An Autonomous Subgram SMA-Based Swimmer

Conor K. Trygstad, Francisco M. F. R. Gonçalves, Néstor O. Pérez-Arancibia · arXiv · 2026

Swima is a bioinspired subgram microswimmer using onboard-powered SMA-wire actuators and computation to achieve autonomous swimming with measured speed, turning, and trajectory tracking.

Moderate AI ConfidenceGood SourceWorking PrototypeReadiness Unknown

Plain English summary

The authors present Swima, a bioinspired 900-mg swimmer propelled by two small, high-work-density shape-memory alloy (SMA) actuators made from SMA wires. They integrate onboard power and computation using a custom printed circuit board (PCB) and an 11-mAh 3.7-V lithium-ion battery, enabling autonomous swimming for more than 18 minutes. Performance is reported as swimming speeds up to 22.4 mm/s, turning rates up to 14°/s, and trajectory following of 0-degree heading references with RMS tracking errors around 6.5° across multiple tests. The abstract states this is the first subgram microswimmer with onboard power, actuation, and computation developed to date.

Why this matters

The abstract claims Swima is the first subgram microswimmer with onboard power, actuation, and computation developed to date. The abstract reports a working microswimmer prototype and performance metrics, but does not provide evidence of commercialization, deployment, or market readiness.

Key findings

  • Bioinspired 900-mg swimmer propelled by two 10-mg high-work-density SMA-wire actuators
  • Onboard power and computation enable autonomous swimming for over 18 minutes
  • Maximum reported speed: 22.4 mm/s (0.56 Bl/s)
  • Maximum reported turning rate: 14°/s
  • Trajectory tracking: ~6.5° RMS error for 0-degree heading reference across multiple tests

Limitations

The abstract does not specify experimental setup details, environmental conditions, power/actuation efficiency, long-term durability, or how performance scales beyond the reported tests.

Publication

Publisher
arXiv
Publication date
June 12, 2026
Research type
Preprint
arXiv
2606.15028
Access
open

Tags

More on Shape-Memory Alloys

See all →
Shape-Memory Alloyspaper· Jul 15, 2026

Functional Shape Recovery Response of Heat-Treated NiTiCu SHAPE Memory Alloy Wire

Heat treatment of NiTiCu shape memory alloy wires enhances their shape recovery response, achieving a high recovery ratio.

Ümit Zeybek · Black Sea Journal of Engineering and ScienceLaboratory Research
Shape-Memory Alloyspaper· Jul 9, 2026

A Deformable Robot Based on Shape Memory Alloy and Phase Change Metal

This research presents a novel deformable robot that integrates Shape Memory Alloys and phase change metals for enhanced adaptability and locomotion.

Junwei Zhang, Bo Yuan +1 · WileyWorking Prototype
Shape-Memory Alloyspaper· Jul 8, 2026

Development of a Directional Vibrator Using Shape-Memory Alloy Wires

This paper proposes a compact directional vibrator using Shape-Memory Alloy wires for enhanced haptic feedback in virtual and augmented reality applications.

Yuto Kawahara, Renke Liu +1 · MDPI AGLaboratory Research
Shape-Memory Alloyspreprint· Jul 3, 2026

Martensitic Transformation in Crystal-Amorphous Superlattices of NiTi Shape Memory Alloy

Crystal-amorphous superlattices in NiTi are simulated to shift martensitic transformation behavior, boosting reversibility and stiffness while raising transformation critical stress.

Bhavna Singh, Shivam Tripathi · arXivSimulation
Shape-Memory Alloyspreprint· Jul 2, 2026

HVAF Spraying of NiTi Coatings: Microstructure, Phase Transformation and Shape Memory Behavior

HVAF spraying enables thick (100–300 µm) NiTi shape-memory alloy coatings on mild steel that can undergo martensitic transformation and exhibit thermal actuation and shape memory effects after annealing.

Sneha Samal, Shrikant Joshi +8 · arXivLaboratory Research
Shape-Memory Alloyspaper· Jun 30, 2026

Omnidirectional Shape Proprioception for Untethered Shape Memory Alloy‐Driven Soft Robotic Arms

This research presents a modular soft robotic arm using shape memory alloys for improved proprioception and adaptability in unstructured environments.

Yiming Ouyang, Zelin Chen +6 · WileyLaboratory Research
Method note: Summaries and ratings on this page are generated by AI from the abstract only. Read the original paper for full context. · Model: gpt-5.4-nano-2026-03-17