Robots are Being Taught to Autonomously Evolve to Survive in Distant Planets

Scientists have been trying to reach a goal of sending robots to other planets to make conditions suitable for humankind. These robots should be durable, adaptable, and recyclable to survive the inhospitable cosmic environment of other planets.

According to an article in The Conversation, roboticists and computer scientists from the University of York, Edinburgh Napier University, Bristol Robotics Laboratory, and Vrije Universiteit Amsterdam collaborated to make such sets of robots produced using a 3D printer.

The team said that the parts of the robots are autonomously assembled and will continually evolve to rapidly adapt to the conditions they are in.

Their work presents a new milestone towards the autonomous robotic ecosystems that aims to help build humanity's future homes far away from Earth.

Using Robots to Explore Exoplanets

The invention of robots has come a long way since humanity's first attempts decades ago. Today, different industries have been employing robots to improve the efficiency and effectiveness of their jobs.

However, designing robots that can work in unknown and inhospitable environments in exoplanets or deep ocean trenches remains a challenge. When designing robots to be used on other planets, they should consider the shape and size of the robot, deciding if it should walk or crawl, and the tools needed to manipulate its environment.

Most importantly, scientists and engineers should consider how these robots will survive extremes of pressure, temperature, and chemical corrosion.

According to TNW's report, citing the article in The Conversation, Darwin's theory of evolution has already solved this problem. Biological evolution may take millions of years but with artificial evolution, scientists have made robots evolved in just a few hours or even in minutes.

ALSO READ: Creators of Sophia the Robot Plan to Mass-Produce Robots this Year

Artificial Evolution

For artificial evolution to work, humans must be removed from the loop. That means robots must autonomously manufacture, assemble, and test themselves.

It is also expected for evolved robots to be capable of sensing their environment and have diverse means of moving, such as wheels, jointed legs, or a mixture of both. Also, the University of York scientists said that some evolution should happen within the hardware wherein an ecosystem of robots evolves in real-time and real space.

With the Autonomous Robot Evolution (ARE) project, it can do all of the above. Through the use of 3D manufacturing, child robots were born that utilize hardware-software evolutionary architecture design. Physical robots will be tested in the real environment, while their digital clones will undergo rapid simulation evolution in a software program.

This novel kind of hybrid hardware-software system introduces a new type of evolution where the new generations are produced from the union of the physical "father" (hardware) and virtual "mother" (software).

Child robots are a product of successful robots that make their "genetic code" available to reproduce and improve future generations. Meanwhile, those robots that are deemed less fit will be hoisted away and recycled into new ones as part of the evolutionary cycle of robots.

Scientists have made advances to the robot by adding new artificial evolutionary algorithms that made the robots able to drive, crawl, and learn to navigate through complex mazes.

Ultimately, the goal of this hybrid technology is for robots to live and work for long periods in challenging and dynamic environments without humans' help.