by Ben Fisher
Posted on November 11, 2015
DJI has launched a standalone computer as part of their developer product range. It’s called the Manifold and it fits comfortably on the DJI Matrice 100 Platform. The DJI Manifold has been designed to specifically support the DJI SDK developer community as an out-of-the box solution. It has a Quad-core, 4-Plus-1™ ARM® with a low-power NVIDIA Kepler™-based GeForce® graphics processor. The embedded Linux computer ships with Ubuntu operating system pre-installed and has 2GB of RAM. The price is £409 ($499), obviously you will need a Matrice 100, currently £2,799 ($3,299) and why not have the guidance system £899 ($999) also. So the sum total for a state of the art flying robot will set you back £4,107 ($4,797).
DJI is not alone in its quest to build the ultimate onboard flying computer. Qualcomm and Intel also have products ready for the drone market. Qualcomm has it’s Snapdragon chip set which powers a majority of high-end smartphones today (so chances are you have this chip in your pocket). Intel has the Edison chip set which 3DR recently announced they are testing for higher levels or computation on their drones.
This extra onboard compute power is used in combination with the flight controller and camera to allow for more autonomy and real-time image processing. In our last post we looked at how MIT students developed an autonomous drone for $1,700 and how the industry is focused on developing sense and avoid technology. It’s no surprise the heavyweights of the computing world have moved in on the drone industry. More onboard processing power opens more possibilities, including sense-and-avoid. Whoever cracks the system first and gains approval from Aviation Authorities the world over opens up a whole new level of demand. This market is said to be worth billions in the near future given the amount of applications there are for this type of technology. Just think how the internet and mobile phones has shaped all of our lives, well, the drone industry is both of these things combined with the freedom of autonomy.
You can also expect some drones to get smaller and more powerful following Moore’s Law (the founder of Intel). There are of course advantages and disadvantages for smaller drones. But one of the main advantages is safety, as Vijay Kumar demonstrates in his TED talk regarding his research. As demonstrated, smaller size means lower inertia, lower inertia means more resistance to collisions making the drones more robust. Perfect if you want to send a swarm of drones to map an area of a disaster zone.
Since DJI has market share of the commercial drone industry, it’s an important step to see an embedded computer ready for the market which has connectivity and the ease of plug-and-play. It keeps DJI at the forefront of the industry as it steps up to be a part of the sense-and-avoid arms race.