State Consistencies for Cyber- Physical System Recovery Fanxin Kong, - - PowerPoint PPT Presentation

State Consistencies for Cyber- Physical System Recovery

Fanxin Kong, S

yracuse Universit y

Oleg S

• kolsky, James Weimer, Insup Lee, Universit y of

Pennsylvania

April 15, 2019

Department of Electrical Engineering and Computer S cience

2

Cyber-Physical Systems

We are living in a Cyber-Physical System world!

3

Security

4

CPS Attack Surfaces

• Cyber attack surfaces

‐ e.g., communication, networks, computers, ...

• Environmental attack

surfaces

‐ e.g., GPS signal, electro- magnetic interference, ...

• Physical attack surfaces

‐ e.g., locks, casings, cables, …

• Human attack surfaces

‐ e.g., phishing, blackmail, …

5

What we study and why?

Target: Sensor Attacks

• The attacker can arbitrarily

change sensor measurements

• environmental attack surfaces
• cyber attack surfaces

Controller Physical system Sensor Actuator Network Malicious signals Malicious packets 30mi/h 100mi/h

6

What we study and why?

Target: Sensor Attacks

• The attacker can arbitrarily

change sensor measurements

• environmental attack surfaces
• cyber attack surfaces

Controller Physical system Sensor Actuator Network Malicious signals Malicious packets

Goal: Resilience

• To ensure control performance

under sensor attacks

7

Ideally…

• Ideally, the system performs (almost) the same as

if there is no attack

• Example: cruise control under a speed sensor attack

Speed sensor attack

8

Outline

• Background
• Review on CPS

recovery

• Roll‐forward recovery
• How well does it work
• S

tate consistencies for CPS recovery

• Consistency definitions
• Evaluation
• Conclusion

9

CPS recovery

Roll-forw ard recovery: Rolling the system to the current tim e by starting from a consistent cyber-physical-state

• Estimated

speed Prediction using historical state

• Example: model-based prediction (ICCPS2018)

10

Scenario: travelling in a straight line

• Testbed: an unmanned vehicle. Each

front wheel is driven by a motor, and each motor has a speed sensor

• Goal: to keep a vehicle travel in a

straight line, i.e., the two front wheels have the same speed

• Controller: a PID controller supervises and controls

the speed difference of the two front wheels

• Attack: the attacker modifies a speed sensor’s

measurements to a constant value

11

How well does it work?

No protection With protection

speed difference speed difference

• -- desired ∆ ―actual ∆

recovery large

The vehicle keeps turning

small

The vehicle travels almost straightly

12

What kind of states is used?

• Cyber-physical st at es: the cyber information

that reflects physical states

• Cyber-physical consist ency: whether the

physical state can be accurately reflected by the corresponding cyber information

We use Consistent Cy ber-Phy sica l Sta tes

Cyber‐physical logic‐consistency Cyber‐physical timing‐consistency Synchronization Freshness

13

A system diagram of CPS

Physical System Controller Physical space Cyber space A cyber‐physical state is denoted as

14

Cyber-Physical Logic-Consistency

The logic‐consistency is confined to values, is NOT enough.

𝑦 𝑦 𝑦 𝑦̅

• sample

sample 𝑗 1 𝑈

• 𝑗𝑈
• 𝑦̅

𝑦̅

15

Cyber-Physical Timing-Consistency

16

(1) Syn-Timing-Consistency (1/ 2)

• sample

sample

• sample

sample

• NO

YES

17

(1) Syn-Timing-Consistency (2/ 2)

• sample

sample actuate actuate

• sample

actuate time

• : NO
• : YES

18

(2) Exp-Timing-Consistency

Calculating the expire tim e time The error of state prediction is unacceptable

19

Evaluation

• Simulator: DC motor speed control using PID controller
• Goal: to keep a vehicle travel at a constant speed
• Scenario: an attack is found out and the system

performs recovery ONCE to predict the current state

20

Violating Logic-Consistency

21

Violating Syn-Timing-Consistency

Current ( ) and speed ( ) have different tim e stam ps

22

Need of Exp-Timing-Consistency

Using older states for recovery resulting in larger drifts

23

Conclusion

• Review on CPS

recovery

• Model‐based roll‐forward recovery
• How well does it work
• S

tate consistencies for CPS recovery

• Defined logic and timing consistencies
• Why the consistencies is needed

Thank you! Q&A