Real-Time Indoor Path Planning Using Object Detection for Autonomous Flying Robots

Unknown closed spaces are a big challenge for the navigation of robots since there no global and pre-defined positioning options in area. One simplest most efficient algorithms, artificial potential field algorithm (APF), may provide real-time those places but fall into local minimum some cases. To overcome this problem to present alternative escape routes robot, possible crossing points buildings be detected by using object detection included path planning algorithm. This study utilized proposed sensor fusion method an improved classification detecting windows, doors, stairs these objects were classified as valid or invalid The performance approach was evaluated simulated environment with quadrotor that equipped camera laser imaging ranging (LIDAR) sensors navigate through unknown space reach desired goal point. Inclusion allows robot from areas where it is congested. has been tested different scenarios based on compared other APF methods. results showed methods reinforced algorithms similar same goals room. For outside current room, traditional quite unsuccessful reaching goals. Even though able targets, gave approximately 17% better than successful example achieving targets can also work discover building between rooms.