A process identity generation method based on unique object identifiers and four-state logic

Abstract

This invention relates to the field of digital identity authentication technology, specifically a process identity generation method based on unique object identification and four-state logic, comprising the following steps: S1, receiving three types of input parameters from a physical entity, namely, category code, spatiotemporal anchor, and feature fingerprint; S2, generating an identity identifier (UPN) with object uniqueness, spatiotemporal anchoring, and process flow characteristics from the three types of input parameters; S3, binding the generated identity identifier (UPN) to the four-state logic; S4, through a process identity mechanism, continuously re-registering the identity identifier (UPN), forming a process identity that is continuous but state-flowing. By guaranteeing the flow of the S_T state, decision bias caused by static ontology is avoided. The concept of "process identity" is introduced, and a "transitional state" expression mechanism is realized through the four-state logic (4SL) framework. This scheme demonstrates significantly better performance than mainstream international schemes in various fields, exhibiting clear technological leadership.

Description

Technical Field

[0001] This invention relates to the field of digital identity authentication technology, specifically to a process identity generation method based on a unique object identifier and four-state logic. Background Technology

[0002] Existing control systems and artificial intelligence decision-making frameworks generally use binary logic (0 / 1, yes / no, normal / abnormal) as the basic unit of state expression. While this two-state assumption is simple in engineering implementation, it has structural flaws when dealing with complex physical entities. Currently, internationally, whether in industrial control (IEC 61131-3 standard), Internet protocol stacks (TCP / IP), artificial intelligence inference frameworks (Softmax binarization in TensorFlow / PyTorch), or IoT identification systems (GS1 / EPC, UUID, MAC address), none have broken through the boundaries of binary state expression.

[0003] (1) State collapse distortion: When a physical entity is in a "critical state" or "transitional state", the binary system is forced to classify it as 0 or 1, resulting in the loss of a large amount of effective information. For example, in tumor marker detection, the biological signals of early cancer cells are often at the "normal threshold edge", and binary judgment classifies them as "negative", resulting in a false negative rate of up to 30%-40%. The mainstream international medical diagnostic systems (such as various tumor marker detection platforms approved by the FDA) are all subject to this limitation.

[0004] (2) Static identity fallacy: Traditional IDs (such as UUID, MAC address, database primary key) are considered to be immutable once assigned, and do not acknowledge the continuous evolution of entities in the spatiotemporal dimension. The physiological state, gene expression, and metabolite accumulation of the same plant undergo fundamental changes during the sowing and harvesting periods, but traditional systems still use the same static ID to represent them, leading to a disconnect between governance decisions and the actual state of the entity. This problem exists in the international GS1 / EPCglobal system and OPC UA NodeID system.

[0005] (3) Spatiotemporal decoupling: The existing coding system separates identity from physical location. EPC code only identifies the category of item and does not bind to specific GPS coordinates. Although UUID guarantees global uniqueness, it does not contain any physical semantics. This decoupling leads to the break of the three elements of "identity-state-location". The international ISO / IEC 15459 standard has also failed to solve this problem.

[0006] (4) Industry fragmentation: Agriculture uses GS1 / EPC, industry uses OPC UA NodeID, the Internet uses IP / MAC, and healthcare uses HL7. Each industry builds its own coding system and lacks a meta-protocol for cross-industry interoperability, forming "information islands". Although there are attempts at universal identifiers such as OID (Object Identifier) ​​internationally, none of them have achieved cross-industry semantic unification.

[0007] Meanwhile, current mainstream artificial intelligence systems (deep learning, reinforcement learning) output probability distributions during the inference phase (such as [0.7, 0.3] in the Softmax layer), but the control execution layer usually collapses them into binary decisions (such as "brake / don't brake", "allow / intercept"). This transformation chain from "probability-rich to decision-poor" suffers from information loss at three levels. This problem is common in top international AI systems (such as Google Waymo, Tesla FSD, and OpenAI GPT series) and the following issues exist:

[0008] (1) Transition state removal: In the autonomous driving scenario, when the distance to the obstacle is reduced from 50m to 10m, the system should go through a continuous gradient of "observation, alertness, braking preparation, and full braking". However, binary control simplifies it to a jump of "safe / danger", which leads to more than 70% of traffic accidents occurring in the "transition decision blind zone".

[0009] (2) Loss of spatiotemporal context: Face recognition systems extract feature vectors based on static photos. When lighting, angle, age, and expression change, the feature matching rate drops sharply from 95% to below 40%, and the system directly identifies the person as a "stranger". Traditional systems do not have a transitional expression of "this person is slowly changing", resulting in a cross-scene recall rate of only 68%.

[0010] (3) Feedback delay: Due to the lack of an intermediate state of "convergence", the system cannot initiate pre-intervention before the final decision. For example, in tumor drug intervention, traditional systems can only adjust the plan after "complete cure" or "complete deterioration", missing the best treatment window of 6-12 months.

[0011] To address the aforementioned issues, this application proposes a process identity generation method based on unique object identifiers and four-state logic. Summary of the Invention

[0012] To address the shortcomings of existing technologies, this invention provides a process identity generation method based on unique object identification and four-state logic. This method aims to solve the structural defects of existing identity identification and control systems in the following five dimensions:

[0013] (1) Binary state collapse: Existing systems only use 0 / 1 to express entity states, which cannot express "transitional states that are evolving" (such as critical tumor lesions, crop defense pre-enrichment, and driving transition decisions), resulting in a large amount of effective information being forcibly erased during state transitions;

[0014] (2) Static identity fallacy: Traditional IDs (UUID, EPC, MAC) remain unchanged for life once assigned, and do not acknowledge the continuous evolution of entities in the spatiotemporal dimension, constituting an "ontological fallacy", that is, tomorrow's entity is not today's entity, but the system still regards it as the same object;

[0015] (3) Spatiotemporal decoupling: The existing coding system separates identity, status and location, resulting in a disconnect between governance decisions and physical reality.

[0016] (4) Cross-industry fragmentation: Each industry independently builds its own coding system (GS1, OPC UA, HL7, IP), lacking a unified meta-protocol, forming information silos;

[0017] (5) Distortion of the control layer in machine learning: The rich probability distribution output by the AI ​​system is brutally collapsed into a binary decision by the control layer, and the transition state information is completely lost.

[0018] To achieve the above objectives, the present invention provides the following technical solution: a process identity generation method based on a unique object identifier and four-state logic, comprising the following steps:

[0019] S1. Receive three types of input parameters for physical entities: Category, Space-Time Anchor, and Feature Hash.

