UPenn Engineering
Department of Computer & Information Science
School of Engineering and Applied Science
University of Pennsylvania
Insup Lee UPenn Lee
Home About Me Group Research Teaching PRECISE News Publications Awards Contact
Current Projects    |    Former Projects    |   


Integrated Static and Dynamic Approaches to High-Assurance for Learning-Enabled Cyber-Physical Systems (DARPA ASSURED AUTONOMY) - Project page

medical devices

The project develops approaches to reason about safety of autonomous vehicles that rely on machine learning for perception and control. Our approach combines design-time techniques with run-time monitoring techniques. One design-time technique is closed-loop verification of deep neural network (DNN) controllers, implemented in a tool VeriSig. Another run-time technique is calculation of robustness measures for classifiers and generation of run-time confidence monitors. Run-time monitoring techniques include detection of anomalous behaviors, evaluation of confidence in the outputs of learning-enabled components, and validation of operational assumptions made at design time. Finally, we develop new argument patterns for the construction of assurance cases for autnomous vehicles, which make use of design-time evidence of safety as well as dynamic evidence produced by monitors discussed above..

Assuring the Safety, Security and Reliability of Medical Device Cyber Physical Systems (NSF CPS) - Project page

medical devices

The objective of this research is to establish a new development paradigm that enables the effective design, implementation, and certification of medical device cyber-physical systems. The approach is to pursue the following research directions: 1) to support medical device interconnectivity and interoperability with network-enabled control; 2) to apply coordination between medical devices to support emerging clinical scenarios; 3) to close the loop and enable feedback about the condition of the patient to the devices delivering therapy; and 4) to assure safety and effectiveness of interoperating medical devices. Novel design methods and certification techniques will significantly improve patient safety. The introduction of closed-loop scenarios into clinical practice will reduce the burden that caregivers are currently facing and will have the potential of reducing the overall costs of health care.

High-Confidence Medical Device Software and Systems (NSF) - Project page


The development and production of medical device software systems is a critical issue as medical device software is increasingly sophisticated and medical devices are networked. Of particular importance is how to ensure such medical device systems are safe and effective. There are three projects that we are pursuing: development of the reference implementation and assurance cases of Generic Patient Controlled Analgesia (GPCA), model-based development of the Pacemaker Challenge, design of Generic Decision Support Architecture (G-CDSA). The latter is based on our experience in building a smart alarm for post CABG surgery patients, a decision caddy for vasospasm risk analysis, models of blood glucose control guidelines and a closed-loop PCA controller.

Security and Privacy-Aware Cyber-Physical Systems (NSF, Intel) - Project page


The project aims to achieve a comprehensive understanding of CPS-specific security and privacy challenges. This understanding will enable us to (1) develop techniqes to prevent security attacks to CPS and to detect and recover from malicious attacks to CPS; (2) develop techniques for security-aware control design by develop attack resilient state estimator; (3) ensure privacy of data collected and used by CPS, and (4) establish an evidence-based framework for CPS security and privacy assurance, taking into account the operating context of the system and human factors.

SPARCS: Synthesis of Platform-aware Attack-Resilient Control Systems (DARPA HACMS) - Project page


The project aims to develop control systems for autonomous vehicles that are resilient to external attacks. Our approach is to combine control-level techniques for controller design and code-level techniques for control task synthesis. Control-level defenses address attacks on the control system, such as attacks on sensors, actuators, communication networks, and computational resources available to the controller. Our control-level defense strategies are based on redundancy within a control loop, as well as new methods for detection, identification, and mitigation of attacks. Code-level defenses prevent injection of malicious code into the operation of the controller itself, achieved through verified code synthesis of control task code.

Real-Time Embedded Systems: Compositional Scheduling Framework (NSF, ONR) - Project page


Real-time systems are ones in which correctness depends not only on logical correctness but also on timeliness. In the real-time systems community, substantial research efforts have concentrated on the schedulability analysis problem, which determines whether timing requirements imposed on the system can be satisfied. However, there is no widely accepted technique that supports the compositionality of timing requirements, i.e., how component-level timing requirements can be independently analyzed, abstracted, and composed into the system-level timing requirements. We have developed a compositional real-time scheduling framework for supporting the compositionality of timing requirements. Our compositional scheduling framework is supported by the CARTS tool.


Relevant Research

High-Confidence Medical Devices: Cyber-Physical Systems for 21st Century Health Care


PRECISE Publications at Scholarly Commons

Copyright © 2011-2019 University of Pennsylvania. All Rights Reserved. Site designed by Academic Web Pages.