Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Multi-agent predictive control barrier functions
In this project, we want to explore possible extensions of predictive control barrier functions to the multi-agent setting. Predictive control barrier functions [1] allow certifying safety of a system in terms of constraint satisfaction and provide stability guarantees with respect to the set of safe states in case of initial feasibility. This allows augmenting any human or learning-based controller with closed-loop guarantees through a so-called safety filter [2] which is agnostic to the primary control objective. As current formulations are restricted to single agents, the goal is to investigate how this formulation can be extended for multi-agent applications and how the interactions between the agents can be exploited in order to reduce computational overhead.
Keywords: predictive control, multi-agent systems, safety filter, control barrier functions
Multi-agent systems exist in various domains, including robotics, autonomous vehicles, and smart grids. Coordinating the actions of multiple agents in such systems while ensuring safety and stability is a fundamental challenge addressed, e.g., through model predictive control strategies in [3]. Predictive control barrier functions (PCBF) offer a solution which inherits the benefits of predictive control methods as well as the properties of control barrier functions [4]. By minimizing constraint violations, it can be ensured that safety of a system is guaranteed for all times if the initial state is safe and that the set of safe states is stabilized if the initial condition does not satisfy the constraints. The PCBF solution is then used in a safety filter optimization problem to filter potentially unsafe inputs which are proposed by an external agent to the system. This research project seeks to extend the capabilities of PCBF in multi-agent systems, possibly considering distributed control architectures where agents operate autonomously with limited communication while achieving common control objectives and respect safety considerations.
Prerequisites:
-Background in System Theory and Model Predictive Control. Advanced Model Predictive Control course is a plus.
[1] Wabersich, Kim Peter, and Melanie N. Zeilinger. "Predictive control barrier functions: Enhanced safety mechanisms for learning-based control." IEEE Transactions on Automatic Control (2022).
[2] Wabersich, Kim Peter, and Melanie N. Zeilinger. "A predictive safety filter for learning-based control of constrained nonlinear dynamical systems." Automatica 129 (2021): 109597.
[3] Conte, Christian, et al. "Distributed synthesis and stability of cooperative distributed model predictive control for linear systems." Automatica 69 (2016): 117-125.
[4] Ames, Aaron D., et al. "Control barrier function based quadratic programs for safety critical systems." IEEE Transactions on Automatic Control 62.8 (2016): 3861-3876.
Multi-agent systems exist in various domains, including robotics, autonomous vehicles, and smart grids. Coordinating the actions of multiple agents in such systems while ensuring safety and stability is a fundamental challenge addressed, e.g., through model predictive control strategies in [3]. Predictive control barrier functions (PCBF) offer a solution which inherits the benefits of predictive control methods as well as the properties of control barrier functions [4]. By minimizing constraint violations, it can be ensured that safety of a system is guaranteed for all times if the initial state is safe and that the set of safe states is stabilized if the initial condition does not satisfy the constraints. The PCBF solution is then used in a safety filter optimization problem to filter potentially unsafe inputs which are proposed by an external agent to the system. This research project seeks to extend the capabilities of PCBF in multi-agent systems, possibly considering distributed control architectures where agents operate autonomously with limited communication while achieving common control objectives and respect safety considerations.
Prerequisites:
-Background in System Theory and Model Predictive Control. Advanced Model Predictive Control course is a plus.
[1] Wabersich, Kim Peter, and Melanie N. Zeilinger. "Predictive control barrier functions: Enhanced safety mechanisms for learning-based control." IEEE Transactions on Automatic Control (2022).
[2] Wabersich, Kim Peter, and Melanie N. Zeilinger. "A predictive safety filter for learning-based control of constrained nonlinear dynamical systems." Automatica 129 (2021): 109597.
[3] Conte, Christian, et al. "Distributed synthesis and stability of cooperative distributed model predictive control for linear systems." Automatica 69 (2016): 117-125.
[4] Ames, Aaron D., et al. "Control barrier function based quadratic programs for safety critical systems." IEEE Transactions on Automatic Control 62.8 (2016): 3861-3876.
-Investigate multi-agent predictive control and control barrier function methods.
-Derive a multi-agent formulation for the predictive control barrier function method for linear dynamical systems.
-Investigate possible distributed control architectures.
-Validate the proposed approaches through simulations and possibly real-world experiments on the group’s miniature race care platform.
-Investigate multi-agent predictive control and control barrier function methods.
-Derive a multi-agent formulation for the predictive control barrier function method for linear dynamical systems.
-Investigate possible distributed control architectures.
-Validate the proposed approaches through simulations and possibly real-world experiments on the group’s miniature race care platform.
Please send your CV and your transcript to: Alexandre Didier adidier@ethz.ch; Dr. Ahmed Aboudonia ahmedab@control.ee.ethz.ch; Dr. Andrea Carron carrona@ethz.ch
Please send your CV and your transcript to: Alexandre Didier adidier@ethz.ch; Dr. Ahmed Aboudonia ahmedab@control.ee.ethz.ch; Dr. Andrea Carron carrona@ethz.ch
