Research on Actuation, Tracking and Closed-loop Control of Bioinspired Magnetic Microrobot

Abstract

Gastrointestinal cancer poses a significant threat to human life and health, with colorectal cancer and stomach cancer exhibiting among the highest fatality rates across all cancers, a trend that has intensified in recent years. Timely diagnosis and treatment are imperative in mitigating the progression of gastrointestinal ailments. The development of efficacious medical examination devices stands to aid medical practitioners in early lesion detection and subsequent interventions to impede disease advancement. Presently, gastroenteroscopy, encompassing tethered and wireless endoscopes, serves as the primary modality for assessing the condition of the gastrointestinal tract.

Conventional endoscopes on the market predominantly operate passively, propelled by the gastrointestinal tract’s peristalsis. However, such endoscopic robots encounter challenges of diminished inspection efficiency and prolonged examination durations. While magnetic capsule endoscopes have been developed by some researchers, their capacity for on-demand biopsy and drug release is limited by their constrained loadbearing capability.

Moreover, real-time position feedback is paramount for intervention and treatment. Traditional vision-based position-tracking methods falter in vivo due to inadequate lighting conditions. Similarly, prevalent in vivo target imaging technologies—such as ultrasound, fluorescence imaging, and radiation-based methods—prove unsuitable for sustained capsule robot position tracking due to their inherent risks to human health and fluctuating tracking accuracy. Recent advances in magnetic sensor array-based tracking techniques have emerged as a preferred means for localizing in vivo targets, leveraging magnetic fields that penetrate the human body without posing health risks. Nonetheless, extant magnetic localization systems and algorithms exhibit shortcomings, including confined positioning ranges and susceptibility to interference from electromagnetic equipment and geomagnetic fields.

Given the limitations inherent in current magnetic capsule robots, this study proposes a capsule-sized magnetic robot inspired by the innovative structure of a caterpillar larva (i.e., Manduca sexta). The devised robot boasts a substantial load capacity exceeding 300 times its own weight. Additionally, a dynamic tracking system is introduced to achieve precision positioning across the human body. Moreover, a multipoint simultaneous tracking algorithm is proposed to mitigate background field noise effects on tracking outcomes.

The bioinspired microrobot, named MiaBot, draws inspiration from the segmented body-wall and viscera structure, as well as the piston-like visceral locomotion mechanism of a caterpillar larva (i.e., Manduca sexta). MiaBot, featuring internal magnetic actuation, integrates a dual-coil array positioned at opposite ends of a plastic skeleton, with a permanent magnet housed within the skeleton to mimic the larva’s segmented structure. The magnet oscillates within the skeleton akin to piston-driven visceral locomotion upon the application of alternating current to the circuit. With its formidable force output, MiaBot demonstrates exceptional cargo-carrying capabilities across various terrains and exhibits superior directional control, swifter turning speeds, and narrower turning radii compared to its counterparts.

Furthermore, MiaBot holds promise for biomedical applications, particularly in accessing challenging tubular environments for tasks like biopsy and drug delivery. The bioinspired design and unique actuation mechanism exemplify a new frontier in microrobotics, particularly for deployment in high-friction environments and confined spaces.

To track the magnetic microrobot in real-time, a novel dynamic tracking solution incorporating a movable sensor array is proposed to maintain consistently high tracking accuracy over an expansive area. In this solution, magnet tracking accuracy is initially optimized within a confined range, with the sensor array repositioned by an external robotic arm should the target microrobot exceed this optimized range. Additionally, a multipoint locating algorithm is introduced to mitigate varying background noise effects, culminating in enhanced magnetic tracking range and accuracy, thereby augmenting microrobot position feedback in medical applications.

The proposed robust magnetic tracking system leverages the magnetic density of multiple sampling points for simultaneous localization, offsetting background noise by subtracting magnetic field values at different positions. This method enables multipoint simultaneous positioning through optimization algorithms, with simulation analyses elucidating the impact of signal-to-noise ratio on localization accuracy. Experimental validation conducted under various geomagnetic noise and permanent magnet environments corroborates the robustness of the proposed method across disparate background noise environments. The envisioned application of this method in wearable systems for tracking magnetic capsule endoscopes holds substantial promise.

In summary, this thesis addresses several limitations encountered by current gastrointestinal microrobots, spanning enhancements in propelling force and workspace, expanded tracking coverage, and enhanced anti-interference capabilities against background noise. The proposed solutions are poised to facilitate the integration of capsule endoscopy robots into clinical settings, enabling the completion of real-world tasks.

Date of Award	29 Aug 2024
Original language	English
Awarding Institution	City University of Hong Kong
Supervisor	Jun LIU (Supervisor)