Dogs have long since been our companions. According to studies, by far the most loyal and intense companionship of all pets is of dogs. They are a part of millions of families and many dogs actually play essential roles in the unit. They can take instructions and learn to do as the owner requires. Hence, they become good guardians of a place and recognize scents, as well as help the elderly with mobility. There are also emotional support dogs in psychiatric facilities. With the age of digitalization, it is only natural that we would soon dig around to find ways to replicate these loving and useful companions.
On the grounds of Edinburgh University, a dog undergoes training. Although there is a lot of kicking, pushing, and shoving involved, calling PETA is unnecessary. This dog happens to be a robot. Thousands of years ago, men came upon the idea that the wild furry beasts might be domesticated. They were tamed and taught the ways of living peacefully with mankind, not only that, they contributed to hunting and taking care of the sheep. Now centuries later, Alex Li from the Advanced Robotics Lab is doing the same. But with a robot dog.
Chip was one of the first lively dogs that were made for playful activities, emotional support, and to imitate the countenance that of a dog especially for catering to younger children. The first police dog was first introduced in 2019. Moreover, Boston Dynamics’ internet famous dog, Spot has also been modified and now has an arm and a certain amount of mobility.
So what is specific about this robot dog?
Alex Li is trying to replicate how children are taught football. Learning takes place in small steps. By specifically teaching individual skills the robots can utilize them in different tasks; falling and getting back up, trotting around, and walking. The tasks were taught in two different neural networks. Eight of these networks are interconnected. Alex Li suggested how a leader could be able to direct and adapt to situations and use skills accordingly.
This is like adaptational learning (learning from the memory of experiences) in human beings and this way the AI might be able to have a memory to perform tasks more efficiently. The robots might be able to learn while they are given different tasks and choose from the set of skills provided to them. Raia Hadsell at the Robotics Laboratory highlighted how it is essential for robots doing work around the house to understand changing environment so that they do not mess up the tasks by staying rigid.
Cover: (Image Credits: popularmechanics.com)
Reported By: Zarrish Ahmed