Predicting Digging Success for Unmanned Aircraft System Sensor Emplacement

We have developed an autonomous, digging, Unmanned Aircraft System (UAS) for sensor emplacement. A key challenge is quickly determining whether or not a particular digging activity will lead to successful emplacement, thereby allowing the system to potentially try another location. We have designed a first-of-its-kind decision-making algorithm using a Markov Decision Process to autonomously monitor the activity of a digging UAS activity to quickly decide if success is likely. Further, we demonstrate through many experimental trials that our method outperforms other decision-making methods with an overall success rate of 82.5%.