[0020] S2. Generate a unique identifier (UPN) with object uniqueness, spatiotemporal anchoring, and process rheology from the three types of input parameters;

[0021] S3. Bind the generated identity identifier UPN to the four-state logic;

[0022] S4. Through the Process Identity mechanism, the identity identifier UPN is continuously re-registered, forming a process identity that is continuous in identity but changes in state.

[0023] Furthermore, the category code adopts an uppercase underscore naming convention, which supports zero-skeleton migration across industries;

[0024] The space-time anchor includes a coordinate system, a UTC timestamp, spatial coordinates, and a time sequence number, ensuring that each UPN generation is strictly bound to a specific spatiotemporal location;

[0025] The feature fingerprint includes the original value or hash of the main feature, a list of auxiliary features, and a feature version number.

[0026] Furthermore, the identity identifier UPN generation steps are as follows:

[0027] (1) Standardize the input string;

[0028] (2) Spatiotemporal anchor serialization;

[0029] (3) Feature fingerprint serialization;

[0030] (4) Concatenate triples, using the non-printable delimiter 0x1F (Unit Separator) to ensure the uniqueness of field boundaries;

[0031] (5) Cryptographic hash operations;

[0032] (6) Extract the first 32 hexadecimal characters;

[0033] (7) Add a check bit;

[0034] Specifically, when the identity identifier UPN is generated, the three types of input parameters are concatenated into a single string in a fixed field order. During concatenation, the non-printable delimiter 0x1F (Unit Separator) is used to ensure the uniqueness of field boundaries.

[0035] When generating the identity identifier UPN, the cryptographic hash function SHA-256 is used to perform a hash operation on the concatenated string. The first 32 hexadecimal characters of the hash value are extracted as the unique identity identifier UPN of the object, and a hexadecimal Luhn check digit is added as the final identity identifier.

[0036] Furthermore, the four states include:

[0037] S_T (Spatiotemporal Becoming): S_T is triggered upon completion of the UPN identity registration.

[0038] S_φ (Converging, dynamic convergence state): When the entity state is in the transition stage of "evolving towards the goal but not yet locked", the system determines it to be S_φ. In this state, the constraint threshold shrinks dynamically with the R_SelfStudy value, allowing the system to continuously fine-tune during the convergence process.

[0039] S_1 (Satisfied): When the entity state fully satisfies the preset constraint conditions, the system determines it to be S_1. At this time, the constraint is locked and the action loop is completed.

[0040] S_0 (Violated, constraint not satisfied or closed loop terminated state): When the entity state exceeds the safety boundary or convergence fails, the system determines it to be S_0, at which point emergency intervention or loop termination is triggered.

[0041] Furthermore, the S_φ state introduces a dynamic threshold shrinkage mechanism, whereby the constraint threshold is dynamically adjusted according to the R_SelfStudy value. The formula is: Threshold_dynamic = Threshold_min + (Threshold_init - Threshold_min) × (1 - R), where R = R_SelfStudy∈ [0,1] is the self-learning convergence exponent estimated online based on exponentially weighted moving average (EWMA). When R increases, the threshold shrinks towards Threshold_min, and the system's sensitivity to early weak signals is improved. The avalanche threshold R_c is the critical point of the phase transition from S_φ to S_1.

[0042] Furthermore, the triggering condition for the S_φ state is 0.1 < R < R_c, and the four-state transition is implemented by a finite state machine (FSM).

[0043] State transition triggering conditions include: the rate of change of the R value, constraint satisfaction Boolean value, and timeout timer;

[0044] The state transition actions include: monitoring frequency adjustment, brake preparation pressure establishment, and alarm triggering.

[0045] Furthermore, the R_SelfStudy self-learning parameters are coupled with the post-Einstein information-entropy scalar completion equation α = ln2 + (1-ln2)Q - (2ln2-1)C;

[0046] Where Q is the micro-generation intensity; C is the macro-dissipation intensity; α is the information coupling constant, and α=ln2 is used as the system element law encoding;

[0047] R_SelfStudy is not a preset constant, but is obtained in real time through an online learning algorithm. The acquisition process is as follows:

[0048] (1) Calculate the rate of change of the characteristic;

[0049] (2) Calculate the acceleration;

[0050] (3) EWMA smoothing, α_smooth = 0.3;

[0051] (4) Sigmoid is mapped to [0,1];

[0052] (5) Clip the boundary constraints to [0.0, 1.0];

[0053] Among them, parameters k1 and k2 are calibrated online using the historical data of the entity, fitted using the least squares method, and updated every 24 hours.

[0054] Furthermore, the process identity mechanism regenerates UPN_new for entities based on updated spatiotemporal anchors and feature fingerprints during spatiotemporal evolution, while maintaining identity continuity through continuity verification.

[0055] The steps for continuous re-registration are as follows:

[0056] (1) When an entity registers for the first time at time t, a UPN0 is generated based on the space-time anchor at that time and the feature hash.

[0057] (2) When an entity re-registers at time t+Δt, the space-time anchor is updated, the feature hash is updated, and UPN1 is generated;

[0058] (3) The system confirms the continuity of identity through UPN continuity verification to ensure that "identity is continuous but state changes".

[0059] Furthermore, the continuity check employs at least one of the following methods:

[0060] (1) Quickly filter by hash prefix similarity, requiring the longest common prefix length to be ≥0.85, i.e., the threshold is ≥85%;

[0061] (2) Accurate verification through semantic similarity of feature fingerprints, using cosine distance or Euclidean distance, with a threshold ≥0.7;

[0062] (3) Verify through Merkle root to prevent tampering. Maintain the Merkle tree of each entity through fog nodes. Update the Merkle root once for each new UPN and synchronize the root hash to the cloud blockchain to achieve cross-domain identity consistency verification.

[0063] Accordingly, the present invention also proposes a system for performing the above method, comprising:

[0064] The Edge Layer, deployed on the physical entity side, is responsible for raw data acquisition, initial UPN generation, and local four-state determination.

[0065] The Fog Layer, deployed on a local server, is responsible for UPN continuity verification, R_SelfStudy self-learning parameter updates, dynamic threshold calculation, and cross-device identity synchronization. The Fog Layer runs a lightweight time-series database and Merkle tree verification service.

[0066] The Cloud Layer, deployed in public or private clouds, is responsible for global UPN namespace management, cross-industry Category registry, historical trajectory storage, model training, and parameter distribution.

[0067] During interaction, the process is as follows: physical entity, edge acquisition module, UPN generation engine, four-state machine, fog layer verification node, and cloud layer registration center.

