Goal
I envision a future where embodied human-centered computing systems, such as diagnostic devices, seamlessly integrate with human environments and biology, continuously evolving to meet individual needs and dynamically responding to changes in their surroundings. Driven by this vision, my goal is to enhance the capabilities of these systems by developing cognitive frameworks that are energy-efficient, robust, trustworthy, and explainable. Unlike existing solutions, my work focuses on creating sustainable sensory-motor cognitive capabilities that enable such systems to learn and adapt in real-time while optimizing energy use, ensuring long-term efficiency and reduced environmental impact. Furthermore, my research is inherently interdisciplinary, drawing inspiration from biological principles to design more adaptable and resilient systems. Central to this is the concept of algorithm-system-hardware co-design, where cognitive algorithms are developed in conjunction with customized system models and energy-efficient hardware architectures. This holistic design strategy fosters tight integration between computational processes and physical components, resulting in more scalable and robust systems than traditional siloed designs.
Current Research Projects
Coming soon…!
Past Research Projects
Hardwre-Software Codesign of Hybrid Microfluidic Biochips for Biomolecular Quantitative Analysis: System Modeling, Synthesis, and Optimization Methodologies
Considerable effort has been devoted in recent years to the design and implementation of microfluidic platforms for biomolecular quantitative analysis. However, today’s platforms suffer from the drawback that they were optimized for sample limited analyses, thus they are inadequate for practical quantitative analysis and the processing of multiple samples through independent pathways. Design optimization techniques for microfluidics have been studied in recent years, but they overlook the myriad complexities of biomolecular protocols and are yet to make an impact in microbiology research. The realization of microfluidic platforms for real-life quantitative analysis requires a new optimization flow that is based on the realistic modeling of biomolecular protocols.
Motivated by the above needs, this project is focused on an optimized and trustworthy transfer of benchtop biomolecular analysis, particularly epigenetic studies, to programmable and cyber-physical microfluidic biochips. In collaboration with Duke Molecular Genetics and Microbiology, Mohamed has transferred gene-expression analysis and epigenetic protocols, e.g., chromatin immunoprecipitation, from bench-scale settings and has streamlined a generic optimization flow for various classes of biomolecular analysis protocols. Adopted optimization methods were based on cyber-physical system integration, real-time systems, CFD simulations, formal methods, modeling of stochastic processes, regression analysis, and more.
Selected Publications: [J4][J7][J10][J13]
Improving Trust in Emerging Microfluidic Biochips-based DNA Forensics
Microfluidics-driven biomolecular analysis can offer remarkable benefits, especially for mission-critical applications such as forensic DNA analysis. Several microfluidic commercial developers, including U.S.-based IntegenX, ANDE, and Lockheed Martin, have already started to roll out prototypes aiming to replace traditional benchtop procedures for DNA forensics, and it is anticipated that design automation and cyber-physical integration will play a significant role in advancing this technology. However, the pressure to drive down costs besides the proliferation of IoT-based connectedness will lead to cheap untrusted microfluidic devices and a multitude of unanticipated privacy violations if preventative measures are not taken. Trustworthy DNA forensic science are particularly relevant to defense applications and broader security needs.
Sensitive information in a microfluidic device can include data collected after processing of the fluids and personally identifying metadata. Irresponsible handling of patient data has led to the breakup of companies in the past, and current device makers would do well to learn from those mistakes. Other issues include trust in the sensor readings themselves; the rise and fall of Theranos, and the invalidation of two years worth of test results set a poor precedent for microuidics diagnostics. The nascent nature of microfluidics in biomolecular quantitative analysis presents an opportunity to incorporate security and trust in such critical applications before it becomes too late to rescue this rising industry from security threats.
Motivated by the above needs, this project is focused on creating the science of secure DNA forensics with the help of microfluidics technology. In collaboration with NYU Center for Cybersecurity (Professor Ramesh Karri’s group) and Duke Molecular Genetics and Microbiology, Mohamed has worked on the assessment of the security implications of emerging forensic flows and has also provided appropriate countermeasures to secure such flows.
Selected Publications: [J2][J3][J19]
The Internet of Microfluidic Things: An Integrative Cyber-Physical System for Large-Scale Microfluidics-Driven Diagnostics
The integration of microfluidics and biosensor technology is transforming microbiology research by providing new capabilities for clinical diagnostics, cancer research, and pharmacology studies. This integration enables new approaches for biochemistry automation and cyber-physical adaptation. Similarly, recent years have witnessed the rapid growth of the Internet of Things (IoT) paradigm, where different types of real-world elements such as wearable sensors are connected and allowed to autonomously interact with each other. Combining the advances of both cyber-physical microfluidics and IoT domains can generate new opportunities for knowledge fusion by transforming distributed local microfluidic elements into a global network of coordinated microfluidic systems.
This research aims to streamline this transformation and it presents a research vision for enabling the Internet of Microfluidic Things (IoMT). To leverage advances in connected Microfluidic Things, the research introduces new perspectives on system architecture, and describe technical challenges related to design automation, temporal flexibility, security, and service assignment. This vision can play a critical role in advancing the responses of clinical healthcare to global pandemics such as COVID-19. It can also be used to support complex cancer research and pharmacology studies.
Selected Publications: [J5][J8][C19]
Hardware-Software Codesign for Real-Time Error Recovery in Cyber-Physical Digital Microfluidic Biochips
Digital microfluidics is a reconfigurable lab-on-chip technology that has achieved remarkable success in miniaturizing point-of-care (POC) quantitative-analysis testing. However, a major stumbling block in the monitoring and controlling of diseases via such POC systems is the lack of reliable diagnostic tests that can recover from unexpected errors. In addition, diagnostic tests, similar to all other quantitative-analysis procedures, are inherently stochastic systems that exhibit complex interactions among their constituent biochemical components. Such characteristics signify the need for real-time error-recovery methods that verify the correctness of on-chip fluidic interactions on-the-fly during bioassay execution.
To add resilience to digital-microfluidic control, Mohamed first introduced an efficient design method for digital microfluidic platforms to support error detection and recovery. The proposed design is based on cyber-physical system integration and it enables real-time monitoring of biochemical reactors (droplets) using capacitive sensors. Mohamed and his colleagues from Duke Microfluidics Lab designed and tested an all-hardware implementation of a cyber-physical microfluidic platform, thus enabling a portable POC setting that is resilient against faults.
Sample Publications: [J1][C1][C2]