Navy researchers start push to improve the autonomy and intelligence of unmanned vehicles

by John Keller

ARLINGTON, Va. - Experts in the U.S. Office of Naval Research (ONR) in Arlington, Va., are launching a project to improve the intelligence and autonomy of unmanned vehicles for use on land, at sea, and in the air. To do this, ONR officials envision several simulation demonstrations aimed at improving existing and future unmanned vehicles and control stations.

Intelligent autonomous vehicles help Navy and Marine Corps commanders to make the most of available manpower, remove sailors and Marines from unnecessary harm, and to increase situational awareness.

Current unmanned vehicles rely heavily on human operators for control, experts say. Significant workload from several skilled operators often is necessary to control even one unmanned vehicle. Even with heavy human workloads, most existing unmanned vehicles are unreliable, have poor survivability, and require considerable time and effort to allow replanning when unexpected situations arise, experts say.

The new project, called Development and Demonstration of Intelligent Autonomy in Unmanned Vehicles (solicitation number 02-024), seeks to add capability to help unmanned vehicles:

• replan to deal with failures, damage, attrition, new threats and targets, changes in no-fly zones or rules of engagement, new intelligence, as well as enabling operators to input new mission plans quickly;
• combine tactics, countermeasures, and maneuvers autonomously or semi-autonomously to avoid and evade threats; and
• work as a team, with multiple sensors, payloads, or weapons, with separate operators monitoring sensor data. Limited-fidelity simulation demonstrations are to mature high-risk technologies.

The dynamic replanning and autono-mous vehicle control capability deals with rapid replanning of route, data collection, and communications. The capability determines the path the vehicle may need to take considering terrain obstacles, the vehicle's ability to maneuver, environmental factors such as weather or ocean currents, threats, and payload capabilities.

In addition, data collection replanning deals with issues such as sensor control and imaging requirements, while communications replanning deals with how and when to transmit data and takes into account issues such as potential line-of-sight link locations, satellite availability, and communications frequencies.

Autonomous threat response will help the vehicle avoid and evade threats, and operate closely to threats, by enabling the vehicle to reason about the world; developing sensors to detect threats is out of the scope of this effort.

Multi-vehicle co operation will focus on distributed co op erative control al gorithms that en able teams of un manned aircraft, ground ve hicles, subma rines, and surf ace vessels to work together.

This capability will focus on distributed algo rith ms for coop eration be tween ve hicles on coupled tasks with strong timing constraints and the possibility of un expected contingencies. Control theory and artificial intelligence ap proach es are appli cable to this effort.

ONR officials say they expect to award as many as four contracts in fiscal year 2003. At first, technologies are to be developed and demonstrated in a simulation environment with a common interface to integrate with simulations such as Joint Integrated Mission Model (JIMM) or Extended Air Defense Simulation (EADSIM), vehicle/sensor simulations, displays, and other simulation interfaces.

Planned software interfaces are Distributed Simulation/High Level Architecture (DIS/HLA), Common Object Request Broker Architecture (CORBA), and User Datagram Protocol/Internet Protocol (UDP/IP). The DIS/HLA interface has the most computational overhead and the UDP/IP interface has the least computational overhead. For that reason, the latter may be most suitable since it will allow offerors to use the interface that is most cost-effective for their project.

Later, contractors may switch to integrated hardware demonstrations to validate simulation results. Proposed technologies should be applicable to several different platforms to include air, ground, undersea, and sea surface vehicles.

Submit technical questions by e-mail to Alan C. Schultz at [email protected], and contractual questions by e-mail to Edwin Little at [email protected]. For more information look up the project solicitation on the World Wide Web at http://www2.eps.gov/spg/USN/ONR/ONR/02-024/listing.html.