Modeling Human Steering Behavior in Teleoperation of Unmanned Ground Vehicles With Varying Speed

Area of Science:

• Robotics and Human-Machine Interaction
• Cognitive Modeling and Human Factors

Background:

• Prior human operator models predicted steering performance for unmanned ground vehicles (UGVs) at constant speeds.
• Extending these models is crucial for varying speed scenarios, incorporating acceleration/deceleration effects and teleoperation-induced time delays.
• A key challenge is parameterizing models for predictive capability without human subject data.

Purpose of the Study:

• To extend a prior human operator model for predicting steering performance in teleoperated UGVs under varying speed conditions.
• To incorporate a far-point speed control model and address challenges of acceleration, deceleration, and time delays.
• To develop a parameterization strategy enabling predictive model capabilities without human subject data.

Main Methods:

• The study adapted the ACT-R cognitive architecture and two-point steering model from previous work.
• A far-point speed control model was integrated to enable predictions for varying speeds.
• Human subject experiments were conducted to validate the extended computational model's performance.

Main Results:

• The parameterized computational model successfully predicted the trend of average lane-keeping error for human subjects.
• The model accurately forecasted the minimum achievable lane-keeping error under varying time delays.
• Validation confirmed the model's ability to capture human steering behavior in dynamic teleoperation scenarios.

Conclusions:

• The extended computational model effectively predicts human steering behavior in teleoperated UGVs operating at varying speeds.
• This model offers a valuable tool for simulation-based development and testing of teleoperated UGV technologies.
• The research facilitates replacing human operators in certain development and testing phases, enabling more efficient UGV system studies.