The World’s First AI Satellite by Intel

0
704

On September 2, Artificial intelligence made its first step into the orbit with Intel providing required the smarts. PhiSat-1 has a hyperspectral-thermal camera and Movidius Myriad 2 Vision Processing Unit (VPU) powered by Intel. This chip may be familiar as it is used in consumer devices like smart cameras and drones.

The PhiSat-1 is one of the satellites that are on the mission to surveil polar ice as well as soil moisture. On top of these, in order to create a federal satellite network, PhiSat-1 ensures an inter-satellite communication system.

Around two-thirds of the Earth’s surface is engulfed with clouds will lead to capture many useless images, save and will be sent over downlink bandwidth, saved again, and then studied by scientist or an algorithm. This may generate a massive amount of data and the capacity of sensors to produce data is increasing but by a factor of three to five per generation. So in order to detect and thrust aside cloudy images that can save around thirty percent of the bandwidth an on-board processing unit is used.

The AI software used on the Intel Myriad 2 was developed by Ubotica (a startup). To venture this chip beyond the planet’s atmosphere was challenging. They had to write a handbook on the chip, its characteristics, and how to perform the test from scratch as they were doubtful regarding the testing.

Another challenge raised was regarding the training. For this, a large amount of data is required to train the AI Algorithm on what exactly is a cloud. As there was no existing data with the team, synthetic data from other missions were used to train.

The first AI satellite came into reality with the help of various organizations across Europe to put required smarts on board that include the system, software integration, and testing. And after the on-ground verifications, the PhiSat-1 entered the sun-synchronous orbit and the history of space.

Currently, the ESA and Ubotica are working on their new project PhiSat-2 which is expected to identify wild-fire, spot rouge ship, climate change, trace soil moisture, and the growth of the crops. If this comes into reality it will be of great help such as to prevent the wild-fire and much more.