| |
Behavioral Cloning and Obstacle Avoiding for Two Different Autonomous Vehicle
| [View]
[Download]
|
| Author(s) |
Ranka Kulic, Zoran Vukic |
| Abstract |
The problem of dynamic path generation for an
autonomous vehicle in environments with unmoving and
moving obstacles is presented. The first vehicle kinematical
and then an complete mathematical model is given by
nonlinear equations describing a 12 state dynamical system
simulated in Matlab or Simulink environment. The goal is
to find the regulator design method for the named
autonomous vehicle in 3D- space in situation with infinite
number of obstacles. To design the vehicle regulator in this
paper the behavioral cloning approach is applied. In
behavioral cloning, the system learns from control traces of
a human operator. The advantage of the named approach
lies in the fact that a complete path can be defined without
using sophisticated symbolical models of obstacles. The
proposed methodology it substantially differs from the
others. The other advantage of the proposed methodology is
the fact meaning that some of the learning algorithms
enable a complete path off-line defining. |
| Download |
| Filename: | T5-601.pdf |
| Filesize: | 435.6 KB |
|
| Type |
Members Only |
| Date |
Last modified 2007-03-12 by System |
| |
 |