Human-Machine Shared Driving Control for Semi-Autonomous Vehicles Using Level of Cooperativeness

Chen Lv (吕辰), Anh-Tu Nguyen, Jagat Jyoti Rath, Chen Lv, Thierry-Marie Guerra, Jimmy Lauber

TL;DR

This paper introduces a novel haptic shared control system for semi-autonomous vehicles that dynamically adjusts the automation's assistance level based on the driver's cooperativeness and workload. By integrating a driver-in-the-loop model with a polytopic LPV control framework, the system aims to effectively manage driver-automation conflicts and enhance safety and comfort during lane keeping.

What Problem Does This Paper Address?

The increasing prevalence of semi-autonomous vehicles (SAE Level 3) presents a significant challenge in managing the interaction between the human driver and the automation, particularly during complex driving tasks like lane keeping. A key issue is the potential for driver-automation conflict, where the two agents may have differing objectives, leading to unpredictable behavior and reduced safety. Existing approaches often struggle to adapt to the dynamic nature of human driving behavior, varying road conditions, and the inherent uncertainties in the system. This paper addresses the need for a robust shared control strategy that can effectively balance the control authority between the human and the automation, ensuring seamless cooperation and minimizing conflicts.

Methods

The core of this work is a new haptic shared control concept that introduces a "level of cooperativeness" indicator. This indicator, derived from the driver's and automation's torque outputs, along with the driver's workload, determines the required level of haptic authority.

Key methodological components include:

  • Driver-in-the-Loop (DiL) Vehicle Model: A comprehensive model is developed, integrating vehicle dynamics (yaw-slip, steering), lane positioning, and a dynamic driver model. This model captures the complex interactions within the driver-vehicle system.
  • Level of Cooperativeness (CI) Index: A novel metric is proposed to quantify the alignment of driving objectives between the human driver and the automation. This index is calculated over a time window based on the torques generated by each agent.
  • Driver Activity Variable (η): This variable combines the cooperativeness index and the normalized driver torque to represent the driver's need for assistance.
  • Assistance Factor (Γ(η)): A dynamic mapping is used to translate the driver activity into an assistance factor, which modulates the automation's torque output.
  • Polytopic Linear Parameter-Varying (LPV) Control: The DiL vehicle model is reformulated into a polytopic LPV framework. This allows for the design of a controller that can handle time-varying parameters like vehicle speed and the assistance factor, while guaranteeing stability and performance (specifically, an $l_\infty$-gain performance).
  • Lyapunov Stability Theory: This theory is employed to ensure the closed-loop stability of the system and to derive the control design conditions.

Datasets

Not discussed in the paper. The validation was performed using a high-fidelity simulation environment (MATLAB-SIMULINK) with a dynamic test track featuring varying curvatures and road friction conditions.

Key Insights and Ideas

  • Quantifying Cooperativeness is Key: The introduction of a "level of cooperativeness" index provides a concrete way to measure how well the driver and automation are working together, moving beyond qualitative descriptions.
  • Dynamic Authority Allocation: The system's ability to dynamically adjust haptic authority based on cooperativeness and workload is crucial for managing conflicts. When conflict arises, the automation reduces its authority, allowing the driver to take over.
  • LPV Control for Time-Varying Systems: The application of polytopic LPV control is well-suited for this problem, as it can effectively handle the inherent time-varying nature of vehicle speed and the adaptive assistance factor.
  • Integrated Driver Model for Realism: Using a detailed driver model that accounts for anticipatory and compensatory behaviors makes the simulation results more representative of real-world driving.

Key Results

The proposed shared control system demonstrated strong performance in simulations:

  • Lane Keeping: Low tracking errors ($|y_L|_{max} = 0.522$ m, $|\psi_L|_{max} = 0.063$ rad) and RMS values ($|y_L|_{RMS} = 0.338$ m, $|\psi_L|_{RMS} = 0.024$ rad) were achieved, indicating effective lane centering.
  • Driving Comfort: Good driver comfort was observed, with controlled steering rates ($|\dot{d}_a|_{max} = 1.686$ rad/s, $|\dot{d}_a|_{RMS} = 0.407$ rad/s) and lateral accelerations.
  • Vehicle Stability: The system maintained vehicle stability, with yaw rates within safe operating limits.
  • Driver-Automation Conflict Management: The system effectively reduced driver-automation conflict, as evidenced by the PRatio and Conflict metrics. The minimum conflict value was found to be greater than the design threshold.
  • Robustness: The controller showed robustness to vehicle parametric uncertainties (mass, inertia) and variations in driver parameters ($K_a$, $K_c$) across different road friction conditions.

Limitations and Future Work

  • Simulation-Based Validation: The current validation is based on simulations. Real-world testing is necessary to confirm the system's performance and safety in actual driving scenarios.
  • Modeling Uncertainty: While robustness to certain uncertainties was evaluated, the paper acknowledges the need for further investigation into control robustness with respect to modeling uncertainties in real-time applications.
  • Extreme Maneuvers: The current evaluation focused on lane keeping. Future work will explore the system's performance during extreme maneuvers like obstacle avoidance and highway merging.
  • Real-time Variable Estimation: For practical implementation, dealing with the estimation of vehicle variables for feedback control (e.g., using LPV observers) will be crucial.

Summary updated

Abstract

This paper proposes a new haptic shared control concept between the human driver and the automation for lane keeping in semi-autonomous vehicles. Based on the principle of human-machine interaction during lane keeping, the level of cooperativeness for completion of driving task is introduced. Using the proposed human-machine cooperative status along with the driver workload, the required level of haptic authority is determined according to the driver’s performance characteristics. Then, a time-varying assistance factor is developed to modulate the assistance torque, which is designed from an integrated driver-in-the-loop vehicle model taking into account the yaw-slip dynamics, the steering dynamics, and the human driver dynamics. To deal with the time-varying nature of both the assistance factor and the vehicle speed involved in the driver-in-the-loop vehicle model, a new ℓ∞ linear parameter varying control technique is proposed. The predefined specifications of the driver-vehicle system are guaranteed using Lyapunov stability theory. The proposed haptic shared control method is validated under various driving tests conducted with high-fidelity simulations. Extensive performance evaluations are performed to highlight the effectiveness of the new method in terms of driver-automation conflict management.
BibTeX
@article{2021human,
  author = {Chen Lv (吕辰) and Anh-Tu Nguyen and Jagat Jyoti Rath and Chen Lv and Thierry-Marie Guerra and Jimmy Lauber},
  title = {{Human-Machine Shared Driving Control for Semi-Autonomous Vehicles Using Level of Cooperativeness}},
  journal = {Sensors},
  year = {2021},
  doi = {10.3390/s21144647},
}
Figure