Modeling of a Large Structured Environment

Abstract

Abstract. AIMS project attempts to link the logistic requirements of an intelligent warehouse and state of the art core technologies of automation, by providing an awareness of the environment to the autonomous systems and vice versa. In this work we investigate a solution for modeling the infrastructure of a structured environment such as warehouses, by the means of a vision sensor. The model is based on the expected pattern of the infrastructure, generated from and matched to the map. Generation of the model is based on a set of tools such as closed-form Hough transform, DBSCAN clustering algorithm, Fourier transform and optimization techniques. The performance evaluation of the proposed method is accompanied with a real world experiment.

Semi-supervised semantic labeling of adaptive cell decomposition maps in well-structured environments

“…Concerning the systematic shortcomings, we plan to improve our results by augmenting a landmark map generated from vision sensor from our previous work [5]. The landmark map generated in [5] models the infrastructure of the environment (pillars) providing a more accurate spatial decomposition of the model.…”

“…The landmark map generated in [5] models the infrastructure of the environment (pillars) providing a more accurate spatial decomposition of the model. Here, we expect to take inspiration from research on hierarchical multi-modal methods for semantic analysis.…”

“…A central element for fulfilling the project objectives is the establishment of a robust and flexible semantic mapping framework. Our previous work [4], [5] focuses on spatial modelling methods well suited for semantic annotation and inference. In this present paper, we address the problem of providing human knowledge to robots for the purpose of semantic mapping.…”

“…Examples of matching a template with the model are; [14] builds a geometric model of the world, and looks for templates (e.g. walls, floors, or objects), [15] proposes an inference model over a 3D correlating objects framework with semantic labels, [16] provides a topological graph of the environment based on the concept of virtual door, [17] embeds semantics in kernels and the map is refined and segmented accordingly, [5] presents semantic labels as graph templates and matches to the connectivity map of the open space, [3] performs a segmentation and semantic labeling of objects in 3D data. The last example is the probabilistic framework for semantic inference proposed by [18] which covers a large variety of techniques on different levels.…”

Sensor based adaptive metric-topological cell decomposition method for semantic annotation of structured environments

“…An important next milestone in the project is the inclusion of visual perception. In [22] we proposed a modeling technique for structured environment by the means of vision sensors. That study starts with mapping a warehouse by its infrastructural landmarks (pillars).…”