5G real-time streaming media transmission control method and system based on link load factor

Abstract

The invention discloses a 5G real-time streaming media transmission control method and system based on a link load factor, and belongs to the technical field of streaming media wireless transmission. In order to solve the technical problems of large queue delay, feedback lag, low link utilization rate and the like in real-time communication in the 5G network, a cooperative control mechanism combining base station side link state sensing, load factor-based ECN marking on a core network UPF side, terminal ACK feedback and sending end code rate adaptive adjustment is mainly adopted. According to the invention, accurate regulation and control of real-time streaming media transmission under a 5G dynamic link can be realized, and the bandwidth utilization efficiency and the service continuity are improved.

Description

Technical Field

[0001] This invention belongs to the field of wireless streaming media transmission technology, specifically relating to a 5G real-time streaming media transmission control method and system based on link load factor. Background Technology

[0002] With the widespread deployment of 5G technology, mobile networks are rapidly evolving towards lower latency and higher bandwidth, giving rise to real-time interactive applications that are highly sensitive to latency and bandwidth, such as video conferencing, cloud gaming, and AR / VR. However, despite the low-latency mechanisms introduced by 5G systems, their radio access network (5G-AN) remains the main bottleneck causing end-to-end latency. Current real-time transmission (RTC) systems mostly rely on end-to-end feedback (such as RTT and ACK) for congestion control, which suffers from feedback lag and untimely response. This is particularly pronounced in scenarios with sudden bandwidth jitter or multi-user concurrency, where problems such as drastic fluctuations in transmission rate and uncontrollable frame-level latency are especially prominent.

[0003] The L4S architecture accelerates the congestion feedback path through ECN marking and a dual-queue mechanism, improving the congestion response capability of some links. However, the original L4S mechanism does not consider wireless channel fluctuations, frame-level data characteristics, and the differentiated needs of multiple users, resulting in insufficient stability in complex environments. Traditional static rate configuration and coarse-grained control methods are also difficult to adapt to dynamic network changes and cannot provide effective guarantees for frame-level QoS indicators within a short time window.

[0004] Therefore, there is an urgent need for an adaptive transmission control mechanism that has link status awareness, can finely adjust the transmission strategy according to the characteristics of frame-level data, and is feasible to deploy, so as to effectively suppress tail latency, improve the smoothness of video streams, and enhance service quality in multi-user scenarios while ensuring high bandwidth utilization. Summary of the Invention

[0005] The purpose of this invention is to address the common technical problems in existing 5G network real-time communication systems, such as large queue delays, delayed congestion feedback, low link bandwidth utilization, and the difficulty of stable operation of traditional low-latency, low-packet-loss service mechanisms in complex wireless channels. This invention proposes a 5G real-time streaming media transmission control method and system based on link load factors to achieve accurate perception and rapid response to network congestion, thereby improving the transmission stability, bandwidth adaptability, and quality of service assurance capabilities of real-time streaming media services in dynamic network environments.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] A 5G real-time streaming media transmission control method based on link load factor includes the following steps:

[0008] 1) The sending end encodes the raw video data to generate encoded video data, which is then transmitted to the receiving end via the 5G core network and 5G base station;

[0009] 2) Collect link status parameters at the 5G base station side, and calculate the link load factor ρ based on bandwidth, queue length, frame interval and frame rate;

[0010] 3) The link load factor ρ is reported to the User Plane Function (UPF) module of the 5G core network. The UPF module uses the token bucket algorithm to map the link load factor ρ to the congestion label probability and sets the Congestion Experience (CE) bit in the Explicit Congestion Notification (ECN) field of the IP header of the video data packet to generate the CE label as a congestion label.

[0011] 4) The receiving end receives video data packets carrying CE tags, parses the CE tag information, and carries the link load factor ρ or CE tag ratio information in the generated ACK message, and sends it to the sending end;

[0012] 5) The sending end receives the ACK message, dynamically adjusts the video encoding bit rate based on the information it carries, calculates the transmission rate, and realizes adaptive control of the video stream.

