Algorithm for Dynamic Robot Trajectory Planning Based on Semantic Object Detection Using a Mivar Expert System
Shen Qiujie, Gong Shengshuo, Anton Kotsenko, О. О. Варламов
Abstract
A route planning algorithm based on semantic object detection technology is proposed to improve the safety of robots in indoor environments. This algorithm combines twodimensional laser scanning data with information on the categories and positions of objects identified by detecting objects in images. Depending on the importance of the object category, the algorithm increases the obstacle area on the map, which allows the robot to avoid important objects (e.g. children) when planning a route and improves the safety of the robot in the presence of people. A corresponding prototype of the data fusion algorithm is implemented in the mivar expert system, which can be used in airports, railway stations, shopping malls and other premises to improve the safety of robots.