Vehicle Teleoperation: Human in the Loop Performance Comparison of Smith Predictor with Novel Successive...

Area of Science:

• Robotics and Control Systems
• Human-Machine Interaction
• Autonomous Vehicle Technology

Background:

• Vehicle teleoperation enables remote control of autonomous vehicles when automation fails, bridging manual and automated driving.
• Key challenges in teleoperation include variable time delays, actuator saturation, and environmental disturbances, which degrade performance.
• Existing predictive methods for delay compensation fail to account for sudden disturbances, leading to inadequate steering and poor path tracking.

Purpose of the Study:

• To introduce and evaluate the Successive Reference-Pose-Tracking (SRPT) approach for vehicle teleoperation.
• To compare the path-tracking performance of SRPT against the Smith predictor in human-in-loop experiments.
• To assess the effectiveness of SRPT in mitigating the impact of delays and disturbances on teleoperation.

Main Methods:

• Developed a novel SRPT method transmitting successive reference poses instead of direct steering commands.
• Generated reference poses using a joystick steering wheel interface.
• Conducted human-in-loop experiments with 18 drivers in a simulation integrating a 14-DOF Simulink vehicle model and Unity game engine, incorporating bidirectional variable delays.

Main Results:

• The SRPT approach demonstrated significant improvements in reference tracking compared to the Smith predictor.
• Human effort required for teleoperation was substantially reduced across all tested scenarios with SRPT.
• SRPT effectively handled challenging conditions including low adhesion, lateral wind, tight corners, and obstacle avoidance.

Conclusions:

• The SRPT approach offers a superior method for vehicle teleoperation, enhancing control accuracy and reducing operator workload.
• Transmitting reference poses is more robust to delays and disturbances than transmitting steering commands.
• SRPT represents a significant advancement for reliable and efficient remote operation of autonomous vehicles.