Multi-Processing and Other Ideas for the Future

Abstract

It has been implicit throughout the earlier chapters of this monograph that Prolog+ is used to control a single image processor. However, there are many occasions when a single view of an object, or scene, does not provide sufficient data to enable a sensible decision to be made about it. Consider the system shown in Figure 9.1, in which there are several cameras, viewing a very complex object. Each of the cameras is connected to its own dedicated image processor. These image processors must, of course, be coordinated together. For example, they must be synchronised, to capture images and to begin processing at the appropriate moment. In addition, their outputs must be combined, even if this is as simple a process as logically ANDing them together. Of course, it is an unnecessary restriction to expect that each slave processor will always run the same program; their actions need to be coordinated at a much higher conceptual level. Can this be achieved by some suitable extension to Prolog+, or is the framework that we have described restrictive in some way? Does Prolog+ need to be modified in order to build intelligent multi-processor systems?