RAMPAGE: Towards Whole-body, Real-Time and Agile Motion Planning in Dynamic Cluttered Environments for Mobile Manipulators
We proposed a framework for mobile manipulators to plan and execute wholebody trajectories in cluttered dynamic environments. Environmental maps are constructed in real time with only on-robot sensors. Extensive benchmark comparisons and experiment demonstrate that the proposed framework can plan and execute collision-free wholebody trajectory at an average speed of 1.86m/s among unknown dynamic obstacles with a maximum speed of 2m/s in simulation. Pick-and-place experimental results further verify our method’s effectiveness and efficiency in the real world.
Jan 29, 2024 :The supplementary materials are added at Supplentary_Material_for_Rampage.pdf to demonstrate that our solver can achieve SOTA performance in terms of solution speed and quality. Python codes for benchmarks are also available in this repository.
Oct 10, 2023 : Supplementary videos for TIE submission are added.
Feb 19, 2024 : Paper is accepted to IEEE Transactions on Industrial Electronics.
Feb 23, 2024 : PDF of paper is released here for early access.
Mar 16, 2024 : This paper is available at IEEE Xplore
Tie_Supplementary_Video_Yuqiang1006_small_small_small.mp4
Several experimental results for the revised manuscript to demonstrate the real-time updated ESDF in dynamic environments
The mobile manipulator perceives the obstacle pointcloud using on-robot lidar with 360 degree FOV. The video shows that both static and dynamic obstacles are well modelled in ESDF.
esdf_sim-2023-12-22_16.08.42.mp4
Dynamic.mp4
Static.mp4
In the pick-and-place experiment, the mobile manipulator perceives the obstacle using on-robot cameras and constructs the ESDF in real-time.
esdf_exp-2023-12-22_15.33.28.mp4
esdf_exp-2023-12-22_15.31.41.mp4
For more details, please contact auyuqiangyang@gmail.com.