A multi-path quic video transmission scheduling method and system for low-altitude cellular handover risk

Abstract

The application discloses a kind of multi-path QUIC video transmission scheduling methods and systems for low altitude cellular switching risk, the wireless measurement information of multiple cellular links is collected frequently on low altitude flight platform, and each path switching risk is predicted using prediction model;The sending end combines switching risk and operating state, dynamically decides the routing and splitting of video stream, reduces sending or empties in-transit data for high-risk path, and preferentially carries key video for low-risk path;Video content is graded and path scoring and traffic allocation are performed, active congestion control and data protection are implemented during switching process;Combined with the feedback of receiving end, the strategy is continuously adjusted, and closed-loop optimization is formed.The application introduces scheduling jointly with link switching context and path state, suppresses high-risk path exposure in advance, reduces healthy path involvement probability, avoids simple fallback leading to stop when multi-path high-risk, effectively reduces video freezing and lag, and improves the continuity and reliability of low-altitude video backhaul.

Description

Technical Field

[0001] This invention relates to the field of video transmission scheduling technology, and in particular to a multi-path QUIC video transmission scheduling method and system for addressing the risks of low-altitude cellular handover. Background Technology

[0002] At present, 5G / 5G-A can support video backhaul services in the low-altitude economy. However, in scenarios involving high-speed movement, frequent switching, multiple links, multiple high-definition video streams, or continuous high-reliability backhaul, mechanisms such as switching risk awareness, multi-path scheduling, and QoE (Quality of Experience) feedback are still needed to improve continuity and reliability.

[0003] As low-altitude terminals such as drones acquire multi-cellular link access capabilities, multi-path transmission can improve bandwidth utilization and link robustness. However, most existing multi-path scheduling methods still rely primarily on link quality indicators such as round-trip time, packet loss rate, queue length, congestion window, or bandwidth for route selection, lacking the ability to perceive the risks inherent in the cellular handover process itself.

[0004] In low-altitude, high-speed mobile scenarios, link states change more rapidly, and cellular handovers are more frequent. When a path is in the process of handover or multiple paths simultaneously enter a high-risk state within a similar time window, traditional solutions tend to continue sending video data to the high-risk path, leading to out-of-order delivery, reassembly delays, and video freezes. Meanwhile, healthy paths may also be indirectly affected by connection-level resource contention, congestion control coupling, or buffer backlog. Simply reverting to single-path transmission can easily cause transmission starvation, resulting in inconsistent video frame output. These problems are particularly pronounced in video services, as they are highly sensitive to short-term timing anomalies. Even minor erroneous decisions by the scheduler within the handover window can trigger continuous zero-frame output, long-tail stuttering, and decreased service continuity.

[0005] Therefore, existing technologies still need to be improved and developed. Summary of the Invention

[0006] The main objective of this invention is to provide a multi-path QUIC video transmission scheduling method and system for addressing low-altitude cellular handover risks. This aims to solve the problem that existing scheduling methods struggle to handle link handover risks when low-altitude flight platforms are transmitting multi-path QUIC video under cellular network coverage. This invention combines handover risk information and operational status information from multiple cellular paths to collaboratively optimize the video stream routing and distribution process. This reduces the probability of video freezing, stuttering, and transmission interruptions under high-risk conditions of single-path handover or multi-path overlap, thereby improving the transmission continuity and service experience of low-altitude video backhaul services.

[0007] To achieve the above objectives, the present invention provides a multi-path QUIC video transmission scheduling method for addressing low-altitude cellular handover risks. The multi-path QUIC video transmission scheduling method for addressing low-altitude cellular handover risks includes the following steps: High-frequency acquisition of wireless measurement information from multiple cellular links is performed on a low-altitude flight platform to obtain wireless measurement information reflecting the link change trend. The trained handover risk prediction model is used to predict the cellular handover risk of each path, and the prediction results are provided to the multi-path QUIC transmitter. When transmitting video data, the multi-path QUIC transmitter makes dynamic decisions on the routing and splitting of video streams by combining the switching risk information and operating status of each path. For paths with a switching risk higher than the preset risk threshold, it reduces the transmission or clears the data in transit in advance. For paths with a switching risk lower than the preset risk threshold, it prioritizes carrying critical video data. The video stream is classified into content categories and path scoring and traffic allocation are performed. Active congestion control and in-transit data protection are implemented when handover is imminent or during handover. By combining the path status and service quality feedback returned by the receiver, the transmission strategy is continuously adjusted to form a closed-loop optimization mechanism for low-altitude cellular handover risks. The transmitter makes adaptive adjustments based on the feedback results and enters the next scheduling cycle.

