Bioregenerative life support systems introduce novel challenges for the development of model-based approaches to their control given the varying characteristic of the biological processes that constitute them. Switching control paradigms provide an alternative to manage such uncertainty by allowing flexibility into the control path, enabling different control modes depending on the situation of the system. This paper presents a perceptionbased approach that combines sensor information to define those conditions and act upon them. Combined sensor information creates sensing spaces in which the operational conditions of the system are found. The decomposition of the sensing spaces into perceptual elements or granules allows for situation assessment, system integration strategies, and the implementation of fail-safe and fail-operational mechanisms -all these critical in a wider range of complex socio-technical systems. This paper proposes the use of intelligent agents based on fuzzy associative memories (FAM's) to decompose sensing spaces into granular structures composed of n-dimensional non-interactive fuzzy sets. Granular structures resulting from such decomposition allow for the incremental development and automation of the system by associating a control task to each operational condition. Furthermore, the real-time information obtained from the membership value of the granules may provide a resource for situational awareness and for the design of new ecological interfaces to enhance humansystem interaction and real-time decision making. The approach presented in this paper is applied to the dynamic model of a reconfigurable aquatic habitat that serves as a small-scale bioregenerative test bed for life support control research. Results show how information generated by the FAM enhances the situation observability of the system.

This paper presents a reactive agent built under the Subsumption Architecture, which purpose is to control a Ball and Beam system. The architecture of the agent is implemented making use of a fuzzy associative memory (FAM) and feedback control laws. The FAM is used to address the action selection problem known for this architecture. The FAM defines different fuzzy conditions. These conditions are associated to feedback control laws. The integrated control signal is obtained by defuzzifying the FAM through the weighted average technique. The timeresponse of the agent is simulated on a model of the Ball and Beam system and validated with an implementation at the laboratory. The results show the behavior of the variables of the system and the dynamic characteristics of the agent. The control signals stay active at every moment and the FAM allows smooth transitions between them, depending on the fuzzy condition of the system. 89978-1-4244-1608-0/07/$25.00 ©2007 IEEE.

This paper presents the design, modeling, and simulation of a reconfigurable aquatic habitat for experiments in regenerative life support automation; it supports the use of aquatic habitats as a small-scale approach to automation experiments relevant to largerscale regenerative life support systems. The habitat consists of a ten-gallon tank with four compartments, containing animal and botanical elements. The water volume serves as the medium through which life-support compounds, like oxygen, are transferred between organisms. A motorized hatch allows reconfiguration of the system to allow or prevent the exchange of gases with the atmosphere, and enables the study of fail-safe automation mechanisms. Sensors and actuators measure and intervene to regulate life support variables in the water. The model serves as an analytical reference for future tests in hardware settings, and to test advanced control architectures and policies that enable the system to operate safely and with increasing levels of autonomy, allowing for human intervention if necessary. The goal of the aquatic habitat is to enable life support control concepts that may be challenging to test in larger-scale life support systems. The mathematical description of the dynamic model of the system is presented in this paper with results from simulations of a distributed control approach applied to the habitat.

No abstract

The automation of bioregenerative life support systems poses challenges for the development of model-based approaches given the varying characteristic of the biological processes that constitute them. Switching control paradigms offer an alternative for the management of such uncertainty by introducing flexibility into the control path and allowing for different control modes depending on the operational conditions of the system. This paper presents a perception-based switching control strategy that makes use of sensor information to define and act upon those conditions. Abundant sensor information gives rise to sensing spaces in which the operational conditions of the system are found. A decomposition of the sensing spaces into perceptual elements allows for automation and integration strategies, and for the implementation of fail-safe and fail-operational mechanisms. This paper proposes the use of agents based on fuzzy associative memories to decompose sensing spaces into granular structures composed of n-dimensional non-interactive fuzzy sets. The granular structures allow for the incremental development and automation of the system by associating a control task to each granule. The method presented in this paper is applied to the dynamic model of a reconfigurable aquatic habitat. The habitat serves as a small-scale bioregenerative test bed for life support control research. The method used in this paper may also enable cognitive resources to enhance human interaction with the system.