**ALLSTEPS: Curriculum-driven Learning of Stepping Stone Skills** **Hung Yu Ling $^{\ast}$, Zhaoming Xie $^{\ast}$, Nam Hee Kim, Michiel van de Panne
University of British Columbia** **Best Paper, [SCA 2020](https://computeranimation.org/program.html#program)**
ACM SIGGRAPH / EG Symposium on Computer Animation)
![](allsteps2.png height="150px" border="1") ![](allsteps3.png height="150px" border="1") ![](allsteps1.png height="150px" border="1") __Abstract__ Humans are highly adept at walking in environments with foot placement constraints, including stepping-stone scenarios where the footstep locations are fully constrained. Finding good solutions to stepping-stone locomotion is a longstanding and fundamental challenge for animation and robotics. We present fully learned solutions to this difficult problem using reinforcement learning. We demonstrate the importance of a curriculum for efficient learning and evaluate four possible curriculum choices compared to a non-curriculum baseline. Results are presented for a simulated human character, a realistic bipedal robot simulation and a monster character,in each case producing robust, plausible motions for challenging stepping stone sequences and terrains. __Paper__ [PDF (8.7 Mb)](https://diglib.eg.org/bitstream/handle/10.1111/cgf14115/v39i8pp213-224.pdf) SCA 2020, final version
[PDF (1.8 Mb)](allsteps-arxiv-small.pdf) ArXiv (small)
[ArXiV page](https://arxiv.org/abs/2005.04323) __Video__ [SCA talk (20 min, YouTube)](https://youtu.be/lMNH4xk9c1I)
[SCA overview (3 min, YouTube)](https://youtu.be/XQ0EKfunVT4)
[submission video (5 min, YouTube)](https://youtu.be/jHNdGH1b2sY)
__bibtex__ ````````````````````````` @inproceedings{2020-ALLSTEPS, title={ALLSTEPS: Curriculum-driven Learning of Stepping Stone Skills}, author={Zhaoming Xie and Hung Yu Ling and Nam Hee Kim and Michiel van de Panne}, booktitle = {Proc. ACM SIGGRAPH / Eurographics Symposium on Computer Animation}, year={2020} } `````````````````````````