A multi-agent q-learning based channel access method for MAC layer in vehicle communication

Abstract

The invention discloses an MAC-layer channel access method based on multi-agent Q learning in vehicular communication. The method comprises the steps that: in a VANETs environment, each vehicular node constructs the joint state-action pair mapping relationship and the combined strategy itself; then, whether a new vehicular node is added into a VANET network or not is judged; if the new vehicular node is added into the VANET network, the new added vehicular node rapidly obtains the action space, the state space and the reward function through transfer learning; hereafter, each vehicular node updates the joint state-action pair relationship and the combined strategy itself; if the new vehicular node is not added into the VANET network, whether a current vehicular node has data needing to be sent or not is judged; if the current vehicular node has the data needing to be sent, the action strategy solution satisfying correlated equilibrium can be determined according to a eCEQ algorithm; an action enabling a multi-agent system to achieve correlated equilibrium finally is selected from an action set; a CW value is determined; and, the vehicular node is accessed to a wireless channel to send data according to the CW value. By means of the MAC-layer channel access method based on multi-agent Q learning in vehicular communication disclosed by the invention, the data sending success probability is increased; the withdrawal time is reduced; and the data packet receiving rate, the end-to-end transmission delay problem and the like can be effectively improved.

Description

technical field [0001] The invention belongs to the technical field of the Internet of Things, and relates to an implementation method of MAC layer channel access based on multi-agent Q learning in vehicle communication. Background technique [0002] Since the invention of the motor vehicle during the second industrial revolution, with the rapid development of the automobile field, the automobile has become an indispensable part of people's modern life. Along with the quickening of people's daily life rhythm, the use of transportation means such as bus, private car is common day by day. While cars bring convenience to people's daily travel, they also cause many problems, such as traffic congestion, environmental pollution, traffic accidents, etc. Among them, traffic congestion has become a serious social problem, which brings many problems to road users, and a large amount of fuel waste and time waste are caused by traffic congestion every year. Not only does people waste ...

AI Technical Summary

Problems solved by technology

[0004] However, the above existing technologies are all improved on the basis of the BEB algorithm. Generally speaking, when the data collides and needs to be avoided, the CW value is multiplied. After the data is successfully sent, the CW will be restored to 15. If there are multiple nodes At the same time, the data is successfully sent, the CW value is restored to 15, and a collision occurs when the data is sent again.

The network load is less considered, and it is not suitable for networks with different load levels, that is, it is not scalable for traffic flows of different densities, and the fairness of channel access has not been effectively improved.

Images