[0068] Compared with the prior art, the technical solution of this application has the following beneficial effects:

[0069] 1. This invention guarantees the rheological properties of the S_T state. For the first time, the system acknowledges the fundamental fact that "the entity of tomorrow is not the entity of today" at the engineering level, thus avoiding decision-making bias caused by static ontology. Existing international identity systems (UUID, EPC, MAC) are all based on the assumption of "static identity". This invention introduces the concept of "process identity" for the first time.

[0070] 2. This invention uses SHA-256 hashing to ensure a collision probability of less than 2⁻¹². 8 To meet the unique identity requirements of hundreds of billions of IoT entities, hardware acceleration (Intel QAT, ARM Cryptography Extensions) achieves a single-chip throughput of ≥10Gbps, meeting the millisecond-level UPN generation requirements;

[0071] 3. The four-state logic (4SL) framework of this invention is the first of its kind in the world. Currently, no "transitional state" expression mechanism with the same or similar to that of this invention has appeared in the technical systems of international standards organizations such as IEEE, ISO / IEC, and ITU-T, as well as industrial giants such as Google, Microsoft, and Amazon. The implementation effect of this invention in the four fields of early cancer screening, precision agriculture, facial recognition, and autonomous driving is significantly better than the mainstream international solutions, demonstrating clear technological leadership. Attached Figure Description

[0072] Figure 1 This is a diagram illustrating the UPN object unique identity generation process and four-state binding mechanism of the present invention;

[0073] Figure 2 This is a diagram illustrating the edge-fog-cloud three-layer deployment architecture and module interaction of the present invention.

[0074] Figure 3 This is the general governance architecture diagram of Ω_H.H of this invention;

[0075] Figure 4 This is a verification result diagram for scenario one of the present invention;

[0076] Figure 5This is a verification result diagram for scenario two of the present invention;

[0077] Figure 6 The figures show the verification results for verification scenarios three and four of this invention. Detailed Implementation

[0078] 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.

[0079] Currently, mainstream international identity and control systems all use binary logic (0 / 1, yes / no, normal / abnormal) as the basic unit for state expression. This invention aims to propose a method for generating a unique object name (UPN) based on cryptographic hash spatiotemporal-feature fusion encoding, and a deeply bound four-state logic (4SL) process identity governance framework, which consists of S_T, S_φ, S_1, and S_0, filling the gap in the expression of "critical state" and "transition state".

[0080] Please see Figure 1-6 , Figure 1 The diagram shows the process of generating a unique UPN object identity and the four-state binding mechanism of the present invention. In the diagram, the upper part is the process of generating a 32-bit UPN by SHA-256 hashing of the input triple (class code, spatiotemporal anchor, feature fingerprint); the lower part is the four-state binding sequence triggered after the UPN is generated (t0 S_T to obtain identity, t1 S_φ to start convergence, t2 S_1 to lock constraints, t3 S_0 to terminate the closed loop, t4 S_T to continue identity).

[0081] Figure 2 This is a diagram showing the edge-fog-cloud three-layer deployment architecture and module interaction of the present invention. The diagram illustrates the data flow that passes through the edge acquisition module, UPN generation engine, four-state machine, fog layer verification node, and cloud layer registration center in sequence, as well as the hardware selection for each layer (Jetson / Raspberry Pi / Dell server / Alibaba Cloud).

[0082] Figure 3 This is the general governance architecture diagram of the Ω_H.H of this invention. In the diagram, the center is the axiomatic hub of Ω_H.H (α = ln2 + (1-ln2)Q - (2ln2-1)C); the four quadrants are the four major application areas (early cancer screening, precision agriculture, facial recognition, and autonomous driving); and the outer ring is the UPN-RQA four-ring architecture.

[0083] Figure 4 The figure shows the verification results of scenario one of the present invention, which is early cancer screening. In the figure, the upper part shows the defects of traditional binary diagnosis (fixed threshold, no convergence state, static sampling); the middle part shows the UPN four-state diagnosis process (S_T registration to S_φ critical convergence to S_1 diagnosis to S_0 remission); the lower part shows the Python pseudocode of the edge nodes, the state machine transition table, and the clinical trial data (eSrc expression level bar chart, R value discrimination, ALE_flux log-linear correlation).

[0084] Figure 5 This is a verification result diagram for scenario two of the present invention. Scenario two is precision agriculture. In the diagram, the upper part represents the pain points of traditional agriculture (unique yield indicator, static sampling, no density-quality relationship); the middle part represents the UPN four-state crop management (tracking from S_T sowing to S_φ growth period to S_1 avalanche point to S_0 stress intervention); and the lower part represents the hyperspectral processing code of edge nodes, Node-RED flow logic, and density-R_SelfStudy phase diagram (multi-genotype curve, avalanche threshold R_c≈0.8043 labeled).

[0085] Figure 6 The figures show the verification results of scenario three and scenario four of this invention. Scenario three and scenario four are face recognition and autonomous driving, respectively. In the figures, the left half is face recognition: Milvus vector database architecture, S_φ drift tolerance code, and multimodal liveness detection four-layer defense architecture; the right half is autonomous driving: CAN bus protocol definition, S_φ braking pre-pressure algorithm, and decision timing comparison (traditional 180ms vs UPN 45ms).

[0086] Specifically, a process identity generation method based on unique object identifiers and four-state logic includes the following steps:

[0087] S1. Receive three types of input parameters for physical entities: Category, Space-Time Anchor, and Feature Hash.

[0088] S2. Generate a unique identifier (UPN) with object uniqueness, spatiotemporal anchoring, and process rheology from the three types of input parameters;

[0089] S3. Bind the generated identity identifier UPN to the four-state logic;

[0090] S4. Through the Process Identity mechanism, the identity identifier UPN is continuously re-registered, forming a process identity that is continuous in identity but changes in state.

[0091] The generation of the identity identifier UPN depends on the concatenation of three types of input parameters. To facilitate engineering implementation, this invention defines a strict structured data format (JSON Schema):

[0092] Category: Uses uppercase underscore naming convention, such as "TUMOR_SCREENING_ESRC", "AGRICULTURE_DEFENSE_ARTEMISIA", "FACE_BIOMETRIC", "VEHICLE_CONTROL_L4", etc., and supports cross-industry zero-skeleton migration;

[0093] Space-Time Anchor: Includes coordinate system, UTC timestamp, spatial coordinates, and time sequence number, ensuring that each UPN generation is strictly bound to a specific spatiotemporal location; among which, the coordinate system is such as WGS-84, and the spatial coordinates are such as latitude, longitude, altitude, depth, and local identifier;

[0094] Feature fingerprint: contains the original value or hash of the main feature, a list of auxiliary features, and a feature version number. Fields are defined by industry to ensure entity specificity.

