Research paper
Bidirectional value learning · Long-horizon risk for rover planning
Backward belief-space learning supplies a long-range risk-aware base policy, updated through forward simulation. Continuous-space rover navigation problems evaluate safety, scalability and planning quality.
Reviewed 2026-11-05 · Source and bibliography checked
- Original title
- Bi-Directional Value Learning for Risk-Aware Planning Under Uncertainty
- Publisher
- Institute of Electrical and Electronics Engineers (IEEE)
- Authors
- Sung-Kyun Kim, Rohan Thakker, Ali-Akbar Agha-Mohammadi
- DOI
- 10.1109/lra.2019.2903259
- Publication date
- 2019-07
- Source language
- en
- Edition
- IEEE Robotics and Automation Letters; 2019-07; DOI 10.1109/lra.2019.2903259
- Access
- Publisher metadata and open author manuscript
- Publication status
- Published
Related resources and engineering notes
Source overview
Backward belief-space learning supplies a long-range risk-aware base policy, updated through forward simulation. Continuous-space rover navigation problems evaluate safety, scalability and planning quality.
Use and review
Institutional abstract and bibliography checked. Risk depends on transition/observation models and costs, rather than establishing real-terrain hazard validation.
Sources & editions
Bi-Directional Value Learning for Risk-Aware Planning Under Uncertainty ↗
Institute of Electrical and Electronics Engineers (IEEE) · IEEE Robotics and Automation Letters; 2019-07; DOI 10.1109/lra.2019.2903259
IEEE RA-L2019-07 author abstract https://publications.ri.cmu.edu/bi-directional-value-learning-for-risk-aware-planning-under-uncertainty.