Abstract
On an ad-hoc network which is consisted of movable nodes, the network topology is dynamic because each node is adaptive. We apply it to the surveillance cameras system at a parking lot. A node with a camera sets on a car. And the node gets the car motion information from the acceleration sensor. In this paper, we discuss the classification of its car motion detection because there are two cases; one case is to drive the car by the driver, and the other case is to move the car by the another person. We will classify the car motion detection into two cases using Neural Networks based on acceleration sensor. And in order to detect the case as soon as possible, we discuss how long we must get the acceleration sensor data.
