The robotic field has been witnessing a progressive departure from classic robotic systems composed of serial/stiff links interconnected by simple rigid joints. Novel robotic concepts, e.g., soft robots, often maintain a series-like structure, but their mechanical modules exhibit complex and unconventional articulation patterns. Research in efficient recursive formulations of the dynamic models for subclasses of these systems has been extremely active in the past decade. Yet, as of today, no single recursive inverse dynamics algorithm can describe the behavior of all these systems. This paper addresses this challenge by proposing a new iterative formulation based on Kane equations. Its computational complexity is optimal, i.e., linear with the number of modules. While the proposed formulation is not claimed to be necessarily more efficient than state-of-the-art techniques for specific subclasses of robots, we illustrate its usefulness in the modeling of different complex systems. We propose two new models of soft robots: (i) a class of pneumatically actuated soft arms that deform along their cross-sectional area, and (ii) a piecewise strain model with Gaussian functions.

翻译：机器人领域正逐步脱离由刚性关节连接传统串行/刚性连杆构成的经典机器人系统。诸如软体机器人等新型机器人概念虽保持类串行结构，但其机械模块展现出复杂且非传统的关节运动模式。过去十年间，针对此类系统子类的递归高效动力学模型研究异常活跃。然而，截至目前，尚无单一递归逆动力学算法能完整描述所有此类系统的行为。本文通过提出基于凯恩方程的新型迭代公式解决了这一挑战。该公式计算复杂度达到最优，即随模块数量线性增长。虽未声称所提公式在特定机器人子类中必然优于现有技术，但本文通过不同复杂系统的建模验证了其实用性。我们提出了两种新型软体机器人模型：(i) 变形沿截面面积发展的气动软体机械臂类，以及(ii) 采用高斯函数的分段应变模型。