Karthik Dantu
Assistant Professor
Computer Science and Engineering
331 Davis Hall
email: kdantu.at.buffalo.edu
twitter: @dkkarthik

I am a tenure-track assistant professor in the Computer Science and Engineering department at University at Buffalo, State University of New York.

Broadly, my research interests are in robotics, embedded systems, and mobile computing. I'm actively looking for undergraduate and graduate students that are interested in pursuing research. If you have a strong background in one or more of the above topics, please drop me an email.

My office hours for Fall 2017 are from 3PM to 5 PM Mondays. Please use those times to come by my office if you want to talk to me. Also, I prefer email for a first conversation.


  • [Aug 2017] I'm teaching CSE 421/521: Operating Systems this fall. There is a piazza discussion forum for it as well.
  • [Jan 2016] The Piazza link to CSE 468/568:Robotics algorithms for Spring 2016 is up.
  • Our new project on sensing on MAVs is in the news [UB News][Smithsonian][IEEE Spectrum][Living Science][NBC News]
  • [May 2015] Our NSF Medium proposal on sensing for MAVs is funded.
  • [Mar 2015] Our proposal on a community for excellence on Sustainable Manufacturing and Advanced Robotic Technologies (SMART) is accepted. Very excited to get started on this multi-disciplinary initiative.
  • The Piazza link to CSE 468/568: Robotics algorithms is up.
  • Our ICRA paper titled "A lightweight vision sensor for MAV guidance" has been accepted. Here is the accompanying video.
  • I am currently looking for masters students interested in working with me on a few embedded systems and robotics projects. If you are skilled in programming, and are able to spare 10-15 hrs a week, drop me an email.
  • Intel has generously donated ten Intel Galileo boards for research. Please drop me an email if you are interested in some embedded projects.
  • I'm teaching a re-designed CSE 468/568: Robotics Algorithms in Spring 2014. I will follow this textbook loosely. It will be an introduction to Robotics from a Computer Science perspective. Topics will include kinematics, perception, and algorithms for localization, planning, and navigation. It will likely include programming assignments to be done using ROS, written assignments, a mid-term, and a final. I expect the course to be fairly intense, both in the material covered, and the programming exercises. With the adequate background, expect to spend an average of about 10-15 hours a week. This course should be of interest to anyone enthusiastic about robotics from CSE as well as other departments in SEAS. If you have issues registering, don't worry. I'm happy to grant access to anyone interested as long as we have space in the class. This also applies to sophomores and juniors as well.

    • Pre-requisites: Linear Algebra, Probability, C++ programming (at an intermediate to advanced level)
    • I am told that it satisfies the AI core course requirement for the CSE Masters program.


My research interests broadly span robotics, sensor networks, embedded systems, and mobile computing. Recently, my research has been in coordination multi-robot systems, specifically micro-aerial vehicle swarms. I am interested in designing scalable distributed algorithms that will enable swarms of MAVs collaboratively accomplish the task at hand.


Coordination In Micro-Aerial Vehicle Swarms [IROS 12] [SenSys 11] [IPSN 12]


Karma: Karma is a framework to program and coordinate micro-aerial vehicle (MAV) swarms. Karma proposes a programming model where the application can be specified as a set of behaviors. Each behavior produces and consumes information on execution. The control flow is specified as dependencies on the presence of such information. Karma also divides the world where the application is being executed into Regions. This allows users to build complex applications from simple behaviors. Our target tracking application is only 50 lines of code.

Simbeeotic: Simbeeotic is an event-driven simulator written in Java to simulate Micro-Aerial Vehicle swarms. It is built on JBullet, a 6 degree-of-freedom physics engine. One can implement control behaviors, attach virtual sensors, and simulate virtual worlds. It is easy to quickly simulate swarm behavior, design custom control algorithms, network behavior, and run repeated simulations. Simbeeotic is open-source and available on github.


Mobile Networking [ROBOCOMM 09] [ICRA 09] [WCNC 03] [ISLPED 02]


Most wireless protocols are designed for static wireless nodes. Every layer in the wireless networking stack implicitly or explicitly adheres to this assumption. Therefore, mobility introduces novel challenges for wireless networking. Further, understanding radio propagation properties allow us to use the radio as a sensor in a multi-robot environment. While there is a fair amount of work in inferring (or at least attempting to infer) range from radio signal strength, there is not much work in using signal strength in combination with mobility to infer other properties. Previously, I have worked on using mobility and signal strength to infer coarse relative bearing in a robot network. I have also worked on improving route stability in a robot network using location and position cues. Finally, I have also worked on power-efficient ways for routing in a robot network both to maximize individual robot lifetime as well as network lifetime.


Embedded Systems [IPSN '05] [Emnets '06] [ICCAD '02]


Modern embedded systems bring unique challenges of their own. I am interested in the challenges that are brought about by the integration of novel sensors, energy limitations, as well as networking of modern embedded systems. Previously, I have helped build a table-top robot called the Robomote that was designed to be a mobile node in a sensor network. We used it to study level set detection in a mobile sensor network. I have also helped study power-awareness issues in embedded systems by adaptive voltage-frequency scaling without loss of QoS for particular applications such as streaming video.


  Recent Publications (citations=1039, h-index=10)

  1. "A Comparison of Deterministic and Stochastic Approaches to Allocating Spatially Dependent Tasks in Micro-Aerial Vehicle Swarms",
    Karthik Dantu, Spring Berman, Bryan Kate, Radhika Nagpal
    In IROS '12: Accepted for publication to IEEE/RSJ International Conference on Intelligent Robots and Systems
    [tl;dr] [Slides] [bib]
  2. "Simbeeotic: A Simulator and Testbed for Micro-Aerial Vehicle Swarm Experiments",
    Bryan Kate, Jason Waterman, Karthik Dantu, Matt Welsh
    In IPSN '12: Proceedings of the 11th International Conference on Information Processing in Sensor Networks (SPOTS Track), Beijing, China, Apr. 16-19, 2012.
    [tl;dr] [Slides] [Source] [bib] [ACM]
    Runner-up for Best Paper Award (SPOTS Track)
    Best Demo Award (IPSN-SPOTS 2012)
  3. "Programming Micro-Aerial Swarms with Karma",
    Karthik Dantu, Bryan Kate, Jason Waterman, Peter Bailis, Matt Welsh
    In SenSys '11: Proceedings of the 9th International Conference on Embedded Networked Sensor Systems, Seattle, Washington, Nov. 1-4, 2011.
    [tl;dr]  [Slides (250 M)]  [Youtube]  [1080P]  [bib]  [ACM]

... Full publication list


The initial template of this homepage was stolen with permission from Prabal Dutta.