[0013] Furthermore, in step 3), the token bucket algorithm is as follows: the number of tokens is dynamically accumulated according to the value of the link load factor ρ. When the accumulated value exceeds the preset threshold, the CE bit is set for the current downlink data packet to mark it, so as to realize the proportional feedback of the link congestion level.

[0014] Further, in step 4), the receiving end counts the number of data packets marked with CE in the received video data packets within one video frame period, and calculates the ratio with the total number of received data packets to estimate the link load factor ρ of the corresponding frame.

[0015] Furthermore, the dynamic adjustment rules for the video coding bitrate in step 5) include:

[0016] When ρ is less than the target threshold Θ, the target coding bit rate Bt is calculated by the following formula: Bt=Bhis×Θ / ρ, where Bhis represents the historical coding bit rate of the previous period;

[0017] When ρ is greater than or equal to Θ and less than 1, according to the compression factor Θ 2 The target coding bit rate Bt is adjusted using the formula: Bt = Bhis × Θ 2 ;

[0018] When ρ equals 1, the sender estimates the queue length based on the receiving rate and the number of flying bytes fed back in the ACK, and performs a fast compression operation to reduce the target coding bit rate;

[0019] When ACK feedback is missing or abnormal, the sending end maintains the encoding bit rate of the previous cycle and uses a sliding window weighted average strategy to smooth bit rate fluctuations.

[0020] Furthermore, in step 5), in a multi-user environment, the link load factor ρ is weighted and adjusted according to user priority. The formula for weighted adjustment is ρi = (λ i *N / Σλ j )×ρ, where ρi represents the weighted link load factor of the i-th user, λ i Σλ represents the priority weight of the i-th user, N represents the total number of users, and Σλ represents the priority weight of the ith user. j This represents the sum of priority weights for all users.

[0021] A 5G real-time streaming media transmission control system based on link load factor includes:

[0022] The video encoding module, deployed at the transmitting end, is used to receive the raw video signal, compress and encode the video data according to the received target encoding bit rate, and output the encoded video data stream.

[0023] The link load factor calculation module is deployed on the 5G base station side to collect link status parameters and calculate the link load factor ρ based on bandwidth, queue length, frame interval and frame rate.

[0024] The ECN tagging module, deployed in the UPF module of the 5G core network, is used to receive the link load factor ρ, map it to the congestion tagging probability using the token bucket algorithm, and tag the CE bit in the ECN field of the IP header of the downlink video data packet.

[0025] The ACK feedback module, deployed at the receiving end, is used to parse the CE mark information in the received downlink video data packets and generate feedback information based on the statistical CE mark ratio. This feedback information is then encapsulated in an ACK message and sent to the sending end.

[0026] The rate control module, deployed at the transmitting end, is used to receive the link load factor ρ or CE flag ratio in the ACK message and dynamically adjust the target coding bit rate of the video coding module based on the target load threshold.

[0027] Furthermore, the UPF module dynamically accumulates the number of tokens based on the value of the link load factor ρ. When the accumulated value exceeds a preset threshold, the CE bit is set on the current downlink data packet to mark it, so as to achieve proportional feedback on the link congestion level.

[0028] Furthermore, the rate control module is used to dynamically adjust the target coding bit rate according to the following rules:

[0029] When ρ is less than the target threshold Θ, the target coding bit rate Bt is calculated by the following formula: Bt=Bhis×Θ / ρ, where Bhis represents the historical coding bit rate of the previous period;

[0030] When ρ is greater than or equal to Θ and less than 1, according to the compression factor Θ 2 The target coding bit rate Bt is adjusted using the formula: Bt = Bhis × Θ 2 ;

[0031] When ρ equals 1, the sender estimates the queue length based on the receiving rate and the number of flying bytes fed back in the ACK, and performs a fast compression operation to reduce the target coding bit rate;

[0032] When ACK feedback is missing or abnormal, the sending end maintains the encoding bit rate of the previous cycle and uses a sliding window weighted average strategy to smooth bit rate fluctuations.

[0033] Furthermore, it also includes a sliding window adjustment module, which is used to perform weighted averaging of the encoding bit rate adjustment behavior of the rate control module within multiple time windows.