[0008] Optionally, the multi-path QUIC video transmission scheduling method for low-altitude cellular handover risks, wherein the high-frequency acquisition of radio measurement information of multiple cellular links on a low-altitude flight platform to obtain radio measurement information reflecting link change trends specifically includes: A 5G measurement module is set up on each cellular interface of the multipath QUIC transmitter to periodically obtain the radio measurement information of the current cell using AT commands. The radio measurement information includes at least the reference signal received power RSRP and the cell identifier PCID. The wireless measurement information is then timestamped to form a measurement sequence ordered by time. The measurement process is performed in a high-frequency manner.

[0009] Optionally, the multi-path QUIC video transmission scheduling method for low-altitude cellular handover risk, wherein the step of predicting the cellular handover risk of each path using a trained handover risk prediction model and providing the prediction results to the multi-path QUIC transmitter specifically includes: The measured sequences are mapped to the actual handover labels to construct time-series classification samples. A recurrent neural network model is used to predict whether a base station handover will occur within a future time window. The model is trained using LSTM and combined with data augmentation to obtain a well-trained handover risk prediction model. The trained handover risk prediction model is deployed as a risk query interface. The risk query interface is used to receive the measurement sequence of a certain cellular path in the most recent period, output the risk value or handover probability of the path being handed over in the subsequent predetermined time window, and provide the prediction result to the multi-path QUIC transmitter.

[0010] Optionally, in the multi-path QUIC video transmission scheduling method for low-altitude cellular handover risks, the risk query interface provides a query service through an HTTP API, and the sending end periodically calls the risk query interface to obtain the risk value or handover probability of each path.

[0011] Optionally, the multi-path QUIC video transmission scheduling method for low-altitude cellular handover risks includes the multi-path QUIC transmitter making dynamic decisions on video stream routing and distribution by combining the handover risk information and operating status of each path during video data transmission. For paths with handover risks higher than a preset risk threshold, transmission is reduced or in-transit data is cleared in advance. For paths with handover risks lower than the preset risk threshold, critical video data is prioritized. Specifically, this includes: When the multi-path QUIC transmitter receives the video stream to be sent, it acquires the switching risk and synchronously acquires the current transmission status information of each path, including at least the round-trip time (RTT), the predicted throughput, the congestion window (CWND), the amount of data in transit (Inflight), the ACK confirmation status, and the changes in path throughput. The transmission status information and handover risk are used together as the joint input for the current scheduling cycle to reflect link quality and link problems; The multi-path QUIC sender updates the risk value or switching probability of each path according to a preset period. The scheduler obtains and updates the switching probability of each path in real time through the HTTP API, and marks high-risk paths according to a preset risk threshold. For paths with a handover risk higher than a preset risk threshold, the scheduler treats high-risk paths as paths to be avoided. For paths with a handover risk lower than a preset risk threshold and a stable state, the scheduler treats low-risk paths as priority paths. If the handover prediction service is unavailable, the scheduler falls back to the basic scheduling module based on MinRTT.

[0012] Optionally, the multi-path QUIC video transmission scheduling method for low-altitude cellular handover risks, wherein the step of content-leveling the video stream and performing path scoring and traffic allocation specifically includes: The multi-path QUIC transmitter distinguishes the video streams to be transmitted according to their importance, treating key frames (I-frames) and reference frames (P-frames) as high-priority data, and B-frames, redundant frames, or data with low timing sensitivity as relatively low-priority data. The scheduler comprehensively considers the switching risk, round-trip time (RTT), throughput prediction, congestion window (CWND), in-flight data volume, and ACK confirmation status of each path to score or rank the paths and generate routing results or traffic splitting ratios for the current scheduling cycle. For paths with a switching risk below a preset risk threshold and a stable status, high-priority video data will be prioritized. For high-risk paths that are not yet completely ineffective, reduce the allocation ratio or allocate only low-priority data.

