Cyber-Physical System Design Automation: A Tale of Platforms and - PowerPoint PPT Presentation

From Electronic Design Automation to Cyber-Physical System Design Automation: A Tale of Platforms and Contracts Pierluigi Nuzzo Ming Hsieh Department of Electrical and Computer Engineering University of Southern California, Los Angeles nuzzo@usc.edu In Honor of Alberto Sangiovanni-Vincentelli International Symposium on Physical Design, San Francisco, April 16, 2019

Cyber-Physical System Design: What Can Go Wrong? 2 Pierluigi Nuzzo, USC

The Quest for the Next Level of Abstraction: System Level Design Courtesy: A. Sangiovanni-Vincentelli 3 Pierluigi Nuzzo, USC

Platform-Based Design Contracts Applications What’s next? 4 Pierluigi Nuzzo, USC

Cyber-Physical System Design: State of the Art Conventional V&V techniques do not scale to highly complex or “Let’s Get Physical: Computer Science Meets Systems,” ETAPS Workshop, 2014 adaptable systems Virtual Experienced Integration architects must Cost System Functional System Verification rely on accrued Optimization Specification Architecture & Validation knowledge and (V&V) heuristics to take risky decisions . . . Size/Power Subsystem Subsystem Optimization Design Testing Power Data & Control Thermal Management Controller Component Component Design Testing P TOLEMY II Networking Sensors A large number of poorly integrated Actuators V ERILOG languages and tools VHDL Physical system (plant) Embedded system (computation) 5 Pierluigi Nuzzo, USC

Learning from Logic Synthesis High level function model Gate library (platform) - Separation of function and architecture t 4 ’ - Common language for functional and architectural level netlists (Boolean t 1 t 3 +fgh logic, NAND2 gate) at 2 +c - Automatic mapping inv(1) nand2(2) d+e b+h Technology Mapping nor(2) nand3 (3) (covering ) aoi21 (3) restructuring restructuring f oai22 (4) xor (5) c d’ e’ nand3(3) f g g and2(3) a d Gate library Function nor3 (3) F e b’ h’ in netlist oai21(3) model in h b netlist a oai21 (3) nand2(2) c inv(1) Mapped design Courtesy: A. Sangiovanni-Vincentelli 6 Pierluigi Nuzzo, USC

Platform-Based Design Application Space: System Specification Performance Reliability Safety System Requirements Synthesis (Optimization) Behavioral and Non-Functional Models LNA LNA Sensors Networks Processors Controllers Actuators Implementation Space: Platform Library [A. Sangiovanni-Vincentelli and A. Ferrari, ‘90] 7 Pierluigi Nuzzo, USC

Automotive Avionics Smart Buildings Platform Platform Design- Instance Space Export Function Platform Mapped Synthetic Biology (Architectural) Space Space Function Platform Instance Instance Platform Function Mapped (Architectural) Space Space Function Platform Instance Instance Mixed-Signal Systems on Chip ASV Triangles 8 Pierluigi Nuzzo, USC

Platform-Based Design With Contracts Application Space: System Specification Performance Reliability Requirement Safety Formalization System Refinement Requirements Rules Synthesis (Optimization) Contracts Behavioral Abstraction and Non-Functional Models Rules Composition Rules LNA LNA Sensors Networks Processors Controllers Actuators Implementation Space: Platform Library 9 Pierluigi Nuzzo, USC

Assume/Guarantee (A/G) Contracts Environment Contracts are Assume-Guarantee v out Gain: 10 pairs v in Component – Component properties are guaranteed under a set of Assumptions: |𝒘 𝒋𝒐 | ≤ 𝟑 Guarantees: 𝒘 𝒑𝒗𝒖 = 𝟐𝟏𝒘 𝒋𝒐 assumptions on the environment – Global properties of systems are System Design derived based on local properties of the components Raclet McMillan Sangiovanni ‘ 12 Misra ‘ 81 Meyer ‘ 92 ‘ 09 Henzinger ‘ 97 ‘ 08 Time Nuzzo ‘ 09 Lamport ‘ 83 Clarke ‘ 98 Benveniste ‘ 08 Henzinger ‘ 01 Software Engineering and Verification 10 Pierluigi Nuzzo, USC

A Rigorous Calculus for Modular and Hierarchical Design Conjunction ∧ System Modular Requirements Requirement verification of Refinement “global” properties ≽ ⊗ of systems out of Composition local properties of Component Component Component components Req. Req. Req. ⊨ Satisfaction Step-wise Component Component Component refinement of large, Design Design Design complex architectures System Design reuse Design 11 Pierluigi Nuzzo, USC

