One of the nicest things about writing this comment every month is that I can do a bit of ‘futurology’ – looking at social, software and hardware trends and extrapolating further – the sort of thing that old buffers do in front of large audiences and get paid enormous amounts for without having to actually deliver on their visions.
What prompted this thought was the article from Mass Motion on pedestrian simulation, which sets up complex scenarios in which thousands of individual agents – each of them set up with their own idiosyncracies, are given a task to get from A to B, usually within a 3D model of an airport, railway station etc., and left to themselves to carry out that task, interacting with all other agents and the obstacles they will meet on their way.
That is not far removed from the tasks that will be set for autonomous vehicles, each of which will have a destination keyed in by a passenger, and a model of the city in which it operates, updated in real-time by the activities of other driverless vehicles. Using similar algorithms to those used by Oasys in Mass Motion, they will be able to find optimum routes to accomplish the journey successfully and in a timely manner.
The spanner in the works is having to share the city space with driven vehicles, and their propensity for random activity. To ensure autonomous traffic flows efficiently within the city, other vehicles will have to be hybrid beasts - autonomy taking over in the city - or banned altogether.
Another thought - when nobody owns a car in the city, and is able to hail an autonomous one or use a more efficient public transport system - what happens to the London Cabbie (whose ‘knowledge’ becomes worthless overnight) - or Uber?
To make sure you get your copy of the Newsletter emailed to you personally, every time, click here to register.