[0013] Optionally, the multi-path QUIC video transmission scheduling method for low-altitude cellular handover risks, wherein the active congestion control and in-transit data protection performed near the handover or during the handover process specifically includes: When a path is identified as a high-risk switching path, the multi-path QUIC sender will no longer inject new data into the high-risk switching path, and will reserve time in advance to drain the Inflight data on the high-risk switching path.

[0014] Optionally, the multi-path QUIC video transmission scheduling method for low-altitude cellular handover risks, wherein the step of continuously adjusting the transmission strategy by combining the path status and service quality feedback returned by the receiver to form a closed-loop optimization mechanism for low-altitude cellular handover risks, and the transmitter adaptively adjusting based on the feedback results and entering the next scheduling cycle, specifically includes: The receiving end continuously receives data streams from multiple paths and returns path status feedback and service quality feedback to the sending end; Based on the feedback results, the multi-path QUIC transmitter updates the splitting ratio, path priority, and video transmission parameters. When the requirements are met, it adjusts the video resolution, bitrate, or GOP transmission strategy, and submits the updated video to be sent back to the scheduler to correct the scheduling strategy for the next cycle.

[0015] Optionally, in the multi-path QUIC video transmission scheduling method for low-altitude cellular handover risks, the path status feedback includes arrival jitter, out-of-order status, and reception throughput, and the service quality feedback includes stuttering, playback continuity, and freeze events.

[0016] Furthermore, to achieve the above objectives, the present invention also provides a multi-path QUIC video transmission scheduling system for low-altitude cellular handover risks, wherein the multi-path QUIC video transmission scheduling system for low-altitude cellular handover risks is used to implement the multi-path QUIC video transmission scheduling method for low-altitude cellular handover risks, and the multi-path QUIC video transmission scheduling system for low-altitude cellular handover risks includes: a multi-path QUIC transmitter, a handover risk prediction model, and a receiver; The multipath QUIC transmitter is used for wireless measurement and acquisition, path status acquisition, scheduling decision-making, traffic splitting control, and video transmission adjustment. The switching risk prediction model is trained using low-altitude flight measurement data and outputs the switching risk of each path through an interface. The receiving end is used to monitor the reception effect and return the path status and QoE feedback to the multipath QUIC sending end.

[0017] In this invention, high-frequency wireless measurement information of multiple cellular links is collected on a low-altitude flight platform to obtain wireless measurement information reflecting the link change trend. A trained handover risk prediction model is used to predict the cellular handover risk of each path, and the prediction results are provided to the multi-path QUIC transmitter. When transmitting video data, the multi-path QUIC transmitter makes dynamic decisions on the routing and distribution of video streams based on the handover risk information and operating status of each path. For paths with handover risk higher than a preset risk threshold, the transmission is reduced or in-transit data is cleared in advance. For paths with handover risk lower than the preset risk threshold, key video data is prioritized. The video stream is classified according to content and path scoring and traffic allocation are performed. Active congestion control and in-transit data protection are performed when handover is imminent or during handover. Based on the path status and service quality feedback returned by the receiver, the transmission strategy is continuously adjusted to form a closed-loop optimization mechanism for low-altitude cellular handover risk. The transmitter makes adaptive adjustments based on the feedback results and enters the next scheduling cycle. This invention integrates link switching context and path operation status into the scheduling decision process, and performs collaborative optimization of video stream routing and distribution. This enables the system to suppress the exposure of high-risk paths in advance before switching, reduce the probability of healthy paths being affected during switching, and avoid overall shutdown caused by simple single-path backoff in high-risk scenarios with multiple overlapping paths. This reduces the probability of video freezing, stuttering and shutdown, and improves the transmission continuity and reliability of low-altitude video backhaul services. Attached Figure Description

