The massive embedding of digital technology like sensors, actuators and microcontrollers in our cars,
toys, musical instruments, and the light and sound systems of a dance performance has change the way we
design the control of our every day artifacts. The control is implemented by programming the microcontrollers
to react to stimuli from the environment with a response that make the artifact function as intended.
Examples include an airbag controller of a car, a remote controller for an audio/video device, and a game
computer. A number of microcontrollers might be included in a system to obtain the desired functionality.
Furthermore, these specialized devices often has to interface to net based system like the internet,
an alarm net or a home net. From a technical point of view a major problem with the design and implementation
of these embedded systems is the simultanious evolution of hardware and software. This include the instrumentation
of the system with sensors and actuators as the programs interface to the environment and the choice of
programming languages, development tools, and runtime platforms. From an observational point of view a
major problem is to make an embedded system behave understandable even in situations that have not been
delt with explicitely during the design and implementation process.
The research themes include traditional software engineering topics like specialized operating systems
for small microcontrollers, platform independence through virtual machines on microcontrollers, programming
of special purpose sensors and actuators, how can a virtual machines be taylored to the peripheral devices
of a particular system, handling of realtime demands, fault tolerance, and power supply.
The research also include themes from Behaviour Based Artificial Inteligence like the concept of autonomous
agents and multiagent systems. When an embedded system is considered from an observational point of view
the behaviour of the embedded system is described in categories different from the categories used to
describe the internal actions of the control program. To bridge this gap between categories the metaphore
of an autonomous agent with a body of sensors and actuators situated in an environment might
be useful during the design phase of embedded systems.
A large portion of the activities in this area include teaching at all levels. Two courses with the titles
Adaptive Robotics and Communicating embedded, and situated agents are given at
Master and PhD level.
One lab has been established to provide facilities for the teaching and reasearch in embedded systems:
The LEGO Lab.
The LEGO Lab is a laboratory for teaching and research in multiagent systems based on simple robot
technology such as the LEGO MindStorms RCX microcontroller. The aim is to investigate massive use of simple
robot technology in multiagent systems for entertainment and artistic expression in physical installations.
The aim is also to investigate the usage of LEGO robotics for teaching at a wide range of levels.
- Ole Caprani, area manager
- Jakob Fredslund, PhD student
- Nis Haller Baggesen, PhD student
- Rasmus B. Lunding, composer
- Line Kramhøft, designer
- Mads Wahlberg, light designer, robot designer