We develop an online probabilistic metric-semantic mapping approach for autonomous robots relying on streaming RGB-D observations in both centralized settings and decentralized multi-robot networks. Notably, the generated maps contain full continuous distributional information about the geometric surfaces and semantic labels (e.g., chair, table, wall), making them appropriate for uncertainty-aware motion planning. Our approach is based on online Gaussian Process (GP) training and inference, which avoids the complexity of GP classification by regressing a truncated signed distance function (TSDF) representation of the regions occupied by different semantic classes. Online regression is enabled through a sparse inducing-point approximation of the GP posterior. To scale to large environments, we further consider spatial domain partitioning via an octree data structure with overlapping leaves. An extension to the multi-robot setting is developed by having each robot execute its own online measurement update and then combine its posterior parameters via a local weighted geometric average with those of its neighbors. This yields a decentralized information processing architecture in which the GP map estimates of all robots converge to a common map of the environment while relying only on local one-hop communication. Our experiments demonstrate the effectiveness of the probabilistic metric-semantic mapping technique in 2-D and 3-D environments in both single and multi-robot settings.
@inproceedings{Zobeidi_GPMapping_IROS20,
author = {E. Zobeidi and A. Koppel and N. Atanasov},
title = {Dense Incremental Metric-Semantic Mapping via Sparse Gaussian Process Regression},
booktitle = {IEEE/RSJ Int. Conf. on Intelligent Robots and Systems (IROS)},
year = {2020}
}
@inproceedings{Guo_IFOM_IROS19,
author = {S. Guo and N. Atanasov},
title = {Information Filter Occupancy Mapping using Decomposable Radial Kernels},
booktitle = {IEEE/RSJ Int. Conf. on Intelligent Robots and Systems (IROS)},
year = {2019},
doi = {https://www.doi.org/10.1109/IROS40897.2019.8968609}
}
@inproceedings{Sun_OccupancyGridFiltering_ACC18,
author = {K. Sun and K. Saulnier and N. Atanasov and G. Pappas and V. Kumar},
title = {Dense 3-D Mapping with Spatial Correlation via Gaussian Filtering},
booktitle = {American Control Conference (ACC)},
year = {2018},
doi = {http://www.doi.org/10.23919/ACC.2018.8431777}
}
@inproceedings{Duong_KernelMapping_ICRA20,
author = {T. Duong and N. Das and M. Yip and N. Atanasov},
title = {Autonomous Navigation in Unknown Environments using Sparse Kernel-based Occupancy Mapping},
booktitle = {IEEE International Conference on Robotics and Automation (ICRA)},
year = {2020},
doi = {https://www.doi.org/10.1109/ICRA40945.2020.9197412}
}