Drone incidents are increasingly present in the news and one of the key challenges to enable new applications such as drone delivery is safety. With its Silicon Eye collision avoidance system Insightness is releasing a technology that will be a stepping stone towards safe and autonomous drones. While existing collision avoidance systems can only detect static obstacles, Insightness’ system has the capability to detect and avoid moving obstacles even if the drone itself is moving.
Not only celebrities like Enrique Iglesias or Marcel Hirscher made bad experiences with drones but today an increasing number of drone incidents are reported. While many of these incidents are due to an operator mistake, some of them represent substantial problems to the next generation of drones that will fly autonomously beyond the line of sight. Autonomous drones have to make sure that they do not collide with static obstacles even at higher altitudes but they also have to take care of moving obstacles such as birds, kites, model airplanes or other drones. But most importantly they have to ensure they do not hit vehicles with people on board such as paragliders, amateur planes, helicopters or airliners.
Existing collision avoidance solutions allow drones to evade static obstacles, but evading moving obstacles when the drone itself is moving remains a challenging task. As shown in the according release video (teaser), Insightness has now developed a collision avoidance system that solves this challenge by using its novel vision sensor. Insightness’ efficient high-speed Silicon Eye vision sensor allows detecting motion within milliseconds even if the sensor itself is moving. Insightness’ collision avoidance system therefore does not only allow evading static obstacles by computing their distance to them, it also allows identifying and avoiding moving threats.
With the launch of its evaluation kit, Insightness now makes this technology available to interested drone and robotics manufacturers. The evaluation kit features not only a miniaturized version of the vision sensor that easily fits on any drone (3.5mmx3.5mm), but it is also shipped with a mini-computer (credit card size) and the necessary software for the depth and motion map computation. The kit, which is only available for business customers, is configured to work out of the box for a fast evaluation of its capabilities.
The spatial awareness that is provided by Insightness’ vision systems does not only enable beyond-line-of-sight flight for drones, it will also play a crucial role in creating immersive and interactive augmented reality experiences. Knowing quickly where things are and how they move without burning a lot of power is essential for embedding virtual content in the real world. Insightness is currently also working on evaluation kits for these applications.
The Silicon Eye® Technology
The Silicon Eye technology, which is also known under its scientific name of dynamic vision sensor (DVS/DAVIS), is a vision sensor technology inspired by the computational principles found in our eyes. The in-pixel data compression results in a low power consumption, low data rate, fast reaction time and allows to operate the vision sensors under extreme light conditions. The technology was developed at the Institute of Neuroinformatics in Zurich and due to its advantages for applications in computer vision, it is getting increasingly popular. While Insightness has direct access to this technology and 10+ years underlying experience, its competitors such as Chronocam or Hillhouse Technology are bound to mimic its advantages.
Insightness was founded in 2014 as a spin-off of the ETH Zurich and the University of Zurich. The company is designing highly efficient vision sensors and is integrating them in novel computer vision solutions for robots and smart devices. The strong team behind Insightness has long-lasting experience with the Silicon Eye technology (incl. multiple PhD theses) and can build on the long-lasting industry experience of its board members and advisors including Jim Lewis (former CEO of Mesa Imaging and multiple other startups) or Albert Bergemont (former VP R&D Maxim Integrated). For the evaluation of its collision avoidance system, Insightness has built a drone in collaboration with Prof. Davide Scaramuzza’s lab at the University of Zurich.
Evaluation Kit Teaser: https://youtu.be/6aGx-zBSzRA
Evaluation Kit Release Video (full length): https://youtu.be/pAI_ZJOD7zQ
Corporate Video: https://www.youtube.com/watch?v=b0bpfoCju3I
Press Picture Kit: http://www.insightness.com/files/press/17-03-rino2/image-set.zip
Christian Brändli, CEO
+41 77 430 32 59