A state identity continuous maintenance method for autonomous intelligent system

Abstract

The application discloses a state identity continuous maintenance method and system for an autonomous intelligent system and a storage medium. The method comprises the following steps: obtaining target state information and corresponding execution intention information, determining a state entity to be continuously evolved and establishing state identity identification information; establishing an identity binding relationship between the state identity identification information and the state entity, forming a continuous identity association; maintaining the identity binding relationship in the state entity migration process, and associating multiple states with different state values with the same state entity. The state identity identification information is different from timestamp and version number information, and is used for identifying the state entity rather than a specific state value. When the state value changes, the entity identity is not determined again due to the change of the state value, so that the system can determine the subsequent state as the same state entity based on the identity binding relationship and the continuous state association condition. The application realizes the continuous maintenance and verifiable attribution of the state entity identity of the autonomous intelligent system in the process of state continuous change and cross-execution domain migration.

Description

Technical Field

[0001] This invention relates to the fields of autonomous intelligent systems, intelligent control, state management and continuous execution technology, and in particular to a method and system for continuously maintaining the state and identity of autonomous intelligent systems.

[0002] More specifically, the present invention relates to a method and system for continuously maintaining the identity of a state entity during the continuous changes, continuous migrations, cross-execution domain migrations, state rollbacks, and state evolutions of the state of an autonomous intelligent system.

[0003] This invention is particularly applicable to autonomous intelligent systems, heterogeneous computing systems, intelligent control systems, industrial control systems, robotic systems, and control systems for complex physical processes that exhibit continuous state evolution characteristics. In some embodiments, this invention can be applied to nuclear fusion control systems for identity preservation, migration identification, and state attribution verification of continuously evolving control state entities. Background Technology

[0004] As artificial intelligence systems gradually evolve from one-time task processing to long-term operation, autonomous decision-making, and continuous execution, the state in the system is no longer just static data at a certain moment, but may continuously change during execution, and multiple continuous states together constitute a continuously evolving state entity.

[0005] For example, during the execution of a continuous task by an autonomous intelligent system, the system state can gradually transition from an initial state to multiple subsequent states. The state parameters may change at different transition stages, and the execution environment, computing resources, execution domain, and control parameters may also change. However, from the perspective of system operation logic, these states may still belong to the same continuously executing state entity.

[0006] Existing systems typically use timestamps, version numbers, sequence numbers, transaction numbers, or other identifying information to distinguish states. Timestamps are primarily used to indicate the chronological order of state occurrence or processing; version numbers are mainly used to indicate changes in data, configuration, or software states; and sequence numbers or transaction numbers are typically used to distinguish different data records or execution processes.

[0007] The above identification methods can distinguish different states to some extent, but their main function is usually to indicate the time, version, or order in which a state occurs, and they cannot directly solve the following problems:

[0008] First, when a state entity changes continuously from a first state to multiple subsequent states, how can we confirm that these states with different state values ​​still belong to the same state entity?

[0009] Second, when the state value changes, how to avoid the identity of the state entity being redefined due to the change in the state value;

[0010] Third, how to maintain the continuous association of the state entity's identity when the state entity migrates from one execution domain to another;

[0011] Fourth, when the system performs a state rollback or recovery operation, how can it be confirmed that the recovered state belongs to the original state entity, rather than another state entity with the same or similar state parameters?

[0012] Fifth, when multiple state entities evolve in parallel, how to avoid confusion between different state entities caused by relying solely on state values, time, or version information;

[0013] Sixth, in autonomous control systems with continuous state evolution characteristics, how to structurally distinguish between "changes in state values" and "maintaining the identity of state entities".

[0014] Especially in autonomous intelligent systems, the system state may simultaneously include execution state, data state, environmental state, control state, execution intent, and other constraint information. As the model, execution resources, or execution domain change, the same autonomous task may generate a large number of different state values. If the state identity is determined solely by the current state value or the current version number, it may be impossible to determine whether different states belong to the same continuously evolving entity.

