NUS OstraBot: Record-Breaking Biohybrid Swimming Robot Powered by Self-Trained Lab-Grown Muscles

🌊 NUS Engineers Pioneer Self-Training for Lab-Grown Muscles

The National University of Singapore (NUS) has marked a significant milestone in biohybrid robotics with the development of a groundbreaking self-training platform for lab-grown skeletal muscle tissues. Biohybrid robots, which integrate living biological components such as cultured muscle cells with synthetic scaffolds or structures, promise actuators that are inherently soft, adaptive, and energy-efficient, particularly at microscales where traditional motors falter. Led by Assistant Professor Tan Yu Jun from the NUS Department of Mechanical Engineering within the College of Design and Engineering, the team addressed a core limitation: the historically low force output of these cultured muscles, which has hindered practical robot performance. 77 76

This innovation draws inspiration from arm-wrestling dynamics, mechanically coupling two muscle tissues via a simple sliding block mechanism. As the cells mature over a week—specifically from day three when spontaneous contractions begin—the tissues pull antagonistically against each other. This creates continuous cycles of shortening and lengthening without any external power, controls, or intervention, effectively allowing the muscles to 'exercise' autonomously. The result? Record-breaking force generation: a maximum of 7.05 millinewtons (mN) and stress of 8.51 mN per square millimeter, over an order of magnitude higher than previous benchmarks for this commercially available skeletal muscle cell line, ensuring reproducibility and affordability.

OstraBot: The Fastest Skeletal Muscle-Powered Swimmer

Integrating these supercharged muscles into a novel robot dubbed OstraBot, the NUS team engineered an ostraciiform swimmer mimicking the locomotion of boxfish—characterized by a rigid body and oscillating twin tails for efficient propulsion. A single self-trained muscle drives the flexible tails, optimized through a physiology-based computational model that simulates the entire actuation chain: from electrical stimulation triggering calcium signaling, to muscle activation, force production, and hydrodynamic thrust. 77

Under 3 Hz electrical stimulation at optimal tail stiffness, OstraBot achieves a swimming speed of 467 millimeters per minute—the fastest ever recorded for any skeletal muscle-driven biohybrid robot. This surpasses identical designs with conventionally cultured muscles by more than threefold, demonstrating tangible gains from the self-training paradigm. Moreover, the robot exhibits precise controllability: speed scales with electrical field strength, and it responds dynamically to acoustic cues, such as starting or stopping on a clap, akin to neural-muscle interfaces in living organisms. Previously, such robots either moved incessantly without regulation or lacked the power for visible responses.

Assistant Professor Tan Yu Jun: Trailblazer in Sustainable Soft Robotics

At the helm is Asst Prof Tan Yu Jun, whose research at NUS Mechanical Engineering centers on intelligent and sustainable materials for soft robotics. Her lab explores self-healing biomaterials and self-adaptive systems that detect defects and initiate repairs, promoting 'biocircular' robots—devices designed for full lifecycle sustainability, from creation to benign degradation. Tan's accolades include being named in MIT Technology Review's Innovators Under 35 Asia-Pacific list and the 'Science, She Says!' award, underscoring her impact. 20

"The purpose of this study was not just to build a faster robot, but to remove a fundamental bottleneck in the field and open the door to high-performance biohybrid systems designed with sustainability in mind," Tan explained. Her vision extends beyond speed, emphasizing long-term stability, energy efficiency, and environmental responsibility—critical for real-world deployment.

Technical Innovations Driving the Self-Training Platform

The self-training process unfolds step-by-step: Day 1-2 involves seeding commercial C2C12 mouse skeletal myoblasts onto gelatin-methacrylate hydrogels patterned for alignment. By day 3, spontaneous twitches emerge as cells fuse into multinucleated myotubes. The antagonistic coupling via the sliding block harnesses these for progressive overload, peaking efficacy around day 5. No nutrients, media changes, or stimuli are needed beyond standard culture conditions, making it scalable.

• Force Amplification: Self-trained muscles yield 7.05 mN peak force vs. ~0.5 mN untrained.
• Stress Metrics: 8.51 mN/mm², enabling meaningful thrust.
• Model Optimization: Finite element simulations predict optimal geometries, reducing trial-and-error.

This plug-and-play actuator paves the way for standardized biohybrid components across labs worldwide.

Singapore's Thriving Ecosystem for Soft Robotics Research

NUS exemplifies Singapore's ascent in soft robotics, ranked among the global top 5 for research output. The Advanced Robotics Centre at NUS fosters interdisciplinary work, from AI-powered microswimmers to manta ray-inspired autonomous underwater vehicles. Complementing this, NTU Singapore advances neuron-inspired controllers for compliant grippers. Government backing amplifies these efforts: In 2026, Singapore committed over S$1 billion (about US$779 million) to public AI research through 2030, alongside S$37 billion for broader R&I, positioning universities as hubs for translating lab breakthroughs to industry.NUS News on OstraBot 77 47

Such investments yield tangible outcomes: NUS spinouts in sustainable materials and bioactuators attract venture capital, while collaborations with A*STAR enhance biomaterial scaling.

Biomedical and Environmental Applications on the Horizon

OstraBot's prowess unlocks applications where rigidity fails. In biomedicine, temporary implants could deliver drugs, clear blockages, or monitor tissues before dissolving—no retrieval surgery required. Environmentally, biodegradable swarms could survey delicate ecosystems like coral reefs or polluted wetlands, powered quietly and sustainably.Nature Communications paper 67

• Drug Delivery: Navigate vasculature with adaptive propulsion.
• Sensing: Integrate lab-grown sensors for real-time data.
• Degradability: Use transient materials for zero-waste ops.

Tan's team eyes full biodegradability next, aligning with Singapore's green tech ambitions.

Challenges Overcome and Remaining Hurdles

Key challenges included harnessing uncontrolled twitches productively and scaling force without compromising viability. The antagonistic trainer resolved this elegantly. Future hurdles: extending lifespan beyond weeks, enhancing nutrient delivery in vivo, and neural integration for sophisticated behaviors. Physiology models will iterate designs, while Singapore's talent pipeline—bolstered by programs like SkillsFuture—supplies skilled researchers.

Career Opportunities in Singapore's Robotics Higher Education

For aspiring engineers, NUS offers PhD scholarships, research assistantships in soft robotics labs. Singapore's ecosystem demands interdisciplinary talent: mechanical engineers versed in biomaterials, AI specialists for control, biologists for tissue engineering. With unemployment low and salaries competitive (median ~S$5,500 for fresh tech grads), it's a prime destination. 56

Programs like NUS Integrated Sustainable Design emphasize real-world impact, preparing graduates for roles in A*STAR, medtech firms, or startups.

Global Context and Singapore's Competitive Edge

While Harvard and EPFL advance hybrid walkers, NUS leads in sustainable swimmers. Singapore's strategic investments—S$1B AI, quantum hubs—ensure agility amid US-China tensions. By 2030, intl student growth to 200k targets robotics as a pillar, fostering diverse teams.

This OstraBot feat positions NUS as a beacon for higher education innovation, blending academia, industry, and policy for societal good.

Photo by Aiper Pool Cleaner on Unsplash

Future Outlook: Toward Autonomous Biohybrid Swarms

Tan envisions swarms of OstraBot variants for collective tasks, with onboard 'homeostasis' for self-repair. Collaborations with NTU on AI controllers could enable learning behaviors. As Singapore eyes Smart Nation 2.0, such research drives economic resilience, job creation, and global leadership in green robotics.