FROM DATA-FLOW TASK TO MULTITASKING - APPLYING THE SYNCHRONOUS APPROACH TO ACTIVE VISION IN ROBOTICS

In this paper, we apply the synchronous approach to real time active v isual three-dimensional (3-D) reconstruction using a camera mounted on a robot end-effector, It illustrates the adequateness of SIGNAL, a sy nchronous data flow programming language and environment, for the spec ification of a system dealing with various domains in control theory a nd computer vision, More precisely, our application consists in the 3- D structure estimation of a set of geometrical primitives using an act ive vision paradigm, At the level of camera motion control, the visual servoing approach is specified and implemented in SIGNAL as a functio n from sensor inputs to control outputs, Furthermore, the 3-D structur e estimation method is based on the ''structure from controlled motion '' approach (constraining camera motion for optimal estimation). Its s pecification is made in parallel to visual servoing, and involves the delay mechanism of SIGNAL for the specification of filters, This recon struction involves to be focused on each object; we thus present a per ception strategy for sequencing several estimations, using task preemp tion and time intervals in SIGNAL, It allows us to consider in an unif ied framework the various aspects of the application: from dataflow ta sk specification to multitasking and hierarchical task preemption, The integration of these techniques is validated experimentally by their implementation on a robotic cell, Merits and drawbacks of the proposed framework with respect to more typical asynchronous approaches are al so discussed.