| Title |
Indoor Environment Recognition of Mobile Robot Using SVR |
| Authors |
Jun-Hong Shim ; Jeong-Won Choi |
| DOI |
http://dx.doi.org/10.5207/JIEIE.2010.24.8.119 |
| Keywords |
Mobile Robot ; Ultra Sonic Sensor ; SVR ; RBF Kernel |
| Abstract |
This paper proposes a new solution about physical problem of autonomous mobile robots system using ultrasonic sensors. An mobile robot uses several sensors for recognition of its circumstance. However, such sensor data are not accurate all the time. A means of removing the noise that sensor data contains constantly, It is possible for simulation to estimate its circumstance based on ultrasonic sensor data by learning algorithm SVR(Support Vector Regression). To use SVR, it is being selected parameter and kernel which are the component of SVR. Selecting the component of SVR, the most accurate parameter data was selected through the tests because it is not existed determined data. In addition, choosing the kernel uses RBF(Radial Basis Function) kernel which is the most generalized. This paper proposes SVR based algorithm to compensate for the above demerits of ultrasonic sensor through the experimentation under three different environments. |