Multisensor Knowledge Systems

Abstract

We describe an approach which facilitates and makes explicit the organization of the knowledge necessary to map multisensor system requirements onto an appropriate assembly of algorithms, processors, sensors, and actuators. We have previously introduced the multisensor kernel system and logical sensor specifications as a means for high-level specification of multisensor systems. The main goals of such a characterization are to develop a coherent treatment of multisensor information, to allow system reconfiguration for both fault tolerance and dynamic response to environmental conditions, and to permit the explicit description of control.In this paper we show how logical sensors can be incorporated into an object-based approach for the interpretation of 3D structure. Considering the inherent difficulties in interpreting general configurations of lines in space, and considering the ubiquitousness of special line configurations in manufactured environments and objects, we advocate the use of computational units tuned to the occurrence of special configurations. The organized use of these units circumvents the inherent difficulties in interpreting general configurations of lines. After a brief examination of the problem of interpreting general configurations of lines in space, a number of computational units are proposed which are naturally derived from angular relations. The process of propagation (which allows interpretation to spread over the image) is also advocated. Such computational units and processes, which are simple and efficient, can be conveniently organized in a rule-based framework where the occurrence of the various special configurations can be tested. The multisensor knowledge system provides such a framework.

<title>Hybrid vector quantization/neural tree network classifiers for speaker recognition</title>

“…Data fusion has been used to combine inputs from multiple sensors in various applications including robotics [14] and handwriting recognition [15]. Numerous methods exist for combining the data of several classifiers.…”