[0015] Therefore, it is necessary to establish a state identity maintenance mechanism that differs from traditional timestamps, version numbers, and sequence numbers, so that state entities can maintain continuous identity associations as state values ​​change continuously, and can verify the ownership of state entities during state transitions, cross-execution domain migrations, rollbacks, and state evolutions. Summary of the Invention

[0016] (a) Technical problems to be solved

[0017] The purpose of this invention is to provide a method and system for maintaining the continuous state identity of autonomous intelligent systems, in order to solve the problem that existing state identification methods are mainly used to represent changes in time, version or sequence, and it is difficult to maintain the continuous association of the identity of the same state entity during continuous changes in state values.

[0018] The present invention further aims to achieve the following: determining the corresponding state identity information when a state entity is established; establishing an identity binding relationship between the state identity information and the state entity; maintaining the identity binding relationship when the state entity undergoes continuous state transitions; ensuring that the same state entity maintains the same identity even when it has different state values ​​at different transition stages; distinguishing different transition stages without changing the state entity's identity through transition version information; verifying the state entity's identity during state transition completion, state rollback, and cross-execution domain transitions; and maintaining a unique association between the state identity information and the corresponding state entity throughout the state entity's lifecycle, preventing the identity from being reassigned to another state entity.

[0019] (II) Technical Solution

[0020] To achieve the above objectives, the present invention provides a method for continuously maintaining the state and identity of an autonomous intelligent system, comprising:

[0021] Obtain the target state information of the autonomous intelligent system and the execution intent information corresponding to the target state information;

[0022] Based on the target state information and the execution intent information corresponding to the target state information, determine the state entity to be continuously evolved, and establish state identity information for the state entity.

[0023] The state identity information is linked to the state entity to form a continuous identity association for the state entity during continuous state evolution.

[0024] During the process of the state entity transitioning from a first state to at least one subsequent state, the identity binding relationship between the state identity information and the state entity is maintained, so that the first state and the at least one subsequent state are still associated with the same state entity when the state value changes.

[0025] The state identity information is distinct from the timestamp information used to represent temporal sequence and the version number information used to represent state version changes. The state identity information is used to identify the state entity, rather than to identify a specific state value. When the state value changes, the identity of the state entity is not re-determined due to the change in the state value, so that the system can determine the subsequent state as the same state entity corresponding to the first state based on the identity binding relationship and the continuous state association condition.

[0026] In one implementation, the state identity information is determined when the state entity is established and remains unchanged throughout the lifecycle of the state entity.

[0027] In one implementation, each state transition has corresponding transition version information. The transition version information is used to distinguish the transition states of the same state entity at different transition stages without changing the state identity information.

[0028] In one implementation, the state identity information is associated with at least one of the state parameter information and execution intent information of the state entity.

[0029] In one implementation, when the state entity undergoes multiple consecutive state transitions, the state identification information is kept uniquely associated with each transition state, so that multiple transition states with different state values ​​are associated with the same state entity.

[0030] Furthermore, in the above method for maintaining continuous state identity, state authentication operations can also be performed:

[0031] After a state entity migrates, the state identity information corresponding to the migrated state to be verified is obtained; the state identity information corresponding to the migrated state to be verified is compared with the state identity information corresponding to the target state entity; the identity attribution of the migrated state to be verified is verified by combining the pre-established identity binding relationship and the continuous state association condition; when the state identity information is consistent and the continuous state association condition is met, the migrated state to be verified is determined to belong to the target state entity; when the state identity information is inconsistent or the continuous state association condition is not met, the migrated state to be verified is rejected from belonging to the target state entity.

[0032] The continuous state association conditions may include, but are not limited to: the existence of state transition relationships, the validity of transition order relationships, the validity of identity binding relationships, the satisfaction of preset continuity conditions between previous and subsequent states, the satisfaction of preset execution constraints in the transition process, or at least a combination of the above conditions.

[0033] (III) Distinguishing between state entities and state values

[0034] One of the core aspects of this invention lies in distinguishing between "state entities" and "state values." In this invention, state values ​​describe the specific state of a state entity at a certain transition stage. A state entity, on the other hand, represents a logical entity that can continuously evolve across multiple states while maintaining its identity association. Therefore, the same state entity can correspond to multiple different state values.

[0035] For example:

[0036]

[0037] Where Eᵢ represents the state instance representation of the same state entity in the i-th stage; Sᵢ represents the state value in the i-th stage; and V0 represents the state identity information of the state entity. This indicates the corresponding migration version information.

