A New Sensory System for Modeling and Tracking Humans Within Industrial Work Cells

This paper introduces a new sensory system for the 2-D detection, tracking, and prediction of human motions. Using the data acquired through the sensory system, a model of a human based on a superquadric modeling technique is obtained. The sensory system requires a minimum computational burden for its measurements and almost no reengineering efforts for the integration with existing systems. The sensory system comprises a floor mat for the repeated measurement of the footprints of a human either standing or walking. An intelligent firmware for inferring the required information and identification of the human orientation and for rendering a real-time human model based on primitive data received from the mat is discussed. Furthermore, the integration of the floor mat in a multisensory system for obtaining a 3-D human model is discussed. Several results are presented to validate the effectiveness of the sensory system.