New machine autonomy to enable military unmanned ground vehicles (UGVs) to reach maximum structural speeds
ARLINGTON, Va. – Computer scientists at the University of Washington in Seattle are continuing a project to develop machine autonomy that enables unmanned ground combat vehicles to maneuver safely over rough off-road terrain at speeds a human driver could achieve.
Officials of the U.S. Defense Advanced Research Projects Agency (DARPA) in Arlington, Va., announced a $10.5 million contract to University of Washington in September for the Robotic Autonomy in Complex Environments with Resiliency (RACER) phase-two program.
Such a vehicle should be able to operate quickly over unstructured off-road terrain at speeds limited not by the autonomy software or processing time, but only by onboard sensor limitations and vehicle mechanical limits. Its speed should be similar to that of a human driver or a tele-operated vehicle. This contract follows an initial $10.4 million RACER contract awarded to University of Washington in 2023 for the RACER program.
In addition to University of Washington, Carnegie Mellon University in Pittsburgh and NASA-Jet Propulsion Laboratory in Pasadena, Calif., have been involved in the RACER program.
The three RACER teams have conducted live on-vehicle tests at the National Training Center at Fort Irwin near Barstow, Calif., and at Camp Roberts, a California National Guard base in Monterey and San Luis Obispo counties in California.
The self-driving car industry is making rapid advances in on-road autonomous mobility of commercial cars, yet military off-road autonomy algorithms and capability development has lagged because of the increased complexity of the off-road problem, DARPA researchers say.
Military off-road must deal with 3D surfaces, hundreds of obstacles, incomplete mapping data, no defined road networks, and no driving rules. RACER seeks to develop machine autonomy algorithms that enable unmanned ground vehicles (UGVs) to operate at high speeds in open and complex terrain, with minimal interventions.
RACER phase-one focused on developing, maturing, and demonstrating algorithms on lightweight tactical all-terrain vehicles (LTATV) in three DARPA-hosted field experiments.
RACER phase-two is maturing technology, and increasing system speeds over long off-road distances and complex maneuver terrain with a decreasing number of human interventions. Phase-two, moreover, is using a combat-size vehicle for field demonstrations, while keeping to the LTATV for algorithm maturation.
DARPA is providing LTATVs equipped with a sensor suite, computing resources, an optional baseline autonomy stack, vehicle and sensor models, and initial data sets.
This brings the total value of the contract to $23.1 million. University of Washington will do the work in Seattle and Ellensburg, Wash, and should be finished by March 2026.
For more information contact University of Washington online at https://racer.cs.washington.edu, or DARPA at https://www.darpa.mil/program/robotic-autonomy-in-complex-environments-with-resiliency.
John Keller | Editor-in-Chief
John Keller is the Editor-in-Chief, Military & Aerospace Electronics Magazine--provides extensive coverage and analysis of enabling electronics and optoelectronic technologies in military, space and commercial aviation applications. John has been a member of the Military & Aerospace Electronics staff since 1989 and chief editor since 1995.