1. admin@techtreasure.net : techtreasure :
Samsung Neuromorphic chip copies memory - Drone superpowers - Tech Treasure
November 27, 2021, 9:04 am

Samsung Neuromorphic chip copies memory – Drone superpowers

Technology Desk
  • Published Sunday, October 24, 2021,
  • 86 Seen
Drone Superpowers
Drone Superpowers

Samsung neuromorphic chips unusual robots the latest developments drone superpowers the robotization of the U.S army and other high-tech news in today’s one episode wheels up.

Samsung together with scientists from Harvard presented a new approach to reverse engineering the brain based on memory chips. The essence of the concept is expressed in just two words copy and paste.

In this case copying of the map of neural connections of the brain is proposed to be carried out with the help of a revolutionary matrix of Nano-electrodes. The scientist would then simply paste this map into a high density three-dimensional solid-state memory network.

Samsung neuromorphic chip copies memory
Samsung neuromorphic chip copies memory

With this approach the authors propose to create a memory chip that will approach the unique computational capabilities of the brain. It will have low power consumption be easy to learn adapt to its environment and even be autonomous and support cognitive functions.

Remember that so far all of this is beyond the reach of current technology Japan’s tetra aviation has announced that it will begin delivering single-seat mk5 vertical takeoff and landing electric aircrafts in late 2022.

The device with a battery capacity of 13.5 kilowatt hours can carry a pilot with a paddle of less than 91 kilograms for distances up to 160 kilometers and hit speeds of 160 kilometers per hour.

The company recently showed another test of the mk5 sn2 prototype in the unmanned mode. There’s currently some orders for the electric plane but only customers with a private pilot’s license will be able to buy one.

Samsung neuromorphic chip
Samsung neuromorphic chip

This is due to the fact that tetra aviation is going to register the device as a homemade experimental model. Engineers of the California institute of technology unveiled an updated Leo robot with two legs and four propellers.

The Leonardo robot uses synchronized control of thrusting propellers and leg joints to realize smooth transitions between flight and walking modes. It follows a smooth flight path to the landing point matches the landing speed to the chosen walking speed and triggers the walking phase when at least one foot touches the ground.

After landing the robot continues walking tracking its trajectory and this allows Leo to skateboard and perform other tricks. The downside of this design is that it has to be very light which reduces the viability of the robot when walking and performing tasks.

Also the thrusters are angled to the hull which reduces their effectiveness. When flying and as a result a 2:5 kilogram robot can walk for three and a half minutes or fly for about 1000 seconds but this project is evolving. Autonomous racing drones have learned to traverse a forested area at 400 kilometers per hour.

Samsung neuromorphic chip copies memory
Samsung neuromorphic chip copies memory

Conventional obstacle avoidance involves collecting data from sensors creating a map based on that data and finally making a plan based on that map. The approach works but it does take time. So robotic engineers at the university of Zurich partnered with intel to develop a system in which the drone could perform a direct comparison of the input from the intel real sense 435 stereo camera with real-time collision-free trajectories.

They were able to do this through simulation. The drone was trained like a real pilot trying to make mistakes gradually gaining speed and experience only in simulations. And the engineers managed to make the transition from simulation to real conditions without losing the quality of training.

Samsung neuromorphic chip copies memory- Drone superpowers
Samsung neuromorphic chip copies memory- Drone superpowers

Researchers from the Massachusetts institute of technology have created a robotic system that could find small objects among haphazardly scattered things. The RF fusion system is a robotic arm with a camera and RF antenna attached to the grip. it combines signals from the antenna with visual data from the camera to locate and retrieve objects. even if the object is buried under a pile of junk and completely out of sight. of course the task of finding lost keys is only an illustration of the system’s capabilities. however in the future it will be used for searching sorting and other tasks in warehouses and beyond.

Darpa will conduct another test of part of the offset program. It’s quite the experiment to simultaneously control a swarm of 200 robots and drones in urban environments. And that only one operator will be responsible for all of their control.

