Robot Learning of Human Trust

The codebase for the paper "Improving Trust Estimation in Human-Robot Collaboration Using Beta Reputation at Fine-grained Timescales", published in the 2025 IEEE Robotics and Automation Letters (RA-L).

---

Paper | Code | ArXiv | Slide | Video | Presentation | ReadMe

---

Authors

Resul Dagdanov, Milan Andrejevic, Dikai Liu, Chin-Teng Lin

Watch YouTube Video

Video 1: Detailed Explanation of the Proposed Framework [YouTube Link]

---

Follow ReadMe File for Experiments and Source Code

---

Citation

@article{dagdanov2025improving,
  author  = {Dagdanov, Resul and Andrejevi{\'c}, Milan and Liu, Dikai and Lin, Chin-Teng},
  title   = {Improving Trust Estimation in Human-Robot Collaboration Using Beta Reputation at Fine-Grained Timescales},
  journal = {IEEE Robotics and Automation Letters},
  year    = {2025},
  volume  = {10},
  number  = {8},
  pages   = {8562--8569},
  doi     = {10.1109/LRA.2025.3585653}
}

---

Teaser

Figure 1: Teaser of the Proposed Framework

---

General Framework

Figure 2: General Framework

---

Autonomous Tiling with Collaborative Robot

Video 2: Robot Executing Tiling Operation Autonomously after Learning from Demonstrations

---

Human Trust Measurement

Figure 3: Measurement of Human Trust Toward a Robot (7-point Likert Scale)

---

Methodology

Figure 4: Illustration of an Iterative Human Trust Modeling Process (Proposed Framework)

---

Data Collection Process

Figure 5: Data Collection by Human Demonstrator in UTS Robotics Institute Lab Environment

---

Reward Function with Maximum-Entropy Optimization

Video 3: Visualization of Robot Decision-Making Policy in ROS Simulation Environment during IRL Optimization