Sensory architectures for biologically inspired autonomous robotics

Engineers have a lot to gain from studying biology. The study of biological neural systems alone provides numerous examples of computational systems t hat are far more complex than any man-made system and perform real-time sen sory and motor tasks in a manner that humbles the most advanced artificial systems. Despite the evolutionary genesis of these systems and the vast app arent differences between species, there are common design strategies emplo yed by biological systems that span taxa, and engineers would do well to em ulate these strategies. However, biologically-inspired computational archit ectures, which are continuous-time and parallel in nature, do not map well onto conventional processors, which are discrete-time and serial in operati on. Rather, an implementation technology that is capable of directly realiz ing the layered parallel structure and nonlinear elements employed by neuro biology is required for power- and space-efficient implementation. Custom n euromorphic hardware meets these criteria and yields low-power dedicated se nsory systems that are small, light, and ideal for autonomous robot applica tions. As examples of how this technology is applied, this article describe s both a low-level neuromorphic hardware emulation of an elementary visual motion detector, and a large-scale, system-level spatial motion integration system.