“The core vision is to transform drug carriers from passive drifters into active, intelligent miniaturized robots that can sense, navigate, and respond on demand,” explains Professor Wang. The review deconstructs this vision into two technological pillars: targeted delivery control and on‑demand release control.
For targeted delivery, the authors survey a diverse arsenal of propulsion strategies. Chemical propulsion – using platinum catalysts or reactive metals to decompose hydrogen peroxide or generate bubbles – offers high speeds but is often limited by biocompatibility. Enzymatic systems, such as urease‑ or catalase‑powered motors, can utilize endogenous fuels (urea or tumor‑overexpressed H₂O₂), moving closer to clinical safety. Physical‑field propulsion is the most mature and clinically promising: magnetic fields enable deep‑tissue penetration and precise steering; acoustic fields provide non‑contact actuation with integrated imaging feedback; and optical systems offer high spatiotemporal control for superficial applications. Biohybrid designs – combining living cells (bacteria, sperm, algae, immune cells) with synthetic components – leverage natural chemotaxis, immune evasion, and tissue‑homing capabilities. “Each propulsion mode has strengths and limitations,” notes Professor Huang. “The key is matching the mechanism to the specific biological barrier – whether it’s the mucus layer, the dense tumor extracellular matrix, or the blood‑brain barrier.”
Navigation is equally critical. The review covers trajectory planning algorithms – from rapidly exploring random trees for static environments to reinforcement learning for dynamic, uncertain in vivo conditions – and closed‑loop feedback controllers that correct real‑time deviations. Importantly, the authors highlight the growing role of autonomous targeting paradigms, where microrobots use “inside‑out” physical intelligence to sense pH gradients, hypoxia, enzyme enrichment, or inflammatory chemokines and migrate chemotactically toward lesions without external guidance. Swarm intelligence further amplifies capabilities: thousands of microrobots can self‑assemble into reconfigurable formations, navigate collectively, and even perform distributed “sense‑and‑respond” tasks – though clinical translation of multi‑agent systems faces unique safety and regulatory hurdles.
On the release front, the review presents a rich toolkit of on‑demand mechanisms. Endogenous triggers (pH, redox, enzymes, osmotic pressure) exploit the pathological microenvironment to spontaneously release drugs at the target site. Exogenous triggers (magnetic, ultrasound, light, mechanical force) enable active, external control of release timing and dosage. Beyond single‑stimulus systems, the authors discuss programmed and logic‑gated release – sequential delivery of multiple drugs, AND‑gate logic requiring simultaneous signals, and even closed‑loop feedback systems where biosensors monitor biomarkers and dynamically adjust drug output. “The ultimate goal is a theragnostic microrobot that senses, decides, and actuates in real time, truly personalizing therapy,” says Professor Wang.
To organize the vast preclinical landscape, the review introduces a four‑tier hierarchy based on barrier traversal capability. Tier I cover local delivery within accessible cavities – gastrointestinal, urogenital, pulmonary – where microrobots have already shown strong therapeutic readouts, such as 90% tumor reduction in bladder cancer models. Tier II addresses systemic circulation, where magnetic steering combined with cell‑membrane camouflage (erythrocyte, platelet, or neutrophil coatings) extends blood half‑life and enables targeted accumulation. Human‑scale validation in hepatic arteries of pigs demonstrated a 1.7‑ to 2.6‑fold increase in target‑lobe delivery, marking a translational milestone. Tier III tackles deep tissue penetration, where helical, microneedle, or enzyme‑degrading robots actively drill through dense extracellular matrix. Tier IV, the most challenging, targets immune‑privileged sites – the brain, retina, testes – requiring trans‑barrier transport without disrupting local homeostasis. Recent examples include fluoroscopy‑guided magnetic fibers delivering doxorubicin to intracranial tumors in minipigs.
Despite this promise, the authors candidly address critical translational barriers. Biosafety remains paramount: metallic components can cause chronic toxicity, and immune evasion strategies (PEGylation, cell‑membrane coating) may trigger anti‑PEG antibodies or complement activation over time. Long‑term degradation kinetics and clearance routes are still poorly defined. System robustness – maintaining precise navigation and controlled release under pulsatile flow, tissue deformation, and sensor drift – demands advanced control algorithms and fail‑safe mechanisms. Regulatory and ethical challenges loom large: how do we validate “emergent” swarm behaviors that are inherently nondeterministic? How do we assign liability for autonomous in‑vivo decisions?
Looking forward, the authors envision a paradigm shift from “invader” designs that resist the biological environment to “symbiotic” platforms that actively harness host physiology – for example, using glucose or oxygen gradients as both fuel and guidance cues. Artificial intelligence will drive closed‑loop autonomy, while digital twins and organ‑on‑a‑chip platforms will accelerate preclinical validation. Near‑term clinical entry is most likely through ex‑vivo or easily accessible cavities (gastrointestinal, bladder, eye), where retrievability and monitoring are feasible. Mid‑ to long‑term strategies will expand to systemic circulation and deep tissues, requiring integrated “micro‑therapeutic units” with onboard sensing, logic, and actuation, alongside harmonized regulatory standards.
“This is not a distant dream,” concludes Professor Huang. “With the convergence of smart materials, AI, and microrobotics, we are building the foundation for a new era of precision medicine – where therapy is delivered not just to the right organ, but to the right cell, at the right time, with the right dose, and with intelligent feedback.”
Authors of the paper include Shihao Zhong, Anping Wu, Shanming Bai, Henwei Huang, Toshio Fukuda, and Huaping Wang.
This work was supported by the National Natural Science Foundation of China under Grants 62573060 and U22A2064, in part by the Beijing Natural Science Foundation under Grant L242023, and in part by the Fundamental Research Funds for the Central Universities under Grant 2025CX01003.
The paper “Autonomous Microrobots for Spatiotemporally Active Therapeutic Delivery and Controlled Release” was published in the journal Cyborg and Bionic Systems on Jun. 29, 2026, at DOI: 10.34133/cbsystems.0617.
Cyborg and Bionic Systems