Adaptation System and Method for Protocol Compatibility Conversion of Heterogeneous Intelligent Agent Systems
By reshaping the input data packets into a continuous data stream in a heterogeneous system and utilizing a double-buffered ping-pong structure and fixed-length transmission, the timing jitter problem caused by the randomness of the input data packet length and interval is solved, ensuring stable data supply and timing consistency for downstream devices.
Patent Information
- Application Number
- CN202611387377.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-08
- Publication Date
- 2026-11-17
AI Technical Summary
Existing technologies cannot effectively eliminate output timing jitter caused by the randomness of input data packet length and arrival time interval in heterogeneous systems, which affects the instruction execution accuracy of downstream intelligent agents with high timing consistency requirements.
A protocol-compatible conversion system for adapting to heterogeneous intelligent systems is adopted. Through a data receiving module, a stream reconstruction module, and a fixed-length transmission control module, the boundaries of the input data packets are broken, and the data is reshaped into a continuous data stream. A double-buffered ping-pong structure and a preset fixed threshold are used for fixed-length transmission to ensure that the length of the output data blocks is constant and the interval is stable.
It achieves a constant output data block length and stable time interval, eliminates the impact of input fluctuations on output timing, reduces system complexity and hardware cost, and is suitable for heterogeneous network environments such as industrial control and autonomous driving.
Smart Images

Figure CN123123433A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of intelligent agent communication and data processing, and more specifically to a system and method for adapting to the protocol compatibility conversion of heterogeneous intelligent agent systems. Background Technology
[0002] In heterogeneous systems containing agents with various communication rates, protocol conversion is a crucial step in achieving device interconnection. Current technologies typically parse and convert input data packets before immediately forwarding them to downstream devices. However, due to the randomness of the length and arrival time intervals of upstream input data packets, direct forwarding causes fluctuations in the time intervals of output data packets. For downstream agents such as industrial actuators and motion control modules, which have high requirements for timing consistency, these timing fluctuations can lead to instruction execution deviations, severely impacting the system's collaborative accuracy.
[0003] To address this issue, existing solutions often employ complex queue scheduling algorithms, priority allocation mechanisms, or service frame reassembly logic for traffic shaping. Among these, double-buffering-based solutions only buffer complete data packets and cannot resolve output timing jitter caused by fluctuations in input packet length. Continuous stream reassembly-based solutions require identifying service frame boundaries, introducing complex parsing logic and uncertain processing delays. Time-triggered fixed-length transmission solutions require global clock synchronization, resulting in high system complexity and inability to adapt to different clock domains in heterogeneous systems. All of these solutions fail to fundamentally eliminate timing jitter caused by input data characteristics and introduce numerous parameters requiring manual tuning, making them difficult to adapt to the actual operating environments of different heterogeneous systems. Therefore, this paper proposes a protocol compatibility conversion system and method for intelligent agent heterogeneous systems to address these problems. Summary of the Invention
[0004] Technical problems to be solved To address the shortcomings of existing technologies, this invention provides a system and method for protocol compatibility conversion of heterogeneous intelligent agent systems, thus solving the problems mentioned in the background section.
[0005] Technical solution To achieve the above objectives, the present invention adopts the following technical solution: an intelligent agent heterogeneous system protocol compatibility conversion system, comprising: a data receiving module for receiving input data packets; a stream reassembly module for converting the format of the input data packets and breaking the original packet boundaries of the converted data, writing it as a continuous data stream into a buffer storage module; a buffer storage module adopting a double-buffered ping-pong structure for alternately storing continuous data streams and accumulating the amount of data written in each buffer; and a fixed-length transmission control module for monitoring the accumulated amount of data currently written into the buffer, and when the accumulated amount of data reaches a preset fixed threshold, switching the writing buffer and sending the data in the full buffer as a fixed-length data block, and automatically clearing the buffer after transmission to prepare for the next write.
[0006] The above scheme reshapes variable-length, variable-interval input data packets into a boundless continuous data stream, and triggers ping-pong switching and fixed-length transmission based solely on the pure physical quantity of the cumulative data volume currently written into the buffer. This fundamentally cuts off the transmission path of input fluctuations to the output end, while simultaneously solving the problems of data loss and instruction truncation under a single-buffer structure, ensuring that the length of the output data block is constant and the transmission interval is stable.
[0007] Preferably, the specific steps for breaking the original packet boundaries and writing the converted data as a continuous data stream to the buffer storage module are as follows: The frame header and trailer of the input data packet are stripped, and only the payload data is extracted; the payload data is seamlessly spliced into a continuous byte stream according to the receiving sequence, and the continuous byte stream is written to the currently active write buffer without retaining the original packet boundaries. The payload length output by the upstream agent in this system is an integer multiple of a preset fixed threshold to ensure that service instructions are not truncated to a fixed length.
