Current research

I am looking for creative, energetic, and hard-working BS, MS, and PhD students with interests in the Internet of Things, edge computing, mobile systems, and wireless and mobile networking to join my lab at Duke University. Please e-mail me your CV, your transcripts, and a brief note about your research interests at maria.gorlatova /at/

11/30/2018: New: we have a number of summer research opportunities in the lab for US-based undergraduate students. Some of the project descriptions are available here. Please e-mail me at maria.gorlatova /at/ by February 28th, 2019 if interested in joining the lab.

I lead the Duke University Intelligent Interactive Internet of Things (I^3T) Lab.

My work is in the cross-disciplinary area currently known as the Internet of Things (IoT). The long-term vision of my research is taking the IoT to the point where intelligent, reliable, adaptive collaborative IoT deployments can be created near-automatically beginning-to-end, from hardware form factor generation to communication support and cloud infrastructure specifications. My research can be seen as covering two key sub-components of this vision: (1) making IoT deployments more capable (e.g., more energy-independent, more intelligent, more readily deployable in commonplace environments, more readily accessible to the users), and (2) automating IoT design decisions currently done via engineering trial and error (e.g., automating the logic behind data rate and protocol selection and task separation decisions).

The current focus of my work is on enabling the next level of intelligence, interactivity, adaptability, cognition, and deployment automation in the Internet of Things with fog and edge computing. I am also interested in new IoT node form factors and new actuators for IoT devices, particularly for smart city-oriented and human-centered IoT deployments. Last but not least, I am captivated by the transformative promise of augmented reality as a technology that helps us interact with the IoT-captured, previously invisible, properties of the world in fundamentally new ways.

Enabling practical augmented reality with edge computing: Current augmented reality deployments have multiple limitations that need to be overcome for augmented reality to become a practical pervasive technology. These limitations include high energy consumption, limited multi-user experiences, and general lack of robustness and intelligence. We are currently exploring a range of approaches for improving augmented reality experiences with edge computing, including making the experiences more adaptive and intelligent, more secure, and enabling advanced communication and networking support for them. My vision for intelligent augmented reality has recently been summarized in a Network World article available here.

Related publications:

  • S. Ahn, M. Gorlatova, P. Naghizadeh, M. Chiang, P. Mittal, Adaptive Fog-based Output Security for Augmented Reality, in ACM SIGCOMM VR/AR Network Workshop, Budapest, Hungary, Aug. 2018. [Paper PDF]

Related media coverage:

Intelligence on the edge: I am excited about the opportunities  associated with bringing advanced intelligence close to the end users with edge and fog computing. However, it is far from trivial to adapt both training and inference algorithms to the constraints of edge systems, and to figure out how to best exploit the advantages of both edge nodes and cloud components of the overall IoT systems. We are currently exploring multiple directions in adapting different machine learning algorithms to collaborative edge/fog conditions. This research will enable the next level of interactivity and cognition in Internet of Things deployments, while also reducing network loads and energy consumption associated with machine learning algorithms.

Related publications:

  • Y. Ruan, L. Zheng, M. Gorlatova, M. Chiang, C. Joe-Wong, The Economics of Fog Computing: Pricing Tradeoffs for Distributed Data Analytics, Fognet and Fogonomics, Wiley, in print, 2018 (invited book chapter).
  • T. Chang, L. Zheng, M. Gorlatova, C. Gitau, C.-Y. Huang, M. Chiang, Demo: Decomposing Data Analytics in Fog Networks, in Proc. ACM SenSys’17, Delft, Netherlands, Nov. 2017. [Demo abstract PDF] [Video of the demo]

Aiding Internet of Things communications and control with smart gateways: Currently, many elements of communications, networking, and control in the Internet of Things are statically pre-configured. We are exploring how smart gateways can aid in automatic protocol selection, bandwidth allocation, functionality placement, automatic management of network resource reservations, and other adaptive on-demand behavior in IoT systems. Among other techniques, we are examining the applications of reinforcement learning in these contexts. This research will lead to increased capabilities and reduced energy consumption in the IoT systems, and will enable supporting the combination of ultra-low-latency and high-bandwidth communications that are required in modern practical IoT deployments.

Related publications:

  • P. Naghizadeh, M. Gorlatova, A. Lan, M. Chiang, Hurts to be Too Early: Benefits and Drawbacks of Communication in Multi-agent Learning, to appear in IEEE INFOCOM, Paris, France, May 2019 (19.7% acceptance rate).
  • H. Inaltekin, M. Gorlatova, M. Chiang, Virtualized Control over Fog: Interplay between Reliability and Latency, to appear in the IEEE Internet of Things Journal, 2019. [PDF available via IEEE Early Access]
  • S. Ahn, M. Gorlatova, P. Naghizadeh, M. Chiang, P. Mittal, Adaptive Fog-based Output Security for Augmented Reality, in ACM SIGCOMM VR/AR Network Workshop, Budapest, Hungary, Aug. 2018. [Paper PDF] [Princeton Engineering video]

IoT applications: The vast majority of my research applies across a broad range of applications and use cases. I am particularly interested in smart city-oriented and human-centered IoT deployments. In smart cities, the challenges associated with collaborative intelligence at scale are particularly pronounced. In human-centered IoT deployments, such as deployments focused on fitness, wellness, and medical applications, intelligence-on-the-edge is on the cusp of enabling in-context real-time performance analysis, activity suggestions, and behavior modification recommendations. Furthermore, both smart city and human-centered IoT applications call for new IoT node form factors and novel approaches to energy and communication resource management in the IoT.

Industry and public engagement: I strongly believe that In applied fields like the Internet of Things, it is particularly important to share research findings with broad technical and non-technical communities. I place emphasis on making code, data, and experimental how-to guides widely available for a wider scientific community, and on developing long-term industry collaborations. I have advised IoT startups in the fitness and wellness space; I am also particularly excited about developing opportunities for seamless transfer of research to industry-wide deployments, such as contributing edge computing-related developments to the EdgeX Foundry open source project.

Related work:

Previous research:

My Ph.D. research focused on developing energy harvesting active networked tags for ubiquitous networking of commonplace objects in the Internet of Things [Dissertation PDF]. This research was recognized with the 2016 IEEE Communications Society Young Author Best Paper Award, the 2011 ACM SenSys Best Student Demonstration Award, and the 2011 IEEE Communications Society Award for Advances in Communications. More information about this work: EnHants, energy harvesting, prototypes and the testbed we developed.

My M.Sc. research and key elements of my industry research focused on security and privacy in wireless and mobile networks. [All publications]