[0038] In the above process, although the state instance representation and state value corresponding to different stages may change, the state entity to which it belongs will not change due to the change of the state value or state instance representation; as long as the pre-established identity binding relationship and continuous state association conditions are met, the multiple state instances of different stages can be determined to belong to the same state entity.

[0039] Therefore, the state identity information in this invention is not used to represent a specific state value, but rather to maintain the identity association of state entities during continuous state evolution.

[0040] (iv) The difference between status identifiers, timestamps, and version numbers

[0041] Traditional timestamps are primarily used to describe when an event or state occurs, for example... It can represent the chronological order in which states appear, but it cannot prove on its own whether S0, S1, S2, and S3 belong to the same state entity. Version numbers are mainly used to indicate state version changes, for example... It can distinguish between different versions, but version changes do not necessarily mean that the identity of the state entity is maintained.

[0042] This invention structurally distinguishes between the identity of a state entity, changes in state values, and changes in migration versions, which can be represented as follows: , , V0 maintains its correspondence throughout the lifecycle of the state entity, while Sᵢ and Mᵢ can change as the state evolves. This forms... ,and This allows the system to simultaneously express "the state has changed" and "the state entity has not had its identity replaced".

[0043] (v) Status Identity Establishment Mechanism

[0044] During the state entity establishment phase, the system acquires target state information and corresponding execution intent information. Target state information may include: one or more state parameters, current execution environment parameters, current control parameters, data state, computing resource state, constraint state, task state, and other information used to describe the target state. Execution intent information may include: target task, execution objective, control purpose, state transition purpose, operation request, execution intent generated by autonomous decision-making, and other information used to describe the expected behavior of the state entity.

[0045] Based on the above information, the system determines the state entity to be continuously evolved and establishes state identity information corresponding to that state entity. In some implementations, the state identity information can adopt a numerical identifier, logical identifier, internal identifier, unique identifier, or other identifier form that can form a unique identity association in the system. This invention does not limit the specific encoding form of the state identity information; the specific encoding method can be selected according to the system scale, computing architecture, reliability requirements, security requirements, and execution environment.

[0046] This invention does not focus on a specific encoding algorithm, but rather on establishing an identity binding between state identity information and state entities, and maintaining this binding relationship during the continuous evolution of the state entities.

[0047] (vi) Generation and anti-tampering mechanism of status identity

[0048] To enable the establishment, storage, and verification of state identity information in a real system, this invention further provides an implementable state identity generation mechanism.

[0049] It should be noted that this invention does not limit the state identification information to a specific encoding algorithm. The state identification information can be generated by one or more of the following: random numbers, pseudo-random numbers, device identifiers, entity initial state summaries, execution intent summaries, system scope identifiers, and other information capable of distinguishing entities.

[0050] In one implementation, when a state entity is established, the system first obtains:

[0051] S0 represents the initial state information of the state entity;

[0052] I0 represents the corresponding initial execution intent information;

[0053] N0 represents a unique random number or random seed generated by the system for the entity;

[0054] D0 represents the system scope, execution scope, or device scope information corresponding to the state entity when it is created.

[0055] The system can standardize the above information and generate an initial state entity summary:

[0056]

[0057] Furthermore, state identity information can be generated based on the entity's unique random number or random seed and the initial state entity digest:

[0058]

[0059] in: This indicates the pre-defined data normalization process; This indicates a summary calculation; This indicates that the generated information will be converted into an identity format that the system can use; V0 represents the status identity information of the status entity.

[0060] The above formula is only used to illustrate one possible implementation method and does not limit the specific identity generation algorithm of the present invention.

[0061] In another implementation, state identification information can be directly formed from randomly generated unique entity identifiers, without including a state value digest. For example:

[0062]

[0063] Subsequently established:

[0064]

[0065] B0 represents the identity binding relationship between the state identity information and the state entity. Therefore, the state identity information does not need to be regenerated as the state value changes.

[0066] To prevent unauthorized modification of the binding relationship between state identity information and state entity, the present invention can further establish integrity protection information for the identity binding relationship.

[0067] In one implementation, an integrity summary is calculated for the identity binding relationship and its key associated information:

[0068]

[0069] Where: V0 is the status identifier; H0 is the status entity summary; I0 is the execution intent association information; Scope0 is the identity scope information. The system saves:

[0070]

[0071] When it is necessary to verify whether the identity binding relationship has been modified, recalculate:

[0072]

[0073] And judge If they match, then the integrity verification of the above identity binding record has passed. If the identity binding record changes, the system will refuse to establish a status identity attribution relationship directly based on the binding record, or it will enter the exception handling process.