[0034] Furthermore, it also includes a multi-user priority strategy module, used to weight and adjust the link load factor ρ according to user priority in a multi-user environment, and the rate control module allocates a higher target coding bit rate to high-priority users; the weighted adjustment formula is ρi = (λ i *N / Σλ j )×ρ, where ρi represents the weighted link load factor of the i-th user, λ i Σλ represents the priority weight of the i-th user, N represents the total number of users, and Σλ represents the priority weight of the ith user. j This represents the sum of priority weights for all users.

[0035] The beneficial effects achieved by this invention are as follows:

[0036] 1. This invention can accurately reflect the congestion level of the wireless downlink by monitoring the link status and estimating the link load factor in real time at the 5G base station, thus avoiding the feedback lag problem of traditional reliance on packet loss or RTT.

[0037] 2. This invention uses the token bucket algorithm to mark data packets with ECN through the 5G core network UPF module, realizing proportional explicit congestion marking of downlink data packets, forming a feedback mechanism that is dynamically coupled with link load, and significantly improving the timeliness and accuracy of congestion signals.

[0038] 3. This invention utilizes the receiver to send back the CE tag via ACK message to achieve frame-level link status feedback, supports the sender to adaptively adjust the encoding bitrate at the frame granularity, realizes fine-grained bitrate control under low latency, and effectively ensures the continuity and image quality stability of streaming media services.

[0039] 4. The present invention uses a non-linear growth and flexible backoff mechanism to regulate the bit rate at the sending end based on the relationship between the link load factor and the target threshold, which can improve the link utilization and avoid packet loss or lag caused by sudden congestion.

[0040] 5. In the event of missing ACK feedback or long-term link congestion, this invention enables a sliding window smoothing mechanism or a queue length estimation strategy to ensure the stability and adaptability of video stream transmission under uncertain feedback conditions.

[0041] 6. This invention supports a dynamic link load allocation strategy under multi-user scheduling. Combined with a priority control mechanism, it can achieve differentiated service guarantees in concurrent scenarios, thereby improving the overall system resource scheduling efficiency and user experience.

[0042] 7. This invention has the advantage of modular deployment. The relevant functions can be implemented through software upgrades based on existing 5G base stations and core network equipment, and has good engineering feasibility and deployment cost advantages.

[0043] 8. This invention is particularly suitable for low-latency, high-bandwidth real-time streaming media service scenarios, including cloud gaming, AR / VR, and high-definition video conferencing, and can effectively control transmission bandwidth consumption while ensuring service quality. Attached Figure Description

[0044] Figure 1 This is a schematic diagram of the architecture of a 5G real-time streaming media transmission and control system based on link load factor.

[0045] Figure 2 This is a flowchart of the link load factor calculation and feedback process in a 5G real-time streaming media transmission and control system.

[0046] Figure 3 This is a graph showing the changes in bit rate and bandwidth under different wired network latency conditions.

[0047] Figure 4 A graph showing the bitrate and bandwidth allocation for video streams of different priorities. Detailed Implementation

[0048] To make the various technical features, advantages, or effects of the present invention more apparent and understandable, detailed descriptions are provided below through embodiments.

[0049] This invention provides a 5G real-time streaming media transmission and control system based on link load factor, which aims to solve the problems of high queue latency, feedback lag, low bandwidth utilization and poor adaptability of traditional L4S mechanisms in existing real-time communication systems in 5G network environments. It is particularly suitable for multimedia interactive scenarios with high requirements for low latency and high throughput, such as cloud gaming, AR / VR, and high-definition video conferencing.

[0050] like Figure 1-2 As shown, this system is deployed in the end-to-end transmission path of the 5G network, including the video transmitting end, 5G base station, 5G core network UPF (User Plane Function) module, and receiving end user equipment. The following functional modules are deployed in each of the above nodes:

[0051] Link load factor calculation module: Deployed in 5G base stations, it is used to estimate the link load factor ρ based on network status parameters such as bandwidth, queue length, current frame rate and frame interval;

[0052] ECN tagging module: Deployed in the 5G core network UPF, it receives the link load factor ρ from the base station, generates congestion tags using the token bucket mechanism, and sets the ECN (Explicit Congestion Notification) bit in the IP header of downlink data packets, where the CE (Congestion Experienced) bit is set to be valid as a congestion status indicator bit.