Vertical Contracts Horizontal Contracts : How to check or enforce compatibility? Vertical Contracts: How to check or enforce consistency between the two levels? Think about the role of design rules in physical design 12 Pierluigi Nuzzo, USC

Electric Power System (EPS) in TerraSwarm “More - Electric” Aircraft 13 Pierluigi Nuzzo, USC

Aircraft Electric Power System Design Design architecture, i.e., the set of Generators Batteries AC Buses DC Buses Rectifiers Transformers Transformers & Rectifiers Contactors Loads and their interconnections … and the control algorithm under safety, reliability and real-time performance requirements Typical requirement: The probability that a critical bus is unpowered for more than 70 ms shall Loads be smaller than 10 -9 … …less than 1 failure per 100,000 years Single Line Diagram modified of operation! from Honeywell Patent “A Contract - Based Methodology for Aircraft Electric Power System Design,” IEEE Access, 2014 “A Platform -Based Methodology with Contracts and Related Tools for the Design of Cyber-Physcal Systems,” Proc. IEEE, 2015 14 Pierluigi Nuzzo, USC

Methodology and Tools: 1. No AC bus shall be Summary simultaneously powered by more than one AC source. 2. The aircraft electric power system shall provide power with the following characteristics: 115 Top-level Specification +/- 5 V (amplitude) and 400 Hz (frequency) for AC loads and 28 +/-2 V for DC loads. C ver/sim C A,syn C C,syn 3. The failure probability at an essential load must be less than 10 -9 during a mission. 4. DC buses shall not be Architecture Static/ unpowered for more than 70 ms. Design Extra-functional Component Discrete Event and Control Hybrid Design Verification and Simulation-Based Continuous Design Space Exploration Time and Hybrid Component and Contract Lower-level Implementation Library “Methodology and Tools for Next Generation Cyber -Physical Systems: The iCyPhy Approach,” P. Nuzzo, A. Sangiovanni- Vincentelli, R. Murray, INCOSE 2015 15 Pierluigi Nuzzo, USC

Aircraft Power System Design with CHASE Logic specification are up to 4,500 literals in size Inconsistent when time is less than 20 ms “CHASE: Contract -Based Requirement Engineering for Cyber- Physical System Design,” P. Nuzzo et al., DATE, 2018 Demonstrated reasoning about temporal properties of networks and integration with Natural Language Processing tools (IBM W ATSON ) 16 Pierluigi Nuzzo, USC

Optimized Selection of Reliable and Cost-Effective Architectures Dreamliner-like power system Application space based on Honeywell patent reproduced in ~4 min Optimization (MILP) Final architecture (topology, routing, mapping) Implementation space (library) Architecture exploration of aircraft air management systems “Optimized Selection of Reliable and Cost -Effective Cyber- Physical System Architectures,” DATE’14 “A Mixed Discrete -Continuous Optimization Scheme for Cyber- Physical System Architecture Exploration,” ICCAD’15

Reasoning About Software and Dynamics: Satisfiability Modulo Convex Programming (SMC) Boolean Constraints SAT SAT + Convex Solvers SMT Solvers Mixed Integer Programming Convex Optimization Convex Constraints Controller Synthesis for Robotic Motion Planning Secure State Estimation [ICCPS’16, TAC 17, TECS 18] [CDC’16, HSCC’17, CDC’17, ICRA’19] 18 Pierluigi Nuzzo, USC

Stochastic Contracts for CPS Design with Uncertainty Expressed in Stochastic Signal Temporal Logic (StSTL) to support probabilistic constraints Balance expressiveness with tractability of verification and synthesis Probabilistic Environment “The battery charge level B shall not be less than 0.3 Model with probability larger than or equal to 0.95” GEN 1 GEN 3 GEN 2 C1 C5 C3 AC Bus 1 AC Bus 2 C2 C4 TRU TRU C6 C7 DC Bus 1 DC Bus 2 C8 C9 Battery 1 Battery 2 C10 C11 Sheddable Non-sheddable Sheddable Non-sheddable DC Loads 1 DC Loads 1 DC Loads 2 DC Loads 2 Stochastic Model of Aircraft Power Battery charge versus time System [TECS 19] (50 simulations) 19 Pierluigi Nuzzo, USC

What’s Next? Compositional (modular, hierarchical) abstractions for CPS design Computational tools for reasoning about the interaction between discrete and continuous models Dealing with uncertainty 20 Pierluigi Nuzzo, USC

Thank you 21