Abstract
Leveraging robotics in the study of biology has been the basis of biorobotics for seventy years and has lead to many insights. The robot has typically been used either as a surrogate for the animal to produce evidence for a model of a behaviour or to interact with the animal in manipulation experiments. However, scientists are finding new ways of applying robotics in biology research and several new fields are emerging.
In this dissertation, I present a novel taxonomy for biorobotics sorting most research into ten different categories and I point out areas that should be explored. Based on the taxonomy, I argue that the ability to handle data fast enough to influence the process producing it is becoming essential in the field and that modern IT infrastructure offers support for such online operation. The dissertation consists of three manuscripts made in interdisciplinary collaboration between biologists and engineers. In the manuscripts, we explore the use of IT infrastructure, in particular microservices and event processing, to support biological research.
We developed a novel architecture for online computational ethology and tested it in case studies of long term sound recordings of wild bats and video recordings of freely-behaving mongoose. We then applied the architecture in the development of a novel virtual reality setup for experimental manipulation of social communication in songbirds. We demonstrated the setup with zebra finches in an effort to control the production of mate-directed song. Finally, we applied the same principles to build a novel neuronal simulator and demonstrated online robot control based on a spiking neural network.
Through contributions to ethology, songbird research and neuroscience, we show that online biorobotics can enable new types of experiments and that leveraging the rapid development in robotics and IT provides new opportunities for studying biology.
In this dissertation, I present a novel taxonomy for biorobotics sorting most research into ten different categories and I point out areas that should be explored. Based on the taxonomy, I argue that the ability to handle data fast enough to influence the process producing it is becoming essential in the field and that modern IT infrastructure offers support for such online operation. The dissertation consists of three manuscripts made in interdisciplinary collaboration between biologists and engineers. In the manuscripts, we explore the use of IT infrastructure, in particular microservices and event processing, to support biological research.
We developed a novel architecture for online computational ethology and tested it in case studies of long term sound recordings of wild bats and video recordings of freely-behaving mongoose. We then applied the architecture in the development of a novel virtual reality setup for experimental manipulation of social communication in songbirds. We demonstrated the setup with zebra finches in an effort to control the production of mate-directed song. Finally, we applied the same principles to build a novel neuronal simulator and demonstrated online robot control based on a spiking neural network.
Through contributions to ethology, songbird research and neuroscience, we show that online biorobotics can enable new types of experiments and that leveraging the rapid development in robotics and IT provides new opportunities for studying biology.
| Originalsprog | Engelsk |
|---|---|
| Status | Udgivet - 14. dec. 2020 |