Team AVR - Control, Vision and Robotics Lab

Event-based Control

From Team AVR - Control, Vision and Robotics Lab
Jump to: navigation, search

Abstract

The classical (time-triggered) discrete time framework of digital controlled systems consists in sampling the system uniformly in time with a constant sampling period. Although periodicity simplifies the design and analysis, it results in a conservative usage of resource, because the control law is computed and updated at the same rate regardless it is really required or not. Alternatively, some works addressed more recently resource-aware implementations of the control law, where the control law is event-driven. Typically, when the system variables do not change significantly, it is desirable to limit sensor sampling and control actuation.

The event-based paradigm is well-motivated for limited-resource systems, in particular for embedded and networked control systems, since it allows to relax the frequency of computations (and consequently the CPU utilization and communications) for the same final performance. However, an event-based control design does not only consists in deciding how to actuate on the system, but also when to actuate. Therefore, an event function determines the time instants, called "events", when the control law needs to be updated. The control signal is kept constant between two successive events. As a consequence, the implementation of such strategies can result in additional challenges, like determining how frequently the control signal needs to be updated and applied such that the stability and performance properties are still guaranteed.

Staf

Contacts

Students

  • Imane Khayour, Master (2016), "Event-Based Control of a 1-DoF Cable-Driven Parallel Robot with Embedded Actuator"
  • Imane Khayour, PhD (since Oct. 2016), "Control of Cable-Driven Parallel Robots with Embedded Actuator"

Partners

  • Nicolas Marchand, GIPSA-lab, Grenoble, France
  • J. Fermi Guerrero-Castellanos, BUAP, Puebla, Mexico

Major facts

  • Part of the program committee for the IEEE International Conference on Event-Based Control, Communication, and Signal Processing (EBCCSP'15 & EBCCSP'16)

Funding

  • Projet interne ICube COREDRONE (Commande robuste et évènementielle de drones), 2016-2017

Main publications

  • B. Boisseau, S. Durand, J.J. Martinez-Molina, T. Raharijaona, N. Marchand. "Attitude Control of a Gyroscope Actuator using Event-Based Discrete-Time Approach" (Best Paper Award). Proceedings of the 1st IEEE International Conference on Event-Based Control, Communication, and Signal Processing (EBCCSP'15), special session on Event-Based Vision and Robotics. 2015.
  • S. Durand, B. Boisseau, J.J. Martinez-Molina, N. Marchand, T. Raharijaona. "Event-Based LQR with Integral Action". Proceedings of the 19th IEEE International Conference on Emerging Technologies and Factory Automation (ETFA'14), International Workshop on Event-Based Systems (EBS'14). 2014.
  • N. Marchand, S. Durand, J.F. Guerrero-Castellanos. "A General Formula for Event-Based Stabilization of Nonlinear Systems". IEEE Transactions on Automatic Control (TAC), Volume 58, Issue 5, pages 1332-1337. 2013.
  • S. Durand, N. Marchand. "Further Results on Event-Based PID Controller". Proceedings of the 10th European Control Conference (ECC'09). 2009.