DTU is the only Danish scientific participant in the EU's flagship
Human Brain Project to map the human brain. An agreement covering the Danish part of the project has just been signed.
Professor Henrik Hautop Lund, DTU Electrical Engineering, will contribute research into how the brain's ability to adapt to changes in its environment can be passed on to robots.
The goal of the Human Brain Project over the next ten years is to map the brain and make this knowledge available to everyone on a number of super computer platforms. The work involves constructing computer models based on detailed knowledge of the brain's signal systems, molecules, and neurons.
"Trying to understand the human brain is a huge challenge. But just as people have mapped the human genome, the aim of the Human Brain Project is to map the human brain, and try to understand the connections between neurons, signals etc. in the brain through simulations on supercomputers," notes Henrik Hautop Lund.
The hope is that diseases such as Parkinson's and Alzheimer's can be cured and prevented in the longer term. And that better knowledge of the chemical processes in the brain might lead to new pharmaceuticals which can repair damage to the brain resulting from blood clots, cerebral haemorrhages or other diseases related to the nervous system.
The Danish contribution to the large Human Brain Project—involving 750 researchers from 24 countries—focuses on improving technology in robots so that they can more closely mimic the human brain.
The vision is to make technology as adaptable as humans
"You cannot look at the brain in isolation. You have to look at it in connection with the surroundings. For example, the cerebellum is where our human motor lies—it remembers how we move. The cerebellum can connect sensory impressions with knowledge of movement, so that we can adapt our movements—and walk on a heaving ship deck or catch a ball that comes flying towards us. We would like to be able to mimic these kind of steady movements using robots, so that prosthetic hands and legs can operate more effectively in future," explains Henrik Hautop Lund.
"Just as people have mapped the human genome, the aim of the Human Brain Project is to map the human brain."
Professor Henrik Hautop Lund, DTU Electrical Engineering
Professor Lund's role in the large Human Brain Project will be to transfer some of the knowledge of the brain the rest of the project gathers, from the computer models, and see how it can best be used modularly in robots in the real world.
"Basically, there's a difference between how engineers react in a laboratory when they develop a robot, and how Mrs Petersen reacts at home in her living room. Our goal with the modular approach is to adapt the robot functioning so that it can read and adapt to any situation—whether it is in a North American living room, an African nomad tent or an wooden house in Asia. This requires new knowledge, which we will find by trialling different models. We will investigate what happens if you change the sensory system—the entire bio-robotic aspect, which covers the interaction between the body and brain."
The research group's vision is to make the technology as good at reading and adapting to changes as the human systems. The first step for DTU researchers is to implement spiking neural networks, which mimic the functioning of the brain, on a neurocomputer controlling a modular robot. They will then conduct various experiments which reflect the different disruptions from the surroundings people encounter each day.