[0074] Therefore, the anti-tampering described in this invention does not require absolute physical immutability, but rather enables unauthorized modifications to be detected through integrity verification.

[0075] (vii) Mechanism for Continuous Maintenance of Status and Identity

[0076] After a state entity is established, the system allows state values ​​to change, for example... During this process, the system does not re-determine the identity of the state entity due to changes in state values. The corresponding state identity relationship remains unchanged. .therefore, This creates a separation between changes in state values ​​and the preservation of the state entity's identity. This separation allows the system to determine that although a subsequent state is no longer a previous state value, it still belongs to the same state entity that was previously established.

[0077] (viii) Migration version information

[0078] In order to distinguish the states of the same entity in different migration stages, the present invention can establish corresponding migration version information for each state migration.

[0079] For example, the system can generate migration version information for multiple migration stages. Each migration version is used to distinguish the corresponding different migration stages. For the same state entity, This means that the state identity information remains unchanged across all migration stages. The same state entity can simultaneously have different state values, different migration versions, and the same state identity. For example:

[0080]

[0081] The migration version describes "which stage of the continuous evolution process we are currently in," while the state identity describes "whether these stages belong to the same state entity." They are functionally distinct.

[0082] (ix) Identity binding relationship

[0083] This invention establishes an identity binding relationship between state identity information and the state entity during the state entity establishment phase. This identity binding relationship can correspond to one or more state parameter information, execution intent information, or other entity association information of the state entity. It can be represented as follows: Where V represents the state identity information, and E represents the state entity. During the continuous evolution of the state entity, Stay valid. Even Changes have occurred, but the status remains unchanged. This ensures the continuity of the identity of state entities across different state values.

[0084] (x) Uniqueness of Status Identity

[0085] Throughout the lifecycle of a state entity, a unique association is maintained between the state identity information and the corresponding state entity. State identity information that has already established a valid identity binding with a state entity is not reassigned to another state entity during the lifecycle of that state entity. Therefore, for two state entities... Its status identification information should meet the following requirements. This avoids identity confusion between multiple parallel state entities. It should be noted that the uniqueness in this invention mainly refers to the uniqueness of identity association within the preset system scope and the life cycle of the state entity, and does not require the use of globally unique coding in an absolute mathematical sense in all systems, all times, or all external environments.

[0086] (xi) Continuous state association conditions

[0087] This invention does not use "identity of state identity information" as the sole criterion for determining state attribution. During state identity verification, the system can also combine pre-established continuous state association conditions for comprehensive judgment. These continuous state association conditions may include: a preset migration relationship exists between preceding and following states; the preceding state is a legitimate migration source for the following state; the migration process satisfies preset state transition constraints; the state identity binding relationship remains valid; the migration version relationship is continuous; the execution intent association relationship is valid; no identity replacement occurs during the state migration process; and the execution domain where the state migration occurs meets preset conditions. For example... Even if E2 carries V0, the system can still further verify whether E0→E1→E2 constitutes an allowed continuous migration chain. Therefore, the authentication of this invention is not a simple string comparison, but a determination of state entity attribution based on identity binding relationship and continuous state association relationship.

[0088] (xii) Cross-execution domain state transition

[0089] In autonomous intelligent systems, the same state entity may migrate from a first execution domain to a second execution domain, for example, D1→D2. When the execution domain changes, the state value, operating environment, and computing resources may all change. This can be represented as:

[0090]

[0091] in ,but The system is thus able to confirm that the state after migrating to the second execution domain still belongs to the original state entity.

[0092] The execution domain can be: different CPU execution domains, different GPU execution domains, different heterogeneous computing nodes, different container or virtual execution environments, different agent operating environments, different controllers, different hardware control modules, or different physical control units.

[0093] (xiii) Authentication during state rollback

[0094] In some autonomous intelligent systems, the system may need to roll back from the current state to a historical state. For example... The system rolls back from E3 to the historical state corresponding to E1. The system can obtain the state identity information corresponding to the target rollback state and compare it with the state identity information of the target state entity. If... If the historical state and the original state entity satisfy the pre-established continuous state association condition, then it can be determined that the rollback state belongs to the original state entity.

