TOKYO, JAPAN:  Mitsubishi Electric unveils plethora of technological developments including an aerial display, compact AI, and cyber attack detection technology that cater across industry verticals. The new technological developments unveiled are:

Aerial Display

The company announces development of an Aerial Display that projects images measuring approximately 56 inches diagonally and projected at least one meter from the beam splitter.

This display technology consists of two key components:  beam splitter and retro-reflective sheet. The beam splitter is an optical device that divides incoming light into reflected light and transmitted light. The retro-reflective sheet is an optical device that reflects incoming light back in the incident direction. A beam splitter and retro-reflective sheet are arranged in a set with a screen. The light that is projected from the screen is reflected by the beam splitter and then subsequently by the retro-reflective sheet. The result is that the light re-converges as an image that appears to be floating in the air.

To realize this technology, Mitsubishi Electric developed an optical simulation program to calculate the optimal arrangement of the screen, beam splitter and retro-reflective sheet to achieve an aerial display measuring approximately 56 inches diagonally projected at least one meter from the beam splitter.

Mitsubishi Electric has also unveiled advanced technologies such as:

Compact AI

MitsubishiElectric develops a small-memory, compact AI that can be implemented in vehicle equipment industrial robots and other machines. The latest development reduces computational costs for inference which involves processes such as identification, recognition and prediction to anticipate unknown facts based on known facts.

A machine-learning algorithm known as deep learning can perform high-level inference, but it requires high computational costs and memory since it employs a deep neural network. Mitsubishi Electric has used more effective network structure and computational models to develop a novel algorithm that realizes a more compact AI with the same inference performance as a conventional AI.

This compact AI technology can be used to reduce costs and perform at high level and high speed inference by establishing a highly secured environment.

Cyber attack detection technology

The latest cyber attack detection technology from Mitsubishi Electric classifies computer virus behavior to about 50 different patterns. The recent report from Symantec throws light on innumerable viruses mushrooming every day. The company’s new technology possesses the capabilities of identifying unknown viruses based on their behavior, thus, plugs data information leakage.

During a cyber attack, the virus has to take several steps, such as infecting a targeted computer, receiving commands from the attacker, investigating the infected computer, and then illegally obtaining access rights to further expand its activities. Each of those steps also has a set of associated behavior patterns. Mitsubishi Electric has identified about 50 behavior patterns, and has defined dedicated log analysis rules for each that allow the system to monitor suspicious activity and accurately detect a virus for preventing information leakage. While the number of new viruses each year grows astronomical, a dozen common behavior patterns are expected to evolve each year. These patters can quickly be added to the detection system, stopping all new viruses that might seek to exploit the pattern.

With conventional technology it is hard to distinguish between a legitimate and malicious cyber attack, Mitsubishi Electric's new technology defines a sequence of behaviors as a cyber-attack scenario. By using correlation analysis to determine whether a particular sequence of activities follows the scenario or not, the technology is able to distinguish between legitimate activities that follow similar patterns and actual cyber attacks.