Short Term Trajectory Prediction

Short-term trajectory prediction focuses on accurately forecasting the movement of objects, particularly vehicles and pedestrians, within a short time horizon. Current research emphasizes improving prediction accuracy in complex and safety-critical scenarios by incorporating risk assessment and intention understanding into models, often employing recurrent neural networks (RNNs) like LSTMs and GRUs, sometimes augmented with convolutional neural networks (CNNs). This work is crucial for advancing autonomous vehicle safety and enhancing applications like redirected walking in virtual reality, where precise movement prediction is essential for seamless user experience and collision avoidance.