Semi-Autonomous In Vivo Computation in Internet of Bio-Nano Things

Abstract

Magnetically assembled bioresorbable nanoswimmers can be used to highlight small tumors, thereby increasing the diagnostic capability of existing medical imaging techniques. Built upon our earlier work, this paper proposes a novel in vivo computational framework for early cancer detection. Engineered nanoswimmers experience a change in their physical properties under the influence of tumor-induced biological gradients. The biologically sensed data by such bio-nano things (nanoswimmers) can either trigger an autonomous target-directed motion or be assisted through external manipulation for steering the swarm towards the target. Previously developed externally manipulable in vivo computation requires constant monitoring of nanoswimmers, introducing positioning and steering errors along with a limit on the swarm size. A parallel approach called autonomous in vivo computation helps to resolve the above drawbacks, but the tumor homing is slow contributing to a higher percentage of pre-detection loss of nanoswimmers. We propose the spot sampling strategy for an autonomous swarm which considers the whole swarm as a single entity for the purpose of its tracking and steering. We show through computational experiments (1) that the proposed semi-autonomous in vivo framework can achieve faster tumor sensitization in complex environments having static and mobile obstacles, and (2) that the spot sampling provides sufficiently precise data to steer the swarm towards the target, saving around 90% of the monitoring resource. Our proposed framework also helps to achieve a large swarm size (number of nanoswimmers) which in return can achieve higher deposition of nanoswimmers on malignant tumors.