According to defense news American companies Northrop groundman and Raytheon will participate in the experiment. Northrop groundman will test the ryze system which allows the operator to set tasks for unmanned systems with the help of sketches on a tablet. Raytheon’s role in the test has not yet been reported.

Drone superpowers
Drone superpowers

Robotire an automotive robotics company and automation startup out of Detroit has received 7.5 million dollars to develop its automatic tire changing system. The company plans to use the money to make its solution more accessible to consumers. Robotire’s patented system reduces tire changing time to 15 minutes.

A sail drone marine drone has been in the middle of a hurricane in the Atlantic ocean and sent developers of video from inside. The drones are designed to collect autonomous oceanographic and atmospheric data and can stay afloat for up to 12 months.

Drone Superpowers
Drone Superpowers

Developers taught the drone to land on a vertical wall and save charge. It seems to be enough to equip the drone with Velcro and prepare the wall. However it’s still necessary to teach the drone to autonomously calculate and perform the landing maneuver on the wall using its cameras inertial unit and computing module.

Two cameras were needed one looked down and monitored the drone’s position. While the other was pointed sideways and monitored the landing pad. Once it understood the position of the pad the drone was able to calculate the landing trajectory. If the landing is successful the engines then turn off.

A U.S civilian drone set a new record for the longest time in air. The drone villa unmanned spent eight days and 500 minutes and 47 seconds above the ground covering almost 20 000 kilometers. At the same time developers say their device can do better, namely to fly for 10 days while carrying up to 3.5 kilograms or transfer almost 70 kilograms at distances that can be covered for 3 days. This drone weighing less than 600 kilograms can reach speeds of 450 kilometers per hour.

The drones were taught to land independently to recharge. so they can work forever without human help. Researchers from Robo work lab at the university of Nevada created the mini hawk. A 3d printed vertical takeoff and landing aircraft for research projects. They then equipped it with an open source flight control system and autopilot software.

The cherry on the cake was the ability of aircraft type drones to combine into large models using simple magnetic mounts on wing tips. as well as their ability to migrate. The latter means the many hawk can land in any suitable place and recharge from solar panels located on its wings and then rise to the air again and continue mission.

Drone superpowers
Drone superpowers

But in china the drones failed, instead of a light snow at the opening of yet another mall viewers saw a rain of glowing drones. The drones rained down directly on buildings cars and people’s heads luckily no one was seriously injured.

U.S engineers unveiled a robot which turns walls into interactive panels by applying metallic conductive traces to them. The device consists of an Arduino board and parts printed on a 3d printer. Instructions for assembling it yourself are available on GitHub we’ll leave you a link in the description.

The robots attach to the wall and can sequentially apply conductive tracks first with metal and then drawings or instructions with regular ink. The desired circuit can be drawn in a simple application and after its implementation on the wall or other surface you just need to add a microcontroller or use the service to make the device work.

Concern automation national engineering corporation and t hunter will create a service to find cyber criminals by analyzing more than 40 public data parameters. In theory the technology will allow a suspect’s email or phone number to be matched against such massive amounts of data such as IP addresses, payment system information, advertising identifiers and other aspects. the goal is to help identify suspects identities.

The marker robotic platform has become the guardian of the Cosmo drome. The robot is now testing autonomous patrol modes perimeter intruder detection identification and interception as well as anti-drone and anti-terrorist attacks.

Drone superpowers
Drone superpowers

Cognitive pilot announced it has signed an agreement with the saint peters burg tractor plant to begin mass production of unmanned tractors in 2022. The Cairo vets tractor will be equipped with autopilot systems based on AI. so far the machines will be able to follow a set trajectory autonomously. automatically control the optimum speed.

Stop in front of obstacles and work in low visibility conditions but in the future it’s planned to install the next version of the autopilot which will allow the tractors to calculate the optimal route and redefine the course by visual landmarks. As well as to analyze the state of the soil and the crops.

Share This Blog

One thought on "Samsung Neuromorphic chip copies memory – Drone superpowers"

  1. Sayad maruf hasan says:

    wow, nice and informative article… need more. thanks

Leave a Reply

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

Read More Blogs
© All Rights Reserved.
Theme Customized BY LatestNews