Comparison of camera configuration for real-time drone route planning in 3D building maze

In this research, we investigate what camera settings are effective for an indoor automatic search system. We recommend installing RGB cameras with depth sensors like the Kinect and show how they should be installed to facilitate searches in indoor environments such as buildings with multiple floors. To validate camera configurations, the RTA∗ algorithm is used for automatic searching and we also measured how fast a drone could move to goal points in a simulation of a 3D-building model. We also studied various patterns of restart points because a drone has limited battery life, which restricts the available flight time. In the experiment, we allowed six batteries and each flight could last 600 seconds. This experiment showed that we should use three cameras positioned on the forward, upward, and backward of a drone to conduct a 3D building floor search because drones can easily rotate in the yaw direction, but cannot rotate in the pitch direction. We also showed that once the drone had returned to its start position for a battery replacement, it should restart from that point for effective searching.