2023 Fiscal Year Annual Research Report
Driver-automation mutual adaptation: modeling, design, and evaluation of haptic interface for cooperative driving tasks
| Project/Area Number |
21K17781
|
|---|
| Research Institution | The University of Tokyo |
|---|
Principal Investigator |
王 正 東京大学, 生産技術研究所, 特任助教 (20837497)
|
|---|
| Project Period (FY) |
2021-04-01 – 2024-03-31
|
| Keywords | HMI / Automated driving / Machine learning / Intelligent vehicles / Driver behavior modeling |
|---|
| Outline of Annual Research Achievements |
Our research focuses on driver-automation mutual adaptation for haptic shared control.
In FY 2023, our research achievements are summarized as follows:
1. A driver model was developed to capture lateral control behavior during driver-vehicle shared control, integrating a gated recurrent unit network for evaluation against state-of-the-art models. The analysis highlights the network's superior accuracy in predicting driver behavior and maintaining acceptable lateral position error, demonstrating robustness in driver-automation shared control systems
2. Based on previous experimental results showing improved lane-keeping performance and collaborative behaviors with shared control strategies, a novel driving simulator study is designed to test a mutual adaptive shared control system with updating trust values. The study verifies the effectiveness of the proposed system through interactive simulation, focusing on improvements in lane-keeping performance, collaborative behaviors, and user satisfaction.
3. A high-fidelity driving simulator experiment was conducted to present a comprehensive comparison between the Personalized Predictive Haptic Steering Assistance System and a Manually Triggered System, evaluating their effectiveness in lane-keeping and lane-changing assistance with different haptic torque strengths. The results showed that both systems reduced lane departure risk and shortened lane-changing time, but they differed in their impact on driver focus, with the Predictive-Strong method maintaining attention better than the Triggered-Strong system.
|
