siddhant

Knowledge / Robotics

Autonomous Robotic Systems

Integrated perception, estimation, planning, control, decision-making, and recovery.

By Siddhant Krishna · Published 2026-10-06 · Updated 2026-10-06

01

Autonomous Stack

A representative autonomous system connects perception, state estimation, prediction, planning, control, and task execution.

sensing → perception → localization → planning → control → actuation

02

Task-Level Planning

Task planning reasons about actions at a higher level than geometric motion planning. For example, an indoor service robot may first decide to navigate to a kitchen, locate an object, grasp it, and deliver it before generating the precise trajectories required for each step.

03

Failure Recovery

  • Detecting perception failures.
  • Replanning when obstacles change.
  • Handling localization loss.
  • Stopping safely under uncertain conditions.
  • Retrying recoverable actions.
  • Escalating to human intervention when autonomous recovery is inappropriate.

References

  1. Thrun, Burgard & Fox, MIT Press.
    https://mitpress.mit.edu/9780262201629/probabilistic-robotics/
  2. LaValle, Cambridge University Press.
    https://lavalle.pl/planning/
  3. Open Robotics.
    https://docs.ros.org/en/rolling/

Related

Contact

Get in Touch

Want to chat? Just shoot me a dm with a direct question on twitter and I'll respond whenever I can. I will ignore all soliciting.