Robot Motion Planning in Learned Latent Spaces

Brian Ichter, Marco Pavone

This paper presents Latent Sampling-based Motion Planning (L-SBMP), a methodology towards computing motion plans for complex robotic systems by learning a plannable latent represen…