250 likes | 263 Views
Safety Verification of Model Helicopter Controller Using Hybrid Input/Output Automata. Sayan Mitra MIT Hybrid Systems: Computation and Control Prague, Czech Republic 2003 Joint work with Yong Wang (U. Beijing), Nancy Lynch, Eric Feron. Verification Techniques. Algorithmic
E N D
Safety Verification of Model Helicopter Controller Using Hybrid Input/Output Automata Sayan Mitra MIT Hybrid Systems: Computation and Control Prague, Czech Republic 2003 Joint work with Yong Wang (U. Beijing), Nancy Lynch, Eric Feron MITLCS
Verification Techniques • Algorithmic • Model checking e.g.[Alur, et al. 95] • Automatic: HyTech • Essentially for finite-state systems, subclass of linear hybrid systems • Over approximating set of unsafe states [Bayen, et al. 02] • Deductive • Invariant assertions, simulation relations e.g. [Manna, Sipma 98] • Can accommodate infinite-state systems: STeP • Requires human effort • User interaction MITLCS
Talk Outline • Introduction٭ • Hybrid I/O Automata definitions • Specification of Quanser • Safety Verification • Conclusions MITLCS
The HIOA Model[Lynch, Segala, Vaandrager 01, 03] • General, mathematical modeling framework. • States, discrete transitions • Trajectories: Maps left closed intervals of time to variable values • Support for decomposing hybrid system descriptions: • External behavior: Models interaction of component with environment. • Composition: Synchronizes external actions, external “flows”; respects external behavior. • Levels of abstraction: Implementation notion • Can incorporate analysis methods from: • CS: Invariants, simulation relations, compositional methods. • Control theory: Invariant sets, stability analysis, robust control. MITLCS
Hybrid I/O Automaton • V = U Y X: Input, output, and internal (state) variables • Q: States, a set of valuations of X • Q : Start states • A = I O H: Input, output, and internal actions • D Q A Q: Discrete transitions • T: Trajectories for V. I O X U Y H MITLCS
Trajectory Axioms and Executions • Set T of trajectories is closed under: • Prefix • Suffix • Countable concatenation • fstate, lstate • Execution fragment: 0a11a22…, where: • Each iis a trajectoryof the automaton and • Each (i.lstate, ai ,i+1.fstate) is a discrete step. • Execution: • Execution fragment beginning in a start state. MITLCS
Model Helicopter System • Manufactured by Quanser • User controllers not necessarily safe, can crash the helicopter on the table. • Supervisory pitchcontroller needed to ensure safety. • Safe operating region • Saturated actuator outputs : Umin or Umax • Must contend with • Sensor errors • Actuator delay MITLCS
Helicopter System Actuator Plant Sensor θ0,θ1 U buffer, u dequeue now, next θ0,θ1 Sample Sample Sample Sample Sample Command(S) Command(S) Command(S) Supervisor UserCntrl Useroutput(Xu) Useroutput(Xu) mode, Xs , S, rt Xu MITLCS
Plant U θ0,θ1 Plant Variables: θ0 :Pitch angle θ1: Pitch velocity Trajectories: evolve:d(θ0) = θ1 d(θ1) = -Ω2cos θ0+ U Input bounds: Umin , Umax Safe Region: S = { s | θmin≤ s.θ0≤ θmax } θ0,θ1 MITLCS
Sensor Discrete transition: Sample(θ0d , θ1d) precondition: now = next and θ0dє [θ0- є0 , θ0+ є0 ] and θ1dє [θ1- є1, θ1- є1] effect: next = next + Δ Trajectories: evolve: d(now) = 1 stopping condition:now = next θ0 ,θ1 Sensor now, next } Nondeterministic choice Sample(θ0d , θ1d) MITLCS
User Controller • Arbitrarily bad user • On receiving Sample, • Useroutput(Xu) • Non deterministic choice, Xuє [Umin, Umax ] MITLCS
Actuator • Actuator delay Ta • modeled as a FIFO queue of Supervisor(User) outputs • buffer: length [Ta / Δ] • Enqueue S received from supervisor • Dequeue u from bufferhead, • u changes discretely • Made into piece-wise continuous output U MITLCS
Modeling Actuator Delay • Ta Currently modeled as a single discrete jump from Umin to Umax after time Ta. • Alternatively • Approximate exponential rise by adding k intermediate values in the buffer, for every command from the supervisor. • Output from buffer will change every Δ/k time. • Model as continuous function Ta MITLCS
Safe Operating Region θ1 S C R U I θ0 θmin θmax Assumption: Cannot cross I in Δ time. MITLCS
Supervisor Sample Supervisor Command(S) mode, Xs , S, rt Userout(Xu) • On receiving sample, computes Xs • If s is above I+then Xs = Umin • If s is below I-then Xs = Umax • On receiving useroutput(Xu), computes S • If mode = userthen • If s is inU then S = Xu • Else mode = supervisor ; S = Xs • If mode = supervisor then • If s is inI then S = Xu ; mode = user • Else S = Xs MITLCS
Safety Verification • Assertional Proofs • Reasoning based on current state of the system • Finding the invariants is challenging • Strengthen statement • Proofs are easy, for proving I • Base case: I • Discrete part: s a s’ є D, show I(s) implies I(s’) • Continuous part: closed τє T, show I(fstate(τ)) implies I(lstate(τ)) MITLCS
Key Lemmas • All trajectories are closed • Any trajectory τє T, ltime(τ) - ftime(τ) ≤ Δ. MITLCS
User mode θ1 S C A0 A1 A2 R AΔ U A0 = R For 0 ≤ t ≤ t’ ≤ Δ At’ At U AΔ I θ0 MITLCS
User mode Safety • Any reachable state in the user mode is within R. • Proof: • Discrete part is easy • Any closed trajectory τє T, if fstate(τ)є At then lstate(τ)є At-ltime(τ). MITLCS
buffer flushed, Supervisor mode kicks in. Cannot go outside R from U, in the user mode Returns to Iand mode switches back to user . mode switches to supervisor, but buffer contains stale user commands. Executions in User and Supervisor modes MITLCS
Supervisor mode Correct input to plant • If s is above I+then last [rt/Δ] entries in buffer are Umin • rt: stopwatch for supervisor mode • Similarly, s is below I-then … Umax Settling phase rt ≤ Ta • Any reachable state is within C • All trajectories starting from within R remains within C • Proof similar to User mode Recovery phase rt > Ta • Any reachable state is within C • Proof: At any point on boundary of C, the vector field points inwards MITLCS
Conclusions • Design of supervisory controller • Controller has been implemented [Ishutkina]. • Specification Language • Demonstration of HIOA framework • Specification • Compositional • Nondeterminism models uncertainties in devices or user inputs. • Purely assertional proofs • Discrete and continuous parts • CS and Control Theory techniques • Current/Future Work • Performance guarantees for mobile computing algorithms • Theorem prover support MITLCS
Thank You.Questions ? MITLCS
Current/Future Work • Incorporate control theory methods: • Invariant sets, Stability analysis using Lyapunov functions, robust control methods. • More examples: • Systems with more complicated discrete behavior and dynamics, e.g. mobile computing, embedded systems. • Develop analysis tools for HIOA programs: • Theorem-provers, automated tools • As extension to IOA toolset MITLCS