Jonathan Eden

Musculoskeletal Static Workspace Analysis of the Human Shoulder as a Cable-Driven Robot

The workspace analysis of cable-driven parallel manipulators has been widely studied, where the cables have been considered as ideal force generators. Due to the differences in actuation dynamics, workspace analysis has not been previously conducted for musculoskeletal systems. In this paper, static workspace analysis is performed on the human shoulder with a Hill-type physiological muscle model. The key characteristic of physiological muscles is that its ability to produce force is dependent on its length. Such type of workspace analysis on musculoskeletal systems as cable-driven manipulators is proposed for the first time. The generated shoulder workspace is validated by comparing the range of motion to that from benchmarks of human data. The significance of considering physiological muscles is demonstrated by comparing the musculoskeletal workspace with that of systems with ideal force generators. The novel formulation provides a new computational approach to perform workspace analysis for a wider range of engineered and biological systems.

Cable Function Analysis for the Musculoskeletal Static Workspace of a Human Shoulder

The study of cable function allows the contribution of particular cables towards the generation of motion to be determined for cable-driven parallel manipulators (CDPMs). This study is fundamental in the understanding of the arrangement of cables for CDPMs and can be used within the design of optimal cable arrangements. In this paper, the analysis of cable function for the musculoskeletal static workspace of a human shoulder is performed. Considering the muscles within the shoulder as state dependent force generators, the set of muscles required in sustaining the gravity force is determined for each workspace pose. As a result, the set of poses that each muscle is responsible for (muscle function) can be computationally determined. By comparing the results to the muscle function from biomechanics studies, it is shown that the results from the proposed cable function analysis are consistent with that reported in the literature of human studies.

On the positive output controllability of linear time invariant systems

This paper considers the output controllability of autonomous linear control systems that are subject to non-negative input constraints. Based on the evaluation of the geometric properties of the system, necessary and sufficient conditions are proposed for the positive output controllability of continuous linear time invariant systems. To aid in the practical evaluation of positive output controllability, additional sufficient conditions are derived for which efficient numerical techniques exist. These conditions are evaluated over a set of numerical examples which support the theoretical results.

Fluid Motion Planner for Nonholonomic 3-D Mobile Robots With Kinematic Constraints

Fluid motion planners are a type of artificial potential field (APF) motion planners that use the differential equations of fluid flow to determine the desired trajectory. The fluid flow approach in motion planning can efficiently produce natural-looking trajectories. However, the differential equations used in previous studies are restricted to motion planning in 2-D environments. In this paper, the fluid flow approach is extended to a motion planning framework for 3-D mobile robots that avoids spheroidal obstacles. Compared with existing APF approaches, kinematic constraints in both speed and curvature are also considered. Possessing the efficiency of 2-D fluid motion planners, the proposed approach is able to plan natural-looking reference trajectories for nonholonomic 3-D mobile robots. The approach is demonstrated through various 3-D example scenarios. The work can be considered as a fundamental framework for 3-D fluid motion planning, where additional kinematic constraints and more complex scenarios can be incorporated.

CASPR-ROS: A Generalised Cable Robot Software in ROS for Hardware

In this paper, the software platform CASPR-ROS is introduced to extend the author’s recently developed simulation platform CASPR. To the authors’ knowledge, no single software framework exists to implement different types of analyses onto different hardware platforms. This new platform therefore takes the advantages of CASPR, including its generalised CDPR model and library of different analysis tools, and combines them with the modular and flexible hardware interfacing of ROS. Using CASPR-ROS, hardware based experiments can be performed on arbitrarily CDPR types and structures, for a wide range of analyses, including kinematics, dynamics and control. The case studies demonstrate the potential to perform experiments on CDPRs, directly compare algorithms and conveniently add new models and analyses. Two robots are considered, a spatial cable robot actuated by PoCaBot units and an anthropomorphic arm actuated by MYO-muscle units.