends them according to the estimated number of lost packets. If it is unknown which packets have been lost, forward error correction network coded packets are sent to compensate for the loss of subsequent information packets.

[0033] The present invention specifically comprises the following steps:

[0034] A method for multipath information packet encoding, scheduling and real-time transmission under delayed feedback conditions, comprising the following steps:

[0035] Step 1: Determine the scheduling window length. Specifically, first determine the scheduling window length based on the different transmission rates of each path in the multipath channel. If the number of paths is Z, and the time required to transmit a data packet determined by the transmission rate of each path is T1, T2, ... T Z , then the length of the scheduling window is the least common multiple of the transmission time of each path, that is, w = lcm(T1, T2, ... T Z This step ensures that all paths can synchronously complete the transmission of their planned packets within a scheduling window period, thereby facilitating the scheduling of multi-path information packets.

[0036] Step 2: Schedule the transmission of data packets within the window. Specifically, at the beginning of the transmission, the sender has N packets waiting to be transmitted, considering that the channel erasure probability of each path is e1, e2, ...e Z And the transmission time is T1, T2, ...T Z , the total number of packets that can be transmitted in a scheduling window is n w ,Only when a single path completes the transmission of an information packet or coding packet can the next packet be transmitted on the path, and the transmission of each path is carried out in parallel;

[0037] Step 3: Packet processing and feedback at the receiving end. Specifically, when the receiving end successfully receives the information packet, it will immediately upload the information packet to the application layer and feedback the reception status of the information packet to the sending end. The feedback information will be delayed by a certain time D1, D2, ..., D Z Arriving at the sender. This step is the key to updating the sender's status in real time so that the sender can adjust the subsequent sending strategy based on the latest situation.

[0038] Step 4: Packet encoding, scheduling, and sending at the beginning of the scheduling window. Specifically, when the sender starts transmitting, if it is the beginning of the scheduling window, the sender first confirms the packet that the receiver has not successfully received through the received feedback information, and then sends the packet to the receiver through the feedback delay D1, D2, ..., D Z And the erasure probability of each path e1,e2,…e Z , the transmitter knows that the feedback delay D1, D2, ..., D Z The number of information packets and coded packets sent within the time interval, according to the path erasure probability e1, e2, ... e Z Obtain the expected number of unknown state packets lost and the expected number of unknown state coded packets received, thereby estimating the total number of unsuccessfully received packets. Encode these estimated unsuccessfully received packets into repair coded packets, and determine the number of packets, repair coded packets, and forward error correction coded packets to be sent, with the total number being n. w , arrange these n according to the receiving time sequence of each sending position in the scheduling window w The sending position of each packet is determined to eliminate the problem of out-of-order reception caused by different transmission rates on multiple paths. The repair code packet is received first, followed by the information packet, and finally the forward error correction code packet. After the scheduling is completed, the sender sends the information packet on the currently available path according to the current schedule.

[0039] Step 5: Scheduling and sending packets within the scheduling window. Specifically, when the sender starts transmission, if it is not the start of the scheduling window, the sender first confirms the packets that the receiver has not successfully received by receiving feedback information from the receiver, and inserts these lost packets into the head of the current sending queue. The lost packets and the remaining unsent packets in the scheduling window are re-arranged in the order of reception time to prioritize the retransmission of lost packets. After completing the scheduling arrangement, the sender sends the packets on the currently available path according to the current scheduling arrangement.

[0040] Step 6: When the receiving end successfully receives the data packet, repeat step 3;

[0041] Step 7: After completing steps 1-6, repeat steps 4-6 until the N information packets that the sender initially needs to transmit are transmitted.

[0042] The scheduling window length w defined in step 1 is determined by the asymmetric packet transmission time T1, T2, ... T Z The lowest common multiple of is determined, w is:

[0043] w=lcm(T1,T2,…T Z ).

