NOTE: The RSN Lab is moving from the University of Minnesota to the University of Texas at Austin CS and Texas Robotics. This is our new website (currently under construction). In addition to the information below, you can visit Prof. Isler’s website or the the current RSN website for more information.

The Robotics, Sensing and Networks Laboratory was established in 2005 with the goal of developing robotic systems that can operate in dynamic and complex environments. Such systems often involve multiple robots, and perform tasks involving sensing and communication. We call them Robotic Sensor Networks.

• Lab members
• Some good times
• Publications
  • Theses
  • Google Scholar
  • DBLP

Research Philosophy

In our lab, we study fundamental problems that arise in RSN tasks and build RSN systems for representative applications.

For example, visibility plays an important role in robotic systems. We study mathematical properties of visibility using both classical methods and learning-based approaches. In the past, we solved a number of classical visibiliy based pursuit-evasion games. We are now investigating learning based belief state representations and reinforcement learning based methods for pursuit-evasion.

In addition to these fundamental results, we deployed field-grade (sometimes even commercial-grade) robotic systems for environmental monitoring, agriculture and home automation applications. Below, you can find a sampling of these projects.

Some of Our Ongoing and Recent Projects

• Agricultural Robotics
  • Yield Mapping
  • Fruit Picking
  • CowBot
  • Row Crops
• Environmental Monitoring
  • Tracking invasive fish with autonomous boats
  • Aerial monitoring
• Uniformly Sampling Trajectories
• Robot Follow-Ahead
• Pursuit-Evasion
• Generative Models for Polygons
• Reactive Motion Planning
• Hierarchical Online Signed Distance Functions
• Cost-to-Go Function Generating Networks for High Dimensional Motion Planning
• Geometrically Grounding LLMs
• Higher Order Function Networks
• FineControlNet: Fine-level Text Control for Image Generation with Spatially Aligned Text Control Injection
• VioLA: Aligning Videos to 2D LiDAR Scans
• RIC: Rotate-Inpaint-Complete for Generalizable Scene Reconstruction
• Real-time Simultaneous Multi-Object 3D Shape Reconstruction, 6DoF Pose Estimation and Dense Grasp Prediction
• Neural Optimal Control using Learned System Dynamics
• Continuous Object Representation Networks: Novel View Synthesis without Target View Supervision
• Category-Level Global Camera Pose Estimation with Multi-Hypothesis Point Cloud Correspondences

Historical Notes

Our logo was designed by Nikhil Karnad.

On our 20th anniversary, we are changing our name from Robotic Sensor Networks Lab to Robotics, Sensing and Networks Lab to reflect the broadening scope of our research.

Thank you for visiting!