Autonomous Maneuver & Mobility (AIMM)

Objectives
Revolutionize AI-enabled systems for autonomous maneuver that can rapidly learn, adapt, reason, and act in multi-domain operations.
• Can we develop resilient autonomous off-road navigation for combat vehicles at operational speed that can autonomously move to a position of advantage against a near peer adversary?
• Can we enable autonomous systems to reason about the environment for scene understanding with the ability to incorporate multiple sources of information and quantify uncertainty?
• Can unmanned vehicles autonomously maneuver on the ground against a near peer adversary as part of an MDO force?

Research Questions

  1. How do we develop autonomous mobility for combat environments?
    • Can we develop resilient autonomous off-road navigation for combat vehicles at operational speed that can independently move to a position of advantage against a near-peer adversary?
    • How can we predict what the environmental impact is on the next autonomous agent’s action?
    • How can autonomous systems learn in the field without violating operational constraints (e.g., safety)?
    • Can we develop scene understanding for robust autonomous behaviors? Can we learn from small data sets that don’t require massive a priori curated data? Can we learn where we should look next to accelerate performance?
    • Can we conduct autonomous route and area reconnaissance in support of a manned force to develop enemy situation and make initial contact with enemy forces? Can we build on the lessons learned from the autonomous car industry for the Army?
     
  2. How do we enable context-aware decision making for autonomous agents?
    • Can we enable autonomous systems to reason about dynamic environments for scene understanding with the ability to incorporate multiple sources of information and quantify uncertainty?
    • Given information on environmental factors and enemy resources, how can we predict the enemy’s intentions, plans, and actions?
    • Can we improve performance by incorporating higher-level decision making and reasoning?
     
  3. How do we integrate autonomous mobility and context-aware decision making to deliver autonomous maneuver designed for MDO environments?
    • Can unmanned vehicles autonomously maneuver on the ground against a near-peer adversary as part of an MDO force?
    • Can we enable autonomous systems to perform in a predictable manner, to support Human-Agent-Teaming by developing trust in the autonomous systems?
    • How do we build the capability to predict and certify autonomous system performance in a target domain/environment when testing and experimentation have been performed in a different domain/environment? How do we differentiate domains?
    • How can unmanned vehicles and Soldiers seamlessly maneuver through dynamic environments of varying complexities?