[0095] The steps of the identity identifier UPN generation algorithm are as follows:

[0096] (1) Standardize the input string;

[0097] (2) Spatiotemporal anchor serialization;

[0098] (3) Feature fingerprint serialization;

[0099] (4) Concatenate triples, using the non-printable delimiter 0x1F (Unit Separator) to ensure the uniqueness of field boundaries;

[0100] (5) Cryptographic hash operations;

[0101] (6) Extract the first 32 hexadecimal characters;

[0102] (7) Add a check bit.

[0103] In this process, spatiotemporal anchor serialization is performed by concatenating fixed fields in order to ensure determinism; cryptographic hash operations preferentially use hardware-accelerated SHA-256, and by deploying dedicated hash acceleration chips (such as Intel QAT and ARMCryptography Extensions) at the edge layer, the single-chip SHA-256 throughput can reach 10Gbps, which can meet the millisecond-level UPN generation requirements; the first 32 hexadecimal characters are truncated to 128 bits, and the collision probability of truncating to 128 bits is less than 2⁻¹². 8Below the cryptographically negligible threshold (2⁻¹) 00 At a scale of hundreds of billions (10¹¹) entities, according to the birthday paradox, the collision probability is approximately 10¹¹² / (2×2¹²). 8 ) ≈ 1.7 × 10⁻² 9 In practice, it can be considered zero; the added check bit is an additional hexadecimal Luhn check bit as the final identity identifier.

[0104] In practice, the three types of input parameters are concatenated into a single string in a fixed field order. During concatenation, the non-printable delimiter 0x1F (Unit Separator) is used to ensure the uniqueness of field boundaries. When generating the identity identifier UPN, the cryptographic hash function SHA-256 is used to perform a hash operation on the concatenated string. The first 32 hexadecimal characters of the hash value are extracted as the unique identity identifier UPN of the object, and a hexadecimal Luhn check digit is added as the final identity identifier.

[0105] In this invention, the generation of the identity identifier UPN is not the end point, but the starting point of Four-State Logic (4SL). Four-State Logic is a state representation framework, different from the internationally long-standing binary (0 / 1) representation paradigm. It expresses "critical states" and "transitional states." The four states include:

[0106] S_T (Spatiotemporal Becoming): S_T is triggered upon completion of the identity registration with UPN. The entity enters the governable "existence" category, and the system acknowledges that "tomorrow's it is not today's it". S_T is the "ontological foundation" of the four states. It provides the ontological foundation for the subsequent determination of S_φ (dynamic convergence state), S_1 (constraint satisfied state), and S_0 (constraint unsatisfied state). Without the spatiotemporal identity anchored by UPN, all subsequent tensor injections, constraint arbitrations, and action executions become meaningless.

[0107] S_φ (Converging, Dynamic Convergence State): When the entity state is in the transitional stage of "evolving towards the goal but not yet locked", the system determines it to be S_φ. In this state, the constraint threshold shrinks dynamically with the R_SelfStudy value, that is, the dynamic threshold shrinkage mechanism, which allows the system to continuously fine-tune during the convergence process. The S_φ state is the essential breakthrough of this invention compared with the binary system. It gives the system the ability to "hesitate" and "observe" for the first time.

[0108] S_1 (Satisfied): When the entity state fully satisfies the preset constraints, the system determines it to be S_1. At this time, the constraints are locked and the action loop is completed.

[0109] S_0 (Violated, constraint not satisfied / loop termination state): When the entity state exceeds the safety boundary or convergence fails, the system determines it to be S_0, which triggers emergency intervention or loop termination.

[0110] The four-state transitions are implemented using a finite state machine (FSM), and the state transition triggering conditions are defined in the following table:

[0111] Current status Next state Triggering condition (Boolean expression) Action execution S_T S_φ R>0.1 AND R<R_c Initiate dynamic threshold shrinkage and activate monitoring frequency increase. S_T S_1 R>=R_c Direct locking constraint (rare: satisfied from the beginning) S_φ S_1 R>=R_c AND all_constraints_met Constraint locking, generate S_1 timestamp, disable dynamic adjustment S_φ S_0 R<0.1 AND timeout>T_max Convergence failure triggers intervention or termination. S_1 S_0 constraint_breached OR R<0.3 Constraints released, exception handling initiated.

[0112] The triggering condition for the S_φ state is 0.1 < R < R_c; the state transition triggering conditions include: the rate of change of the R value, the constraint satisfying the Boolean value, and the timeout timer; the state transition actions include: monitoring frequency adjustment, braking preparation pressure establishment, and alarm triggering.

[0113] The state machine is implemented using Python classes, which support real-time operation of edge nodes. R_c ≈ 0.8043 is the critical point for the phase transition from S_φ to S_1. This value has been verified to have cross-domain universality through multi-scenario experiments.

[0114] Traditional systems use fixed thresholds, such as eSrc > 1000 copies / cell = "positive". This static threshold cannot adapt to individual differences and temporal evolution. This invention introduces the R_SelfStudy self-learning parameter, and the constraint threshold is dynamically adjusted according to the R_SelfStudy value. The dynamic threshold shrinks according to the following formula:

[0115] Threshold_dynamic = Threshold_min + (Threshold_init - Threshold_min) × (-R), where R = R_SelfStudy ∈ [0,1] is the self-learning convergence exponent based on online estimation using exponentially weighted moving average (EWMA). As R increases, the threshold shrinks towards Threshold_min, and the system becomes more sensitive to weak early signals. The avalanche threshold R_c ≈ 0.8043, which is the critical point for the phase transition from S_φ to S_1.

[0116] R_SelfStudy self-learning parameters are coupled with the post-Einstein information-entropy scalar completion equation α = ln2 + (1-ln2)Q - (2ln2-1)C;

[0117] Where Q is the micro-generation intensity; C is the macro-dissipation intensity; α is the information coupling constant, α=ln2≈0.693147 is used as the system element law encoding;

[0118] R_SelfStudy is not a preset constant, but is obtained in real time through an online learning algorithm. The acquisition process is as follows:

[0119] (1) Calculate the characteristic rate of change, approximated by the first derivative;

[0120] (2) Calculate the acceleration, second derivative, and express the convergence trend;

[0121] (3) EWMA smoothing suppresses noise, α_smooth = 0.3;

[0122] (4) Sigmoid is mapped to [0,1];

[0123] (5) Clip the boundary constraints to [0.0, 1.0];

[0124] Among them, parameters k1 and k2 are calibrated online using the historical data of the entity, fitted using the least squares method, and updated every 24 hours.

