Design of a soft robotic endoscope with enhanced bending and AI-based prediction

Abstract Minimally invasive surgery (MIS) increasingly benefits from soft robotic platforms that offer enhanced compliance and safer human–robot interaction. However, one of the most important challenges of design is still to achieve larger bending at low actuation pressure while maintaining structural integrity. This paper introduces a proposed novel soft robotic endoscope aimed at maximizing bending performance and reducing needed pressure in complex surgeries below the maximum limit of blood pressure, which is 0.24 bar (Boutouyrie in Hypertension 34(4):475–480, 1999), to enhance patients’ safety and reduce their trauma. The proposed design demonstrates significant improvements in both bending angle and operating pressure compared to existing architecture. Finite element modeling was utilized to evaluate the performance of the novel design under several test cases. Then, machine learning techniques were implemented to develop an Artificial Intelligence (AI) model aimed at predicting the most effective parameters needed for testing the design. Instead of implementing Finite Element Analysis for each new design, a highly accurate model will now predict bending angle, twisting angle, bending speed and twisting speed for the proposed design under any input pressure conditions. Experimental validation was developed by fabricating the module and testing it to validate the proposed design’s performance. The results show that the proposed design achieves higher bending angles at same pressures, compared to previous designs, under the specified design conditions.