[0008] This preferred solution completely strips away the protocol encapsulation structure and seamlessly splices the payload, eliminating the interference of input data packet boundaries on the subsequent reassembly process. This allows the buffer storage module to deal only with pure byte streams, providing the necessary prerequisite for subsequent fixed-length truncation based on pure physical quantities and avoiding the timing uncertainties introduced by preserving packet structure in traditional fragmentation mechanisms. Simultaneously, system-level constraints ensure the integrity of business instructions, eliminating the need to introduce complex frame judgment logic within the conversion system.
[0009] Preferably, the preset fixed threshold is set as follows: the preset fixed threshold is a constant value preset based on the minimum number of bytes received by the downstream receiver or the number of bytes required for a fixed processing cycle, and remains constant during system operation.
[0010] This preferred approach ensures that the transmission rhythm is precisely matched with the downstream processing capacity by anchoring the threshold to the inherent physical characteristics of the downstream device rather than the network state. At the same time, the constant threshold eliminates any additional timing disturbances that may be introduced by any dynamic adjustment mechanism, thus guaranteeing the absolute determinism of the output timing.
[0011] Preferably, the specific steps for switching the write buffer and the send buffer are as follows: when the cumulative amount of data in the current write buffer reaches a preset fixed threshold, the write buffer is switched to another idle buffer within the same clock cycle, and a full buffer data send operation is triggered at the same time. After the send is completed, the buffer is cleared and its corresponding counter is reset.
[0012] This preferred solution achieves equivalent atomic operations at the hardware level through a ping-pong switching mechanism, ensuring the continuity of data stream writing and the independence of each output data block. It prevents historical residual data from causing cumulative errors in subsequent transmission timing and completely solves the data loss problem caused by transmission and writing conflicts under a single-buffered structure.
[0013] Preferably, the decoupling relationship between the transmission time interval of the fixed-length data block and the input data packet is implemented as follows: when the system byte processing rate is not lower than the downstream bus transmission rate, the transmission time interval of the fixed-length data block is determined only by the constant physical transmission rate of the downstream bus and a preset fixed threshold, decoupled from the length and arrival time interval of the input data packet. The average data rate of the upstream agent applicable to this system is strictly matched with the downstream transmission rate, ensuring that the buffer will not experience long-term overflow or underload.
[0014] This preferred scheme clarifies that the only determining factor for the output timing is the constant physical parameters within the system. It mathematically proves the independence of the output timing from the input characteristics, providing a predictable and verifiable stable data supply cycle for downstream timing-sensitive devices.
[0015] Preferably, the specific steps for the cumulative data volume in the monitoring buffer storage module are as follows: the cumulative number of bytes currently active in the buffer is read directly through a hardware register, without the need for software calculation or protocol analysis.
[0016] This preferred solution utilizes the atomic read characteristic of hardware registers to obtain the cumulative data volume, avoiding the uncertain delays caused by software polling, interrupt response, or complex algorithm calculations, ensuring the real-time and deterministic nature of trigger determination, and further reducing the system implementation complexity and timing jitter risk.
[0017] Preferably, the specific steps for format conversion of the input data packet are as follows: performing protocol parsing on the input data packet, extracting the target control command, and converting the target control command into a payload in the target protocol format.
[0018] This preferred approach limits protocol conversion to the necessary operations of extraction and format mapping, eliminating redundant processing unrelated to stream reassembly, ensuring that the converted data can be directly used as part of a continuous byte stream, and supporting the simplicity and efficiency of the overall architecture.
[0019] Preferably, the specific steps for sending the data in the buffer storage module as a fixed-length data block are as follows: sending the fixed-length data block to a downstream receiving end that is sensitive to timing consistency, wherein the fixed-length data block is used to drive the downstream receiving end to perform a synchronization action.
[0020] This preferred solution clarifies that the application target of this system is time-sensitive equipment. The constant cycle time of the fixed-length data block directly serves the synchronous execution requirements of the equipment, reflecting the deep coupling between the technical solution and the specific application scenario.
[0021] Furthermore, this invention also provides a method for adapting to heterogeneous intelligent agent system protocol compatibility conversion, comprising: receiving an input data packet; converting the format of the input data packet and breaking the original packet boundary of the converted data, writing it as a continuous data stream into the currently active write buffer; storing the continuous data stream in the currently active write buffer and accumulating the amount of data written; monitoring the accumulated amount of data in the currently active write buffer, and when the accumulated amount of data reaches a preset fixed threshold, switching the write buffer and sending the data in the full buffer as a fixed-length data block, and after the sending is completed, clearing the buffer and resetting its counter.
