Robotic Active Capsule Endoscope (RACE): advanced design, fabrication and control

该项目将研究开发两种先进的比赛原型，有线和无线版本，能够在下胃肠(GI)轨道模型中向前、向后和旋转运动。建议的原型将在审查当前原型的局限性的基础上设计和制造，使用玛丽女王创新(QMI)的概念验证资金制造，以确保在QMI指定的专利律师‘Kilburn and Strode’的指导下获得专利。目前的样机都有基本的控制机构，需要设计先进的闭环控制方法，以基于实时位置和姿态参数传感的比赛样机的平稳导航。将研究两种方法来获得胶囊的实时位置和方向：(I)插入胶囊内的加速度计和(Ii)透视。作为第一步，将开发和实施比例积分和导数(PID)控制方法，以建立基于所获得的参数的胶囊反馈控制的可行性。然而，由于整个系统被认为是高度非线性的，非线性控制方法有望获得更好的性能。因此，将研究一种基于人工智能(AI)的控制方法，如基于模糊逻辑的控制方法。为了减少组织损伤的可能性，将研究(I)胶囊模型和(Ii)胶囊与胃肠道表面相互作用动力学模型的非线性模型预测控制(NMPC)方法。

The project will investigate the development of two advanced RACE prototypes, wired and wireless versions, capable of forward, backward and rotational movements in a lower Gastrointestinal (GI) track phantom. The proposed prototypes will be designed and fabricated on the basis of reviewing the limitations of the current prototypes, fabricated using Queen Mary Innovation (QMI) Proof-of-Concept funding, to secure patents as guided by the QMI assigned patent attorney 'Kilburn and Strode'. The current prototypes are facilitated with basic control mechanism and advanced closed-loop control methodologies are required to be designed for smooth navigation of the RACE prototypes based on real-time position and orientation parameters sensing. Two methods will be investigated to obtain real-time position and orientation of the capsule: (i) accelerometer inserted inside the capsule and (ii) fluoroscopy. As a first step, a Proportional Integral and Derivative (PID) control method will be developed and implemented to establish the feasibility of feedback control of the capsule based on the obtained parameters. However, as the overall system is supposed to be highly nonlinear, a nonlinear control approach is expected to achieve better performance. As such, an artificial intelligence (AI) based such as fuzzy-logic based control approach will then be investigated. To reduce the possibility of tissue damage a nonlinear model predictive control (NMPC) approach incorporating (i) the model of the capsule and (ii) the model of the interaction dynamics between the capsule and GI tract surface will be investigated.