Embedded Controller Rehab-Robot

Marcus Vinicius Zerbini Canhaço
University of São Paulo
Sao Carlos, Brazil
See full Bio Contact
Entry date: 02-Apr-2013
Final Submission: 01-Jul-2014
The improvement of embedded systems adopted by robotic applications has been increasing the demand for smaller size, greater mobility, lower power consumption, modularity, reliability and availability of a real-time operating system. Robot manipulators designed for rehabilitation therapy in humans are a source of several challenges related to achieve the security requirements and control performance.

Project details

Recent improvement in embedded systems technology fits well to the demand of robotic applications, namely, smaller size, greater mobility, lower power consumption, modularity, reliability and compatibility with real-time operating systems. Robotic systems that are designed for human rehabilitation therapy are a source of several additional challenges like safety requirements and a cooperative behavior of the motor controllers.

The objective of this project is to develop a structure that integrates the controller rehabilitation wrist robot (RWT) to a computer game based HMI. To this end, an embedded system was conceived, that relies on:

  • Windows CE operating system;
  • CANOpen industrial communication protocol to connect servo drivers and controller;
  • Toradex ® Colibri T20 Tegra 2 board;
  • Maxon servomotor and a EPOS driver.


The EPOS offers CAN and a serial communication interfaces for the low level control systems. A CAN network will be installed between the embedded systems and the robot servo drivers. [Fig.1]

Communication Structure

The project is based on the hypothesis that an improvement on the therapy process and consequently a better recovery of the patient´s motion capacity may be achieved due to an appropriate integration of robot and game technologies to the patients.

The robot system is constituted by a forearm bracket and one stabilizer handle for the hand (joystick). The DC servomotor is connected to the robot joint and it´s encoder allows the acquisition of the  joint angle history during the therapy session. This data also allow us to calculate velocity and acceleration values. These informations provide the physicists and therapists with a better picture of the patient condition.  

The control strategy is conceived to work in different conditions. If the patient is not able to produce the requested movements, the manipulator helps him/her providing additional torque in the same direction of the movement. This is called a collaborative operation mode. If the patient needs to improve muscle strength the controller may be configured and used to resist to the patient movements.

This project aims to improve interaction between patient, physicists and therapists, since the physicists can better monitor the work performed by the therapist with the patient. Recording procedures also open new possibilities, such as off-line data analysis and interpretation of the patient progress. All the movements, efforts and important data are recorded in a global database which can be accessed for later statistical analyses of one or several patients.


Nevertheless, rehabilitation procedures are very boring. To overcome this drawback we propose the additional development and integration of a computer game running on the Colibri T20 Tegra Board. The game should run inside the embedded system. The game’s purpose is to increase the patient motivation. The game is structured with: movements, amplitude, frequency and efforts. It allows the therapists to apply a more deterministic motion protocol.


The use of robotic technologies, data analyses will also help the therapist to focus on the therapy process itself, increasing his/her autonomy and improving the effectiveness. 

Comments Hide