Algorithm to eliminate traffic jams for the automotive Internet
Automobile manufacturers are testing the peer-to-peer communication systems between cars via WiFi, on which V2V (vehicle-to-vehicle) communication networks will be built. Such networks can drastically improve the situation on the roads, if the cars start to transmit information to each other about traffic congestion, traffic jams, traffic lights, etc. At intersections, cars can check their trajectories, exchanging data in real time. Creating a V2V network is crucial for normal autopilot operation, which in the future will replace live drivers and should be coordinated with other cars.
Technically, a V2V network can be raised right now, but the main problem is software. What are the optimal algorithms and optimal vehicle speed in a stream? How should robots-cars coordinate their movement if there are most of them on the road? Scientists from Seoul University who published a scientific paper Dissolution of traffic jam via additional local interactions ( PDF ) are working on one aspect of this problem. They created an algorithm for V2V, which guarantees a very fast elimination of traffic jams through the exchange of information and instantaneous change in the speed mode of all cars before and after traffic jam.
In recent years, many mathematical models have been developed and a lot of experimental data have been collected on the results of measuring traffic patterns. Scientists have come to a consensus that there are three main patterns:
Hyun Keun Lee and Beom Jun Kim from Seoul University have proposed a simple algorithm for automating such a transition. With the help of cellular automata, they have compiled a model of road traffic and introduced into it the factor of information exchange on the current speed between cars. Scientists also took into account the setting of "optimistic" and "defensive" driving strategies. In the first case, the driver keeps the distance from the vehicle in front less than the safety standards require, and with a “defensive” strategy, on the contrary, the distance is too large.
')
The simulation of the situation showed that if, when approaching the traffic jam of all drivers, to switch to a protective strategy, the flow rate decreases. At the same time, those cars that leave the traffic jam should do it quickly using an optimistic strategy. Thus, the basic condition for sucking the cork is observed - and it quickly disappears.
This is an interesting and simple algorithm for how to eliminate traffic jams using a V2V network. True, for the most efficient operation of the algorithm, it is desirable that the vehicles be controlled in automatic mode and can instantly switch to the required driving mode and change their speed in accordance with the command. Live drivers can not react so quickly and clearly to the incoming information.
However, it is possible to imagine that the partial autopilot and V2V radio communication will appear in the next generation of cars. It is also necessary that these algorithms be taken as a single standard and supported by autopilot systems from all manufacturers.
On the other hand, for effective resorption of traffic jams, it is not necessary that 100% of the cars act in accordance with the described algorithm, rather fewer of them. What kind of "intelligent" machines required in the stream still needs to be calculated.
It remains to hope that with the automation of traffic in the future, traffic jams will completely disappear from our roads, and driving alone will bring pleasure.
via Technology Review
Technically, a V2V network can be raised right now, but the main problem is software. What are the optimal algorithms and optimal vehicle speed in a stream? How should robots-cars coordinate their movement if there are most of them on the road? Scientists from Seoul University who published a scientific paper Dissolution of traffic jam via additional local interactions ( PDF ) are working on one aspect of this problem. They created an algorithm for V2V, which guarantees a very fast elimination of traffic jams through the exchange of information and instantaneous change in the speed mode of all cars before and after traffic jam.
In recent years, many mathematical models have been developed and a lot of experimental data have been collected on the results of measuring traffic patterns. Scientists have come to a consensus that there are three main patterns:
- free flow, in which the density of vehicles is low enough and each of them can drive at the highest possible speed;
- synchronized flow in which a high density of vehicles causes drivers to slow down a little;
- traffic jam - the speed drops to zero and the movement occurs by the method of short-term jerks.
Hyun Keun Lee and Beom Jun Kim from Seoul University have proposed a simple algorithm for automating such a transition. With the help of cellular automata, they have compiled a model of road traffic and introduced into it the factor of information exchange on the current speed between cars. Scientists also took into account the setting of "optimistic" and "defensive" driving strategies. In the first case, the driver keeps the distance from the vehicle in front less than the safety standards require, and with a “defensive” strategy, on the contrary, the distance is too large.
')
The simulation of the situation showed that if, when approaching the traffic jam of all drivers, to switch to a protective strategy, the flow rate decreases. At the same time, those cars that leave the traffic jam should do it quickly using an optimistic strategy. Thus, the basic condition for sucking the cork is observed - and it quickly disappears.
This is an interesting and simple algorithm for how to eliminate traffic jams using a V2V network. True, for the most efficient operation of the algorithm, it is desirable that the vehicles be controlled in automatic mode and can instantly switch to the required driving mode and change their speed in accordance with the command. Live drivers can not react so quickly and clearly to the incoming information.
However, it is possible to imagine that the partial autopilot and V2V radio communication will appear in the next generation of cars. It is also necessary that these algorithms be taken as a single standard and supported by autopilot systems from all manufacturers.
On the other hand, for effective resorption of traffic jams, it is not necessary that 100% of the cars act in accordance with the described algorithm, rather fewer of them. What kind of "intelligent" machines required in the stream still needs to be calculated.
It remains to hope that with the automation of traffic in the future, traffic jams will completely disappear from our roads, and driving alone will bring pleasure.
via Technology Review
Source: https://habr.com/ru/post/128620/
All Articles