Conclusion

Autonomous vehicles have gone beyond the prototype stage, and promise to bring radical changes in many areas. Using these vehicles to smooth traffic flow or alleviate disruptions is a response to major challenges: instability in high-density traffic increases the risk of accidents, driving stress and fuel consumption, and reduces the average throughput of stretches of road.

The strategies associated with tracking models can be implemented in driving software, provided sufficient data are available, and can be used to return to a stationary state, even if this is normally highly unstable.

This illustrates the complementary nature of physical models and information processing.

Nevertheless, we are moving towards neural network learning.

We can also look to biomimicry. With ants, as density...