[0022] The above method achieves the same technical concept at the method level through the corresponding steps and processes of the system. That is, by breaking the continuous stream reassembly of the input packet boundary, double-buffered ping-pong switching, and unique fixed-length triggering based on the cumulative data volume, the influence of input fluctuations on output timing is completely eliminated, while ensuring the integrity of business instructions and guaranteeing the timing consistency of data received by downstream intelligent agents.
[0023] Beneficial effects The present invention has the following beneficial effects: 1. This adaptive intelligent agent heterogeneous system protocol compatibility conversion system and method breaks the original boundary of the input data packet and writes the converted data as a continuous data stream into the dual-buffered storage module. Combined with the ping-pong switching and fixed-length transmission mechanism triggered by the current cumulative data volume written to the buffer reaching a preset fixed threshold, it solves the problem of the impact of input data packet length differences and arrival interval fluctuations on output timing. At the same time, it avoids the problems of data loss and instruction truncation under the single-buffered structure, and realizes that the output data block length is constant and the time interval between adjacent output data blocks depends only on the constant physical transmission rate of the downstream bus and the preset threshold, thus providing a stable data supply rhythm for downstream devices.
[0024] 2. This invention solves the problems of high system implementation complexity and difficult parameter tuning caused by traditional solutions that rely on complex queue scheduling, dynamic priority allocation or feedback adjustment algorithms, by relying solely on the cumulative data volume that can be directly read from hardware registers as the sole trigger criterion, without introducing any additional control variables or dynamic adjustment mechanisms. It achieves extreme simplification of control logic, reduces the demand for computing resources, and helps to reduce hardware costs and improve system reliability.
[0025] 3. Through the above-mentioned mechanism of decoupling output timing and input characteristics, combined with reasonable system-level constraints, this invention solves the problem of maintaining absolute timing consistency between devices with different communication rates in heterogeneous intelligent agent systems. It realizes predictable and zero-jitter data supply for timing-sensitive devices such as industrial actuators and motion control modules, and is suitable for heterogeneous network environments such as industrial control and autonomous driving with a large number of steady-state operations.
[0026] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description
[0027] Figure 1 This is a structural diagram of the protocol compatibility conversion system for heterogeneous intelligent agent systems adapted to this invention; Figure 2 This is a flowchart illustrating the protocol compatibility conversion method for heterogeneous intelligent agent systems adapted to this invention. Detailed Implementation
[0028] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0029] Example 1
[0030] like Figure 1 As shown, this embodiment provides a protocol compatibility conversion system and data stream processing method for adapting to heterogeneous intelligent agent systems. This system, as the core hub connecting upstream heterogeneous intelligent agents and downstream time-sensitive devices, mainly includes a data receiving module, a stream reassembly module, a buffer storage module, and a fixed-length transmission control module in its hardware architecture. It should be understood that although... Figure 1 The specific connection order of each module is shown, but in other embodiments, the physical layout or integration method of each module can be adjusted as long as the logical processing order of the data flow is satisfied. For example, the stream reassembly module and the buffer storage module can be integrated into the same FPGA chip.
[0031] The data receiving module is used to receive input data packets. Specifically, as the system's entry point, it is responsible for adapting to the physical interfaces and link-layer protocols of various upstream intelligent agents with different communication rates. In actual operation, input data packets typically have variable lengths, and their arrival time intervals fluctuate randomly due to network load. After completing basic physical layer verification, the data receiving module transparently transmits the raw data packets to the next level without performing any content-based caching or scheduling processing, thus ensuring the integrity and real-time performance of the raw data stream.
[0032] The stream reassembly module is used to convert the format of input data packets and break the original packet boundaries of the converted data, writing it as a continuous data stream into the buffer storage module. This is the core feature that distinguishes this invention from the packet forwarding mechanism in the prior art. Specifically, the specific steps for breaking the original packet boundaries of the converted data and writing it as a continuous data stream into the buffer storage module are as follows: The frame header and frame trailer of the input data packet are stripped, and only the payload data is extracted; the payload data is seamlessly spliced into a continuous byte stream according to the receiving sequence, and the continuous byte stream is written to the currently active write buffer without retaining the original packet boundaries. The payload length of the upstream intelligent agent output in this system is an integer multiple of a preset fixed threshold to ensure that the service instructions are not truncated to a fixed length. In this process, the stream reassembly module first parses the heterogeneous protocol, extracts the pure service data, i.e., the payload, and completely discards the frame header, frame trailer, and check fields used for transmission control. Subsequently, regardless of the payload length of the previous data packet, the payload of the current data packet is written directly after it without inserting any separators, length markers, or packet sequence numbers in between. For example, if the previous packet extracted 100 bytes and the current packet extracted 50 bytes, then in the currently active write buffer, these 150 bytes represent a continuous memory region, making it impossible to distinguish from physical storage where the previous packet ends or the next packet begins. This unstructured processing method frees subsequent stages from the discreteness of the input packets, laying the necessary foundation for pure physical quantity fixed-length truncation. Since the payload length of each independent service instruction output by the upstream agent is an integer multiple of a preset fixed threshold W, after seamlessly splicing all payloads according to the receiving time sequence, any continuous segment of length W in the continuous byte stream will necessarily contain an integer number of complete service instructions. There will be no situation where a single instruction is split into two fixed-length data blocks, fundamentally ensuring that the downstream receiver can correctly parse and execute all instructions.
