Many people may not know, but Mitsubishi Motors is just one of many parts of a large corporation. Although the corporation itself is also called the “Mitsubishi Motors Corporation” (MMC), Mitsubishi Motors of course takes care of the automotive side of things. At the same time, many of MMC subsidiaries tend to design technology that may one day find its way into Mitsubishi Motors vehicles, and one of the most well-known would be Mitsubishi Electric for their strides in self-driving technology and artificial intelligence, some of which we saw during the 2019 Tokyo Motor Show. After a series of events and technologies involving artificial intelligence, a team within Mitsubishi Electric known as Maisart has developed a new type of artificial intelligence that could make GPS navigation in vehicles sounds more organic and human-like.

What does human-like mean? To talk about this, let’s discuss another system developed by Maisart. This team, Maisart is short for “Mitsubishi Electric's AI creates the State-of-the- ART in technology”, and in early January 2020, released reports on a new human-machine interface designed to improve upon the modern day smart assistant. A smart assistant is like the Amazon Alexa or Google Home, and Maisart designed a system or artificial intelligence that could understand commands more intuitively by adding to and upgrading a database of information, and then making services related to those words immediately available. For instance, someone could say “Play chillstep music”, and that alone will cause the smart assistant to play music without an awake command. How that plays out for smart machines always listening will be an interesting argument.

Now, with that in mind, let’s take a look at another system designed by Maisart that does something similar. Called Scene-Aware Interaction acts much like self-driving car technology today. Autonomous vehicles tend to use LiDAR systems to create a wireframe or 3D map of its surroundings for navigating the roads. Scene-Aware Interaction collects information from images and video captured with cameras, audio information recorded with microphones, and localization information measured with LiDAR. Then, by utilizing another technology by Mitsubishi Electric, Attentional Multimodal Fusion technology, Scene-Aware Interaction breaks all of the information down into their appropriate categories based on level of importance, followed by the ability to use appropriate word selections for describing scenes with high accuracy.

What does all this mean? In terms of a GPS system, Scene-Aware Interaction acts like a virtual assistant, and instead of giving directions like “Take a right turn in one mile,” the driver will get something more intuitive and organic, like “Take a right turn on the corner of X and Y”. The kind of directions someone real would give you and takes less mental computing. This technology will also act as a real-time assistant that can alert the driver to pedestrians crossing, as is illustrated in the above image. It’s kind of like an artificially intelligent GPS nav system instead of a voice and a satellite.

Whether or not we’ll see anything like this in vehicles anytime soon, there’s really no clue. Is it even practical is a better question – this is just one step beyond those directions you get from strangers where, “If you hit the barn you’ve gone too far.” It’s not very helpful, but it is more human. What do you think about having more organic dialogue with your GPS system? Join the discussion on NowCar social media.

Photo Source/Copyright: Mitsubishi Electric