General quantitative design framework for multi-rate LCAs enabling DeSS

Develop a general quantitative design framework for multi-rate layered control architectures that achieve diversity-enabled sweet spots, rigorously capturing tradeoffs across layers and providing design procedures that coordinate slow high-level planning with fast low-level feedback.

Background

The paper argues that diversity in tasks and hardware across layers can create diversity-enabled sweet spots (DeSS), where the system attains near-optimal speed and accuracy through multi-rate coordination.

Using sensorimotor control as a case paper, the authors show how slow accurate planning (vision) and fast reflex layers can be combined, and assert that a unifying quantitative framework for designing such multi-rate LCAs is still lacking.