[0044] The total number of packets that can be transmitted within a scheduling window in step 2 is n w is the length of the scheduling window w and the transmission time of each path T1, T2, ... T Z OK, n w for:

[0045]

[0046] In step 4, the known number of unacknowledged packets u and the erasure probability of each path e1, e2, ... e are obtained through feedback information in each scheduling window. Z and feedback delays D1, D2, ..., D Z The estimated number of unacknowledged packets is recorded as For the transmitter, each path has a feedback delay of D1, D2, ..., D Z The number of unacknowledged packets sent within a period is known and is denoted as k1, k2, ..., k Z , from which we can derive the estimated number of unacknowledged packets, namely for

[0047]

[0048] Step 4: The erasure probability of each path e1, e2, ... e Z , the length of the scheduling window w and the transmission time of each path T1, T2, ...T Z Determine the number of information packets, repair coding packets, and forward error correction coding packets sent, which are:

[0049] Repair code package:

[0050] Information Packet:

[0051] Forward error correction coding packet:

[0052] Reference will now be made in detail to exemplary embodiments, examples of which are illustrated in the accompanying drawings.

[0053] Figure 1 1 is a schematic diagram illustrating an example of a transmission and encoding process at the transmitting end and a decoding process at the receiving end under the delayed feedback condition mentioned in the first embodiment of the present invention.

[0054] There are a total of Z = 3 paths in the multipath channel. The fastest path packet transmission time is recorded as one time slot, that is, the first path. The transmission time of the other two paths is 2 time slots and 4 time slots. For the convenience of observation and simulation, the feedback information delay of each path is D1, D2, ..., D ZFor each path, the packet erasure probability is [0.1, 0.2, 0.3].

[0055] Before transmission begins, set system parameters according to application requirements, including: scheduling window length lcm (T1, T2, ... T Z )=4, the total number of packets that can be transmitted in a scheduling window is n w =7, feedback delay D=2, and the number of forward error correction coding packets in a scheduling window is f=1.

[0056] At the beginning of the first time slot, no information packets need to be retransmitted. The sender schedules the 7 information packets and coding packets that need to be sent within the scheduling window so that these 7 information packets and coding packets can reach the receiver in order under the condition of asymmetric path transmission rate, and adds f = 1 forward error correction coding packets to the end of the sending queue of the scheduling window. After the scheduling is completed, the sender immediately sends x1, x3 and c 1～6 On three different paths, packet x1 reaches the receiving end, which then passes it to the application layer and sends a confirmation signal to the sending end.

[0057] At the beginning of the second time slot, the transmitter receives no feedback. It sends x2 along path 1 as originally scheduled. At the end of the second time slot, the receiver successfully receives x2 and uploads it, but x3 is erased by the channel, sending feedback to the transmitter.

[0058] At the beginning of the third time slot, the sender receives no feedback. The sender transmits packet x4 on path 1 and x6 on path 2 as originally scheduled. Packet x6 reaches the receiver, which passes it to the application layer and sends confirmation to the sender. However, packet x6 is dropped by the channel.

[0059] At the beginning of the 4th time slot, the transmitter does not receive any feedback information. The transmitter sends x5 on path 1 according to the original schedule. At the end of the 4th time slot, the receiver successfully receives x4 and c 1～6 , the receiving end feeds back information confirmation to the sending end.

[0060] At the beginning of the 5th time slot, the sender arrives at the beginning of the second scheduling window. At this time, the sender learns from the feedback that packet x5 has been lost, and estimates that another packet will be lost in the unknown reception area. Therefore, the sender adds a repair code packet to the head of the sending queue and adds f = 1 forward error correction code packets to the end of the sending queue in this scheduling window. The sender schedules the 7 information packets and code packets that need to be sent in the scheduling window so that these 7 information packets and code packets can reach the receiving end in order under the condition of asymmetric path transmission rate. After the scheduling is completed, the sender immediately sends c 1～6 , x8 and c 1～12 On three paths respectively. Packet c 1～6 Arrives at the receiving end and sends feedback to the sending end for confirmation.

[0061] At the beginning of the sixth time slot, the transmitter sends packet x7 on path 1 according to the original schedule. Packet x7 is deleted by the channel.

