Robotic Cars

I talked earlier about the challenges of transportation as the world population grows. More people, more cars, more materials to support those people needing to go from where it’s produced to where it is consumed. Motor cars are a wonderful technology. They’ve revolutionized the way we live, but they are very dangerous machines. Every year, over one million people are killed by motor cars and more than a hundred times that number are injured. Globally, this is costing the planet nearly half a trillion dollars every year.

For high-income countries, that’s nearly 2% of their gross national product. In low or middle-income countries, that number is smaller, but it’s growing. As a cost of development is that a larger proportion of your gross national product is being lost through traffic injuries and fatalities. If we look at the 10 leading causes of death in 2002, road traffic injuries were at number 10. By the year 2030, it’s predicted they will have grown to number eight. So while we’ve done great things for the health of pregnant women and reduced the incidence of tuberculosis, the road traffic injuries, the car is going to kill more and more people on the planet and that’s a sad thing.

Here is some more statistics which show by region the percentage of gross national product that is lost due to traffic injuries and fatalities. Now, motor cars are driven by human beings and human beings are the primary cause for the motor vehicle accidents and the fatalities of 10 million accidents in the United States each year. Nine and a half million of those accidents are due to the fault of the drivers. I would rather like the quote here.

Interestingly, most drivers overestimate their own ability. I particularly like this quote. It suggests that most American students, 93% of them believe that they are above average drivers and of course, that’s an impossibility. The bottom line is human beings are not very good drivers even if they believe that they are. Human beings cause accidents which lead to a large number of injuries and deaths.

This graph here shows the number of fatalities per 10 thousand vehicles as a function of time and we can see that over the last 50 years, there’s been steady improvement in reducing the number of fatalities per car and it’s due to improvements in car design, campaigns to stop people driving under the influence of alcohol, the introduction of seatbelts and airbags and so on.

While the general improvement is quite welcome, the sad thing is that there is this plateau, the rate of change of improvement has really fallen off and so we’re stuck on this floor. It’s not at zero. There is this finite number of accidents that’s going to happen as a fraction of the number of vehicles on the road.

One consequence of the imperfection of human drivers is the way that the human-driven cars interact with one another. The dynamics of this system is very, very complex and studied by mathematicians and traffic engineers. As we increase the density of cars on the road, the flow rate actually starts to fall off; this is a situation that we’ve all experienced personally.

If I look out and down from my office window, I see a busy expressway and this is a view of the expressway when it’s actually quite congested and not moving particularly quickly. So the drivers in this situation are thinking that the traffic is not flowing, that the road is jammed up, but from my eagle eye view, I can see that there’s an awful lot of road that’s got no cars on it and this is another problem with human drivers, because they’re not very good, they’ve got quite slow reaction times really, we need to leave large gaps in between the cars.

It would be much more efficient if we could pack more cars into all of that empty space. Human beings with their poor reaction times mean that we can’t do that. If we did that, we’ll just have tons more accidents. But perhaps, if human beings weren’t driving the cars, if we had self-driving cars or robotic cars such as depicted here by one of the Google self-driving cars, then perhaps things could be better. That would be the situation for autonomous cars that were doing some platooning. That is we have a group of eight cars travelling very closely together and moving down the road. With this sort of technology, we would be able to fit a lot more cars on the road. They travel more smoothly. They travel more quickly. We could get more cars down existing roads without having to make them any wider.