It is hard to make the drones fly faster even with the simplest obstacles. However, results were, engineers, have to engage them in either repairing work or replacing them completely. MIT engineers have made their effort in developing a structure. Hence with that, any vehicle can experience a virtual environment being flying in empty space. “Flight Goggles”, the name given to the method, will minimize the crash numbers in educating. However, it can also serve as the environment where researchers can give VR training to them.
Sertac Karaman, an associate professor of aeronautics and astronauts, thinks this to be the game changer. Next week, Karaman along with his colleagues is going to submit details about the methodology at IEEE.
Hence this marks the International Conference on Robotics and Automation. Other members are Thomas Sayre-McCord, Amado Antonini, Winter Guerra, Austin Brown, Jasper Arneberg and Guilherme Cavalheiro. Counting co-authors are Dave McCoy, Fabian Riether, Sebastian Quilter, Yunus Terzioglu, and Ezra Tal. Luca Carlone of Laboratory for Information and Decision Systems counted as co-author. Hence, Yajun Fang of Computer Science and Artificial Intelligence Laboratory of MIT was also part of the team. Moreover, Alex Gorodetsky of Sandia National Laboratories was also a co-author.
A new robo-sport stood as an inspiration for Sertac. He was in the drone competitive racing which was through a remote. However, the vehicles had to cross many complex paths of doors, windows and many other obstacles. The initial thought that strike Sertac was that whether an autonomous drone could pilot faster?
Karaman claims of entering the racing competition of drone in the next 2-3 years. As per him, the human player can lose in front of the pilotless body. But for that, the team needs to design an all-new routine.
Nowadays, autonomous drones get coaching through a physical task. Aircraft pilot in large open space with boundaries around. Though drone crash needs repairing or replacing but it also adds up to cost and development delays. However, these drones aim at gathering information about the surroundings. But if you need something to fly fast, the method needs to trust on. As per the developer despite flying in an empty room, the craft will experience a different room.
However, the device can process the virtual image at 90 frames per second. Hence it is much more than the capabilities of the mankind. With the addition of supercomputer and camera and measurement unit, the task was possible. This entire are amalgamate in a small, printed 3D nylon with reinforced carbon fiber aircraft frame.
To make the device capable of facing any situation the developers set out some experiments. Mentioning one is the experiment in which the vehicles passed through the twice bigger virtual window. Though the room was empty but the instructors set out images of a living room. In order to make the craft learn to go in air, developers used a navigation algorithm.
However, the information came from motion-capture cameras of the facility. Hence the crashes didn’t leave much impact over the time and the cost of the development. Moreover, setting out the final test, the craft, with an onboard camera, passed through the test facilities actual window. More than eight flights, the researchers use navigation algorithm through real windows for 119 times. However, human intervention or crashing as require for only six times.
Talking about the virtual structure of VR training, Sertac said that it is highly adaptable. Hence the arrangement gives the instructors the provision to create an environment as per their wish. Moreover, the team is pairing up with Stata Center to create actual building replicas. Hence, the sessions also help in testing the new sensors or existing sensor’s specification. It will provide them with an idea of how to handle the pilotless instrument when they go up faster. However, flying through the mankind will now be possible for the pilotless craft.
Hence he thought to divide the real test facility into two, one with the piloting aircraft. However, the other half will include man with the motion-capture suit on. Hence it can see the man as it flies. Due to VR training even if it crashes it will not create much harm as everything is virtual. However, he said that there are scoop for many experiments to carry on. Hence NVIDIA Corporation, MIT Lincoln Laboratory and U.S. Office of Naval Research supported the research in part.
Share your views on the VR training of Drone with us. Or even you can connect with us if you have any additional information. We always look forward to hearing from you!