[en] These studies were designed to draw attention to signal handling techniques which have become applicable to the automatic control of reactors as a result of advances in microelectronics. The underlying concepts being explored are the extent to which more information may be obtained from existing sensors, whether changes in the state of a system may be acted upon automatically and the acceptability of stochastic methods of state estimation. If a given state may arise in a number of different ways, automatic action necessarily involves diagnosis of causes if any but the simplest actions, e.g. scram, is required. Up to now this diagnosis has been made by an operator, though a number of schemes intended to identify the primary cause of an abnormal state or to guide an operator to the most easily accessible safe state are in advanced states of development, e.g. at Halden, Westinghouse and GRS-Garching. The activity decribed is continuing, current emphasis being upon diagnostics using approaches based upon artificial intelligence concepts, system optimisation and quantification of the benefits of networking