[0053] ACK feedback module: Deployed at the receiving end device, it is used to parse the ECN tag information in the downlink data packet and feed it back to the sending end in the uplink ACK message. This feedback link ensures the closed-loop transmission of congestion status on the end-to-end path.

[0054] Rate control module: Deployed at the video transmitter, it is used to dynamically adjust the target bit rate of the video encoder based on the link load factor information or CE tag ratio carried in the received ACK;

[0055] Sliding window adjustment module: This is an optional module used to smooth bitrate changes and improve video quality stability.

[0056] The system's operation flow is as follows:

[0057] 1. The video stream enters the 5G access network from the sending end, passes through the 5G core network UPF module, and is forwarded to the receiving end user via the 5G base station.

[0058] The 2.5G base station monitors the status of each user's independent queue, allocated bandwidth, queue length, and frame characteristics in real time. Based on these indicators, the link load factor ρ is estimated. ρ is a normalized quantity that reflects the current level of link congestion and its value ranges from (0,1).

[0059] 3. The link load factor ρ is reported to the 5G core network UPF module, which uses the token bucket algorithm to map ρ to the ECN tag probability and sets the tag for downlink video data packets, that is, sets the CE bit in the ECN field of the IP header to 1 to indicate the current congestion status.

[0060] 4. After receiving a data packet with the CE tag, the receiving device counts the number of tags and encapsulates the tag ratio (i.e., the proportion of CE bits) in an acknowledgment message (ACK) and sends it back to the sending end.

[0061] 5. The rate control module at the video transmitter calculates and updates the target coding bitrate based on the link load factor ρ or ECN flag ratio in the ACK feedback (i.e., the CE bit ratio fed back by the receiver) and the transmission rate of the previous frame, and dynamically adjusts the transmission rate to achieve adaptive control of the real-time video stream.

[0062] The specific bit rate adjustment strategy is as follows:

[0063] (1) If the link load factor ρ is lower than the target threshold Θ, i.e. ρ<Θ, it means that the link is in a low load state. The sender will linearly amplify the historical bit rate Bhis: Bt=Bhis×Θ / ρ.

[0064] (2) If ρ is higher than the target threshold Θ, i.e. ρ>Θ, it means that the system judges the link to be in a congestion trend, and the sender will back off according to the target backoff coefficient Θ. 2 Backtracking from historical bit rates: Bt = Bhis × Θ 2 ;

[0065] (3) If ρ remains equal to 1, it indicates a completely congested state. The sender will estimate the current queue length by using the bandwidth estimation and the number of flying bytes in the ACK, and perform bit rate compression to clear the base station queue.

[0066] (4) If the ACK feedback is abnormal or lost, the sending end maintains the rate of the previous cycle and starts the sliding window weighted average strategy to smooth short-term rate fluctuations, ensuring the robustness and consistency of the system.

[0067] To avoid image quality degradation or playback stuttering caused by frequent bitrate fluctuations, the system introduces a sliding window adjustment module, which performs weighted smooth adjustment based on the bitrate change trend within multiple recent time windows, thereby improving the consistency of video playback while ensuring response speed.

[0068] This system supports a multi-user priority control mechanism. Considering the characteristic of 5G base stations maintaining an independent queue for each user, when multiple concurrent video streams exist, the system can set different link load adjustment strategies based on user priority. For example, for high-priority users, a relatively lower link load factor ρ is allocated to guide the transmitter to allocate higher coding bit rate resources to them, thereby achieving differentiated service guarantees.

[0069] This system also supports the following auxiliary mechanisms:

[0070] Startup phase rate prediction mechanism: During the connection initialization or before feedback is established, a conservative growth model is used for rate prediction to avoid link impact caused by sudden bit rate spikes.

