Artificial Intelligence and Machine Learning for Autonomous Military Vehicles
Abstract
Autonomous vehicles are becoming reality for civilian applications. In the form of intelligent driving assistance, the vehicle autonomy of the third level (smart cruise control, pedestrian recognition, automatic braking, blind zone sensors, rare cross-traffic alerts, collision avoidance, etc.) has been available for commercial and private vehicles for number of years. The autonomy of the fourth and fifth level (supervised autonomy and full unsupervised autonomy) are currently in trials. Despite a substantial progress in this area in civilian applications, autonomy for military vehicles is still quite a challenging task. The main distinctions of military autonomous vehicles are: off road operation, unknown terrain for operation, and a possibility of complete re-routing in the open space. This environment requires different algorithms and environmental awareness for intelligent autonomy controls than those used for civilian applications in the industry. Specifically, the tasks of advanced and current terrain awareness, detection of impassible routes, determination of passible alternative routes and vehicle re-routing in the open space, and optimal vehicle control for a given terrain condition and vehicle need to be solved. The presented work describes recent progress in solving some of such challenges. The results indicate that some of the challenges can be successfully solved by machine learning and artificial intelligence algorithms, thus, providing a substantial aid in manual driving of military vehicles.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 2020
- Accession Number
- AD1106872
Entities
People
- Garett Hoch
- Jacob Desmond
- James Lever
- Jordan Bates
- Mark Bodie
- Michael Parker
- Sally Shoop
- Sergey Vecherin
- Taylor Hodgdon
Organizations
- Engineer Research and Development Center