Military Drone Lessons for Safer Self-Driving Cars
This analysis highlights critical lessons from U.S. military unmanned aerial vehicle (UAV) operations that self-driving car companies are currently overlooking. Author Missy Cummings, a former Navy fighter pilot and researcher, argues that autonomous vehicles frequently fail in routine scenarios like construction zones, necessitating remote human supervision. However, current industry practices, such as employing remote operators in the Philippines, ignore decades of military expertise regarding latency and interface design. The article details two forms of supervisory control: teleoperation, which is highly sensitive to communication delays, and remote assistance, which offers higher-level guidance. A key lesson involves latency; early military attempts at long-distance teleoperation resulted in accident rates sixteen times higher than manned missions due to signal delays and human neuromuscular lag. By examining thirty-five years of UAV challenges, the piece urges the autonomous vehicle industry to adopt proven military standards for remote supervision to enhance safety and reliability, rather than repeating historical errors in control station design and operator training.
Wire timeline
Military Drone Lessons for Safer Self-Driving Cars
This analysis highlights critical lessons from U.S. military unmanned aerial vehicle (UAV) operations that self-driving car companies are currently overlooking. Author Missy Cummings, a former Navy fighter pilot and researcher, argues that autonomous vehicles frequently fail in routine scenarios like construction zones, necessitating remote human supervision. However, current industry practices, such as employing remote operators in the Philippines, ignore decades of military expertise regarding latency and interface design. The article details two forms of supervisory control: teleoperation, which is highly sensitive to communication delays, and remote assistance, which offers higher-level guidance. A key lesson involves latency; early military attempts at long-distance teleoperation resulted in accident rates sixteen times higher than manned missions due to signal delays and human neuromuscular lag. By examining thirty-five years of UAV challenges, the piece urges the autonomous vehicle industry to adopt proven military standards for remote supervision to enhance safety and reliability, rather than repeating historical errors in control station design and operator training.
IEEE Spectrum