[0018] Figure 1 This is a flowchart of a preferred embodiment of the multi-path QUIC video transmission scheduling method for low-altitude cellular handover risks of the present invention; Figure 2This is a schematic diagram illustrating the principle of transmission scheduling in a preferred embodiment of the multi-path QUIC video transmission scheduling method for low-altitude cellular handover risks of the present invention. Figure 3 This is a flowchart illustrating the interaction between three endpoints to complete the entire method in a preferred embodiment of the multi-path QUIC video transmission scheduling method for low-altitude cellular handover risks according to the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer and more explicit, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0020] The preferred embodiment of the multi-path QUIC video transmission scheduling method for low-altitude cellular handover risks described in this invention, such as... Figure 1 and Figure 2 As shown, the multi-path QUIC video transmission scheduling method for low-altitude cellular handover risks includes the following steps: Step S10: High-frequency acquisition of wireless measurement information of multiple cellular links on a low-altitude flight platform to obtain wireless measurement information reflecting the link change trend.

[0021] Specifically, a 5G measurement module is set up on each cellular interface of the multipath QUIC transmitter to periodically acquire the current cell's wireless measurement information using AT (Modem Control Command) commands. The wireless measurement information includes at least the Reference Signal Received Power (RSRP) and the Physical Cell ID (PCID). The wireless measurement information is then timestamped to form a time-ordered measurement sequence. The measurement process is performed at a high frequency, for example, sampling can be performed at a frequency of 100ms or higher, so as to capture the rapidly changing pre-handover information during low-altitude flight as much as possible.

[0022] Step S20: Use the trained handover risk prediction model to predict the cellular handover risk of each path, and provide the prediction results to the multi-path QUIC transmitter.

[0023] Specifically, during the model training phase, the measurement sequences (i.e., the low-altitude flight RSRP and PCID change sequences obtained in step S10) are mapped to real handover labels to construct time-series classification samples. A recurrent neural network model is then used to predict whether base station handover will occur within a future time window. LSTM (Long Short-Term Memory) networks combined with data augmentation can be used for training to obtain a trained handover risk prediction model. Since there is a significant imbalance between handover and non-handover samples, the loss weight of handover samples is further increased using Focal Loss and a positive sample weighted binary cross-entropy loss function (BCE with Logits Loss) to enhance the early warning capability for handover events. This step is part of the pre-deployment training process and can be completed on a ground server, edge server, or R&D workstation.

[0024] The trained handover risk prediction model is deployed as a risk query interface. This interface receives the measurement sequence of a cellular path over a recent period and outputs the risk value or probability of handover occurring on that path within a predetermined time window. The prediction result is then provided to the multi-path QUIC transmitter. The handover risk prediction model provides query services via an HTTP (Hypertext Transfer Protocol) API (Application Programming Interface). The multi-path QUIC transmitter can periodically call this interface to obtain the HO (Handover) prediction probability for each path.

[0025] Step S30: When transmitting video data, the multi-path QUIC transmitter makes dynamic decisions on the routing and splitting of the video stream by combining the switching risk information and operating status of each path. For paths with a switching risk higher than the preset risk threshold, the transmission is reduced or the data in transit is cleared in advance. For paths with a switching risk lower than the preset risk threshold, key video data is carried first.

[0026] Specifically, when the multi-path QUIC transmitter receives the video stream to be sent, it acquires the handover risk and simultaneously acquires the current transmission status information of each path, including at least the round-trip time (RTT), the predicted throughput, the congestion window (CWND), the in-flight data, the ACK (acknowledgment) status, and the path throughput changes. The transmission status information and the handover risk are used together as the joint input of the current scheduling cycle to reflect link quality and link problems, that is, to reflect the two different dimensions of "whether the link quality is good" and "whether the link is about to have a problem".

[0027] The multi-path QUIC transmitter updates the risk value or switching probability of each path according to a preset period, and the scheduler (i.e. Figure 2 MPQUIC in the example obtains and updates the switching probability of each path in real time through the HTTP API, and marks high-risk paths according to a preset risk threshold, such as 0.5 as the judgment threshold in one example.

[0028] For paths with a handover risk higher than a preset risk threshold, the scheduler treats high-risk paths as paths to be avoided. For paths with a handover risk lower than a preset risk threshold and a stable state, the scheduler treats low-risk paths as priority paths. If the handover prediction service is unavailable, the scheduler falls back to the basic scheduling model based on MinRTT (Minimum Round-Trip Time).

