|
Autonomous vehicles should be connected. Autonomous vehicles may be combining data from cameras, onboard sensors, and lidars, they may be intelligent, learning, and adapting to each situation; but, if they are not connected, then they are not very smart.
Active safety functions have been taken about as far as they can go. To advance safety further, connectivity is needed. Communications can be seen as a different type of sensor, providing the nonline-of-sight information that radars cannot.
During fog, where the camera cannot see, communications can relay information from the vehicle in front, and, similarly, it can relay information about the driver’s intentions, which the lidar is unable to detect.
Detecting a human in the dark by using machine learning is difficult, and the algorithms can be fooled. Incidents that seem straightforward to humans may mislead the automated vehicles and vice versa—getting drunk, sleepy, and distracted seems irrational to hazard classification algorithms.
Uber and Tesla are self-driving but not connected and cooperating. Could the accidents have been prevented if the autonomous vehicles were connected?
Full article: IEEE Vehicular Technology Magazine, Volume 13, Number 3, September 2018 |
;){kind=small}
