New generation of radar systems enabled with cognition
Simon Haykin · 2010
The analogy between the visual brain and radar has fueled Professor Simon Haykin's interest in studying the human brain for the past 25 years. He will address the formulation of the perceptionaction cycle that embodies five basic properties: Perception, Memory, Attention, Intelligence, and Language. Realizing that the language in radar information processing is different from that in the brain, it is natural to demand the first four properties of human cognition for a radar system to be cognitive. To satisfy this requirement, a hierarchical cognitive radar will be described that exploits local and global feedback loops around the receiver and transmitter, which are supplied with three distinct memory units: perceptual memory, executive memory, and action-perception matching memory. In so doing, the whole system is coordinated with its environment in a powerful way. Through the use of clever learning procedures, a hierarchical perception-action cycle is built. The stage is now set to present experimental results, which demonstrate how the tracking resolution capability of a radar system is progressively improved as the memory units in the system are added, one by one. Future research challenges include: • Improvements to the cubature Kalman filter in the receiver for perception of the environment. • Approximate dynamic programming in the transmitter for action to control the environment. • Bayesian target detection. • Multiple target surveillance. • Cognitive phased array radar, providing the ultimate in performance.