http://arstechnica.com/science/2016/01/m...eir-minds/
EXCERPT: [...In social physics...] Most pedestrian models are reasonably simple. Pedestrians are particles that are driven by some force to go in a direction; they don’t collide with each other because there is a repulsive force between them keeping them apart. [...] However, you can use any number of different physical models to study pedestrian interactions.
Unfortunately, pedestrian models are not very well tested against data. Most experiments involve paying university students to walk along corridors and through doors under highly artificial conditions. In part, this is because it has been very difficult to obtain data from natural settings, where you need to track individual pedestrians as they walk through some area of interest.
[...] From the university data, he obtained a quarter of a million trajectories [...] Corbetta found that for most pedestrians, a fairly simple model was pretty good. It correctly predicted that as pedestrians walked the corridor, their trajectory would not be a straight line. Instead, it was a curve with some fluctuations about an average position. For many pedestrians, the model worked well. What the model failed to predict well were the changes in speed along the direction of the corridor. And one notable failure stood out: people who change their mind, make a U-turn, and return from whence they came. Since most models use a kind of driving force to give pedestrians a destination, the model does not allow them to change their mind....
EXCERPT: [...In social physics...] Most pedestrian models are reasonably simple. Pedestrians are particles that are driven by some force to go in a direction; they don’t collide with each other because there is a repulsive force between them keeping them apart. [...] However, you can use any number of different physical models to study pedestrian interactions.
Unfortunately, pedestrian models are not very well tested against data. Most experiments involve paying university students to walk along corridors and through doors under highly artificial conditions. In part, this is because it has been very difficult to obtain data from natural settings, where you need to track individual pedestrians as they walk through some area of interest.
[...] From the university data, he obtained a quarter of a million trajectories [...] Corbetta found that for most pedestrians, a fairly simple model was pretty good. It correctly predicted that as pedestrians walked the corridor, their trajectory would not be a straight line. Instead, it was a curve with some fluctuations about an average position. For many pedestrians, the model worked well. What the model failed to predict well were the changes in speed along the direction of the corridor. And one notable failure stood out: people who change their mind, make a U-turn, and return from whence they came. Since most models use a kind of driving force to give pedestrians a destination, the model does not allow them to change their mind....