Insect-scale Autonomy Workshop

What principles, algorithms, and system designs enable reliable autonomy when resources are tightly constrained?

Full-day Invited talks + posters/demos + panel Best Poster/Demo Award

Quick info

Focus: Bio-inspired, resource-constrained Autonomy across perception, design, control, and edge AI.

Venue: IROS 2026 (Pittsburgh)

Date: Oct 1, 2026 (Full Day Workshop)

Time: 8:30 AM — 5.30 PM

Submission deadline: Sep 1, 2026, 11:59 PM AoE

Submit on OpenReview

Insect-scale autonomy workshop banner

Overview

Advances in robotics have made it possible to build increasingly small and agile systems, but true autonomy at insect-scale remains constrained by sensing, computation, power, actuation, and packaging. This workshop focuses on identifying principles, algorithms, and system designs that enable reliable autonomy under tight resource limits.

Focus Area

Autonomy at insect scale is a coupled systems problem: sensing, mechanics, computation, and control all compete for the same mass, power, memory, and latency budget. We structure the workshop around three interacting focus areas.

Resource-aware perception and learning

A. Resource-aware perception and learning

Perception and learning methods that extract task-relevant information from limited sensing and data while operating under severe energy, memory, bandwidth, and latency constraints.

Key question

How can a tiny robot extract and learn the task-relevant information it needs when sensing, memory, bandwidth, and compute are all scarce?

Mechanics, actuation, design, and control

B. Mechanics, actuation, design, and control

Co-design of morphology, actuation, compliant mechanics, passive dynamics, and feedback control to produce agile behavior at small scales, where conventional robot design assumptions often fail.

Key question

How should morphology, actuation, sensing, and control be co-designed so that mechanics carries part of the autonomy burden?

On-board compute, edge AI hardware, and deployment

C. On-board compute, edge AI hardware, and deployment

Embedded architectures and system software for running perception, estimation, planning, learning, and control on-board under strict energy, memory, latency, thermal, and communication budgets.

Key question

What is the minimum on-board compute and sensing stack that can deliver reliable closed-loop autonomy under strict energy, latency, and mass constraints?

Format

The workshop will feature invited keynotes, contributed spotlight presentations, posters, and a panel on evaluation practices, benchmarks, and autonomy targets.

All accepted Paper Track submissions will be presented as posters. A subset of accepted papers will also be selected for short spotlight presentations during the workshop. A Best Paper Award will be presented. A Best Poster Award may also be presented, subject to the number and quality of eligible contributions.

Contact

Correspondence: Chahat Singh and Kaushik Jayaram.

Workshop Sponsor

Sponsors

Edit this page on GitHub