[0029] In the dual-path implementation, it can be further divided into a stable state, a single-path high-risk state, and a multi-path simultaneous high-risk state, so that different transmission control strategies can be adopted subsequently. If the switching prediction service is unavailable, the scheduler falls back to the basic scheduling mode based on MinRTT (MinRTT is a default scheduler of MPQUIC that only selects the path with the smallest RTT as the transmission path) to ensure continuous system operation.

[0030] Step S40: Perform content classification on the video stream and implement path scoring and traffic allocation. Perform active congestion control and in-transit data protection during or near handover.

[0031] Specifically, the multi-path QUIC transmitter distinguishes the video streams to be transmitted according to their importance, treating key frames (I-frames, intra-coded frames) and reference frames (P-frames, predictive-coded frames) as high-priority data, and B-frames (bidirectionally predictive-coded frames), redundant frames, or data with low timing sensitivity as relatively low-priority data. Subsequently, the scheduler comprehensively considers the switching risk, round-trip time (RTT), predicted throughput, congestion window (CWND), in-flight data volume, and ACK confirmation status of each path to score or rank the paths, and generates routing results or traffic splitting ratios for the current scheduling cycle.

[0032] For paths with a switching risk below the preset risk threshold and a stable status, high-priority video data is prioritized; for high-risk paths that are not yet completely failed, the allocation ratio is reduced, or only low-priority data is allocated, in order to improve the continuity of overall video transmission and the QoE of the receiving end.

[0033] When a path is identified as a high-risk handover path, the multi-path QUIC sender stops injecting large amounts of new data into it. Instead, it allocates time in advance to purge Inflight data from the high-risk path, reducing the probability of buffer overflow, packet loss, and retransmissions after the handover. Simultaneously, the sender increases the load on healthy paths (paths with stable quality and low handover risk), allowing another path to fill the transmission gap if one path is temporarily frozen. The purpose of this step is to take action before the actual handover occurs, utilizing handover prediction results, rather than passively recovering after packet loss, timeouts, and congestion window contraction.

[0034] Step S50: Combining the path status and service quality feedback returned by the receiver, continuously adjust the transmission strategy to form a closed-loop optimization mechanism for low-altitude cellular handover risks. The transmitter will make adaptive adjustments based on the feedback results and enter the next scheduling cycle.

[0035] Specifically, the receiving end (i.e., the video receiving end) continuously receives data streams from multiple paths and returns path status feedback and service quality feedback to the sending end; the path status feedback may include arrival jitter, out-of-order status, receiving throughput, etc., and the service quality feedback may include stuttering, playback continuity, freeze events or other QoE indicators.

[0036] Based on feedback, the multi-path QUIC transmitter updates the splitting ratio, path priority, and video transmission parameters. When requirements are met, it adjusts the video resolution, bitrate, or GOP transmission strategy, and resubmits the updated video to be transmitted to the scheduler for further adjustments to the scheduling strategy for the next cycle. This forms a closed-loop control process of "risk perception—scheduling decision—feedback reception—continued adjustment."

[0037] This invention provides a multi-path QUIC video transmission scheduling method for addressing low-altitude cellular handover risks. It is implemented collaboratively by a low-altitude flight platform transmitter, a handover risk prediction module, and a video receiver. The low-altitude flight platform transmitter collects radio measurement information from multiple cellular links, obtains the operational status of each path, performs video stream classification and traffic allocation, and transmits video data via multi-path QUIC according to the scheduling results. The handover risk prediction module outputs the handover risk of each path based on the radio measurement sequence. The video receiver statistically analyzes the path status and service quality information on the receiving side and feeds it back to the transmitter so that the transmitter can continue to adjust its transmission strategy. The core of this scheme is that cellular handover risk information is input as an explicit context into the multi-path QUIC scheduler, enabling video stream routing and distribution to be jointly decided based on both handover risk and path operational status, rather than solely on latency or throughput.

