Learning Sensor Based Risk Map Augmentation for Risk Aware UAS Operation

Unmanned Aerial Systems (UAS) have become increasingly popular and been identified as a good platform for range of tasks from surveillance inspection to delivery maintenance. In many these applications systems operate in environments that are frequented by people or contain sensitive infrastructure which the operation UAS thus poses physical risk terms damage case vehicle failure psychological privacy risks would make their less acceptable. To increase use it is important they can take into account when determining navigation strategies. While this sometimes be done based on prior information, such street building plans cities, priori information often not complete, making essential representations augmented real time sensor information. This paper presents an approach map augmentation uses learned identification aerial pictures fuse additional with data dynamically changing allows effective re-planning