Robotic Security Proposal Study
Existing Vulnerabilities
- sensor data spoofing
- tamper controller input
- ROS vulnerability: auth, DoS, BoF, etc.
- A summary:
Objectives and Specific Aims
The objective of this proposal is to design a secure and efficient multi-robot system, providing comprehensive protections for the workloads’ data and executions
Systematization of knowledge about vulnerabilities in robot applications: understand the problems through literature review; develop tools for security scanning.
1.1 Characterize the vulnerabilities in robot workloads
1.2 Develop software tools to evaluate and analyze robot workloads
Property-based attestation for multi-robot systems.
2.1 Identify security properties for robot applications.
2.2 Design and develop architecture for property-based attestation.
2.3 Design a blockchain-based decentralized attestation protocol.
Information flow tracking in multi-robot systems: detect security violations
3.1 Design and develop a two-level static data flow tracking methodology.
3.2 Design and develop a three-level dynamic control flow tracking technique.
Policy enforcement in multi-robot systems.
4.1 Identify security policies for robot applications.
4.2 Design and develop a policy-based execution enforcement methodology.
4.3 Design and develop a policy-based data encryption methodology.
Risk Analysis and Security Policy Enforcement
Summary
Identify and categorize three kinds of interaction risks; detect and mitigate these risks through security policies; Evaluation
Function Interaction Analysis
- Categorization: Perception, Planning, Control, Drivers, Others
- App analysis: Remote Control, 2D/3D mapping, Navigation, SLAM (map/localization), Face/Person Detection
Risk Analysis
- General Risks
- Robot Specific Risks
- App Specific Risks
Mitigation Methodology
- Coordination Node
- Security Policy
Evaluation
- Risk Identification
- Risk Mitigation: (1) CN analysis (2) Policy Selection (3)
- Overhead
Evaluation Metrics
Exploration
time, cost (sub of distance travelled by each robot), efficiency (information over cost), map completeness, map quality
Time and map completeness seems to be two counter metrics, where limiting one will cause impact on the other.
Navigation
Number of successful tasks, path length, time, collisions, obstacle clearance, etc.
Security related, trajectory towards goal, smoothness of trajectory.
Questions for Multi-robot System
Materials
Multi-robot system course: http://u.cs.biu.ac.il/~yehoshr1/89-689/
multi-robot system example
- Multiple robot car path search: http://wiki.ros.org/tuw_multi_robot
- IEEE: Localization Strategy Based on Multi-Robot Collaboration for Indoor Service Robot Applications
- L-ALLIANCE: TASK-ORIENTED MULTI-ROBOT LEARNING IN BEHAVIOR-BASED SYSTEMS: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.17.9595&rep=rep1&type=pdf
- A multi-robot system discussion on ROS forum: https://answers.ros.org/question/41627/multi-robot-systems-framework-in-ros/
- Multi TIAGo Simulation: http://wiki.ros.org/Robots/TIAGo/Tutorials/Multi_TIAGo_simulation