Microrobots offer immense potential for biomedical applications due to their small size and maneuverability, enabling minimally invasive in situ diagnosis and treatment, particularly targeted drug delivery. While previous research has focused on precise motion control often relying on external visual feedback, integrating self-perception remains a challenge. Self-sensing microrobots, however, offer unique advantages for local diagnosis, leveraging the higher biomarker concentrations near diseased tissue compared to diluted bodily fluids. This allows for noninvasive targeting of disease sites for early detection. Beyond fixed-point detection, microrobots can patrol the body for abnormalities and facilitate in situ disease mapping, continuously monitoring microenvironmental changes. While fluorescence sensing has been used, its tissue penetration depth is limited. This paper proposes a wireless self-sensing AI microrobot utilizing local magnetic field enhancement in electromagnetic imaging for noninvasive in vivo monitoring. The microrobot comprises a sensor head (a tunable RF coil) and a magnetic tail for external field-driven actuation. The thin-film-electrode sensor, constructed as an inductor-capacitor circuit, enhances local signals, transmits data wirelessly without onboard power, and responds to environmental changes such as pH and temperature alterations, thus offering a new sensing method for enhanced in vivo monitoring.

Existing microrobotics research emphasizes precise motion control using external visual feedback and control. Embedding self-perception for local diagnosis in precision medicine remains challenging. Fluorescence sensing, while offering high resolution, is limited by its shallow tissue penetration (less than 0.5 mm for single-photon excitation and less than 1.5-2 mm for multiphoton excitation). The need for a new sensing method for deeper penetration and real-time in vivo signal reading motivates the development of the self-sensing AI microrobot presented in this study.

The proposed wireless self-sensing AI microrobot consists of a sensor head acting as a tunable radiofrequency (RF) coil and a magnetic tail for external magnetic field actuation. The sensor head's thin-film-electrode sensing device is an inductor-capacitor (LC) circuit created using through holes, enhancing local signals for wireless transmission without onboard power. The working range is determined by the sensor circuit structure, with the number of interdigital electrodes influencing capacitance and resonance frequency. Environmental conductivity affects the enhancement of the local RF field by altering the circuit resistance. Changes in the surrounding environment, particularly pH and temperature, influence the sensor circuit's resistance and resonance frequency, providing real-time environmental signal monitoring. When the sensor's resonance frequency matches the imaging device's operating frequency, bright spots appear at the microrobot's location in the image. The fabrication process involves sequential deposition and etching of layers (Al2O3, SiO2, Si3N4, SiNx) on a Si substrate using 3D rolling-up technology to create the helical tail. The sensor head's interdigitated electrodes are fabricated on a SiO2 substrate. The RLC circuit's resonance frequency (fs) and quality factor (Qs) are calculated using the standard equations: fs = 1/(2π√(LsCs)) and Qs = (1/Rs)√(Ls/Cs). The surrounding environment's impact is modeled by considering the equivalent resistance (Rs) and capacitance (Cs) of the interdigitated electrodes, affected by environmental conductivity and permittivity. High-frequency structure simulator (HFSS) was used to simulate the electrical characteristics of the interdigital electrodes, examining impedance changes with varying frequencies and environmental conductivity.

The study successfully demonstrated a self-sensing AI microrobot capable of wireless signal transmission and environmental response without an onboard power source. The microrobot's interaction with external electromagnetic imaging equipment significantly enhances the local RF magnetic field, achieving a depth and spatial resolution superior to existing techniques. The enhancement effect was demonstrated to be modulated by the surrounding environment, reaching up to ~560 times. The microrobot's movement, controlled by an external magnetic field acting on its magnetic tail, was precise and reliable. The fabrication process allowed for the creation of functional microrobots, showcasing the feasibility of producing microrobot swarms. The sensor circuit’s response to changes in pH and temperature confirmed its ability to monitor various biochemical parameters. The use of high-frequency structural simulation (HFSS) validated the design and functionality of the interdigitated electrodes, showing that impedance changes corresponded to changes in the surrounding environment, demonstrating the capacity for real-time environmental monitoring.

The successful development of this self-sensing AI microrobot addresses the limitations of existing microrobotic technologies by integrating self-perception and wireless signal transmission capabilities. The enhanced sensitivity and deep tissue penetration achieved through local magnetic field enhancement significantly improve the capabilities for in vivo monitoring. The minimally invasive nature of this technology opens exciting possibilities for early disease detection and diagnosis, particularly in areas previously inaccessible to conventional medical imaging methods. The ability to monitor changes in pH and temperature offers a powerful tool for tracking microenvironmental changes associated with disease progression. The swarm-like fabrication process paves the way for large-scale production and application. The findings demonstrate a significant advance in the field of microrobotics and have potential applications in various biomedical fields.

This research presents a novel self-sensing AI microrobot for noninvasive wireless monitoring. The combination of local magnetic field enhancement and a passive wireless sensing mechanism offers improved sensitivity and deep tissue penetration. The ability to monitor environmental changes and the scalability of the fabrication process demonstrate the significant potential of this technology for in vivo diagnostics and therapeutic applications. Future research could focus on exploring additional sensing modalities, improving the microrobot's mobility, and integrating more sophisticated control algorithms.

The current study primarily focused on proof-of-concept demonstrations in simulated environments. Further in vivo studies are required to fully validate the microrobot's performance in complex biological tissues. The long-term stability and biocompatibility of the microrobot materials need further investigation. Additionally, the range of detectable biochemical parameters could be expanded by exploring different sensor designs and materials.