Design Of A Modular Distributed Control System For Robotic Exoskeletons

Design of a Modular Distributed Control System for Robotic Exoskeletons
L.F. Rossi1,2, A. Forner-Cordero1 and F.J. Ramirez-Fernandez2 1. Biomechatronics Laboratory. 2. Microelectronics Laboratory
Escola Politécnica. Universidade de São Paulo 05508-010 São Paulo/SP, Brazil E-mail: {luisf.rossi, aforner}@usp.br, jramirez@lme.usp.br

Abstract—In this paper a modular distributed controlarchitecture for robotic exoskeletons is presented. On a research environment, an exoskeleton can be used for experimental studies on several fields, such as motor control, assistive technologies or ergonomy. However, robotic exoskeletons are usually very expensive and time demanding to build, leading to a limited number of existing prototypes. A modular open system may enable the use of a singleexoskeleton device for different several simultaneous researches leading to a better resource utilization. A first step in the development of an open modular exoskeleton is the design of a distributed and modular control architecture for these devices. The presented architecture was divided into tree modules: central processor, nodes and network system. Although this approach is not completely novel, itis innovative on the fact that these three modules follow a standardization in order to enable direct replacement without modification on the other modules. The network system is the only application independent and a custom network was designed in order to be simple to implement without requiring advanced technical knowledge and without a loss of performance. In addition to the advantage ofbeing flexible and easy to modify , the presented architecture enables the implementation of bio-inspired hierarchical control routines, thus being suitable to validate theories and mathematical models. Keywords-component; exoskeletons; distributed control; real time systems; biomechatronics; modular system;

As there is no standardization on development of exoskeletons, there are no off-the-shelfcommercial electronics and mechanical parts available. Therefore, a lot of time is spent adapting existing hardware, and developing a custom system. In addition to this effort, due to the lack of commercial parts, prototypes are usually rather expensive. This makes it very impractical to build several systems. This limitation severely reduces the number of different researches that can be takensimultaneously. This is further aggravated if the exoskeleton is based on non-flexible architecture, leading to a very restricted range of applications. However, despite this problem, little has being said on the literature about systems developed for multiple research fields, with a modular architecture. While some show some degree of flexibility, it is usually a secondary benefit generated bydesign choices that aimed at some sort of performance improvement. In addition to the better resources utilization, a modular and flexible architecture, when based on a good standardization, can enable easy exchange of knowledge among different research groups. This would be very interesting in cooperative projects as each laboratory could specialize on certain aspects of scientific ortechnological problems, leading to a higher efficiency of the time utilization. Finally, this sort of standardization would also lead to more scalable and complex systems. This paper presents a flexible and modular control system architecture for powered exoskeletons. It is based on a distributed system with a custom designed communication network specific for complex robotic systems. The presentedarchitecture also enables the implementation of hierarchical bio-inspired control routines, being suitable for research environments. The rest of this paper is organized as follow: Section II briefly presents a few related works. Section III describes the system designed, explaining the modular division, and detailing each module operation. Section IV briefly presents a implementation of the presented...