Rovio mobile webcam, Pioneer P3-AT mobile
robot, Octomap by K. Wurm et al,
MS Kinect
MSc
dissertation proposal 2014/2015
Vision
Based Navigation for Humanoid Robots
Introduction:
This project is framed on a CMU Program project
proposal whose aim is to research and develop a novel Robotic Assistance
Platform designed to support healthy lifestyle, sustain active aging, and
support those with motor deficits. It involves research universities and SMEs from
Objectives:
IST is involved in developing navigation and
control algorithms for mobile platforms. These platforms consist of tether free
robotic wheeled robots, which must be capable of indoor navigation in
structured environments (e.g. hospital floors, daycare centers). Due to the
limited communications capabilities, most processing must be done onboard which
implies a focus of research on efficient methods and algorithms for obstacle
detection, trajectory planning and map based navigation. The work will be implemented
on the robot VIZZY consisting of a humanoid head and torso mounted on a segway platform. This platform will be used to evaluate the
performance and reliability of the obtained solution.
Detailed Description:
This work will focus on autonomous robot
navigation in indoor scenarios. The student will be acquainted with existing 3D
obstacle detection algorithms and navigation solutions, and will extend these
so as to have the VIZZY robot navigating autonomously inside corridors.
Although current solutions involve only vision, the integration with range data
(kinect) is suggested.
The work is broken into the following subgoals which will need to be completed:
LOCALIZATION
For this work it can be assumed that a known
map is provided. The agent will need to continuously localize itself with
respect to this map, taking into account its motor commands
OBSTACLE DETECTION
Although a map is provided, some dynamic
objects will not appear no the map. These include people wandering in the
corridors, and other possible inanimate obstacles. The agent needs to be aware
of its surroundings and plan routes which will effectively avoid these
obstacles or, as a last resort, stop to avoid collision.
NAVIGATION
The agent is expected to be able to navigate towards
any point on the map by generating a suitable trajectory and following it.
Expected Results:
The students will develop theoretical
navigation, computer vision and signal processing. This project might be
integrated in a dynamic, multidisciplinary and international team and will have
the opportunity to contact with the best research teams and SME’s
in the field. The final result must consist in the demonstration of an
application illustrating the successful achievement of the theoretical and
practical aspects of the work.
Observations:
--
More MSc dissertation proposals on Computer and
Robot Vision in:
http://omni.isr.ist.utl.pt/~jag