[0062] At the beginning of the 7th time slot, the transmitter sends packets x9 and x on path 1 and path 2 respectively. 11 . Information packets x9 and x 11 When it reaches the receiving end, the receiving end uploads it to the application layer and sends feedback to the sending end for confirmation.

[0063] At the beginning of the 8th time slot, the transmitter sends x on path 1. 10 At the end of the 8th time slot, the receiver successfully receives x 10 、x 11 and c 1～11 , the receiving end feeds back information confirmation to the sending end.

[0064] In the above time slots 1-8, the average end-to-end transmission delay is 6.3 time slots, and the transmission efficiency is 1, that is, the receiving end does not receive any redundant information.

[0065] Figure 2 This is a schematic diagram of generating a coded packet according to an embodiment of the present invention. If the transmitter determines based on feedback that a coded packet needs to be sent, the packets within the coding window are encoded using random linear network coding. Specifically, if the coding window contains w data packets, the coded packet is generated by multiplying the packets within the coding window by random coding coefficients and then summing them. The coding coefficients are generally selected from a finite field.

[0066] Figure 4This is a schematic diagram of the simulation performance of the multipath erasure probability of an embodiment of the present invention. The three schemes in the figure are the scheme proposed by the present invention, the retransmission scheme, and the reordered retransmission scheme. The reordered retransmission scheme is based on the retransmission scheme. The transmitting end obtains the reception and decoding status of the information packet through feedback, sends the information packet in the order of reception within the scheduling window, and updates the scheduling in real time based on the feedback information. Specifically, for the convenience of observation and simulation, the feedback delay D1, D2, ..., D of the fixed Z = 3 paths is fixed. Z For D = 15 time slots and the packet erasure probability of each path is [0.1, 0.2, 0.3], set the erasure probability increase coefficient e0 of each path to 0 to 0.1 with a step size of 0.02, and observe the average end-to-end delay of sending N = 100 packets. Figure 4 The results show that the average end-to-end delay of the three schemes increases with the increase of the erasure probability amplification factor e0. The reordering and retransmission scheme, which reorders packets within the scheduling window, has significantly better end-to-end delay than the retransmission scheme, thus verifying the improvement of reordering on end-to-end delay performance. The proposed scheme also has better end-to-end delay than the reordering and retransmission scheme when using a priori and a posteriori error correction mechanism.

[0067] Figure 5 This is a schematic diagram of the simulation performance of multipath feedback delay in an embodiment of the present invention. The three schemes in the figure are the scheme proposed by the present invention, the retransmission scheme, and the reordering and retransmission scheme. Specifically, when the erasure probability of each path of Z = 3 paths is fixed to [0.1, 0.2, 0.3], the erasure probability increase coefficient e0 of each path is fixed to 0. For the convenience of observation and simulation, the feedback delays D1, D2, ..., D Z The delays are all set to D = 0 time slots to D = 30 time slots with a step size of 5 time slots, and the average end-to-end delay of sending N = 100 information packets is observed. Figure 5 Results show that the average end-to-end delay of all three schemes increases as the feedback delay D increases. The reorder-retransmit scheme, which reorders packets within the scheduling window, achieves significantly better end-to-end delay than the retransmit scheme, validating the improvement of reordering on end-to-end delay performance. When the feedback delay D is small (D ≤ 10), the proposed scheme does not utilize feedback information promptly and only performs scheduling at the beginning of each scheduling window. As a result, its end-to-end delay performance is lower than that of the reorder-retransmit scheme and comparable to that of the retransmit scheme. When the feedback delay D is high (D > 10), the proposed scheme uses an a priori / posteriori error correction mechanism to predict errors within the feedback delay and obtain a more accurate estimate of packet loss, resulting in superior end-to-end delay performance compared to the reorder-retransmit scheme.