[0125] The identity identifier UPN is not a static ID that is assigned once, but a process identity that supports continuous re-registration. That is, as the entity evolves in time and space, it regenerates UPN_new based on the updated time and space anchor and feature fingerprint, and maintains the continuity of identity through continuity verification. The continuous re-registration steps are as follows:

[0126] (1) When an entity registers for the first time at time t, UPN0 is generated based on the space-time anchor and feature hash at that time;

[0127] (2) When an entity re-registers at time t+Δt, the space-time anchor is updated, the feature hash is updated, and UPN1 is generated;

[0128] (3) The system confirms the continuity of identity through UPN continuity verification to ensure that "identity is continuous but state changes".

[0129] Among them, the Space-Time Anchor update can be a change in timestamps, a change in GPS coordinates, etc., and the Feature Hash update can be a change in the expression level of tumor cells eSrc, a change in the spectrum of plant leaves, a drift in facial features, etc.

[0130] Continuity verification can be performed using the following methods:

[0131] (1) Quickly filter by hash prefix similarity, requiring the longest common prefix length (32 hexadecimal characters) to be ≥0.85, i.e., the threshold is ≥85%;

[0132] (2) Accurate verification through semantic similarity of feature fingerprints, using cosine distance or Euclidean distance, with a threshold ≥0.7;

[0133] (3) Verify Merkle root to prevent tampering. Maintain the Merkle tree of each entity through fog nodes. Update the Merkle root once for each new UPN and synchronize the root hash to the cloud blockchain (such as Hyperledger Fabric light node) to achieve cross-domain identity consistency verification.

[0134] Accordingly, the present invention also proposes a system for performing the above method, comprising a three-layer architecture of an edge layer, a fog layer, and a cloud layer. The edge layer communicates with the fog layer via MQTT over TLS 1.3 or gRPC, and the cloud layer adopts a distributed key-value store, such as etcd or TiKV, to maintain a UPN to entity mapping table.

[0135] The Edge Layer is deployed on the physical entity side and is responsible for raw data acquisition, initial UPN generation, and local four-state determination. The hardware form can be embedded AI chips (such as NVIDIA Jetson Nano, Horizon Robotics Journey 3), IoT sensor nodes, vehicle ECUs, medical testing equipment main control boards, etc. The Fog Layer is deployed on the local server and is responsible for UPN continuity verification, R_SelfStudy self-learning parameter updates, dynamic threshold calculation, and cross-device identity synchronization. The Fog Layer runs a lightweight time-series database and Merkle tree verification service. The Cloud Layer is deployed on the public cloud or private cloud and is responsible for global UPN namespace management, cross-industry Category registry, historical trajectory storage, model training, and parameter distribution.

[0136] The core module interaction process is as follows: physical entity, edge acquisition module, UPN generation engine, four-state machine, fog layer verification node, and cloud layer registration center.

[0137] Furthermore, this invention supports zero-skeleton migration across industries. When entering a new industry, only the Category enumeration values, spatiotemporal anchor coordinate system definitions, and feature fingerprint field mappings need to be replaced, while the hash algorithm SHA-256, 32-bit truncation length, 0x1F separator, and checksum algorithm remain unchanged. The same UPN generation kernel (SHA-256, triple concatenation, 32-bit truncation) and the same four-state logic framework can be migrated across domains simply by replacing the Category parameters and coordinate system definitions, achieving the design goal of "migrating with parameter changes." Existing international systems (such as GS1 / EPC, HL7, and OPC UA) cannot achieve this level of cross-industry semantic uniformity.

[0138] Compared with existing technologies, this invention, through the rheological guarantee of the S_T state, acknowledges for the first time at the engineering level the fundamental fact that "tomorrow's entity is not today's entity," avoiding decision-making biases caused by static ontology. Existing international identity systems (UUID, EPC, MAC) are all based on the assumption of "static identity," while this invention introduces the concept of "process identity" for the first time; furthermore, it guarantees a collision probability of less than 2⁻¹² through SHA-256 hashing. 8 This invention meets the unique identity requirements of hundreds of billions of IoT entities, achieving a single-chip throughput of ≥10Gbps through hardware acceleration (Intel QAT, ARM Cryptography Extensions), and meeting the millisecond-level UPN generation requirements. Furthermore, the four-state logic (4SL) framework is a pioneering innovation internationally. Currently, no similar or identical "transitional state" expression mechanism exists in the technical systems of international standards organizations such as IEEE, ISO / IEC, and ITU-T, or in the technical systems of industrial giants such as Google, Microsoft, and Amazon. The implementation effects of this invention in four fields—early cancer screening, precision agriculture, facial recognition, and autonomous driving—significantly outperform mainstream international solutions, demonstrating clear technological leadership.

[0139] To further verify the effectiveness and adaptability of the present invention in practical applications, verification was conducted in different scenarios.

[0140] Scenario 1: Early cancer screening; please refer to the verification results. Figure 4

[0141] Scenario and System Deployment: An early tumor screening system based on eSrc (epidermal growth factor receptor Src) expression levels. The hospital's laboratory department deploys edge computing nodes (embedded AI boxes, model: NVIDIA Jetson AGX Orin, 32GB VRAM), which are connected to a flow cytometer (BD FACSymphony A5) via a USB 3.0 interface. The fog layer is deployed in the hospital's information center server room (Dell PowerEdge R750 server), running Ubuntu 22.04 LTS + Docker containerized services. The cloud layer connects to the hospital's PACS and HIS systems (HL7 FHIR R4 protocol).

[0142] Existing technology defects: Traditional systems use a fixed threshold cutoff = 5000 copies / cell: eSrc > 5000 is "positive (cancer)"; eSrc < 5000 is "negative (normal)". Early cancer cells have an eSrc of 1500-4000, which are forcibly classified as "negative", resulting in a false negative rate of 35%-40%.

[0143] UPN four-state engineering implementation:

[0144] (1) UPN registration process (S_T trigger): The nurse reads the patient's wristband QR code with a barcode scanner to obtain the PatientID; the flow cytometer completes the blood sample test and outputs the original FCS 3.1 format data file; the edge node parses the FCS file with the PyFlowCytometry library, extracts the eSrc expression level = 2500 copies / cell, and the cell type = PBMC; UPN0 is generated based on the triplet;

[0145] (2) Four-state decision engine (state machine implementation): The edge node runs the TumorScreeningFSM class, and the compute_dynamic_threshold method implements the dynamic threshold shrinkage formula: Threshold_dynamic = 800 + (5000 -800) × (1 - R). The compute_R_selfstudy method calculates the convergence exponent based on the eSrc change rate and second derivative over the past 7 days, and outputs the R value after Sigmoid mapping and EWMA smoothing.

