The work in the lab is oriented towards the development of biomimetic sensors to support intelligent interactions with the environment by autonomous systems. This interdisciplinary research is directed at the study of biological sensory systems with the aim of extracting a better understanding of the engineering solutions nature, after a very long evolutionary optimization process, has come up with. We maintain that the regions of the design space explored by nature are sufficiently different from the ones explored by human engineers that this approach can function as a source of interesting, new, ideas for engineered systems.
Because biological organisms and their environments form tightly coupled interacting systems in which all components: environmental characteristics and dynamics, sensory morphology, peripheral and central neural processing and behavioural patterns play a significant role our research is carried out at three levels simultaneously: the morphology and mechanics, the signal processing, and the behavioural strategies of the model animal systems. Extraction of significant environment information is considered an emergent property from processing going on at all three levels.
The novel insights gained from this work are made more broadly available either through the publication of scientific articles or through the development of patented technologies that can be licensed. In collaboration with several international partners we develop and carry out national and international projects.
A broad field of research topics are studied in the field of (i) bat bio sonar, (ii) robot sonar technology, (iii) human spatial hearing.
Bats (order Chiroptera) are one of the largest and most successful orders of mammals around the world. This success is partly explained by the exceptional perceptual system, i.e. bio sonar, bats use to get around in the dark. We study bat bio sonar (link naar CILIA) by building models, both conceptual, implemented as software simulations, and physical, implemented as elektromechanical replicas, models.
The intelligence and thus autonomy of robotic systems is to a large extent determined by their perceptual systems. Using a broad range of sensors guarantees autonomous operation in the broadest range of robot applications. We study how the insights gained from biosonar (link naar Chiroping) can be mapped onto high-performance robotic sonar systems. We develop novel sonar prototypes and integrate these prototypes into tailored robot control architectures that are designed to make optimal use of the information provided by the robotic sonar.
Human spatial hearing describes the capability of humans to localize sound sources in their environment. With the rise of virtual reality systems, the need for reproducing realistic auditory stimuli has grown proportionally. We apply the techniques we developed for studying how bats extract spatial information from the echoes returned by the environment to human spatial hearing.
Accurate morphological models extracted from micro-CT images of real bat heads are used as input to acoustic simulation software to quantify the physical filtering that occurs during generation/reception of sounds. This methodology allows to reconstruct and study the spatial cues present in the echo signals as received by bats.
The novel sonar prototypes, both electronic and mechanical subsystem design and development is done in-house, are mounted onto commercially available robotic platforms (robot arms and mobile robots) and together with the tailored signal processing and robot controllers tested in real-world applications.
We have developed an information theory framework to quantify the information content of specific spatial cues extracted from the acoustic signals picked up by binaural hearing systems.