[0095] Therefore, this invention can be used to help avoid misidentifying another state with the same or similar state parameters as the original state entity. It should be noted that the continuous maintenance of state identity is used to solve the problem of attribution to "which state entity to roll back to"; the correctness of the rollback data, the integrity of the state, and the security of the rollback execution can be further guaranteed by other state verification, constraint control, or proof mechanisms.

[0096] (xiv) Parallel evolution of multiple state entities

[0097] In autonomous intelligent systems, multiple state entities with the same or similar execution intentions may exist simultaneously. For example... as well as ,even though This does not mean This invention enables multiple parallel state entities to maintain independent identities by establishing separate state identity information, for example... This enables the system to distinguish different state entities during parallel execution, task replication, branch evolution, or multi-agent cooperation.

[0098] (XV) Implementation Examples of Autonomous Intelligent Systems

[0099] In an autonomous intelligent system, an autonomous task can form a state entity that needs to evolve continuously. For example, when an autonomous intelligent system receives a target task: "Complete the target operation under the conditions of satisfying the current system constraints," the system first obtains the target state and the execution intention. and establish V0 represents the state identification information. Subsequently, after one decision and execution, the state changes. At this point, the system does not recreate a new state entity, but maintains... After further implementation Thus, the following is formed:

[0100]

[0101] The system is therefore able to identify that the above multiple states belong to the same continuously evolving entity.

[0102] (xvi) Example of Continuous Maintenance of State Entity Identity in Nuclear Fusion Control System

[0103] 16.1 Overview of Implementation Examples

[0104] This embodiment applies the present invention to a nuclear fusion control system. Nuclear fusion control systems are typical continuous-state evolution control systems. The plasma state, magnetic field configuration, heating state, fuel supply state, constraint parameters, and safety control state may all continuously change over time. In such systems, the control system does not deal with only one static state but continuously processes... The resulting continuous state evolution process necessitates a distinction between "a change in the current control state" and "an identity replacement of the controlled state entity." This embodiment utilizes state identity identification information V0 to ensure that the continuously changing nuclear fusion control states maintain the same state entity identity.

[0105] 16.2 Establishment of Nuclear Fusion Control State Entity

[0106] The nuclear fusion control system can acquire the target plasma control state and corresponding control intent. Target state information may include: plasma density, plasma temperature, plasma position, plasma shape, plasma current, magnetic field parameters, heating power, fuel injection parameters, confinement state, diagnostic state, safety boundary parameters, and other control state parameters. Execution intent information may include: maintaining the target plasma state, adjusting the plasma shape, adjusting the plasma position, adjusting the heating power, adjusting the fuel supply, suppressing abnormal states, achieving preset control objectives, and other control intents.

[0107] The system determines the nuclear fusion control state entity to be continuously evolved based on the target state information and the corresponding execution intention information, for example... Where S0 is the initial nuclear fusion control state, I0 is the corresponding control intention, and V0 is the state identity information corresponding to the entity in the control state.

[0108] 16.3 Continuous Evolution of Nuclear Fusion States

[0109] During nuclear fusion, the plasma state continuously changes. Throughout these changes, the system maintains its state identification information unchanged, i.e., V0=V1=V2=V3, thus forming:

[0110]

[0111] Although the state values ​​at the different stages mentioned above can be different from each other, for example However, as long as the pre-established identity binding relationship and continuous state association conditions are met, the system will still determine that the states of the different stages belong to the same nuclear fusion control state entity.

[0112] 16.4 Abnormal states and rollbacks in nuclear fusion

[0113] During nuclear fusion control, if an abnormal state is detected, the system may need to revert to a previous safe control state. For example... E3 corresponds to an abnormal state, and the system chooses to roll back to the state corresponding to E1. At this point, the state identification information carried by the rollback state can be verified. And further verification That is, the rollback state belongs to the continuous evolution relationship of the original state entity.

[0114] If the state identity information carried by the state to be rolled back is inconsistent with the target state entity, or if the pre-established continuous state association conditions cannot be met, the system can refuse to use that state as the rollback state of the original state entity. This reduces the risk of incorrect state attribution due to similar state values, historical data confusion, or changes in the execution domain.

