The goal of the project was to predict the motion of an autonomous vehicle and the surrounding agents given their trajectory for the past one second. For this, we used rasterized images as an input to a CNN baseline. To the given parameters, we added the velocities along x and y direction. This helped predicting instantaneous velocity of the agent at every time step. In addition, the baseline was modified by adding an LSTM decoder to study their impact on predictions. Instantaneous velocity was an added parameter for the model and predicting instantaneous velocities can be used to improve the motion prediction of the agents and can facilitate agent interaction and cooperative driving
AmeyDeshpande97/Motion-Prediction-of-Agents-in-the-Vicinity-of-Self-Driving-Car
The goal of the project was to predict the motion of an autonomous vehicle and the surrounding agents given their trajectory for the past one second. For this, we used rasterized images as an input to a CNN baseline. To the given parameters, we added the velocities along x and y direction. This helped predicting instantaneous velocity of the agent at every time step. In addition, the baseline was modified by adding an LSTM decoder to study their impact on predictions. Instantaneous velocity was an added parameter for the model and predicting instantaneous velocities can be used to improve the motion prediction of the agents and can facilitate agent interaction and cooperative driving
Python