A decentralized distributed autonomous artificial intelligence system and a high-availability self-healing implementation method thereof

Abstract

This invention discloses a decentralized distributed autonomous artificial intelligence system, relating to the fields of artificial intelligence and distributed computing. The system comprises six modules: spiking neural network, causal reasoning, P2P node management, full-state persistence, autonomous operation and maintenance, and security protection. This invention achieves decentralized networking of intelligent agents, fault self-healing, state persistence, autonomous operation and maintenance, security protection, and autonomous optimization. It can operate stably for a long time without human intervention and is suitable for scenarios such as intelligent robots, industrial control, mission-critical systems, and large-scale intelligent clusters. This invention solves the technical pain points of traditional AI systems, such as susceptibility to crashes, lack of self-healing, dependence on central servers, low security, and inability to operate autonomously.

Description

Technical Field

[0001] This invention belongs to the technical fields of artificial intelligence, distributed computing, autonomous intelligent agent control and high availability systems. Specifically, it relates to an artificial intelligence architecture and its implementation method that has autonomous operation and maintenance, state persistence, fault self-healing, decentralized networking, security protection and autonomous optimization capabilities. Background Technology

[0002] With the development of artificial intelligence technology, traditional AI systems have significant shortcomings in long-term stable operation, decentralized collaboration, autonomous operation and maintenance, fault recovery, and security protection. Current mainstream AI systems generally rely on centralized server deployment, failing to achieve distributed autonomous operation. A single point of failure can lead to a complete service interruption. Furthermore, traditional AI lacks system-level self-healing capabilities; it cannot automatically recover its operational state after process crashes, file corruption, or power outages, requiring manual intervention for restart and reconstruction. In addition, existing AI lacks autonomous operation and maintenance capabilities, unable to autonomously monitor system resources, perform overload protection and safe restarts, and lacks security protection mechanisms for core models, weights, and system files, making them vulnerable to tampering, deletion, and malicious damage. Therefore, it cannot meet the high reliability, high security, and long-term continuous operation requirements of critical mission scenarios. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a decentralized distributed autonomous artificial intelligence system that achieves decentralized networking, full-state persistence, fault self-healing, autonomous operation and maintenance, security protection and autonomous optimization, enabling the artificial intelligence system to operate stably, securely and continuously for a long time without human intervention.

[0004] The system of this invention includes: a system control unit, a spiking neural network processing module, a causal inference engine module, a P2P distributed node management module, a full-state persistence and self-healing recovery module, a system autonomous operation and maintenance module, a security protection and file protection module, and a peripheral interface layer.

[0005] The spiking neural network processing module is responsible for processing sensory signals, dynamic decision-making, and behavioral output; the causal reasoning engine module is responsible for constructing causal relationships of events, recording time-series information, and realizing interpretable reasoning; the P2P distributed node management module realizes automatic node discovery, online state maintenance, and distributed collaboration through broadcast protocols; the full-state persistence and self-healing recovery module is responsible for periodically generating full system snapshots and completing state reconstruction in case of anomalies; the system autonomous operation and maintenance module is responsible for resource monitoring, overload adjustment, and secure restart; and the security protection module realizes file verification, illegal file isolation, and core data protection through whitelists.

[0006] The beneficial effects of this invention are as follows: it achieves decentralized operation without a central authority, avoiding single points of failure; it has self-healing capabilities, automatically recovering operation after an anomaly; it supports full-state persistence, without losing operational data and historical experience; it has autonomous operation and maintenance capabilities, requiring no manual intervention; it has system-level security protection, preventing core files from being tampered with or deleted; and it supports cross-platform deployment, making it widely applicable to intelligent robots, industrial control, mission-critical systems, and large-scale intelligent clusters. Attached Figure Description

[0007] Figure 1 is a schematic diagram of the overall system architecture of the present invention; Figure 2 is a schematic diagram of the full-state persistence and fault self-healing process; Figure 3 is a schematic diagram of the P2P decentralized node self-discovery and synchronization process; Figure 4 is a schematic diagram of the system autonomous operation and maintenance and resource monitoring process; Figure 5 is a schematic diagram of the security protection and file protection process. Detailed Implementation

[0008] The present invention will be further described in detail below with reference to the accompanying drawings.

