Opportunistic Intermittent Control with Safety Guarantees for Autonomous Systems

Control schemes for autonomous systems are often designed in a way that anticipates the worst case in any situation. At runtime, however, there could exist opportunities to leverage the characteristics of specific environment and operation context for more efficient control. In this work, we develop an online intermittent-control framework that combines formal verification with model-based optimization and deep reinforcement learning to opportunistically skip certain control computation and actuation to save actuation energy and computational resources without compromising system safety. Experiments on an adaptive cruise control system demonstrate that our approach can achieve significant energy and computation savings.

[1]  David Silver,et al.  Deep Reinforcement Learning with Double Q-Learning , 2015, AAAI.

[2]  Rupak Majumdar,et al.  From Iteration to System Failure: Characterizing the FITness of Periodic Weakly-Hard Systems , 2019, ECRTS.

[3]  Jin Jiang,et al.  Fault-tolerant control systems: A comparative study between active and passive approaches , 2012, Annu. Rev. Control..

[4]  Ashish Tiwari.  Approximate Reachability for Linear Systems , 2003, HSCC.

[5]  Mahesh Viswanathan,et al.  Analyzing Real Time Linear Control Systems Using Software Verification , 2015, 2015 IEEE Real-Time Systems Symposium.

[6]  Alberto Bemporad,et al.  An algorithm for multi-parametric quadratic programming and explicit MPC solutions , 2003, Autom..

[7]  Daniel Krajzewicz,et al.  Recent Development and Applications of SUMO - Simulation of Urban MObility , 2012 .

[8]  Wenchao Li,et al.  Exploring weakly-hard paradigm for networked systems , 2019, DESTION@CPSIoTWeek.

[9]  Ufuk Topcu,et al.  Safe Reinforcement Learning via Shielding , 2017, AAAI.

[10]  David Q. Mayne,et al.  Constrained model predictive control: Stability and optimality , 2000, Autom..

[11]  Luigi Chisci,et al.  Systems with persistent disturbances: predictive control with restricted constraints , 2001, Autom..

[12]  Sebastian Junges,et al.  Safety-Constrained Reinforcement Learning for MDPs , 2015, TACAS.

[13]  Goran Frehse,et al.  Formal Analysis of Timing Effects on Closed-Loop Properties of Control Software , 2014, 2014 IEEE Real-Time Systems Symposium.

[14]  Nathan Fulton,et al.  Safe Reinforcement Learning via Formal Methods: Toward Safe Control Through Proof and Learning , 2018, AAAI.

[15]  Paulo Tabuada,et al.  Computing Robust Controlled Invariant Sets of Linear Systems , 2016, IEEE Transactions on Automatic Control.

[16]  J. Löfberg.  Minimax approaches to robust model predictive control , 2003 .

[17]  Shengchao Qin,et al.  Hierarchical Model Predictive Control for Multi-Robot Navigation , 2016, IJCAI.

[18]  Yongduan Song,et al.  Adaptive Fault-Tolerant PI Tracking Control With Guaranteed Transient and Steady-State Performance , 2017, IEEE Transactions on Automatic Control.

[19]  Wenchao Li,et al.  Formal verification of weakly-hard systems , 2019, HSCC.

[20]  Arthur G. Richards,et al.  Robust constrained model predictive control , 2005 .

[21]  Jianglin Lan,et al.  A new strategy for integration of fault estimation within fault-tolerant control , 2016, Autom..

[22]  Olivier Bournez,et al.  Approximate Reachability Analysis of Piecewise-Linear Dynamical Systems , 2000, HSCC.

[23]  Parameswaran Ramanathan,et al.  A Dynamic Priority Assignement Technique for Streams with (m, k)-Firm Deadlines , 1995, IEEE Trans. Computers.

[24]  David Q. Mayne,et al.  Invariant approximations of the minimal robust positively Invariant set , 2005, IEEE Transactions on Automatic Control.