[0115] (XVII) System Implementation Method

[0116] This invention also provides a continuous state and identity maintenance system for autonomous intelligent systems, comprising:

[0117] The state acquisition module is used to acquire the target state information of the autonomous intelligent system and the execution intent information corresponding to the target state information.

[0118] The identity determination module is used to determine the state entity to be continuously evolved based on the target state information and execution intention information, and to establish state identity identification information for the state entity.

[0119] The identity binding module is used to establish an identity binding relationship between status identity information and status entities;

[0120] The identity preservation module is used to maintain the identity binding relationship between the state identity information and the state entity during the process of the state entity transitioning from a first state to at least one subsequent state;

[0121] The authentication module is used to verify the consistency between the state identity information corresponding to the state to be verified and the state identity information corresponding to the target state entity when the state transition is completed, the state is rolled back, or the state transition is across execution domains. It also determines whether the state to be verified belongs to the target state entity by combining the identity binding relationship and the continuous state association condition.

[0122] The modules described above can be implemented using software, hardware, firmware, or a combination of both. These modules can be deployed on the same computing device or on multiple interconnected computing devices.

[0123] (18) Beneficial Effects

[0124] Compared with the prior art, the present invention has at least the following beneficial effects:

[0125] This invention separates state value changes from identity changes, enabling the separation of state entity identity from state value changes.

[0126] Maintain entity identity throughout continuous evolution. By binding state identity information with the identity of state entities, the same state entity can maintain its identity association across multiple transition states.

[0127] Unlike timestamps and version numbers, which are primarily used to describe time sequence or state versions, the state identity identifier of this invention is used to describe the state entity itself, thereby providing state identification capabilities in different dimensions.

[0128] Supports cross-execution domain migration. Even if a state entity migrates from one execution domain to another, it can still be identified as belonging to the same state entity through a persistent state identity.

[0129] Supports state rollback authentication. During state rollback, state identity identifiers and continuous state association conditions can be used to verify whether the rollback state belongs to the target state entity.

[0130] Supports the differentiation of parallel state entities. Multiple state entities with the same or similar execution intentions can maintain their independent identities, reducing state confusion during parallel evolution.

[0131] It supports continuous state evolution tracing. By associating multiple transition stages with the same state identity, an identity association basis can be formed for the continuous evolution process of state entities.

[0132] This invention is applicable to complex continuous control systems. It is not only suitable for autonomous intelligent systems, but also for heterogeneous computing, industrial control, robotics, complex physical process control, and other complex control systems with continuous state evolution characteristics.

[0133] Identity binding integrity protection. An integrity verification mechanism ensures that the state-identity binding relationship can be verified, preventing unauthorized identity replacement or alteration of the binding relationship.

[0134] Identity generation feasibility. A specific process for generating state identity information is provided, enabling the continuous state identity maintenance mechanism to be established and implemented in a real system. Attached Figure Description

[0135] Figure 1 This is a schematic diagram of the overall process of the method for continuously maintaining the identity of a state entity according to the present invention.

[0136] Figure 2 This is a schematic diagram illustrating the relationship between state entities, state values, and state identity identifiers in this invention.

[0137] Figure 3 This is a schematic diagram illustrating the identity preservation process during the continuous evolution of state entities in this invention.

[0138] Figure 4 This is a schematic diagram illustrating the relationship between the status identity identifier, migration version information, and status value of this invention.

[0139] Figure 5 This is a schematic diagram of the status authentication method of the present invention.

[0140] Figure 6 This is a schematic diagram illustrating the continuous maintenance of identity during cross-execution domain state transitions in this invention.

[0141] Figure 7 This is a schematic diagram of the identity verification process during the state rollback process of this invention.

[0142] Figure 8 This is a schematic diagram illustrating the continuous maintenance of the identity of state entities in the nuclear fusion control system of the present invention.

