There is an increased interest in the application of control to combustion. The objective is to optimize combustor operation. monitor the process and alleviate instabilities and their severe consequences. One wishes to improve the system performance, for example by reducing the levels of pollutant emissions or by smoothing the pattern factor at the combustor exhaust. In other cases, the aim is to extend the stability domain by reducing the level of oscillation induced by coupling between resonance modes and combustion. As combustion systems have to meet increasingly more demanding air pollution standards, their design and operation becomes more complex. The trend towards reduced NOx, levels has led to new developments in different fields. Automotive engines and gas turbine combustors are considered in this article. In the first case. complex exhaust aftertreatment is being applied and dedicated engine control schemes are required to ensure and maintain high pollutant conversion efficiency. For gas turbines, premixed combustors, which operate at lower local temperatures than conventional systems have been designed. In both cases, monitoring and control of the operating point of the process have to be achieved with great precision to obtain the full benefits of the NOx reduction scheme. For premixed combustors operating near the lean stability limit, the flame is more susceptible to blowout, oscillation or flashback. Research is now carried out to reduce these dynamical problems with passive and active control methods. In addition to a broad range of fundamental problems raised by Active Combustion Control (ACC) and Operating Point Control (OPC), there are important technological issues. This paper contains a review of some facets of combustion control and focuses on the sensors that take or could take part to combustion control solutions. The current status of ACC and OPC is presented together with the associated control concepts. The state of the art in sensors is reviewed and their applicability is evaluated. Research efforts in combustion diagnostics are to a certain extent devoted to the development of sensors for control applications. The objective of such developments differs from that which is pursued when one wishes to perform detailed measurements on a laboratory scale experiment. The sensor system should not necessarily provide quantitative measurements because relative data are already useful for control purposes. This change of orientation will be discussed and illustrated by examples of current interest. It is concluded that development in control will depend critically on the availability of sensors and on their reliability, robustness, immunity to noise and capacity to operate in a harsh environment. Research is needed on the fundamentals of ACC and OPC but it should also address the more technical aspects of the problem. (C) 2002 Elsevier Science Ltd. All rights reserved.