Speed optimization and reliability analysis of a self-propelled capsule robot moving in an uncertain frictional environment

Abstract

The dynamics of a self-propelled capsule robot for small-bowel endoscopy driven by its internal vibro-impact excitation is studied in this paper. Due to its complex anatomy, the frictional environment in the small bowel is uncertain, so this work aims to maintain the progression of the robot at a desired velocity in the presence of such an uncertainty by using a new optimisation method. The optimisation method consists of the Six Sigma and the Multi-Island Genetic algorithms, and its reliability analysis is carried out with the consideration of parametric and environmental uncertainties by using the Monte Carlo algorithm. In total, five different motions of the capsule, including fast, slow, forward, backward and hovering, are optimised. Extensive numerical studies show that the five desired motions can be fulfilled by various combinations of system and control parameters. Experimental verification is also carried out by using a prototype of the capsule robot to demonstrate the efficacy of the proposed method. A mismatch between the numerical optimisation and the experimental results for the backward motion of the prototype was observed. However, optimisations for forward and hovering motions show good agreements with experimental observations. Potentially, the proposed approach can be used for optimising various progressive robots in different scales with multiple control objectives and constraints.