[0038] Furthermore, such as Figure 3 As shown, the entire process of the multi-path QUIC video transmission scheduling method for low-altitude cellular handover risks includes: S101, Low-altitude flight platform (transmitter) locally collects wireless measurement information from multiple cellular paths; S102. Training a switching risk prediction model based on low-altitude flight measurement data; S103. Deploy the trained handover risk prediction model as a queryable service and output the handover risk corresponding to each path through the HTTP API for the scheduler to make decisions. S104. The sending end obtains the running status information of multiple paths in real time; S105. Determine the current scheduling status based on the handover risk and path status, obtain the handover risk value of each path, and jointly determine the current scheduling input with the path status. S106. Perform content classification on the video stream and execute path scoring and traffic allocation, calculate the transmission strength of each path, and determine the path allocation for a specific frame. S107. Perform active congestion control and in-transit data protection during or near handover. S108. The receiving end feeds back path status and service quality information; S109. The sending end makes adaptive adjustments based on the feedback results and enters the next scheduling cycle.

[0039] This invention achieves its goals by constructing a cross-layer collaborative mechanism of "state awareness - handover prediction - transmission scheduling - feedback adjustment." First, high-frequency acquisition of wireless measurement information from multiple cellular links is performed on a low-altitude flight platform to obtain state data reflecting link change trends, such as signal strength and base station identification. Then, a trained handover risk prediction model is used to predict the cellular handover risk of each path, and the prediction results are provided to the multi-path QUIC transmitter. When transmitting video data, the multi-path QUIC transmitter dynamically decides on video stream routing and distribution based on the handover risk information of each path, as well as operational status such as round-trip delay, throughput, and congestion window. For paths with high handover risk, the system proactively reduces transmission or preemptively clears in-transit data. For paths with relatively stable quality, critical video data is prioritized, thereby reducing packet loss, retransmission, video freezing, and transmission interruption caused by cellular handover. Furthermore, the system can continuously adjust transmission strategies based on path status and service quality feedback returned by the receiver, forming a closed-loop optimization mechanism for low-altitude cellular handover risk, thereby improving the continuity and reliability of low-altitude video backhaul services.

[0040] Furthermore, such as Figure 2 and Figure 3 As shown, based on the above-mentioned multi-path QUIC video transmission scheduling method for low-altitude cellular handover risks, the present invention also provides a multi-path QUIC video transmission scheduling system for low-altitude cellular handover risks. The multi-path QUIC video transmission scheduling system for low-altitude cellular handover risks is used to implement the above-mentioned multi-path QUIC video transmission scheduling method for low-altitude cellular handover risks. The multi-path QUIC video transmission scheduling system for low-altitude cellular handover risks includes: a multi-path QUIC transmitter, a handover risk prediction model, and a receiver. The multipath QUIC transmitter is used for wireless measurement acquisition, path status acquisition, scheduling decision-making, traffic splitting control, and video transmission adjustment, and is responsible for executing steps S101, S104, S105, S106, S107, and S109; the handover risk prediction model (i.e. Figure 2 The path prediction service module is used to train using low-altitude flight measurement data and outputs the switching risk of each path through the interface, and is responsible for executing steps S102 and S103; the receiving end is used to monitor the receiving effect and return the path status and QoE feedback to the multi-path QUIC sending end, and is responsible for executing step S108.

[0041] In summary, this invention provides a multi-path QUIC video transmission scheduling method and system for addressing low-altitude cellular handover risks. The method includes: high-frequency acquisition of wireless measurement information from multiple cellular links on a low-altitude flight platform to obtain wireless measurement information reflecting link change trends; prediction of cellular handover risks for each path using a trained handover risk prediction model, and provision of the prediction results to the multi-path QUIC transmitter; during video data transmission, the multi-path QUIC transmitter dynamically decides on video stream routing and distribution based on the handover risk information and operating status of each path; for paths with handover risks higher than a preset risk threshold, reducing transmission or prematurely clearing in-transit data; for paths with handover risks lower than the preset risk threshold, prioritizing the transmission of critical video data; performing content classification and path scoring and traffic allocation for the video stream; implementing active congestion control and in-transit data protection during or near handover; continuously adjusting the transmission strategy based on path status and service quality feedback returned by the receiver, forming a closed-loop optimization mechanism for low-altitude cellular handover risks; and the transmitter adaptively adjusting based on the feedback results and entering the next scheduling cycle. This invention integrates link switching context and path operation status into the scheduling decision process, and performs collaborative optimization of video stream routing and distribution. This enables the system to suppress the exposure of high-risk paths in advance before switching, reduce the probability of healthy paths being affected during switching, and avoid overall shutdown caused by simple single-path backoff in high-risk scenarios with multiple overlapping paths. This reduces the probability of video freezing, stuttering and shutdown, and improves the transmission continuity and reliability of low-altitude video backhaul services.

