2016, Tweeting cars: Driver Behavior Monitoring System

What is the idea behind the project?

The ascent of 5G technologies has opened many opportunities to explore in direct Device-to-Device communication. This can be a great opportunity to bypass the previously proposed methods of using the driver’s cell phone as a mode of transmission as this raises many questions of putting privacy at risk and makes users wary of participating due to the same fears. The idea is to use the communication and process data in two stages, on the fly as well as long term pattern tracking.

Each vehicle shall have a 5G communication based computing system, that can record the position as well as basic data that is tracked for the dashboard screen – acceleration patterns, braking patterns, top speeds, maneuvering frequencies, and try to map that with GPS as well. Basically, we’ll be trying to collect data for driving pattern tracking. This can then be broadcasted on low power 5G frequencies.

This can be used to address one of the biggest questions of the present era of autonomous cars - “Can autonomous and manually driven cars coexist on same roads safely?”

What is the essence of doing this?

The aim is to observe the driving pattern of a driver over the current trip and on basis of various parameters categories it under the categories of Safe Driver or Unsafe Driver, moreover, the unsafe drivers have been observed on the basis of their driving patterns if it is recommended to be drive behind them or ahead of them for safety. This information is then broadcasted continuously, and as any autonomous car comes closer to a manually driven car, it can listen to this broadcast and then improvise on its trajectory according to that data. This can assist in safer driving for autonomous cars as statistically, most of the accidents involving autonomous cars have been because of the manual drivers’ mistake not because of the autonomous cars’ mistake.

What did the demo include?

The demo was based on observing the driving pattern of a driver and broadcasting it. During the demo, SML world environment was used to simulate a car with various driving patterns, while a screen showed the broadcasted driver behaviour.

Further scope of the project with 5G!

The aim is to make trajectory decisions collaborative, this could be achieved through the above-proposed broadcasting system. This, if successful will make the future self-driving cars more practical and lesser estimations would be involved. 5G makes it more practical because of the low latency (1 ms of less) of these networks. Moreover, this can serve as a Distributed Data Collection system as well, now that we have eliminated the possibility of using the user’s cell phone as the channel for data transmission, there is a need of finding a way to get this data for further analysis without troubling the user. The cars can share their encrypted data dumps with each other and this way everyone keeps collecting other’s data, now with the self-driving cars one of the necessity would be to have traffic signals and speed limit signs also communicative, these are the points that we can have as data collection points and collect these data dumps. 5G makes it more exciting because of the high-speeds of these technologies of the order of 1 GBPS. There are a lot of things that can be predicted from this collected data including finding the unsafe and safe zones, in fact, pollution data collection could be done this way as well.

This if executed well can bring some new trends and dimensions in the way self-driving cars are being approached these days.