Self driving cars may have become a reality very recently but the quest for making such cars has continued since decades. Details of one such self driving car has surfaced which had come into existence in 1989.
ALVINN (Autonomous Land Vehicle In a Neural Network) was the result of eight years of military-funded research at CMU’s robotics institute. It can be considered the forefather of today’s self-driving cars.
The ALVINN implementation processed data from two sources, the video camera and the laser range finder. It was a perception system which learnt to control the NAVLAB vehicles by watching a person drive. Each input unit was fully connected to a layer of five hidden units which were in turn fully connected to a layer of 30 output units.
It used neural network which means the vehicle was “narrowly intelligent” and made decisions without the need for a map, the researchers hoped to build an autonomous driving system that could have been more adaptable across a variety of conditions.