[0071] Feedback interruption fault tolerance mechanism: When ACK feedback is lost, the system operation continuity is maintained by estimating the relationship between the rate and the number of flying bytes;

[0072] Queue clearing mechanism: When congestion is detected to be ongoing, the system enters queue clearing mode and actively compresses the coding rate to release buffer space;

[0073] Fast response mechanism: The system supports subframe-level rate adjustment, which enables adjustment feedback and response to be completed within each video frame, greatly shortening the response cycle.

[0074] This invention also provides a 5G real-time streaming media transmission control method based on link load factor, applicable to low-latency multimedia transmission scenarios in 5G network environments. This method can be deployed between the video transmitter, 5G base station, 5G core network, and receiver, and includes the following steps:

[0075] 1. Link load factor estimation: On the 5G base station side, information such as queue length, bandwidth usage, frame rate and frame interval of each user is collected in real time, and the load factor ρ of the current link is calculated based on the above parameters to reflect the link congestion status.

[0076] 2. Link load factor reporting and mapping: The link load factor ρ is reported to the User Plane Function (UPF) module of the 5G core network. The UPF module uses the token bucket mechanism to map ρ to the congestion probability and sets the ECN (Explicit Congestion Notification) flag for downlink video data packets based on the probability. The CE bit in the IP header is set to indicate the current congestion level.

[0077] 3. ECN tag transmission and reception: Data packets with ECN tags are sent from the 5G core network to the receiving device via the access network. The receiving end parses the ECN tag information in the data packets and counts the proportion of CE bits in the received packets.

[0078] 4. Feedback Construction and Reporting: The receiving end carries the parsed link load factor ρ or CE ratio information in the generated ACK packet and reports this feedback to the video sending end.

[0079] 5. Encoding Rate Adjustment: After receiving feedback, the video transmitter dynamically adjusts the encoder's target bit rate based on the link load factor ρ and the current target threshold Θ, according to the following strategy:

[0080] When ρ < Θ, the bit rate is amplified proportionally.

[0081] When ρ>Θ, according to the compression factor Θ 2 Reduce bit rate;

[0082] If ρ remains 1 or the CE ratio continues to rise, predict the queue length and further compress the bit rate to clear the queue.

[0083] If the ACK feedback is interrupted, maintain the rate of the previous cycle and adjust it using a sliding window smoothing strategy.

[0084] 6. Sliding Smoothing and Multi-User Optimization: Optionally, the video transmitter performs a weighted average based on the historical rate adjustment trend within the sliding window to smooth bit rate fluctuations. Simultaneously, in a multi-user environment, the link load factor is adjusted differentially according to user priority using a weighted method. The formula for weighted adjustment is: ρi=(λ i *N / Σλ j )×ρ, where ρi represents the weighted link load factor of the i-th user, λ i Σλ represents the priority weight of the i-th user, N represents the total number of users, and Σλ represents the priority weight of the ith user. j This represents the sum of the priority weights of all users, thereby guiding the sender to allocate bandwidth resources to high-priority users first.

[0085] Figure 3This paper demonstrates the bit rate adaptive performance of the algorithm under different wired network latency conditions. The experimental scenario simulates a typical 5G home access environment, with three wired network link latencies: 5ms, 15ms, and 25ms, corresponding to the red, blue, and green curves, respectively. The orange curve represents the actual available wireless bandwidth in this scenario. As shown in the figure, under different latency conditions, all bit rate curves exhibit good bandwidth tracking capability, closely matching the actual bandwidth change trend, proving that the feedback mechanism based on the link load factor proposed in this invention has strong dynamic response capability. Specifically, under the 5ms latency condition, the algorithm achieves the best accuracy and smoothness in rate control, with the bit rate curve almost closely following the actual bandwidth. Under the 15ms and 25ms latency conditions, the system still achieves effective rate regulation, but the increased feedback latency leads to greater fluctuations in the bit rate curve. Particularly noteworthy is that even under the high latency condition of 25ms, the algorithm maintains high responsiveness, rapidly adjusting the output bit rate when network bandwidth changes abruptly, avoiding overload or congestion caused by feedback latency, and demonstrating good latency robustness. Figure 3 This further verifies the broad adaptability of the method of the present invention under low to medium-high network latency conditions, indicating that it can be applied to various practical deployment scenarios, including wired access and hybrid network environments.