[0042] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal that includes that element.

[0043] Of course, those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.). The program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The computer-readable storage medium can be a memory, magnetic disk, optical disk, etc.

[0044] It should be understood that the application of the present invention is not limited to the examples above. Those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. A multi-path QUIC video transmission scheduling method for addressing low-altitude cellular handover risks, characterized in that, The multi-path QUIC video transmission scheduling method for addressing low-altitude cellular handover risks includes: High-frequency acquisition of wireless measurement information from multiple cellular links is performed on a low-altitude flight platform to obtain wireless measurement information reflecting the link change trend. The trained handover risk prediction model is used to predict the cellular handover risk of each path, and the prediction results are provided to the multi-path QUIC transmitter. When transmitting video data, the multi-path QUIC transmitter makes dynamic decisions on the routing and splitting of video streams by combining the switching risk information and operating status of each path. For paths with a switching risk higher than the preset risk threshold, it reduces the transmission or clears the data in transit in advance. For paths with a switching risk lower than the preset risk threshold, it prioritizes carrying critical video data. The video stream is classified into content categories and path scoring and traffic allocation are performed. Active congestion control and in-transit data protection are implemented when handover is imminent or during handover. By combining the path status and service quality feedback returned by the receiver, the transmission strategy is continuously adjusted to form a closed-loop optimization mechanism for low-altitude cellular handover risks. The transmitter makes adaptive adjustments based on the feedback results and enters the next scheduling cycle.

2. The multi-path QUIC video transmission scheduling method for low-altitude cellular handover risk as described in claim 1, characterized in that, The high-frequency acquisition of wireless measurement information from multiple cellular links on a low-altitude flight platform to obtain wireless measurement information reflecting link change trends specifically includes: A 5G measurement module is set up on each cellular interface of the multipath QUIC transmitter to periodically obtain the radio measurement information of the current cell using AT commands. The radio measurement information includes at least the reference signal received power RSRP and the cell identifier PCID. The wireless measurement information is then timestamped to form a measurement sequence ordered by time. The measurement process is performed in a high-frequency manner.

3. The multi-path QUIC video transmission scheduling method for low-altitude cellular handover risks according to claim 2, characterized in that, The process of using a trained handover risk prediction model to predict the cellular handover risk for each path and providing the prediction results to the multi-path QUIC transmitter specifically includes: The measured sequences are mapped to the actual handover labels to construct time-series classification samples. A recurrent neural network model is used to predict whether a base station handover will occur within a future time window. The model is trained using LSTM and combined with data augmentation to obtain a well-trained handover risk prediction model. The trained handover risk prediction model is deployed as a risk query interface. The risk query interface is used to receive the measurement sequence of a certain cellular path in the most recent period, output the risk value or handover probability of the path being handed over in the subsequent predetermined time window, and provide the prediction result to the multi-path QUIC transmitter.

4. The multi-path QUIC video transmission scheduling method for low-altitude cellular handover risk according to claim 3, characterized in that, The risk query interface provides query services through an HTTP API. The sending end calls the risk query interface periodically to obtain the risk value or switching probability of each path.

