7 Platform one Uncrewed Systems Technology | June/July 2024 US research agency DARPA has been testing autonomous operations on heavy tracked vehicles, writes Nick Flaherty. The DARPA Robotic Autonomy in Complex Environments with Resiliency (RACER) programme has successfully tested autonomous movement on a new, much larger fleet vehicle – a significant step in scaling up the adaptability and capability of the underlying RACER algorithms. The RACER Heavy Platform (RHP) vehicles, weighing 12 t, are 20 ft-long, skid-steer, tracked vehicles, based on a M5 base platform from Textron. This was developed and used in US Army campaigns of learning for robotic combat vehicle requirements, but it has been fitted with autonomous hardware and software stacks by Carnegie Robotics. The RHPs complement the 2 t, 11 ft-long, wheeled RACER Fleet Vehicles (RFVs) already in use, which have the same Ackermann steering as a car. “Having two radically different types of vehicles helps us advance towards RACER’s goal of platform agnostic autonomy in complex, mission-relevant, off-road environments that are significantly more unpredictable than on-road conditions,” said Stuart Young, the RACER programme manager. “For Phase 2, adding the combatscale RHP robot supports porting and performance demonstration of RACER autonomy stacks at multiple scales concurrently, while moving between highly varied terrains.” RACER’s second phase included the first testing of the heavy vehicles and testing of the RFV fleet vehicles by teams from the University of Washington and NASA’s Jet Propulsion Laboratory. “Our Phase 2 off-road, average autonomous speed goals are higher at lower intervention rates, and both RFVs, plus now RHPs, allow RACER to show adaptability and resiliency of autonomous software at multiple, platform-agnostic, ground robot scales in an array of complex, military-relevant environments,” said Young. “As we also add tactics-based autonomy, we see all of these together as vitally important to army and marine needs in robotic vehicle programmes of record that are closely tracking RACER, and which represent possible transition opportunities for the programme.” Using fully-autonomous RFVs, RACER showed autonomous movement within a 15 square mile terrain that included highly diverse ground vegetation cover, trees, bushes, rocks, slopes and obstructed ditches during the day and at night. The two RACER teams had not previously operated in, nor been exposed to, sensor data sets of the training areas and they were not given any practice time before starting the official courses. The teams successfully completed more than 30 autonomous runs on courses varying from three to 10 miles in length, achieving over 150 autonomous, unoccupied miles at speeds of up to 30 mph. A heavy RHP operated for more than 30 miles in an autonomous-routefollowing-mode over similarly complex terrain to test low-level autonomous control, collect sensor data sets, assess mobility and refine operations. “RACER’s early Phase 2 activities, both with Experiment 4 performance successes in difficult, new-to-theprogramme, military-relevant terrain in Texas, as well as recent incorporation of RHP as a fleet platform, is setting the tone for the programme to achieve tougher autonomous manoeuvre goals while showing autonomy resiliency and adaptability to new environments on any robot at any scale,” said Young. Autonomous vehicles Autonomous experiment on heavy tracked vehicles The tracked autonomous ground vehicle, based on the Textron M5, in tests (Image courtesy of DARPA)
RkJQdWJsaXNoZXIy MjI2Mzk4