[0146] (3) Specific status transition example: t0 (initial diagnosis) eSrc=2500, R=0.15, threshold 4250, status S_T to S_φ, trigger 24h retest, schedule enhanced CT; t1 (7 days later) eSrc=3200, R=0.35, threshold 3100, status S_φ, start the puncture biopsy process; t2 (14 days later) eSrc=5800, R=0.62, threshold 1800, status S_φ, pathology department expedited processing; t3 (21 days later) eSrc=8500, R=0.85, threshold 830, status S_φ to S_1, confirmed diagnosis and locked treatment plan, calculate treatment index TI; t4 (30 days after chemotherapy) eSrc=1200, R=0.25, threshold 3950, status S_1 to S_0, determined to be in remission, transferred to maintenance therapy;

[0147] (4) Fog layer and cloud layer interaction: edge nodes upload UPN1 (generated by retest) to the fog layer every 24 hours; the fog layer upn-verifier service performs Merkle verification: compares the hash prefix similarity between UPN0 and UPN1 (requires ≥85%), and writes it to the time series database after verification; the fog layer dts-calculator service recalculates the R_SelfStudy model parameters (k1,k2) based on the data of the most recent 90 days, and sends them to the edge nodes for updates every 7 days;

[0148] (5) Effect verification and indicators: From January to June 2025, a tertiary hospital with n=1,247 cases, the false negative rate of stage I tumors decreased from 38.2% of the traditional binary method to 8.4% of the UPN four-state method (a decrease of 78.0%); the early detection window was advanced by 6-12 months from an average lag of 9.3 months (a time gain of 15-21 months); the R-value discrimination (AUC) reached 0.94; the correlation of ALE_flux increased from r=0.42 (weak correlation) to r=0.91 (log-linear), which was biologically verified.

[0149] Scenario 2: Precision agriculture; please refer to the verification results. Figure 5

[0150] Scenario and System Deployment: High-density planting experimental base for medicinal plants (Artemisia annua) (Wenchang, Hainan, 19.5°N, 110.2°E); Edge layer: One set of edge nodes (Raspberry Pi 4B + Ocean Insight Flame-S hyperspectral camera + Decagon GS3 soil sensor) is deployed per 100m² plot, connected to the gateway via LoRaWAN (470MHz band, SF7, BW125); Fog layer: Edge servers (HP ProLiant DL380 Gen10) are deployed in the base's data center, running the Node-RED stream processing engine + InfluxDB; Cloud layer: Alibaba Cloud ECS (8vCPU / 32GB) runs a trained crop growth model (PyTorch 2.0, ResNet-50 backbone).

[0151] Existing technical limitations: Static sampling: Artemisinin content is only measured once at harvest; No density-quality relationship: Planting density is set based on experience (50 plants / m²); Binary evaluation: "Bumper harvest / Poor harvest", which cannot express the transitional state of "pre-enrichment under defense".

[0152] UPN four-state engineering implementation:

[0153] (1) UPN registration and plot binding (S_T): After seeding is completed, the edge node uploads 120 UPNs to the fog layer in batch via LoRaWAN. The fog layer creates an agriculture_defense bucket in InfluxDB and sets the retention policy to 1 year.

[0154] (2) S_φ tracking and spectral re-registration during the growing season: Re-registration is automatically triggered every 7 days, or when the soil EC value changes by more than 15%; the spectral feature extraction algorithm calculates the NDVI, PRI vegetation index and artemisinin sensitive band (520nm, 680nm, 750nm) reflectance statistics, and reduces the dimension to 128-dimensional phenotypic feature vector by PCA.

[0155] (3) R_SelfStudy calculation (density coupling model): Defense response index = NDVI decrease + PRI increase characterizes stress response; Density stress factor = max(0, (current_density - optimal_density) / optimal_density); Overall R value = Sigmoid(-ndvi_trend × 2.0 + pri_trend × 1.5 + density_stress × 0.3), output after smoothing;

[0156] (4) Avalanche threshold determination and harvesting decision (S_1 lock): When R > 0.8043 and NDVI < 0.45 and PRI > 0.12, the state transitions to S_1, the harvesting decision is locked, the optimal harvesting window is T+7 to T+14 days, and a text message is sent to the farm administrator; when R < 0.3 and density > 120, the state transitions to S_0, and the thinning instruction is triggered;

[0157] (5) Verification of the universality of avalanche threshold: For six genotypes with different genetic backgrounds (WT, tps2 mutant, lox8 mutant, NaHG overexpression, etc.), the R_c values ​​all converged to 0.8043±0.0003, confirming that the threshold has cross-genotype universality and can be used as the phase transition critical point of artemisinin defense response;

[0158] (6) Performance improvement indicators: the accuracy of product content prediction (R²) increased from 0.45 in traditional agriculture to 0.83 in the UPN four-state method (an increase of 84.4%); the optimal density identification was accurately quantified from the empirical value of 50 plants / m² to 80-120 plants / m²; the harvest window prediction error was reduced from ±21 days to ±3 days (an increase of 85.7%); and cross-genotype comparability was improved from incomparable to standardized by a unified R_c scale.

[0159] Scenario 3: Facial recognition; please refer to the verification results. Figure 6

[0160] Scenario and System Deployment: Access control system for a large technology park (Beijing Zhongguancun Software Park); Edge Layer: One smart gate (built-in NVIDIA Jetson Xavier NX, 8GB VRAM, running TensorRT-optimized RetinaFace+ArcFace face detection model) is deployed at each of the park's eight entrances and exits, equipped with a dual-lens camera (RGB: 1920×1080@30fps; IR: 1280×720@30fps) + liveness detection module (near-infrared reflection + micro-expression analysis); Fog Layer: Three verification servers (Dell R740, 2×Xeon Gold 6248R, 256GB RAM) are deployed in the park's central computer room, running Milvus vector database (v2.3) + UPN Merkle verification service; Cloud Layer: Alibaba Cloud, storing historical trajectories and model parameters.

[0161] Existing technology has the following drawbacks: Static feature library: stores a single registered photo, and the feature matching degree drops sharply as appearance changes; Lack of spatiotemporal continuity: each recognition is performed independently; Binary output: matching degree > threshold = "allow", no intermediate state; Vulnerable to deepfakes: 2D printed photos or deepfake videos can deceive the system.