[0086] Figure 4 The effectiveness of the algorithm in allocating bandwidth to video streams of different priorities was further evaluated in a shared link environment. The experimental environment was set to a fixed 15ms wired network latency, simulating two simultaneously active 5G video stream users: one a high-priority video stream (priority_2, blue curve), and the other a normal-priority video stream (priority_1, red curve). The orange curve represents the actual available wireless bandwidth. As observed in the graph, throughout the nearly 10-second test, the high-priority video stream consistently achieved a significantly higher bit rate, averaging approximately twice that of the normal-priority stream. Simultaneously, the sum of the bit rates of the two video streams remained close to the upper limit of the actual available bandwidth, demonstrating that the system achieved high link utilization during resource scheduling. Furthermore, when available bandwidth experienced sudden changes (e.g., at 2, 4, and 6 seconds), both video streams quickly adjusted their bit rates. The priority scheduling strategy enabled the high-priority video stream to maintain a relatively high quality of service even with insufficient bandwidth. This phenomenon indicates that the bandwidth control strategy based on the priority parameters proposed in this invention is effective and reliable in multi-user scenarios, achieving a good balance between quality of service (QoS) assurance and fairness. Figure 4 This clearly reflects the differentiated service capabilities of the present invention under the multi-stream concurrent scheduling strategy, and further demonstrates its practical application potential in multi-user 5G application environments, such as remote conferencing, multi-user cloud gaming and mixed reality collaboration.

[0087] Figure 3 and Figure 4 Experimental data and analysis fully demonstrate that the present invention achieves high responsiveness and stability of bit rate control under different network latency conditions; it has good bandwidth tracking capability and strong adaptability even in environments with large feedback latency; under the condition of multiple users sharing network resources, it can achieve efficient bandwidth allocation and QoS guarantee based on priority; while improving link utilization efficiency, it ensures the service continuity and quality priority of critical business flows.

[0088] Although the present invention has been disclosed above with reference to embodiments, it is not intended to limit the present invention. Appropriate modifications or equivalent substitutions made by those skilled in the art to the technical solutions of the present invention should be covered within the protection scope of the present invention, which is defined by the claims.

Claims

1. A 5G real-time streaming media transmission control method based on link load factor, characterized in that, Includes the following steps: 1) The sending end encodes the raw video data to generate encoded video data, which is then transmitted to the receiving end via the 5G core network and 5G base station; 2) Collect link status parameters at the 5G base station side, and calculate the link load factor ρ based on bandwidth, queue length, frame interval and frame rate; 3) The link load factor ρ is reported to the User Plane Function (UPF) module of the 5G core network. The UPF module uses the token bucket algorithm to map the link load factor ρ to the congestion label probability and sets the Congestion Experience (CE) bit in the Explicit Congestion Notification (ECN) field of the IP header of the video data packet to generate the CE label as a congestion label. 4) The receiving end receives video data packets carrying CE tags, parses the CE tag information, and carries the link load factor ρ or CE tag ratio information in the generated ACK message, and sends it to the sending end; 5) The sending end receives the ACK message, dynamically adjusts the video encoding bit rate based on the information it carries, calculates the transmission rate, and realizes adaptive control of the video stream.

2. The method as described in claim 1, characterized in that, In step 3), the token bucket algorithm is as follows: the number of tokens is dynamically accumulated based on the value of the link load factor ρ. When the accumulated value exceeds the preset threshold, the CE bit is set on the current downlink data packet to mark it, so as to realize the proportional feedback of the link congestion level.

3. The method as described in claim 1, characterized in that, In step 4), the receiving end counts the number of data packets marked with CE in the received video data packets within one video frame period, and calculates the ratio with the total number of received data packets to estimate the link load factor ρ of the corresponding frame.

