A fault-tolerant model of wireless sensor-actuator network

In a wireless sensor and actuator network (WSAN), a group of sensor nodes, actuators, and actuation devices is geographically distributed and linked by wireless networks. Sensor nodes gather information for art event occurring in the physical world and send the sensed values to actuator nodes. Actuator nodes perform appropriate actions on actuation devices on receipt of sensed values front sensor nodes. Sensor nodes are a low cost, low-powered device with limited energy, computation, and wireless communication capabilities. Sensor nodes may suffer from arbitrary faults. Furthermore, wireless communications are unreliable due to collision and noise of a wireless channel and shortage of power of sensor nodes. Reliable efficient communication among sensor nodes, actuator nodes, and actuation devices is required in the presence of node and network faults. In order to realize the reliability and efficiency, we newly propose a multi-actuator/multi-sensor (MAMS) model where each sensor node sends sensed values to multiple actuator nodes and each actuator node receives sensed values from multiple sensor nodes for each event occurring in an event area. Even if some number of sensor nodes and actuator nodes are faulty and messages sent by sensor nodes and actuator nodes are lost, a required action can be performed on actuation devices. Actuator nodes are required to receive some messages from some number of sensor nodes to make a decision on what actions to be performed on actuation devices. Here, multiple actuator nodes may perform actions on receipt of sensed values. Multiple redundant executions of an action on each device have to be prevented and conflicting actions on each actuation device from multiple actuators have to be serialized. In this paper, we discuss a method to reliably and non-redundantly perform actions. On the other hand, there is one primary actuator node and multiple backup actuator nodes in the passive model. Only the primary actuator node performs actions. In this paper, we present a semi-passive coordination (SPC) model where not only primary but also backup actuator nodes receive messages from sensor nodes. Only one actuator node performs an action on an actuation device and a backup actuator node takes over the primary one if the primary gets faulty. In addition, multiple actuator nodes issue actions to each actuation device. We discuss how to resolve redundant execution of an action on an actuation device and serialize conflicting actions from multiple actuator nodes.