[0033] The buffer storage module employs a dual-buffered ping-pong structure to alternately store continuous data streams and accumulate the amount of data written into each buffer. Specifically, the buffer storage module can be physically implemented using two independent sets of dual-port RAM, FIFO queues, or circular buffers, with a total capacity twice the preset fixed threshold. Its key characteristic is that data write operations are treated as appending to the byte stream, without involving any packet-level metadata management. Each buffer maintains an independent high-precision hardware counter that reflects the total number of bytes written but not yet sent in the corresponding buffer in real time. Because the data exists as a continuous stream, this accumulated data amount is a purely physical scalar, containing no application-layer semantics, allowing subsequent triggering decisions to be completely independent of software protocol stack intervention. During system operation, only one buffer is always in an active write state, while the other buffer is either in a pending send or empty state.
[0034] The fixed-length transmission control module monitors the cumulative data volume in the currently active write buffer. When the cumulative data volume reaches a preset fixed threshold, it immediately switches the write buffer to another idle buffer and sends the data in the full buffer as a fixed-length data block. After transmission, the buffer is automatically cleared and its corresponding counter is reset. This mechanism achieves complete decoupling of output timing and input characteristics, while solving the data loss problem caused by transmission and write conflicts in a single-buffer structure. The preset fixed threshold is not arbitrarily selected or dynamically adjusted, but rather a constant value preset based on the minimum received byte unit at the downstream receiver or the number of bytes required for a fixed processing cycle, remaining constant throughout system operation. Simultaneously, the effective payload length output by the upstream agent must be an integer multiple of this preset fixed threshold to ensure that business instructions are not truncated. For example, if the bus communication cycle of the downstream industrial actuator is fixed at 1ms, and 256 bytes of control instructions must be processed within this cycle to complete a full interpolation operation, then the preset fixed threshold is hard-coded or configured to be 256 bytes, and the length of each control instruction output by the upstream agent is an integer multiple of 256 bytes. This threshold setting and system constraint, based on the inherent physical properties of upstream and downstream devices, ensures a precise match between output rhythm and execution capability, avoiding additional timing noise and business errors introduced by dynamic fluctuations in thresholds or instruction truncation.
[0035] In the execution logic of sending and resetting, the switching of the write buffer and the triggering of sending are completed within the same clock cycle, ensuring the continuity of data flow writing and preventing data loss. The operation of clearing the corresponding buffer and resetting its counter after sending is completed before the buffer is reactivated for writing, ensuring that the initial state of each buffer is blank. The sending completion signal of the fixed-length sending control module is generated by the physical layer controller of the downstream bus. The time to send W bytes of data is strictly equal to W divided by Q2, a constant physical constant without any random fluctuations. Therefore, the timing of clearing the buffer and resetting the counter is completely deterministic, without introducing additional timing jitter. This ping-pong switching mechanism achieves equivalent atomic operations at the hardware level, ensuring that the generation of each output data block is independent and stateless, preventing the cumulative error of historical data from affecting subsequent timing. Since writing and sending are performed in two independent buffers, the writing operation is not interrupted during the sending process and does not conflict with the sending operation in terms of resources, completely resolving the timing contradictions under the single-buffer structure.