5. The multi-path QUIC video transmission scheduling method for low-altitude cellular handover risk according to claim 3, characterized in that, When transmitting video data, the multi-path QUIC transmitter dynamically decides on the routing and splitting of the video stream by combining the switching risk information and operating status of each path. For paths with a switching risk higher than a preset risk threshold, it reduces the transmission of data or prematurely clears data in transit. For paths with a switching risk lower than the preset risk threshold, it prioritizes carrying critical video data, specifically including: When the multi-path QUIC transmitter receives the video stream to be sent, it acquires the switching risk and synchronously acquires the current transmission status information of each path, including at least the round-trip time (RTT), the predicted throughput, the congestion window (CWND), the amount of data in transit (Inflight), the ACK confirmation status, and the changes in path throughput. The transmission status information and handover risk are used together as the joint input for the current scheduling cycle to reflect link quality and link problems; The multi-path QUIC sender updates the risk value or switching probability of each path according to a preset period. The scheduler obtains and updates the switching probability of each path in real time through the HTTP API and marks high-risk paths according to a preset risk threshold. For paths with a handover risk higher than a preset risk threshold, the scheduler treats high-risk paths as paths to be avoided. For paths with a handover risk lower than a preset risk threshold and a stable state, the scheduler treats low-risk paths as priority paths. If the handover prediction service is unavailable, the scheduler falls back to the basic scheduling module based on MinRTT.

6. The multi-path QUIC video transmission scheduling method for low-altitude cellular handover risks according to claim 5, characterized in that, The process of content-based classification of video streams and execution of path scoring and traffic allocation specifically includes: The multi-path QUIC transmitter distinguishes the video streams to be transmitted according to their importance, treating key frames (I-frames) and reference frames (P-frames) as high-priority data, and B-frames, redundant frames, or data with low timing sensitivity as relatively low-priority data. The scheduler comprehensively considers the switching risk, round-trip time (RTT), throughput prediction, congestion window (CWND), in-flight data volume, and ACK confirmation status of each path to score or rank the paths and generate routing results or traffic splitting ratios for the current scheduling cycle. For paths with a switching risk below a preset risk threshold and a stable status, high-priority video data will be prioritized. For high-risk paths that are not yet completely ineffective, reduce the allocation ratio or allocate only low-priority data.

7. The multi-path QUIC video transmission scheduling method for low-altitude cellular handover risk according to claim 6, characterized in that, The implementation of proactive congestion control and in-transit data protection during or near handover is specifically as follows: When a path is identified as a high-risk switching path, the multi-path QUIC sender will no longer inject new data into the high-risk switching path, and will reserve time in advance to drain the Inflight data on the high-risk switching path.

8. The multi-path QUIC video transmission scheduling method for low-altitude cellular handover risk according to claim 7, characterized in that, The transmission strategy is continuously adjusted by combining the path status and service quality feedback returned by the receiving end, forming a closed-loop optimization mechanism for low-altitude cellular handover risks. The sending end adaptively adjusts based on the feedback results and enters the next scheduling cycle, specifically including: The receiving end continuously receives data streams from multiple paths and returns path status feedback and service quality feedback to the sending end; Based on the feedback results, the multi-path QUIC transmitter updates the splitting ratio, path priority, and video transmission parameters. When the requirements are met, it adjusts the video resolution, bitrate, or GOP transmission strategy, and submits the updated video to be sent back to the scheduler to correct the scheduling strategy for the next cycle.

9. The multi-path QUIC video transmission scheduling method for low-altitude cellular handover risk according to claim 8, characterized in that, The path status feedback includes arrival jitter, out-of-order status, and receiving throughput, while the service quality feedback includes stuttering, playback continuity, and freeze events.

10. A multi-path QUIC video transmission scheduling system for low-altitude cellular handover risks, characterized in that, The multi-path QUIC video transmission scheduling system for low-altitude cellular handover risk is used to implement the multi-path QUIC video transmission scheduling method for low-altitude cellular handover risk as described in any one of claims 1-9. The multi-path QUIC video transmission scheduling system for low-altitude cellular handover risk includes: a multi-path QUIC transmitter, a handover risk prediction model, and a receiver. The multipath QUIC transmitter is used for wireless measurement and acquisition, path status acquisition, scheduling decision-making, traffic splitting control, and video transmission adjustment. The switching risk prediction model is trained using low-altitude flight measurement data and outputs the switching risk of each path through an interface. The receiving end is used to monitor the reception effect and return the path status and QoE feedback to the multipath QUIC sending end.

Images