4. The method as described in claim 1, characterized in that, The dynamic adjustment rules for the video coding bitrate in step 5) include: When ρ is less than the target threshold Θ, the target coding bit rate Bt is calculated by the following formula: Bt=Bhis×Θ / ρ, where Bhis represents the historical coding bit rate of the previous period; When ρ is greater than or equal to Θ and less than 1, according to the compression factor Θ 2 The target coding bit rate Bt is adjusted using the formula: Bt = Bhis × Θ 2 ; When ρ equals 1, the sender estimates the queue length based on the receiving rate and the number of flying bytes fed back in the ACK, and performs a fast compression operation to reduce the target coding bit rate; When ACK feedback is missing or abnormal, the sending end maintains the encoding bit rate of the previous cycle and uses a sliding window weighted average strategy to smooth bit rate fluctuations.

5. The method as described in claim 1, characterized in that, In step 5), in a multi-user environment, the link load factor ρ is weighted and adjusted according to user priority. The formula for weighted adjustment is ρi = (λ i *N / Σλ j )×ρ, where ρi represents the weighted link load factor of the i-th user, λ i Σλ represents the priority weight of the i-th user, N represents the total number of users, and Σλ represents the priority weight of the ith user. j This represents the sum of priority weights for all users.

6. A 5G real-time streaming media transmission control system based on link load factor, characterized in that, include: The video encoding module, deployed at the transmitting end, is used to receive the raw video signal, compress and encode the video data according to the received target encoding bit rate, and output the encoded video data stream. The link load factor calculation module is deployed on the 5G base station side to collect link status parameters and calculate the link load factor ρ based on bandwidth, queue length, frame interval and frame rate. The ECN tagging module, deployed in the UPF module of the 5G core network, is used to receive the link load factor ρ, map it to the congestion tagging probability using the token bucket algorithm, and tag the CE bit in the ECN field of the IP header of the downlink video data packet. The ACK feedback module, deployed at the receiving end, is used to parse the CE mark information in the received downlink video data packets and generate feedback information based on the statistical CE mark ratio. This feedback information is then encapsulated in an ACK message and sent to the sending end. The rate control module, deployed at the transmitting end, is used to receive the link load factor ρ or CE flag ratio in the ACK message and dynamically adjust the target coding bit rate of the video coding module based on the target load threshold.

7. The system as described in claim 6, characterized in that, The UPF module dynamically accumulates the number of tokens based on the value of the link load factor ρ. When the accumulated value exceeds a preset threshold, the CE bit is set on the current downlink data packet to mark it, so as to achieve proportional feedback on the link congestion level.

8. The system as described in claim 6, characterized in that, The rate control module is used to dynamically adjust the target coding bit rate according to the following rules: When ρ is less than the target threshold Θ, the target coding bit rate Bt is calculated by the following formula: Bt=Bhis×Θ / ρ, where Bhis represents the historical coding bit rate of the previous period; When ρ is greater than or equal to Θ and less than 1, according to the compression factor Θ 2 The target coding bit rate Bt is adjusted using the formula: Bt = Bhis × Θ 2 ; When ρ equals 1, the sender estimates the queue length based on the receiving rate and the number of flying bytes fed back in the ACK, and performs a fast compression operation to reduce the target coding bit rate; When ACK feedback is missing or abnormal, the sending end maintains the encoding bit rate of the previous cycle and uses a sliding window weighted average strategy to smooth bit rate fluctuations.

9. The system as described in claim 6, characterized in that, It also includes a sliding window adjustment module, which is used to perform weighted averaging of the encoding bit rate adjustment behavior of the rate control module in multiple time windows.

10. The system as described in claim 6, characterized in that, It also includes a multi-user priority strategy module, which is used to adjust the link load factor ρ according to user priority in a multi-user environment, and the rate control module allocates a higher target coding bit rate to high-priority users. The formula for weighted adjustment is ρi=(λ i *N / Σλ j )×ρ, where ρi represents the weighted link load factor of the i-th user, λ i Σλ represents the priority weight of the i-th user, N represents the total number of users, and Σλ represents the priority weight of the ith user. j This represents the sum of priority weights for all users.

Images