Research Analyzer
← Back ICRA 2026

MinNav: Minimalist Navigation for Active Tiny Aerial Robots

Aniket Patil, Mandeep Singh, Uday Girish Maradana, Nitin J Sanket

PDF

AI summary

Key figure (auto-extracted from paper)
MinNav enables tiny aerial robots to autonomously navigate unstructured environments with static/dynamic obstacles and unknown gaps using only a monocular camera and onboard computation, achieving a 70% real-world success rate.
optical flow uncertainty estimation active perception tiny aerial robots monocular navigation obstacle avoidance

Problem

Tiny aerial robots lack the computational power and sensors for traditional 3D depth-based navigation, while existing monocular optical flow methods struggle with dynamic obstacles, unknown gaps, and require prior scene knowledge.

Approach

The authors propose MinNav, a lightweight navigation stack that fuses optical flow and its uncertainty with an active exploratory control policy to classify scene motifs and guide the robot toward free space without prior environmental knowledge.

Key results

  • First unified monocular framework handling static obstacles, dynamic obstacles, and unknown gaps without prior knowledge
  • Lightweight multi-scale neural network (2.8M params) running in ~68ms on an embedded Jetson Orin Nano
  • 70% overall success rate across diverse real-world and simulation experiments
  • Performance on par with depth-based methods while requiring orders of magnitude less computation

Why it matters

Enables reliable, low-cost autonomous flight for resource-constrained tiny aerial robots in unstructured environments where GPS and heavy sensors are impractical.

Abstract

Navigation using a monocular camera is pivotal for autonomous operation on tiny aerial robots due to their perfect balance of versatility, cost and accuracy. In this paper, we introduce MinNav, a navigation stack based on optical flow and its uncertainty to fly through a scene with static and dynamic obstacles and unknown-shaped gaps without any prior knowledge of the scene components and/or their locations/ordering. We further improve success rate by using the activeness of the robot to move around in an exploratory way to find obstacles and navigate. We successfully evaluate and demonstrate the proposed approach in many real-world experiments in various environments with static and dynamic obstacles and unknown-shaped gaps with an overall success rate of 70%. To the best of our knowledge, this is the first solution to tackle all the aforementioned navigation cases without prior knowledge using a monocular camera. Our approach is on par in performance with depth based methods with factors of magnitude less computation required and can readily run onboard tiny aerial robots. SUPPLEMENTARY MATERIAL The accompanying video, supplementary material, code and dataset can be found at https://pear.wpi.edu/ research/minnav.html.

Index terms

Aerial Systems: Perception and Autonomy Perception-Action Coupling Vision-Based Navigation

Related papers