Development of new models of system – level diagnosis using methods and tools for building of artificial intelligence systems
Artificial Intelligence Systems
Теми:
1. System level diagnosis layered approach to fault tolerance for distributed applications
System level diagnosis using by message-passing computing nodes in high-performance platforms for distributed applications is considered. Specific application techniques are employed in implementation of fault tolerance by the application programs for distributed systems. The key innovation in the design is the implementation and utilization for fault tolerance the asynchronous communication between application processes and the cluster manager.
This mechanism allows the asynchronous delivery of messages, known as signals, to application processes. Signals can be initiated by the program manager or by application processes, and they can be atomically broadcasted to all processes of the application task or sent to the cluster manager.
2. Creation of a communication tree for system level diagnosis in wireless sensor networks
Some algorithms that creates a tree routing structures and maintains it for wireless sensor networks are considered. The diagnosis routing paths connecting nodes to the root are locally reconfigured against crash faulty nodes when information is delivered from sensor nodes to the control observer.
Energy efficiency and scalability are provided for the reconfigurations by using only locally available relational information among neighbor nodes, that does not need global maintenance throughout the tree. For routing around crash fault nodes, a communication tree structure connecting sensor nodes to the base station (sink or root) is dynamically reconfigured during information dissemination. The performance of the algorithms is compared to the single path with repair routing scheme
3. Self fault diagnosis in distributed sensor networks using dynamical methods
One major research focus in wireless sensor networks in the past decades has been to diagnose the sensor nodes to identify their fault status.
A distributed self fault diagnosis algorithm using neighbor coordination is considered to identify both hard and soft faulty sensor nodes in wireless sensor networks. The algorithm is distributed (runs in each sensor node), self diagnosable (each node identifies its fault status) and can diagnose the most common faults like stuck at zero, stuck at one, random data and hard faults.
In this algorithms each sensor node uses the fault diagnosis algorithms and evaluates states for dynamic topology networks in which sensor nodes join and leave the network during the diagnosis time.