Space-time scaling laws for self-triggered control

Feedback control laws have been traditionally treated as periodic tasks when implemented on digital platforms. However, the growing complexity of systems calls for efficient implementations of control tasks that reduce resource utilization while keeping desired levels of performance. In this paper we drop the periodicity assumption in favour of self-triggered strategies for the execution of control laws. Such strategies determine the next execution time based on the current state of the plant. Under the self-triggered policy, the inter-execution times scale in a predictable manner: a scaling of the state of the plant entails a scaling in the inter-execution times. This property allows us to derive a simple formula for the next execution time guaranteeing performance. We illustrate the proposed techniques on the control of a jet engine compressor.