[0036] To rigorously verify the timing decoupling effect of this embodiment from a mathematical perspective, the byte processing rate of the stream reassembly module is defined as Q1 (bytes per second), the constant physical transmission rate of the downstream bus is defined as Q2 (bytes per second), and the preset fixed threshold is defined as W (bytes). Since W is a constant value and the stream reassembly module adopts a fully hardware pipelined design, the processing delay of each byte is fixed at one clock cycle. Therefore, Q1 is determined by the hardware clock frequency and data bit width, and is a constant. Q2 is determined by the physical layer standard of the downstream bus, and is also a constant. When the system design satisfies Q1 greater than or equal to Q2, the data processing speed will not become a bottleneck, and the transmission time interval E of the output data block (in seconds) is entirely determined by the downstream transmission rate, which can be expressed as E equal to W divided by Q2. As can be seen from the above formula, the output time interval E is determined only by the internal constant parameters W and Q2 of the system. Let the length of the input data packet be R (in bytes), and the arrival time interval be T (in seconds). Regardless of how R fluctuates between 64 and 1500 bytes, and regardless of how T jitters within the millisecond range, as long as the average data rate of the upstream agent equals the downstream transmission rate (W divided by E), and the total capacity of the buffer storage module is 2W, long-term overflow or underload will not occur, and E will remain constant. The total capacity of the buffer storage module in this system is 2W, which can absorb short-term fluctuations in the upstream data rate not exceeding W. As long as the instantaneous peak of the upstream data rate does not exceed twice the downstream transmission rate, and the average rate equals the downstream transmission rate, the buffer will not overflow or underload. For extreme fluctuations exceeding this range, which are system-level flow planning problems, this solution does not perform additional processing to maintain the purity of the control logic. The dimensions of each parameter are such that the dimension of E equals the dimension of W divided by the dimension of Q², that is, bytes divided by bytes per second equals seconds, which conforms to the physical meaning of time units. This normalization logic proves that the present invention, through the combination of a double-buffered ping-pong structure, continuous flow recombination, and fixed threshold truncation, successfully reshapes a variable-length, variable-interval input flow into a fixed-length, equally interval output flow, thereby physically cutting off the transmission path of input disturbances to the output end.
[0037] Furthermore, this embodiment also provides a method for adapting to the protocol compatibility conversion of heterogeneous intelligent agent systems, such as... Figure 2 As shown, this method corresponds to the system architecture described above, ensuring full disclosure at the method level. The method includes the following steps: Step S100: Receive input data packets. This step corresponds to the function of the data receiving module, completing the access of heterogeneous data. Step S200: Convert the format of the input data packets and break the original packet boundaries, writing the converted data as a continuous data stream into the currently active write buffer. This step corresponds to the core processing logic of the stream reassembly module, achieving destructured reassembly of the data stream by stripping frame headers and trailers and seamlessly splicing the payload. Step S300: Store the continuous data stream in the currently active write buffer and accumulate the amount of data written. This step corresponds to the physical storage and counting function of the buffer storage module, providing a purely physical basis for fixed-length triggering. Step S400: Monitor the accumulated data amount in the currently active write buffer. When the accumulated data amount reaches a preset fixed threshold, switch the write buffer and send the data in the full buffer as a fixed-length data block. After sending, clear the buffer and reset its counter. This step corresponds to the control logic of the fixed-length transmission control module, ensuring absolute consistency of the output timing through ping-pong switching and atomic reset.
[0038] Through the collaborative work of the above systems and methods, this embodiment achieves high-precision timing shaping with only simple hardware logic without relying on complex queue scheduling or intelligent analysis algorithms. It is particularly suitable for steady-state heterogeneous intelligent agent interconnection scenarios with extremely high deterministic requirements.
[0039] Example 2
[0040] In this embodiment, the specific implementation of protocol conversion and the hardware monitoring mechanism for cumulative data volume are further described in detail to demonstrate the determinism and efficiency of the present invention in its underlying physical implementation.
[0041] Regarding the specific steps for format conversion of the input data packet, this embodiment adopts the following method: the input data packet is parsed according to the protocol, the target control command is extracted, and the target control command is converted into a payload in the target protocol format.
[0042] Specifically, the stream reassembly module employs a fully hardware pipelined architecture to implement protocol conversion. Each processing stage corresponds to a fixed clock cycle, ensuring that the processing latency of each byte is completely consistent. The pipeline depth of the stream reassembly module is designed based on the most complex supported protocols. For simple protocols, redundant pipeline stages only perform data pass-through processing without executing any operations. In this way, regardless of the protocol used by the input data packet, the total latency of each byte from entering the stream reassembly module to being output to the buffer storage module is strictly equal to the pipeline depth multiplied by the clock cycle, ensuring that the byte processing rate Q1 remains a constant and does not fluctuate due to differences in protocol complexity. Faced with diverse data packets sent by upstream heterogeneous intelligent agents, the stream reassembly module, according to the pre-defined protocol stack specifications, sequentially completes the stripping of routing information, checksum verification, removal of encapsulation headers and footers, and extraction and format mapping of target control commands at different stages of the pipeline. In this process, a crucial processing principle is destructured processing and constant latency.
[0043] Regardless of the complexity of the upstream input protocol structure, and regardless of the field reassembly involved in the conversion process, the data ultimately written to the buffer storage module must be completely restored to a continuous byte sequence in physical form, and the total latency of each byte from input to output remains constant. This means that the data entering the buffer no longer carries any application-layer semantic boundaries, packet length identifiers, or type tags; at the circuit level, they are simply represented by a continuous arrangement of high and low voltage levels, and the output rate of the data stream is entirely determined by the pipeline clock, unaffected by differences in the input data packet structure. This constant-latency pipeline design ensures that the byte processing rate Q1 of the stream reassembly module remains constant and does not fluctuate due to changes in protocol parsing complexity.