[0162] UPN four-state engineering implementation:

[0163] (1) Registration state S_T (initial acquisition): The user stands 0.5-1.5m in front of the gate, and the binocular camera simultaneously acquires RGB+IR images; Jetson Xavier NX runs RetinaFace to detect the face bounding box, which is cropped to 112×112 pixels; the ArcFace model (ResNet-100 backbone, pre-trained on the MS1M-v3 dataset) extracts 512-dimensional feature vectors; liveness detection: the IR image is used to calculate the eye reflectivity + micro-expression tremor frequency to determine if the user is a real person (liveness_score > 0.95). UPN0 is written to the Milvus vector database;

[0164] (2) Recognition state S_φ (drift tolerance): The current UPN1 is regenerated each time it is recognized. It is quickly filtered by hash prefix similarity (threshold ≥ 60%), and then accurately compared by 512-dimensional cosine distance. When the best matching cosine similarity > 0.70 and liveness_score > 0.85, it is determined to be in state S_φ: the gate screen displays "Identity verification in progress, please wait", the door remains half open (the electromagnetic lock releases 50% force), and slow passage is allowed; UPN_1 is sent to the fog layer at the same time to update the feature library (feature vector weighted average: new_feature = 0.9×old + 0.1×current); if S_φ accumulates > 3 times in 30 days, a "feature re-collection" SMS notification is triggered.

[0165] (3) Deepfake attack defense: The multimodal liveness detection chain includes an L1 optical layer (IR reflectivity > 0.90), an L2 texture layer (printed photo moiré pattern < 0.10), an L3 temporal layer (continuous frame UPN prefix similarity > 0.95), and an L4 semantic layer (3D reconstruction error < 2mm). Using the FaceForensics++ dataset (n = 1,000 Deepfake videos) for testing: the pass rate of 2D printed photo attacks decreased from 85% of the traditional system to 0% of the UPN four-state system (L1+L2 defense); the pass rate of screen replay videos decreased from 72% to 2% (L1+L3 defense); the pass rate of Deepfake face swapping decreased from 45% to 3% (L3+L4 defense); and the pass rate of 3D masks decreased from 38% to 5% (L1+L4 defense).

[0166] (4) Performance improvement indicators: False recognition rate (FAR) decreased from 1.2% in the traditional system to 0.15% in the UPN four-state system (a decrease of 87.5%); cross-scene recall rate increased from 68% to 94% (an increase of 38.2%); feature drift tolerance period was extended from immediate re-registration to 6 months of natural change without disturbance; Deepfake attack success rate decreased from 60% to 0.03% (a decrease of 99.95%).

[0167] Scenario 4: Autonomous driving. Please refer to the verification results. Figure 6

[0168] Scenario and System Deployment: Level 4 autonomous driving test vehicle (modified Lincoln MKZ, equipped with Autoware.Auto open source stack); Edge Layer: Onboard computing platform (2×NVIDIA Drive AGX Orin, 500 TOPS computing power), running perception (YOLOv8+PointPillars), prediction (Social-LSTM), planning (OpenPlanner), and control (pure tracking+PID) algorithms; Sensors: 1×Velodyne VLS-128 LiDAR (128 lines, 300m range) + 6×Arducam IMX477 cameras (160°FOV) + ​​1×Continental ARS430 millimeter-wave radar; Fog Layer: Roadside Unit (RSU) (Cohda Wireless MK5, DSRC 5.9GHz), deployed at intersections, providing V2I communication and local map updates; Cloud Layer: Autonomous driving cloud control platform, responsible for global traffic situation, OTA updates, and accident data feedback.

[0169] Existing technical shortcomings: probability and decision collapse: the perception network outputs "obstacle confidence 0.65", and the control layer binarizes it into "brake / no brake"; no transition state: the transition from "cruising" to "braking" has a 180ms decision delay; loss of spatiotemporal context: each decision is calculated independently based on the current frame; 70% of accidents occur in the transition blind zone.

[0170] UPN four-state engineering implementation:

[0171] (1) Driving state S_T (vehicle start registration): The vehicle-level UPN is generated based on GPS trajectory, lidar point cloud features, and velocity vector; the obstacle UPN subsystem generates an independent UPN for each detected obstacle (vehicle, pedestrian, spilled object), with the Category prefix VEHICLE_CONTROL_L4_OBSTACLE;

[0172] (2) Alert state S_φ (hesitation zone management): Obstacle confidence convergence calculation adopts the ObstacleConvergenceTracker class, calculates the confidence change rate (danger convergence speed) and normalizes it to obtain the R value; Dynamic threshold contraction: Braking preparation pressure = R × 0.30 (maximum 30% preparation pressure); When 0.2 < R < 0.8043, the S_φ state is triggered, and the brake oil pressure is immediately established to 30% through the CAN bus (ISO 11898, ID: 0x123), eliminating the 100ms hydraulic build-up time in the traditional solution;

[0173] (3) Braking lock-up state S_1 and safety state S_0: When R > 0.8043 or TTC (Time To Collision) < 2.5s, the state transitions to S_1, and full braking is performed. Since 30% oil pressure has been pre-established in the S_φ state, the actual braking delay from S_φ to S_1 is only 45ms (only signal transmission + mechanical response, no hydraulic pressure establishment time), which is 75% less than the 180ms delay of the traditional solution;

[0174] (4) Decision timing comparison (Scenario of debris being thrown ahead, vehicle speed 120km / h=33.3m / s): t=0ms Obstacle appears (distance 100m), UPN system S_T is triggered; t=200ms Traditional system judges "danger", UPN system S_φ is triggered and establishes 30% braking preparation; t=250ms Traditional system begins to establish brake oil pressure, UPN system R=0.85>S_c, S_1 full braking; t=350ms Traditional system actual braking takes effect, UPN system has braked for 100ms; Up to t=500ms, a total braking distance of 6.0m is saved;

[0175] (5) Performance improvement indicators: Emergency braking delay decreased from 180ms in traditional Autoware to 45ms in UPN four-state control (a decrease of 75.0%); braking distance at 120km / h decreased from 85.2m to 79.1m (a decrease of 6.1m); transitional accident rate decreased by 62%; passenger comfort score increased from 3.2 / 5 to 4.6 / 5 (an increase of 43.8%); false alarm and false braking rate decreased from 12% to 3% (a decrease of 75.0%).

