When I was starting out in my career I retained my interest in technology, electronics and so on and one day stumbled upon a truly fascinating book – a book I recommend to people even today because of its in-depth yet readable coverage of AI and robotics – the book is The Thinking Computer – Mind Inside Matter. This book is about AI and LISP and stuff one would expect but its very readable, not overly theoretical – it also has a fascinating chapter about a 1966 robot project setup at Stanford – the robot was named Shakey.
The truly fascinating thing about Shakey was that it could solve real generic problems – it could be tasked with locating blocks of specified color, and moving around them or moving the blocks, traversing corridors and so on. What was impressive was that it could be given just a goal and then a powerful planning system would backtrack from the goal to the present situation.
If it was possible to get from the goal back to the start then the task was deemed to be achievable else it was unachievable. During the planning operation the planner created a series of distinct actions and accumulated these. If the task was achievable then these actions were then executed in reverse by the robot and the goal was reached – I found this truly fascinating.
I still consider this a truly good book, very accessible and although its now a little dated its still a great description of the Shakey project and worth getting for that alone.