Data association in a world model for autonomous systems

Abstract

This contribution introduces a three pillar information storage and management system for modeling the environment of autonomous systems. The main characteristics is the separation of prior knowledge, environment model and sensor information. In the center of the system is the environment model, which provides the autonomous system with information about the current state of the environment. It consists of instances with attributes and relations as virtual substitutes of entities (persons and objects) of the real world. Important features are the representation of uncertain information by means of Degree-of-Belief (DoB) distributions, the information exchange between the three pillars as well as creation, deletion and update of instances, attributes and relations in the environment model. In this work, a Bayesian method for fusing new observations to the environment model is introduced. For this purpose, a Bayesian data association method is derived. The main question answered here is the observation-to-instance mapping and the decision mechanisms for creating a new instance or updating already existing instances in the environment model

Semantic world modeling using probabilistic multiple hypothesis anchoring

“…One more alternative is introduced by [15]. Even though the term anchoring is not used, their system can be interpreted as an anchoring algorithm.…”

Multimodal saliency-based attention for object-based scene analysis

“…Therefore, the multimodal classification and fusion at each level of the knowledge hierarchy is done bottom-up whereas the selection of suitable classification algorithms is done in a top-down fashion. The basis for this exploration and analysis is an object-based world model (see [31]), which follows the approach described in [35]. A notable feature of the chosen object analysis approach is that it facilitates the dynamic adjustment of object-specific tracking parameters, e.g.…”

Three pillar information management system for modeling the environment of autonomous systems

“…Observed entities (sensory information) are mapped to instances with attributes and relations in the environment model. This data association problem has been addressed in detail in [1].…”

“…In the following, the basic ideas of the probabilistic formalization of the environment model is introduced, see also [1]. As already mentioned, the environment model mainly consists of a set of instances.…”

“…This paper proposes an intelligent, universal information management system, whose basic concept was already introduced in [1]: the three pillar information system. As reflected by its name, the three pillar information system consists of three pillars: prior knowledge, environment model, and sensory information.…”