Search this site
Embedded Files
Valmiki Kothare
  • Home
  • CV
  • Resume
  • Research
    • Robotics
    • Biological Neural Nets
  • Projects
    • Shipbot
    • Calorie-Expending Machine
    • D3VO Implementation
    • Dense SLAM
    • Inverse Compositional Object Tracker
    • Object Tracking Data Labeler
    • random_wallpaper
    • LED Matrix
    • FPV Drones
    • 3D Printed Rocket
Valmiki Kothare
  • Home
  • CV
  • Resume
  • Research
    • Robotics
    • Biological Neural Nets
  • Projects
    • Shipbot
    • Calorie-Expending Machine
    • D3VO Implementation
    • Dense SLAM
    • Inverse Compositional Object Tracker
    • Object Tracking Data Labeler
    • random_wallpaper
    • LED Matrix
    • FPV Drones
    • 3D Printed Rocket
  • More
    • Home
    • CV
    • Resume
    • Research
      • Robotics
      • Biological Neural Nets
    • Projects
      • Shipbot
      • Calorie-Expending Machine
      • D3VO Implementation
      • Dense SLAM
      • Inverse Compositional Object Tracker
      • Object Tracking Data Labeler
      • random_wallpaper
      • LED Matrix
      • FPV Drones
      • 3D Printed Rocket

Shipbot

Final Demo Encore.MOV
Demonstration of navigation capabilities of the Theseus autonomous Shipbot.

For my undergraduate capstone project and as part of the Shipbot competition, my team and I were tasked with designing a fully autonomous robot capable of manipulating valves and switches in a simulated ship environment. As the software lead for my team, I independently developed and integrated the perception (LiDAR), localization (AMCL), navigation(A*), and control(PID) software, working in conjunction with electrical and mechanical engineers to realize the autonomous capabilities I had implemented in Robot Operating System (ROS) on a fully custom mobile robot (which we lovingly named Theseus). The ROS system diagram for the autonomous protocol I designed and implemented is depicted below. This experience challenged me to work under the constraints of a physical robot and contend not only with software and algorithmic challenges but also with motion model uncertainty and system instability introduced by non-ideal environments. In the end, we were able to autonomously localize and navigate to each station in the environment in the order presented by an arbitrary mission file. Future work for this project would include improving the valve and switch manipulation protocol, which did not perform reliably due to our reliance on LiDAR alone for localization and therefore prevented us from fully achieving the goal of robust, autonomous environment manipulation. Check out our Shipbot project website here. 

EmailPhone numberGitHubLinkedIn
Google Sites
Report abuse
Page details
Page updated
Google Sites
Report abuse