[0044] Based on this, regarding the specific steps for monitoring the cumulative data volume in the buffer storage module, this embodiment directly reads the cumulative number of bytes currently active and written to the buffer through hardware registers, without the need for software calculation or protocol analysis.
[0045] like Figure 1 As shown, each buffer in the buffer storage module is equipped with an independent dedicated hardware register. In practice, this hardware register can be a status register within the corresponding FIFO controller, or an independent hardware counter that increments synchronously with the write pointer of the storage unit. The value of each register is automatically maintained by the corresponding hardware logic, reflecting in real-time and accurately the total number of bytes written but not read from the corresponding buffer. When the fixed-length transmission control module needs to determine whether to trigger transmission, it only needs to directly read the current value of the register corresponding to the currently active write buffer via the hardware bus and compare it with a preset fixed threshold. This process is completed entirely within the hardware domain, without consuming any CPU instruction cycles, and without requiring operating system intervention for memory traversal or data packet parsing.
[0046] This hardware-level monitoring method has significant technical advantages over traditional software polling or interrupt statistics mechanisms. In software implementations, acquiring the amount of data in the buffer typically requires function calls, context switching, and even traversing linked lists. The time consumption of these operations fluctuates randomly due to system load, interrupt priority, and cache hit rate, leading to uncertainty in the transmission trigger time and introducing timing jitter that is difficult to eliminate. This invention, however, utilizes the direct read capability of hardware registers to solidify the monitoring action into a deterministic electrical signal transmission process, with a constant and predictable nanosecond-level physical constant response delay.
[0047] This not only achieves real-time monitoring with zero latency, but also guarantees absolute determinism of trigger determination from the physical level, ensuring that the transmission interval of the output data block strictly follows the mathematical law derived in Embodiment 1, unaffected by any random factors at the software level. It should be understood that although this embodiment uses hardware registers as the preferred implementation, in other embodiments, any equivalent monitoring unit constructed using dedicated ASIC circuits or FPGA internal logic resources that can achieve deterministic counting functionality without software intervention should be covered within the scope of protection of this invention.
[0048] Example 3
[0049] In this embodiment, the timing consistency effect of the present invention in actual operation and its non-obviousness relative to the prior art will be described in detail in conjunction with a specific heterogeneous intelligent agent application scenario. The application scenario shown in this embodiment is intended to verify the robustness of the technical solution in complex industrial environments and is not intended to limit the scope of protection of the present invention.
[0050] Regarding the application scenarios of the system, this embodiment is set as follows: the system is applied in a steady-state heterogeneous network environment where the average data rate of the upstream intelligent agent is strictly matched with the downstream transmission rate and the effective payload length of the upstream output is an integer multiple of a preset fixed threshold, as a gateway or protocol conversion node.
[0051] Specifically, in a typical industrial automated production line, the upstream connection may be a motion control agent or PLC master station based on standard Ethernet. The communication protocol is TCP / IP, and the data packet length varies between 64 and 1500 bytes depending on the number of encapsulated instructions. However, the length of each independent control instruction is always a multiple of 256 bytes, and the average data output rate of the upstream agent is strictly equal to the transmission rate of the downstream bus. Due to network load, the time interval between data packets arriving at the gateway may exhibit random jitter within the range of 1 to 10 milliseconds.
[0052] Downstream, these devices connect to timing-sensitive industrial actuators, such as servo drives or high-speed I / O modules. These devices typically use real-time industrial bus protocols like EtherCAT, CANopen, or PROFINETIRT, requiring a strictly constant communication cycle of 1ms and processing 256 bytes of control commands per cycle. The protocol compatibility conversion system for heterogeneous intelligent systems provided in this invention is deployed between these two types of heterogeneous networks, acting as a hub for transparent transmission and timing shaping, converting the upstream non-deterministic data stream into precisely predictable timing data for the downstream. For scenarios where downstream devices have multiple fixed processing cycle requirements, multiple independent systems can be deployed, each corresponding to different fixed thresholds. Each system operates independently without interference, ensuring its own timing determinism.
[0053] In this application scenario, the fixed-length data block sending mechanism is as follows: a fixed-length data block is sent to the downstream receiving end that is sensitive to timing consistency. This fixed-length data block is used to drive the downstream receiving end to perform synchronization actions.
[0054] The synchronous action here refers not only to the receipt and confirmation of data, but more importantly, to the control calculations and physical execution within the downstream actuator based on fixed cycles. For example, in the motion control scenario of a multi-axis CNC machine tool, the downstream receiver needs to obtain precise position or speed command values within each fixed communication cycle to perform isochronous interpolation calculations. If the data blocks input to the actuator are of varying lengths or their arrival times drift, the buffer management logic within the actuator will generate uncertain processing delays, leading to phase deviations between axes, and consequently causing distortion of the machining trajectory or mechanical vibration.
