High-Performance Autonomous Behaviors for Next-Generation Naval Systems in Communication-Denied Envi
Abstract
The frontiers of disaggregated autonomy are being pushed to address new challenges. One major challenge is operating in a communicat,ion-denied environment. For example, an adversary may jam communications in a team of drones searching for a target, which reduces t,he drones situational awareness and makes them easier to evade. Unfortunately, most multi-agent coordi- nation algorithms predicate, their success on the ability to communicate when needed. And these algorithms are often developed in laboratories with offboard inf,rastructure, such as WiFi, which provides the assumed capabilities. However, real-world settings lack such infrastructure, and the d,enial of communications also denies access to any external sources of information, including other agents. Thus, under communication, denial, the majority of multi-agent collaboration strate- gies cannot be used, though this leaves open the vital question of what s,uch systems should do.Communication is a source of feedback in multi-agent systems, and its loss means that feed- back loops must be, driven by information from other sources. Communication-denied agents must use only (i) onboard sensors to gather information, (ii), onboard computers to process sensor data and make motion plans, and (iii) onboard actuators to act on what they sense and compute.,These onboard behaviors are constrained by runtime, which is limited by mission duration, and energy, which is limited by onboard fu,el or batteries. This project takes a realistic perspective on com- munication, by directly addressing the fact that it can be denie,d, and thus we also take a realistic perspective on the autonomy techniques that must replace communication-dependent behaviors. The,refore, this YIP project will design practical autonomy algorithms for communication- denied environments by addressing the practica,l limits on capabilities and resources that agents face in those environments.We will do so along two complementary lines of inquiry,: (i) How should each agent spend its resources among onboard sensing, computation, and control capabilities? All three use time and, energy, and mission runtime and onboard energy are limited. This project will optimize across time and energy usage in a unified wa,y, which will identify how agents should spend resources to gather, process, and act on information to attain the best possible perf,ormance under communi- cation denial. (ii) What resources are required to attain a desired level of performance? We will quantify th,e minimum resources required for a mission to meet a desired performance specifica- tion under communication denial. This will both,provide criteria for the quantity of resources to allocate at planning time, as well as certificates of feasibility or infeasibility, for allocated resources to meet a desired specification.This YIP project will develop new theory, algorithms, and experimental impl,ementations to address the challenges of communication-denied autonomy. It will also launch thePIs career- long goal of developing,theoretically grounded autonomy formalisms that succeed under realistic, non-laboratory conditions.Approved for Public Release
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jul 13, 2022
- Source ID
- N000142212435
Entities
People
- Matthew Hale
Organizations
- Office of Naval Research
- United States Navy
- University of Florida