Telefonica is fuelling the hype as we motor towards MWC with connected car announcements alongside Spanish automotive giant Seat.
In an early effort to drive traffic towards its stand, Telefonica has carpooled with Seat to give the green light to three new innovations in the connected vehicles race. While there are sceptics who would want to curb autonomous vehicles enthusiasm, the duo is racing towards a happy middle-ground with three assisted driving use cases.
Firstly, the team will introduce pedestrian detection capabilities, which will allow traffic lights to sense the presence of pedestrians with thermal cameras, before relaying this information onto cars in the nearby area. Display panels will be able to inform the driver of potential risks on the road.
Secondly, connected bicycles equipped with a precise geolocation will notify vehicles in the area when the rider decides to turn right. The bikes will be detected by ultra-wideband beacons placed along the road, and should there be a risk of collision, the driver in the car will once again be notified.
While both these ideas will be powered by edge-computing, the final usecase will rely on direct communication interface. Should visibility be particularly low, stationary vehicles would detect moving vehicles, emergency lights would be turned on while the driver would, again, be notified on the display board.
These usecases might not be on the same level as the glories of autonomous vehicles, but there is a satisfactory amount of realism on display. Autonomous vehicles are not going to be on our roads for a long-time, and while that does not mean we should not continue to fine tune the technology, there has to be a focus on improving road safety today. This is exactly what is being done here.
Another similar concept is being developed in MIT. Here, an AI application analyses the way pedestrians are walking to understand whether there might be any risks. This sort of analysis is something we all do subconsciously, but a very useful and important addition to the connected car mix.
Using lidar and stereo camera systems, the AI estimates direction and pace, but also takes pose and gait into consideration. Pose and gait not only inform the pace and direction, but also give clues to future intentions. For instance, if someone is glancing over their shoulder, it could be an indication they are about to step into the road.
Looking further into the future, when autonomous taxis might be a real thing, this could also be incredibly useful. Of course, the simplest way to hail a taxi in this futuristic age will be through an app, but if the vehicle can see and understand an outstretched arm is a signal for a taxi, it would be a useful skill to incorporate into the AI.
All of these ideas are not only relevant for the long-term ambitions of the automotive industry but also very applicable today. Connectivity and AI can be incredibly beneficial for human-operated vehicles, especially with the advancements of edge-computing and leaning on the high bandwidth provided by 5G. Not everything has to be super-futuristic, and it’s nice to see a bit of realism.