Space, the new frontier powered by Machine Learning.

We are explorers. Since the dawn of man, exploration has been at the heart of our survival strategy as a species. We have managed to colonize almost every corner of our planet, living in seemingly inhospitable conditions by adapting and creating technology. The knife, fire, shelter, farming, and clothing, are so intrinsic to the human experience that we forget that it’s all technology that had to go through the whole RnD process.

Ever since the eery beeping of Sputnik was heard by people all over the world, the race to colonize space has been on. Man set foot on the moon in 1969, and things seemed to slow down in the public eye, but world governments and private companies have been slowly and steadily working behind the scenes, developing the technology that would be needed to control space and establish a presence. All the recent technological advancements of the last few decades especially in miniaturization and solid-state technology can be credited to this silent space race. The proof is in the smartphone/laptop that you are using to read this right now.

Space exploration is in the news again and this time startups are taking the lead. SpaceX and Blue Origin are just the more well known of thousands of next-generation privately owned space companies from all over the world. From space transportation and logistics and asteroid mining to deep space and terrestrial imaging, there are startups tackling every aspect of space exploration.

The main driver for this new boom in the space industry seems to be a combination of extreme miniaturization of hardware using solid-state manufacturing techniques, and the recent boom of super-efficient AI’s created using Machine Learning.

Machine Learning is changing everything. Robots and algorithms can be trained to do things that even we don’t know exactly how to do. The lack of explicit instructions makes the system an approximate one but one which far outperforms written instructions when it comes to doing tasks that are so complex that writing instructions for it is physically impossible. Machine Learning tackles the main reason rocket science is considered hard. The sheer number of dynamic variables that need to be considered with millisecond accuracy.

Whether it is self-landing reusable rockets, automatic docking or efficient orbital mechanics, it is not hard to see how Machine Learning is going to be one of the main drivers for making space accessible to everyone.

 

-Triman S Bhullar

Leave a Reply

Your email address will not be published. Required fields are marked *