Trajectory planning method for automated vehicles based on risk prediction in mixed traffic environment

Behavior trajectory prediction research of adjacent vehicles based on LSTM

“…After that, the encoding result of the neighbor vehicle is input into the convolution machine interaction layer and attention layer simultaneously. The former directly outputs the context information of the neighbor vehicle interaction, the latter performs weight evaluation by inputting the neighbor vehicle encoding information and the subject vehicle encoding information, and finally completes the result prediction of the driving intention and trajectory through the softmax layer, and finally uses the decoder to output the trajectory [5].…”