[0068] It will be understood that the present invention is described by way of some embodiments, and it will be appreciated by those skilled in the art that various changes or equivalent substitutions may be made to these features and embodiments without departing from the spirit and scope of the present invention. In addition, under the teachings of the present invention, these features and embodiments may be modified to adapt to specific circumstances and materials without departing from the spirit and scope of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are intended to be protected by the present invention.

Claims

1. A method for multipath information packet encoding, scheduling and real-time transmission under delayed feedback conditions, characterized in that: The following steps are involved: Step 1: Before the start of transmission, the length of the scheduling window is determined according to the different transmission rates of each path in the multipath channel. That is, the number of paths is set to Z, and the time for each path to transmit a packet is T1, T2, ... T Z , then the length of the scheduling window is the least common multiple of the transmission time of each path, that is, w = lcm(T1, T2, ... T Z ); Step 2: When the transmission starts, the sender has N packets waiting to be transmitted, and the channel erasure probability of each path is e1, e2, ... e Z , the transmission time is T1, T2, ... T Z , the total number of packets transmitted in a scheduling window is n w ,Only when a single path completes the transmission of an information packet or coding packet can the next packet be transmitted on the path, and the transmission of each path is carried out in parallel; Step 3: When the receiving end successfully receives the information packet, it immediately uploads the information packet to the application layer and feeds back the feedback information of the information packet to the sending end. The feedback information of each path is fed back after the feedback delay D1, D2, ..., D Z Arrives at the sending end; Step 4: When the sender starts transmitting, if it is the beginning of the scheduling window, the sender first confirms the packets that the receiver has not successfully received through the received feedback information, and then uses the feedback delay D and the erasure probability of each path e1, e2, ... e Z , estimate the total number of unsuccessfully received packets, encode these estimated unsuccessfully received packets into repair coded packets, and decide the number of packets, repair coded packets, and forward error correction coded packets to be sent, the total number is n w , arrange these n according to the receiving time sequence of each sending position in the scheduling window w The sending position of each packet is determined to eliminate the problem of out-of-order reception caused by different transmission rates of multiple paths. The repair code packet is received first, followed by the information packet, and finally the forward error correction code packet. After the scheduling is completed, the sender sends the information packet on the currently available path according to the current scheduling. Step 5: When the sender starts transmission, if it is not the start of the scheduling window, the sender first confirms the packets that the receiver has not successfully received through the received feedback information, and inserts these lost packets into the head of the current sending queue. The lost packets and the remaining unsent packets and coded packets in the scheduling window are rescheduled for transmission according to the order of reception time, so as to give priority to the retransmission of the lost packets. After completing the scheduling arrangement, the sender sends the packets on the currently available path according to the current scheduling arrangement. Step 6: When the receiving end successfully receives the data packet, repeat step 3; Step 7: After completing steps 1-6, repeat steps 4-6 until N information packets are transmitted.

2. The method for multipath information packet encoding, scheduling and real-time transmission under delayed feedback according to claim 1, characterized in that: The total number of packets that can be transmitted within a scheduling window in step 2 is n w is the length of the scheduling window w and the transmission time of each path T1, T2, ... T Z OK, n w for:

3. The method for multi-path information packet encoding, scheduling and real-time transmission under delayed feedback according to claim 2, characterized in that: In step 4, the known number of unacknowledged packets u and the erasure probability of each path e1, e2, ... e are obtained through feedback information in each scheduling window. Z and feedback delays D1, D2, ..., D Z The estimated number of unacknowledged packets is recorded as For the transmitter, each path has a feedback delay of D1, D2, ..., D Z The number of unacknowledged packets sent within a period is known and is denoted as k1, k2, ..., k Z , from which we can get the estimated number of unacknowledged packets, namely for 4. The method for multi-path information packet encoding, scheduling and real-time transmission under delayed feedback according to claim 3, characterized in that: Step 4: The erasure probability of each path e1, e2, ... e Z , the length of the scheduling window w and the transmission time of each path T1, T2, ...T Z Determine the number of information packets, repair coding packets, and forward error correction coding packets sent, which are: Repair code package: Information Packet: Forward error correction coding packet:

Images