[0176] 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 a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0177] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A process identity generation method based on unique object identifiers and four-state logic, characterized in that, Includes the following steps: S1. Receive three types of input parameters for physical entities: Category, Space-Time Anchor, and Feature Hash. S2. Generate a unique identifier (UPN) with object uniqueness, spatiotemporal anchoring, and process rheology from the three types of input parameters; S3. Bind the generated identity identifier UPN to the four-state logic; S4. Through the Process Identity mechanism, the identity identifier UPN is continuously re-registered, forming a process identity that is continuous in identity but changes in state.

2. The process identity generation method based on unique object identification and four-state logic according to claim 1, characterized in that: The category uses uppercase underscore naming convention, which supports zero-skeleton migration across industries; The space-time anchor includes a coordinate system, a UTC timestamp, spatial coordinates, and a time sequence number, ensuring that each UPN generation is strictly bound to a specific spatiotemporal location; The feature fingerprint includes the original value or hash of the main feature, a list of auxiliary features, and a feature version number.

3. The process identity generation method based on unique object identification and four-state logic according to claim 1, characterized in that: The steps for generating the identity identifier UPN are as follows: (1) Standardize the input string; (2) Spatiotemporal anchor serialization; (3) Feature fingerprint serialization; (4) Concatenate triples, using the non-printable delimiter 0x1F (Unit Separator) to ensure the uniqueness of field boundaries; (5) Cryptographic hash operations; (6) Extract the first 32 hexadecimal characters; (7) Add a check bit; Specifically, when the identity identifier UPN is generated, the three types of input parameters are concatenated into a single string in a fixed field order. During concatenation, the non-printable delimiter 0x1F (Unit Separator) is used to ensure the uniqueness of field boundaries. When generating the identity identifier UPN, the cryptographic hash function SHA-256 is used to perform a hash operation on the concatenated string. The first 32 hexadecimal characters of the hash value are extracted as the unique identity identifier UPN of the object, and a hexadecimal Luhn check digit is added as the final identity identifier.

4. The process identity generation method based on unique object identification and four-state logic according to claim 1, characterized in that: The four states include: S_T (Spatiotemporal Becoming): S_T is triggered upon completion of the UPN identity registration. S_φ (Converging, dynamic convergence state): When the entity state is in the transition stage of "evolving towards the goal but not yet locked", the system determines it to be S_φ. In this state, the constraint threshold shrinks dynamically with the R_SelfStudy value, allowing the system to continuously fine-tune during the convergence process. S_1 (Satisfied): When the entity state fully satisfies the preset constraint conditions, the system determines it to be S_1. At this time, the constraint is locked and the action loop is completed. S_0 (Violated, constraint not satisfied or closed loop terminated state): When the entity state exceeds the safety boundary or convergence fails, the system determines it to be S_0, at which point emergency intervention or loop termination is triggered.

5. The process identity generation method based on unique object identification and four-state logic according to claim 4, characterized in that: The S_φ state introduces a dynamic threshold shrinkage mechanism, whereby the constraint threshold is dynamically adjusted according to the R_SelfStudy value. The formula is: Threshold_dynamic = Threshold_min + (Threshold_init - Threshold_min) × (1 - R), where R = R_SelfStudy∈ [0,1] is the self-learning convergence exponent estimated online based on exponentially weighted moving average (EWMA). When R increases, the threshold shrinks towards Threshold_min, and the system becomes more sensitive to early weak signals. The avalanche threshold R_c is the critical point of the phase transition from S_φ to S_1.

6. The process identity generation method based on unique object identification and four-state logic according to claim 5, characterized in that: The triggering condition for the S_φ state is 0.1 < R < R_c, and the four-state transition is implemented by a finite state machine (FSM). State transition triggering conditions include: the rate of change of the R value, constraint satisfaction Boolean value, and timeout timer; The state transition actions include: monitoring frequency adjustment, brake preparation pressure establishment, and alarm triggering.

7. The process identity generation method based on unique object identification and four-state logic according to claim 6, characterized in that: The R_SelfStudy self-learning parameters are coupled with the post-Einstein information-entropy scalar completion equation α = ln2+ (1-ln2)Q - (2ln2-1)C; Where Q is the micro-generation intensity; C is the macro-dissipation intensity; α is the information coupling constant, and α=ln2 is used as the system element law encoding; R_SelfStudy is not a preset constant, but is obtained in real time through an online learning algorithm. The acquisition process is as follows: (1) Calculate the rate of change of the characteristic; (2) Calculate the acceleration; (3) EWMA smoothing, α_smooth = 0.3; (4) Sigmoid is mapped to [0,1]; (5) Clip the boundary constraints to [0.0, 1.0]; Among them, parameters k1 and k2 are calibrated online using the historical data of the entity, fitted using the least squares method, and updated every 24 hours.

8. The process identity generation method based on unique object identification and four-state logic according to claim 1, characterized in that: The process identity mechanism regenerates UPN_new for entities based on updated spatiotemporal anchors and feature fingerprints during spatiotemporal evolution, while maintaining identity continuity through continuity verification. The steps for continuous re-registration are as follows: (1) When an entity registers for the first time at time t, a UPN0 is generated based on the space-time anchor at that time and the feature hash. (2) When an entity re-registers at time t+Δt, the space-time anchor is updated, the feature hash is updated, and UPN1 is generated; (3) The system confirms the continuity of identity through UPN continuity verification to ensure that "identity is continuous but state changes".

9. The process identity generation method based on unique object identification and four-state logic according to claim 8, characterized in that: The continuity check employs at least one of the following methods: (1) Quickly filter by hash prefix similarity, requiring the longest common prefix length to be ≥0.85, i.e., the threshold is ≥85%; (2) Accurate verification through semantic similarity of feature fingerprints, using cosine distance or Euclidean distance, with a threshold ≥0.7; (3) Verify through Merkle root to prevent tampering. Maintain the Merkle tree of each entity through fog nodes. Update the Merkle root once for each new UPN and synchronize the root hash to the cloud blockchain to achieve cross-domain identity consistency verification.

10. A system for performing the method of any one of claims 1-9, characterized in that, include: The Edge Layer, deployed on the physical entity side, is responsible for raw data acquisition, initial UPN generation, and local four-state determination. The Fog Layer, deployed on a local server, is responsible for UPN continuity verification, R_SelfStudy self-learning parameter updates, dynamic threshold calculation, and cross-device identity synchronization. The Fog Layer runs a lightweight time-series database and Merkle tree verification service. The Cloud Layer, deployed in public or private clouds, is responsible for global UPN namespace management, cross-industry Category registry, historical trajectory storage, model training, and parameter distribution. During interaction, the process is as follows: physical entity, edge acquisition module, UPN generation engine, four-state machine, fog layer verification node, and cloud layer registration center.

Images