Multitask Learning for Scalable and Dense Multilayer Bayesian Map Inference

Abstract

This article presents a novel and flexible multitask multilayer 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 intralayer and interlayer 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 back-end 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.

References in corpus (21)

• ORB-SLAM2: an Open-Source SLAM System for Monocular, Stereo and RGB-D Cameras
• Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation
• Past, Present, and Future of Simultaneous Localization And Mapping: Towards the Robust-Perception Age
• An Overview of Multi-Task Learning in Deep Neural Networks
• Multi-Task Learning for Dense Prediction Tasks: A Survey
• Voxblox: Incremental 3D Euclidean Signed Distance Fields for On-Board MAV Planning
• Learning multiple visual domains with residual adapters
• Multi-Task Learning with Deep Neural Networks: A Survey
• Multi-Task Learning as Multi-Objective Optimization
• Volumetric Instance-Aware Semantic Mapping and 3D Object Discovery
• Which Tasks Should Be Learned Together in Multi-task Learning?
• SegMap: Segment-based mapping and localization using data-driven descriptors
• Improvements to Target-Based 3D LiDAR to Camera Calibration
• Gaussian Process Autonomous Mapping and Exploration for Range Sensing Mobile Robots
• Bayesian Spatial Kernel Smoothing for Scalable Dense Semantic Mapping
• Trace Norm Regularised Deep Multi-Task Learning
• Energy-based Legged Robots Terrain Traversability Modeling via Deep Inverse Reinforcement Learning
• TartanAir: A Dataset to Push the Limits of Visual SLAM
• Gaussian Processes Semantic Map Representation
• Hypermap Mapping Framework and its Application to Autonomous Semantic Exploration
• Sparse Bayesian Inference for Dense Semantic Mapping

Cited by in corpus (4)

• Energy-based Legged Robots Terrain Traversability Modeling via Deep Inverse Reinforcement Learning
• Real-Time Metric-Semantic Mapping for Autonomous Navigation in Outdoor Environments
• Traversability analysis with vision and terrain probing for safe legged robot navigation
• CARE: Confidence-rich Autonomous Robot Exploration using Bayesian Kernel Inference and Optimization