Autonomous Vehicles

What are Autonomous Vehicles doing mixed up with Legislation and Defence Contracts?

Autonomous vehicles

Legislation is too detailed to handle with the Conscious Mind – so is driving. You leave the house to drive to work – you turn the key, and seemingly instantly, your workplace appears, without you having any recollection about how you got there (it is more embarrassing when you were meant to go somewhere else).

When you come to a traffic light – currently green – you are calculating whether you will have time to cross the intersection before the light turns red – you are making predictions into the future based on your position, speed, the position and speed of the vehicle behind you (if it is a heavy truck, will it stop in time if you stop), and the flow of traffic in front of you – you are doing a lot of unconscious figuring.

Through practice and experience, task performance improves when actions become more automated and there is less of a requirement for conscious intervention. With driving, it may be the case that through experience, fewer conscious resources are required to control the vehicle as driving skill becomes automatic and this frees up resources to allocate visual attention to other parts of the scene.

Why Do This?

For an AV to mingle with, and compete with, human drivers means that we need to emulate the abilities of the human’s Unconscious Mind – a formidable task. Using English as the training interface to the machine confers major benefits – we can understand what the machine has been told, and we can tell it complex things in a very expressive language, where separate processes – visual, aural, prediction – can be readily bound together.



See the blog post Is It Time to Rethink Autonomous Kit? for more arguments about using English as the training language for Autonomous Vehicles (AV). The post was triggered by Apple abandoning its AV project.