DreamLander: A 6-DOF World Model based Reinforcement Learning Planetary Lander
DreamLander: A 6-DOF World Model based Reinforcement Learning Planetary Lander
This paper introduces DreamLander, a model-based reinforcement learning algorithm based on DreamerV3 designed for optimal guidance and control in 6-degree-of-freedom planetary entry descent landing. The approach uses state sensor observations to achieve high accuracy, with simulated landings within 2 meters of the target zone and velocity norms below 0.3 m/s, demonstrating sample efficiency and generalizability across scenarios.
The authors present DreamLander, a model-based reinforcement learning algorithm built on DreamerV3, which achieves landing within 2 meters of the target zone and a velocity norm below 0.3 m/s in simulated 6-DOF entry descent landing scenarios; claims regarding real-world deployment or fuel efficiency are not independently verified.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
Reported numbers
landing accuracy
2
View original evidence
based approach creates a sample efficient and highly generalizable model that can be applied to a variety of scenarios whilst maintaining a high degree of accuracy, landing within 2 meters of the designated landing zone, and with velocity norm less than 0.3 meters per second. DreamLander makes use of state sensor observations, making progress toward End-to-End solutions of the Entry Descent Landing problem.
Open source S4velocity norm
<0.3
View original evidence
based approach creates a sample efficient and highly generalizable model that can be applied to a variety of scenarios whilst maintaining a high degree of accuracy, landing within 2 meters of the designated landing zone, and with velocity norm less than 0.3 meters per second. DreamLander makes use of state sensor observations, making progress toward End-to-End solutions of the Entry Descent Landing problem.
Open source S4DOF
6
View original evidence
DreamLander: A 6-DOF World Model based Reinforcement Learning Planetary Lander
Open source S1DOF
6
View original evidence
d control. Model based reinforcement learning is proving to be an important tool in simplifying the highly complex process of generating optimal guidance trajectories and control. In this work, we present DreamLander, a state of the art model based reinforcement learning algorithm based on DreamerV3, to generate optimal guidance and control for a realistic 6 degree of freedom entry descent landing problem. The model
Open source S3
- paper: not_reported
What remains unknown
Not established in the collected evidence: Environment, Control, Data origin.
Reported performance applies to the described task. It does not establish general autonomy or deployment readiness.
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
based approach creates a sample efficient and highly generalizable model that can be applied to a variety of scenarios whilst maintaining a high degree of accuracy, landing within 2 meters of the designated landing zone, and with velocity norm less than 0.3 meters per second. DreamLander makes use of state sensor observations, making progress toward End-to-End solutions of the Entry Descent Landing problem.
Open source S4
DreamLander: A 6-DOF World Model based Reinforcement Learning Planetary Lander
Open source S1
d control. Model based reinforcement learning is proving to be an important tool in simplifying the highly complex process of generating optimal guidance trajectories and control. In this work, we present DreamLander, a state of the art model based reinforcement learning algorithm based on DreamerV3, to generate optimal guidance and control for a realistic 6 degree of freedom entry descent landing problem. The model
Open source S3
Source：arXiv Robotics — research abstracts · arxiv.org