Multi-Task Learning for Scalable and Dense Multi-Layer Bayesian Map Inference скачать в хорошем качестве

Multi-Task Learning for Scalable and Dense Multi-Layer Bayesian Map Inference

Lu Gan, Youngji Kim, Jessy W Grizzle, Jeffrey M Walls, Ayoung Kim, Ryan M Eustice, Maani Ghaffari IEEE Transactions on Robotics, 2022. Paper: https://arxiv.org/abs/2106.14986 Code: Self-supervised traversability labeling: https://github.com/ganlumomo/traversa... Multi-task learning deep network: https://github.com/ganlumomo/mtl-segm... Multi-layer Bayesian mapping: https://github.com/ganlumomo/MultiLay... Abstract: This paper presents a novel and flexible multi-task multi-layer Bayesian mapping framework with readily extendable attribute layers. The proposed framework goes beyond modern metric-semantic maps to provide even richer environmental information for robots in a single mapping formalism while exploiting inner and inter-layer correlations. It removes the need for a robot to access and process information from many separate maps when performing a complex task, advancing the way robots interact with their environments. To this end, we design a multitask deep neural network with attention mechanisms as our front-end to provide heterogeneous observations for multiple map layers simultaneously. Our backend runs a scalable closed-form Bayesian inference with only logarithmic time complexity. We apply the framework to build a dense robotic map including metric-semantic occupancy and traversability layers. Traversability ground truth labels are automatically generated from exteroceptive sensory data in a self-supervised manner. We present extensive experimental results on publicly available datasets and data collected by a 3D bipedal robot platform and show reliable mapping performance in different environments. Finally, we also discuss how the current framework can be extended to incorporate more information such as friction, signal strength, temperature, and physical quantity concentration using Gaussian map layers. The software for reproducing the presented results or running on customized data is made publicly available.