[0143] Explanation of reference numerals in the attached figures

[0144] <![CDATA[E0,E1,E2,…]]> State instance representation of the same state entity at different stages <![CDATA[S0,S1,S2,…]]> State values ​​at different stages <![CDATA[V0]]> Status identification information <![CDATA[M0,M1,M2,…]]> Migration version information for each state transition <![CDATA[D1,D2]]> Execution domain B Identity binding relationship <![CDATA[R0]]> Identity binding record

Claims

1. A method for continuously maintaining the state and identity of an autonomous intelligent system, characterized in that, include: Obtain the target state information of the autonomous intelligent system and the execution intent information corresponding to the target state information; Based on the target state information and the execution intent information corresponding to the target state information, determine the state entity to be continuously evolved, and establish state identity information for the state entity. The state identity information is linked to the state entity to form a continuous identity association for the state entity during continuous state evolution. During the process of the state entity transitioning from a first state to at least one subsequent state, the identity binding relationship between the state identity information and the state entity is maintained, so that the first state and the at least one subsequent state are still associated with the same state entity when the state value changes. The state identity information is distinct from the timestamp information used to represent temporal sequence and the version number information used to represent state version changes. The state identity information is used to identify the state entity, rather than to identify a specific state value. When the state value changes, the identity of the state entity is not re-determined due to the change in the state value, so that the system can determine the subsequent state as the same state entity corresponding to the first state based on the identity binding relationship and the continuous state association condition. When the state identity information corresponding to the subsequent state is consistent with the state identity information corresponding to the target state entity and the continuous state association condition is met, the subsequent state is determined to belong to the target state entity; when the state identity information is inconsistent or the continuous state association condition is not met, the subsequent state is refused to be associated with the target state entity.

2. The method according to claim 1, characterized in that, The state identity information is determined when the state entity is established and remains unchanged throughout the lifecycle of the state entity.

3. The method according to claim 1, characterized in that, Each state transition of the state entity has corresponding transition version information. The transition version information is used to distinguish the transition states of the same state entity at different transition stages without changing the state identity information.

4. The method according to claim 1, characterized in that, The status identity information is associated with at least one of the status parameter information and execution intent information of the status entity.

5. The method according to claim 1, characterized in that, When the state entity undergoes multiple consecutive state transitions, the state identity information is maintained as a unique association between each transition state, so that multiple transition states with different state values ​​are associated with the same state entity.

6. The method according to claim 1, characterized in that, Also includes: Obtain the status identity information corresponding to the migration status to be verified; The status identity information corresponding to the migration state to be verified is compared with the status identity information corresponding to the target state entity, and the identity attribution of the migration state to be verified is verified in combination with the pre-established identity binding relationship and continuous state association conditions. When the state identity information corresponding to the migration state to be verified is consistent with the state identity information corresponding to the target state entity, and the continuous state association condition is met, it is determined that the migration state to be verified belongs to the target state entity. If the state identity information corresponding to the migration state to be verified is inconsistent with the state identity information corresponding to the target state entity, or if the continuous state association condition is not met, the migration state to be verified shall be refused to be associated with the target state entity.

7. The method according to claim 1, characterized in that, The identity binding relationship between the state identity information and the state entity remains unique throughout the lifecycle of the state entity, and the state identity information is not reassigned to another state entity.

8. The method according to claim 1, characterized in that, The continuous state association condition is used for at least one of the following operations: State entity identification during cross-execution domain state transitions; State entity authentication during state rollback; Tracing the state evolution process; Distinguishing identities during the parallel execution of multiple state entities.

9. A continuous state and identity maintenance system for autonomous intelligent systems, characterized in that, include: The status acquisition module is used to acquire the target status information of the autonomous intelligent system and the execution intent information corresponding to the target status information; The identity determination module is used to determine the state entity to be continuously evolved based on the target state information and the execution intent information corresponding to the target state information, and to establish state identity identification information for the state entity. The identity binding module is used to establish an identity binding relationship between the state identity identification information and the state entity, so as to form a continuous identity association of the state entity in the continuous state evolution process; An identity retention module is used to maintain the identity binding relationship between the state identity information and the state entity during the process of the state entity transitioning from a first state to at least one subsequent state, so that the first state and the at least one subsequent state are still associated with the same state entity when the state value changes. The authentication module is used to verify the consistency between the state identity information corresponding to the state to be verified and the state identity information corresponding to the target state entity when the state transition is completed, the state is rolled back, or the state transition is across execution domains. It also determines whether the state to be verified belongs to the target state entity by combining the identity binding relationship and the continuous state association condition.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, the processor performs the method according to any one of claims 1 to 8.

Images