Vehicle trajectory prediction method based on global attention and state sharing

Abstract

The invention discloses a vehicle trajectory prediction method based on global attention and state sharing. The solution method comprises the following steps that a GAS-LED trajectory prediction model of a codec LSTM model with a global attention mechanism and state sharing is used; in the GAS-LED track prediction model, a state sharing mechanism with an encoder and a decoder is adopted to reduce the calculation workload, and meanwhile, two parallel calculation GAS-LED track prediction models are adopted to output prediction of the transverse lane changing behavior and the longitudinal driving distance of the vehicle in parallel; in the track prediction task of the lane level, the lane where the vehicle is located is focused on, and the GAS-LED track prediction model outputs corresponding prediction results for the transverse lane change and the longitudinal driving distance; and historical information of the current vehicle and the surrounding vehicles is used as the input of the GAS-LED trajectory prediction model 2, and then the two GAS-LED trajectory prediction models are used in parallel to obtain more output results convenient to predict. Through the scheme, the purpose of high-precision prediction is achieved, and the method has very high practical value and popularization value.

Description

technical field [0001] The invention belongs to the technical field of automatic driving, and in particular relates to a vehicle trajectory prediction method based on global attention and state sharing. Background technique [0002] Existing research on trajectory prediction can be roughly divided into two categories. They are rule-based and learning-based trajectory prediction algorithms, respectively. Rule-based forecasting algorithms simulate traffic flow models by applying traffic rules, while learning-based forecasting algorithms employ machine learning models (recently deep models) to make predictions based on historical trajectories of moving objects. [0003] Rule-based trajectory prediction: Rule-based trajectory prediction algorithms mainly apply traffic rules to simulate traffic flow models. For example, when there is no vehicle ahead on the right, the vehicle will move to the right lane; if there is no vehicle ahead, the vehicle will accelerate. The simulation...

AI Technical Summary

Problems solved by technology

However, the training process of the aforementioned networks is usually time-consuming due to the large number of parameters and hyperparameters

This will prevent the self-driving vehicle from updating the model parameters in time

Furthermore, using a single model to predict lateral and longitudinal information reduces the accuracy of lane level predictions, leading to unreliable maneuvering decisions

[0007] 2. Sudden braking and inappropriate lane changes, both of which are also responsible for most ghost traffic jams

However, in inappropriate lane changes there are: 1) lack of proper framework; 2) uncertainty in the driving behavior of surrounding vehicles; 3) various vehicle maneuvers; resulting in inappropriate lane changes not being well resolved

[0008] 3. The existing LCS system mainly focuses on the safety and comfort of driving, but ignores the impact of the lane changing operation of the autonomous vehicle on the surrounding traditional vehicles, which may lead to traffic congestion

Images