Dynamic platoon formation method under mixed autonomous vehicles flow

1. A dynamic platoon formation method under a mixed autonomous vehicles flow, wherein a platoon request is sent by performing a communication between connected autonomous vehicles (CAVs) on a multi-lane road section, and determining a target lane for platoon and a platooning scheme are determined by:

obtaining road information and dynamic information of all vehicles, wherein the road information comprises a lane distribution, and the dynamic information of all vehicles comprises location information and speed information of each human driven vehicle (HDV) and each CAV;

selecting lanes as candidate lanes for platoon in turn;

performing a following process for each candidate lane for platoon, wherein the following process comprise: obtaining a compatible decision set after constructing a decision space for each CAV; for each compatible decision in the compatible decision set, simulating execution of the compatible decision and predicting the location information and the speed information of all the vehicles after the execution of the compatible decision; constructing a decision space for each CAV again and obtaining a compatible decision set; and continuously cycling the process to generate a decision tree of the candidate lane for platoon, wherein a root node of the decision tree is a traffic flow state at a current moment at which the platoon request is sent; each branch represents a system state transfer after the execution of each selected compatible decision; a terminal node represents a platoon formation; each branch from the root node to the terminal node represents a decision sequence in a process of the platoon formation; and an optimal decision sequence is selected according to a predetermined evaluation index; and

obtaining respective optimal decision sequences corresponding to all the candidate lanes for platoon, and selecting an optimal decision sequence according to the predetermined evaluation index as a final platoon execution scheme, wherein the candidate lane for platoon corresponding to the final platoon execution scheme is a final target lane for platoon.

2. The dynamic platoon formation method under the mixed autonomous vehicles flow according to claim 1 , wherein the decision space for the CAV comprises at least one of:

direct lane change (LC), which means that a current CAV directly changes to an adjacent lane at an original travel speed;

merge with an existing platoon (MERGE), which means that the current CAV is located in the target lane and only needs to merge with an existing platoon in the target lane;

accelerate lane change (ALC) or decelerate lane change (DLC), which means that the current CAV accelerates or decelerates to change to an adjacent lane; and

decelerate gap provide (DGP), which means that the current CAV decelerates to provide a suitable gap for other vehicles to allow other CAVs to change lane.

3. The dynamic platoon formation method under the mixed autonomous vehicles flow according to claim 1 , wherein the compatible decision set comprises a plurality of compatible decisions, each of the plurality of compatible decision comprises decisions from decision spaces for one or more CAVs, and the decisions do not conflict with each other.

4. The dynamic platoon formation method under the mixed autonomous vehicles flow according to claim 1 , wherein the predetermined evaluation index is a time to platoon (TTP), which is a duration from a moment of platoon request to a moment of platoon formation.

5. The dynamic platoon formation method under a mixed autonomous vehicles flow according to claim 1 , wherein the predetermined evaluation index is a distance to platoon (DTP), which is a distance traveled by a leading vehicle sending the platoon request from a moment of platoon request to a moment of platoon formation.