[0055] This invention outputs strictly fixed-length data blocks, ensuring that the amount of data read and processed by the downstream receiving end is completely consistent in each cycle. This guarantees perfect alignment between data processing time and physical execution cycle, achieving true hardware-level synchronization.
[0056] To visually demonstrate the timing decoupling effect of this invention, the actual runtime timing waveform characteristics of the system in the aforementioned scenario are described below. On the upstream input side, the length of the input data packets received by the data receiving module varies randomly between 64 bytes and 1500 bytes, and the arrival time interval between adjacent packets fluctuates irregularly between 1 millisecond and 10 milliseconds. However, after boundary stripping by the stream reassembly module and continuous accumulation by the buffer storage module, the data block output by the fixed-length transmission control module is strictly reshaped into a fixed length of 256 bytes. More importantly, the transmission time interval of the output data block exhibits high constancy, always locked at 1 ms. This complete separation of output timing from input characteristics is based on the mathematical relationship derived in Embodiment 1 above: the transmission time interval of the fixed-length data block is determined only by the constant physical transmission rate of the downstream bus and a preset fixed threshold, decoupled from the length and arrival time interval of the input data packets.
[0057] As long as the system is in a steady-state operation, regardless of upstream network congestion or sudden changes in packet size, the output transmission clock remains locked at the physical constant E equal to W divided by Q². This deterministic timing output provides downstream timing-sensitive devices with a data supply benchmark as precise as a local crystal oscillator.
[0058] Based on this, it is necessary to explain why this invention employs a preset fixed threshold, a double-buffered ping-pong structure, and system-level constraints, rather than the dynamic adjustment strategy commonly used in existing technologies, in order to clarify the inventiveness and non-obviousness of this solution. In existing heterogeneous network gateway designs, to address traffic fluctuations and service instruction integrity issues, engineers typically employ dynamic threshold adjustment algorithms based on network load feedback, complex priority queue scheduling mechanisms, or service frame reassembly logic, attempting to optimize both transmission efficiency and instruction integrity through intelligent adaptation.
[0059] However, in steady-state heterogeneous agent interconnection scenarios with stringent requirements for timing consistency, this dynamic adaptation becomes the root cause of timing jitter. First, the dynamic adjustment algorithm and frame reassembly logic themselves consume computational resources and introduce processing latency, which fluctuates with changes in system load and data packet structure, directly undermining the determinism of output timing. Second, tuning dynamic parameters is extremely difficult; given the diverse characteristics of devices in heterogeneous networks, it is hard to find a set of universal adaptive parameters to guarantee stability under all operating conditions. Finally, while priority queues can guarantee bandwidth for high-priority data, they cannot eliminate the processing time dispersion caused by differences in the length of the data packets themselves.
[0060] In contrast, this invention takes a different approach. By introducing reasonable system-level constraints, including upstream and downstream rate matching and an upstream effective payload length that is an integer multiple of a threshold, it transforms the complex dynamic problem into a simple static problem. Combined with a double-buffered ping-pong structure and a control strategy triggered by a constant threshold and purely physical quantities, it completely resolves the contradiction between timing jitter and instruction integrity. Although system-level constraints limit the scope of application from the perspective of general data transmission, these constraints are naturally present and easily satisfied in many steady-state heterogeneous system scenarios such as industrial control and autonomous driving. By anchoring the triggering conditions to a physical constant that does not change with the environment—that is, a fixed number of bytes—and using a double-buffered structure to resolve timing contradictions at the hardware implementation level, this invention fundamentally cuts off the path of external disturbances propagating to internal timing, achieving absolute timing stability that complex dynamic algorithms in the prior art cannot achieve. This design approach—simplifying the problem through system constraints and then solving the core pain points with minimal hardware logic—is not something that those skilled in the art would easily conceive of when facing general data forwarding problems. Rather, it is a targeted solution for the specific technical problem of synchronous control of steady-state heterogeneous intelligent agents, possessing significant substantial characteristics and progress.
[0061] For buffer underload issues caused by temporary upstream power outages, this solution adheres to the core design principle of not introducing additional control variables and does not add any timeout triggering mechanisms within the system. Upstream power outages are system-level fault states and should be handled by upper-layer applications through methods such as heartbeat detection. This system is only responsible for providing absolutely deterministic timing shaping services when data input is available. This design trade-off results in extremely simple control logic and absolute timing determinism, fully meeting the operational requirements of steady-state industrial systems.
