AreaRead the video transcript expand here
Two sentient beings meet here.
This robotic dog can at least sense and overcome obstacles.
It was taught to him by his master and inventor during a hike in the Swiss Alps.
Takahiro Miki, doctoral student Robotic Systems Lab Zurich:
»The path was quite steep and exhausting even for us.
But the robot got over all the obstacles.
We never had to help him.«
So far, robots have been dependent on maps that they calculate from their environment.
But in difficult terrain, this method reaches its limits.
The Swiss robot dog, however, can feel the resilience of the ground and can even get through snow and tall grass.
Conditions where other robots usually fail.
He has learned to listen to his “body feeling”.
Takahiro Miki, doctoral student at the Robotic Systems Lab Zurich:
»If he takes tactile information into account, he realizes that his map is wrong.
Then he only trusts his sense of touch.«
According to the inventor, the main difficulty lay in bringing vision, arithmetic and touch into the right relationship.
The hike helped collect information for the self-learning algorithm.
Takahiro Miki, doctoral student at the Robotic Systems Lab in Zurich:
“With this control, the robot can get along in areas that are dangerous or too far away for humans.
In damaged nuclear power plants, disaster areas - or even on foreign planets.«
Those are pretty good prospects.