Abstract
Active continuous control of systems using physical sensors are hindered by presence of corrupting noise in discretized measurements which degrades performance. This paper presents a systematic development of organized sensor networks that fuses the dynamic implementation of parallel and sequential network architectures with conventional multi-sensor filtering techniques to improve performance through collaborative network enhancement of noise suppression and sampling rate. The classical control of an inverted pendulum using vision sensors is presented as an illustrative example. Simulation results suggest that an organized sensor network with dynamic throttling outperforms a static network while minimizing overall sensor utilization.