Cognitive anticipation cellular automata model: An attempt to understand the relation between the traffic states and rear-end collisions
Kamal Jetto,Zineb Tahiri,Abdelilah Benyoussef,Abdallah El Kenz
DOI: https://doi.org/10.1016/j.aap.2020.105507
Abstract:We have investigated the accident's statistics of Europe and North America that are provided by the UN. This investigation has shown that accidents due to the traffic represent around 50 % of the total number of accidents every year. Among them, rear-end collisions hold a 20 % share. These numbers display the fact that the interaction between drivers can be held responsible of those incidents. In this respect, we have explored the reasons behind the conflict situations that may be responsible of the occurrence of rear-end collisions by the mean of a cognitive psychology based cellular automata model. Indeed, through field experiments performed by an embedded camera, we have extricated a psychological cognitive process of anticipation. We have defined the latter as the tendency of drivers to accelerate based on the history of their predecessor. Then, we have exploited the tools of the physics of traffic by which we have developed a CA-model that take into consideration this process. As a result, we were able to generate those incidents' situations. By considering two types of drivers: conservative who respect the learned information about the safe manoeuvres but make mistakes or aggressive who violate those secure processes, we have proved the complexity of the relationship between the states of the traffic flow and the drivers' behaviours. In fact, we have shown that rear-end collisions are a result of the anticipation as a response of the drivers to the traffic conditions: the congestion. Moreover, we have also highlighted an improvement of the flow in the congested state up to 11 % due to the anticipation, but that can only be achieved through vehicle-to-vehicle communication. Finally, we have investigated the hot spots. We have found that the traffic perturbations, that generate those hot spots and can be responsible of collisions, are more likely to be located away in the downstream direction. The distance between the two locations depends on the traffic density. This difference between the positions of the traffic perturbation and the hot spot has showcased the complexity, in time and space, of the transmission and the reception of deceleration information by the drivers.