[0062] It should be understood that although this embodiment uses an industrial motion control scenario as an example for detailed description, the technical solution of the present invention is also applicable to other steady-state heterogeneous intelligent agent interconnection scenarios that require absolute timing alignment across rates and protocols, such as sensor fusion nodes in autonomous vehicle networks and distributed protection terminals in smart grids. As long as there is a contradiction between the nondeterminism of the upstream data source and the timing sensitivity of the downstream execution end in the application scenario, and the system constraints of upstream and downstream rate matching and upstream payload length regularization can be met, the protocol-compatible conversion mechanism based on double-buffered ping-pong structure, continuous flow recombination, and fixed threshold triggering provided by the present invention can play its unique timing shaping value.
[0063] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "include," "contain," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus 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 apparatus.
[0064] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.
[0065] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.
[0066] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.
Claims
1. A protocol compatibility conversion system for adapting to heterogeneous intelligent agent systems, characterized in that: include: The data receiving module is used to receive input data packets; The stream reconstruction module is used to convert the format of the input data packet and break the original packet boundary of the converted data, writing it as a continuous data stream into the buffer storage module. The buffer storage module adopts a double-buffered ping-pong structure to alternately store continuous data streams and accumulate the amount of data written in each buffer. The fixed-length transmission control module is used to monitor the cumulative amount of data in the currently active write buffer in the buffer storage module. When the cumulative amount of data reaches a preset fixed threshold, the write buffer is switched and the data in the full buffer is sent as a fixed-length data block. After the transmission is completed, the buffer is cleared and its corresponding counter is reset.
2. The adaptive intelligent agent heterogeneous system protocol compatibility conversion system according to claim 1, characterized in that, The specific steps for breaking the original packet boundaries of the converted data and writing it as a continuous data stream into the buffer storage module are as follows: stripping the frame header and frame tail of the input data packet and extracting only the payload data; seamlessly splicing the payload data into a continuous byte stream according to the receiving time sequence, and writing the continuous byte stream into the currently active write buffer in a manner that does not retain the original packet boundaries.
3. The protocol compatibility conversion system for adapting heterogeneous intelligent agent systems according to claim 2, characterized in that, The preset fixed threshold is set as follows: the preset fixed threshold is a constant value preset based on the minimum number of bytes received by the downstream receiver or the number of bytes required for a fixed processing cycle, and remains constant during system operation.
4. The adaptive intelligent agent heterogeneous system protocol compatibility conversion system according to claim 3, characterized in that, The specific steps for switching the write buffer and sending the data in the full buffer as a fixed-length data block are as follows: When the cumulative amount of data in the currently active write buffer reaches a preset fixed threshold, the write buffer is switched to another idle buffer within the same clock cycle, and the data transmission operation of the full buffer is triggered at the same time. After the transmission is completed, the buffer is cleared and its corresponding counter is reset.
5. The protocol compatibility conversion system for adapting heterogeneous intelligent agent systems according to claim 4, characterized in that, The decoupling relationship between the transmission time interval of the fixed-length data block and the input data packet is implemented as follows: When the system byte processing rate is not lower than the downstream bus transmission rate, the transmission time interval of the fixed-length data block is determined only by the constant physical transmission rate of the downstream bus and the preset fixed threshold, and is decoupled from the length and arrival time interval of the input data packet.
6. The protocol compatibility conversion system for adapting heterogeneous intelligent agent systems according to claim 5, characterized in that, The specific steps for monitoring the cumulative data volume in the buffer storage module are as follows: the cumulative number of bytes currently written to the buffer is read directly through the hardware register, without the need for software calculation or protocol analysis.
7. The protocol compatibility conversion system for adapting heterogeneous intelligent agent systems according to claim 6, characterized in that, The specific steps for format conversion of the input data packet are as follows: the input data packet is parsed according to the protocol, the target control command is extracted, and the target control command is converted into a payload in the target protocol format; the stream reassembly module is implemented with a full hardware pipeline architecture, and the processing delay of each byte is fixed at one clock cycle to ensure that the byte processing rate is a constant.
8. The protocol compatibility conversion system for adapting heterogeneous intelligent agent systems according to claim 7, characterized in that, The specific steps for sending the data in the buffer storage module as a fixed-length data block are as follows: the fixed-length data block is sent to the downstream receiving end that is sensitive to timing consistency, and the fixed-length data block is used to drive the downstream receiving end to perform synchronization actions.
9. A method for protocol compatibility conversion to adapt to heterogeneous intelligent agent systems, characterized in that, include: Receive input data packets; The input data packet is format-converted, and the converted data is broken from the original packet boundary and written as a continuous data stream into the currently active write buffer. Store the continuous data stream in the currently active write buffer and accumulate the amount of data written. Monitor the cumulative data volume in the currently active write buffer. When the cumulative data volume reaches a preset fixed threshold, switch the write buffer and send the data in the full buffer as a fixed-length data block. After the sending is completed, clear the buffer and reset its counter.
