• real-time multi-transducer networks
• pattern recognition
• neural networks
• multi-task data processing
• life-like artifact
• learning
• imitation
• human-machine interface
• data analysis
• biomimetics
• robot-based treatment method
• sensory fusion
• social attention

The advances in sensor and actuator technology foster the use of large multitransducer networks in many different fields. The increasing complexity of such networks poses problems in data processing, especially when high-efficiency is required for real-time applications. In fact, multi-transducer data processing usually consists of interconnection and co-operation of several modules devoted to process different tasks. Multi-transducer network modules often include tasks such as control, data acquisition, data filtering interfaces, feature selection and pattern analysis. Heterogeneous techniques derived from chemometrics, neural networks, fuzzy-rules used to implement such tasks may introduce module interconnection and co-operation issues. To help dealing with these problems the author here presents a software library architecture for a dynamic and efficient management of multi-transducer data processing and control techniques. The framework’s base architecture and the implementation details of several extensions are described. Starting from the base models available in the framework core dedicated models for control processes and neural network tools have been derived. The Facial Automaton for Conveying Emotion (FACE) has been used as a test field for the control architecture.