[0009] After system startup, the core modules are first initialized, including the spiking neural network structure, causal inference engine, P2P communication port, status snapshot directory, file protection whitelist, and resource monitoring thresholds. The system's central control unit coordinates all modules to ensure collaborative operation.

[0010] The P2P distributed node management module periodically sends heartbeat messages via UDP broadcast, which include the node's unique identifier, running status, and service port. It also listens for messages from other nodes in the network and automatically maintains the list of online nodes, achieving decentralized networking and status synchronization.

[0011] The full-state persistence module takes a full snapshot of the neural network weights, causal inference graph, node information, running status, optimization history and experience data at fixed intervals and stores them in binary serialization to the local persistent medium.

[0012] The system monitors its operational status, file integrity, and resource usage in real time. When a process anomaly, file corruption, or resource overload is detected, the self-healing module automatically reads the latest valid snapshot, verifies data integrity, and then reconstructs the neural network structure, inference model, node connections, and operational status, restoring the system to its pre-abnormal working state to continue executing tasks.

[0013] The system's autonomous operation and maintenance module obtains CPU, memory, and disk usage in real time. When these usages exceed preset thresholds, it initiates a load reduction strategy and performs a safe restart in case of severe overload, ensuring long-term stable operation of the system.

[0014] The security protection module periodically scans the protected directory to verify whether files are in the whitelist or legal path, and performs backup isolation and cleanup on illegal files to prevent core code, models and weights from being maliciously tampered with, deleted or replaced.

[0015] This invention does not rely on a centralized server and can achieve distributed collaboration, fault self-healing, autonomous operation and maintenance, and security protection without human intervention. It is suitable for scenarios that require high reliability, high security, and long-term continuous operation, such as intelligent robots, autonomous driving, industrial control, mission-critical systems, and large-scale intelligent clusters.

Claims

1. A decentralized, distributed, autonomous artificial intelligence system, characterized in that, include: The system comprises a central control unit, a spiking neural network processing module, a causal inference engine module, a P2P distributed node management module, a full-state persistence and self-healing recovery module, a system autonomous operation and maintenance module, and a security protection and file protection module. The system achieves persistent runtime data through full-state snapshots, decentralized node collaboration through P2P protocols, automatic crash recovery through a self-healing mechanism, self-management through autonomous operation and maintenance, and core data security protection through file protection.

2. The system according to claim 1, characterized in that, The P2P module uses UDP broadcast to achieve automatic node discovery, automatic online status, offline detection, and status synchronization.

3. The system according to claim 1, characterized in that, The self-healing recovery module can automatically rebuild the full running state from a snapshot after a process crash, file corruption, or power failure and restart.

4. The system according to claim 1, characterized in that, The autonomous operation and maintenance module monitors CPU and memory resources in real time and supports overload protection and safe restart.

5. The system according to claim 1, characterized in that, The security protection module uses a whitelist mechanism to verify core files, automatically clean up illegal files, and isolate and back them up.

6. The system according to claim 1, characterized in that, The system supports cross-platform operation and can be deployed on x86 servers, embedded devices, and intelligent robot hardware.

7. The system according to claim 1, characterized in that, The full-state persistence module supports timed snapshot storage of neural network weights, causal inference graphs, node information, and running status.

8. A highly available, self-healing implementation method for decentralized distributed artificial intelligence, characterized in that, Includes the following steps: Start the system and initialize all modules; start P2P broadcasting to automatically discover nodes in the network and complete network formation; Periodically take snapshots of the system's full operating status and persist them. Real-time monitoring of system operation status and file integrity; automatic reading of the latest valid snapshot and system reconstruction when an anomaly is detected; initiation of resource monitoring to achieve overload protection and autonomous operation and maintenance; Start file protection to ensure the security of core data.

Images