Sensorimotor representation and knowledge-based reasoning for spatial exploration and localisation

Abstract

We investigate a hybrid system for autonomous exploration and navigation, and implement it in a virtual mobile agent, which operates in virtual spatial environments. The system is based on several distinguishing properties. The representation is not map-like, but based on sensorimotor features, i.e. on combinations of sensory features and motor actions. The system has a hybrid architecture, which integrates a bottom-up processing of sensorimotor features with a top-down, knowledge-based reasoning strategy. This strategy selects the optimal motor action in each step according to the principle of maximum information gain. Two sensorimotor levels with different behavioural granularity are implemented, a macro-level, which controls the movements of the agent in space, and a micro-level, which controls its eye movements. At each level, the same type of hybrid architecture and the same principle of information gain are used for sensorimotor control. The localisation performance of the system is tested with large sets of virtual rooms containing different mixtures of unique and non-unique objects. The results demonstrate that the system efficiently performs those exploratory motor actions that yield a maximum amount of information about the current environment. Localisation is typically achieved within a few steps. Furthermore, the computational complexity of the underlying computations is limited, and the system is robust with respect to minor variations in the spatial environments.

Information-Driven Active Audio-Visual Source Localization

“…This approach is inspired by the recent development of theories which reject the assumption of perception as a mostly passive generalization of the properties of coincident stimuli. In particular, the development of theories of the embodied mind in empirical psychology and philosophy of nature [ 16 ], the proposal of sensorimotor systems in biologically-inspired research in computer science [ 17 ], the reinterpretation of the role of motor actions as the binding element between the senses in multisensory research [ 18 ], as well as active perception approaches in robotics [ 19 ] all suggest that multisensory perception is an inherently active process. In practice, possible fields of application of the proposed system include social and rescue robotics as well as automatic camera control systems for video conferences.…”

Orientation and metacognition in virtual space.

“…This resonates with the vast amount of evidence pointing to the supportive effects of active physical movement and idiothetic cues (Klatzky et al, 1998; Wraga et al, 2004), as well as notions of embodiment (Mallot & Basten, 2009; Wilson, 2002). How we perceive our environment, and how we interact with it, is affected fundamentally by basic sensorial information derived directly from our physical immersion in the environment (Zetzsche, Wolter, & Schill, 2008; Zetzsche et al, 2009). Accordingly, there are natural limits to what VR without actual motion can achieve as a tool for research in spatial cognition (Taube, Valerio, & Yoder, 2013).…”

Evidential Combination of Ontological and Statistical Information for Active Scene Classification

“…In the future, we plan to integrate the presented scene classification system into a mobile agent (Zetzsche et al, 2008). Not only does this provide the system with a strong prior due to the agent's past observations, the mobility would also ease the problem of only sensing parts of a scene.…”