PosterPDF Available

Use a Motion Capture System to Control a Humanoid Robot


This poster is about the implementation of a motion capture system in a telepresence robot. For the motion capture process, the hardware used was the perception neuron of Noitom Ldt. And the open source program MyRobotLab to control the microcontrolers that control the servos. For the robot structure it was used the Inmoov.
I. Introduction
II. System Architecture
III. Biovision Hierarchy Format IV. Conclusion
The Axis Neuron require a lot of CPU power process
because in the tests was used an Intel Core i7-
4700MQ, in average it uses about 45%of the CPU,
and it was necessary to reduce the frame rate of the
Axis Neuron because the UART communication can
not work as fast as required for update the position of
the servo motors. In order to take advantage of the
precision of the Perception Neuron it is necessary to
use a more precise robot or improve the mechanisms
of the InMoov.
Fig. 1. Person wearing the Perception Neuron
and controlling the robot
The Motion capture systems (mo-cap) is a branch that has taken a part in
the field of telepresence, and the need of control a robot using a high level
precision Capture System. The Perception Neuron from Noitom
proposed in this work, as suit based in a set of neuron sensors that are
Inertial Measurement Unit (IMU) with 9 DOF and this system provide better
performance and accuracy than other solutions. Moreover is proposed the
use of My Robot Lab (MRL) [1] and an InMoov based Robot [2], both are
open-source projects that usually works together, MRL which is written in
java and provide modules to control the InMoov Robot but does not have
the module to control it with a mo-cap system.
Fig. 2. System Architecture Diagram Fig. 3. Axis Neuron: gets the data from the Perception Neuron Suit
and stream it over TCP/IP using Biovision Hierarchy Format(BVH)
V. References
[1] “My robot lab.” [Online]. Available:
[2] “Inmoov project.” [Online]. Available:
[3] “Axis neuron software.” [Online]. Available:
[4] M. Meredith and S. Maddock, “Motion capture file formats explained,”
Department of Computer Science, University of Sheffield, vol. 211, pp.
241244, 2001.
Diego Calderón, Carlos Girard & Ali Lemus {diegocdl, prospektrgirard, alilemus}
Alan Turing Research Laboratory, FISICC, Universidad Galileo
Fig. 4. A simplified BVH format example to illustrate data transmitted
over TCP/IP between the Axis Neuron and My Robot Lab
ResearchGate has not been able to resolve any citations for this publication.