Autonomous Droplet Navigation via Model-Based Reinforcement Learning

arXiv cs.LG Papers

Summary

This paper presents a model-based reinforcement learning approach for autonomous droplet navigation in microfluidic platforms, achieving reliable navigation through complex geometries with zero-shot transfer from simpler paths.

arXiv:2609.16369v1 Announce Type: new Abstract: Precise manipulation of liquid droplets underpins lab-on-a-chip platforms for diagnostics, chemical synthesis, and biological assays. Yet autonomous droplet transport through confined geometries of varying complexity remains an open challenge. Droplets exhibit contact-angle hysteresis, deformability, and capillary pinning, which make their response to actuation nonlinear and history dependent, that classical controllers and pre-programmed trajectories cannot cope in multi-turn environments. Here we demonstrate autonomous navigation of a liquid droplet through geometries of increasing complexity on a gravity driven (Labyrinth) platform using model-based reinforcement learning. A thin silicone oil film reduces contact-line pinning while two-axis tilt supplies the gravitational driving force, and an overhead camera tracks the droplet in real time. An offline-trained policy discovers effective tilt strategies from limited physical interaction data, without simulation or analytical droplet models. The system operates under partial observability, as oil-film thickness, instantaneous contact angle, and droplet deformation state remain hidden from the controller. Despite these challenges, the learned policy achieves reliable navigation across straight, right-angle, and curved-arc paths, including outside-corner geometries. We further demonstrate that a policy trained on a simpler geometry transfers to complex ones, succeeding zero-shot on right-angle and staircase paths and reaching full success on a curved arc with a fifth of the training data. The findings suggest promising avenues for enabling droplet based microfluidic systems to serve as intelligent chemical laboratories.
Original Article
View Cached Full Text

Cached at: 09/16/26, 08:49 AM

# Autonomous Droplet Navigation via Model-Based Reinforcement Learning
Source: [https://arxiv.org/abs/2609.16369](https://arxiv.org/abs/2609.16369)
[View PDF](https://arxiv.org/pdf/2609.16369)

> Abstract:Precise manipulation of liquid droplets underpins lab\-on\-a\-chip platforms for diagnostics, chemical synthesis, and biological assays\. Yet autonomous droplet transport through confined geometries of varying complexity remains an open challenge\. Droplets exhibit contact\-angle hysteresis, deformability, and capillary pinning, which make their response to actuation nonlinear and history dependent, that classical controllers and pre\-programmed trajectories cannot cope in multi\-turn environments\. Here we demonstrate autonomous navigation of a liquid droplet through geometries of increasing complexity on a gravity driven \(Labyrinth\) platform using model\-based reinforcement learning\. A thin silicone oil film reduces contact\-line pinning while two\-axis tilt supplies the gravitational driving force, and an overhead camera tracks the droplet in real time\. An offline\-trained policy discovers effective tilt strategies from limited physical interaction data, without simulation or analytical droplet models\. The system operates under partial observability, as oil\-film thickness, instantaneous contact angle, and droplet deformation state remain hidden from the controller\. Despite these challenges, the learned policy achieves reliable navigation across straight, right\-angle, and curved\-arc paths, including outside\-corner geometries\. We further demonstrate that a policy trained on a simpler geometry transfers to complex ones, succeeding zero\-shot on right\-angle and staircase paths and reaching full success on a curved arc with a fifth of the training data\. The findings suggest promising avenues for enabling droplet based microfluidic systems to serve as intelligent chemical laboratories\.

## Submission history

From: Rajneesh Anand \[[view email](https://arxiv.org/show-email/ec35ee4e/2609.16369)\] **\[v1\]**Mon, 14 Sep 2026 21:31:28 UTC \(7,158 KB\)

Similar Articles