Pedstrian Trajectory Prediction Based on LSTM

Pedestrian trajectory prediction is a key technology in Intelligent Driving that can provide valuable guidance for subsequent Intelligent Driving path decisions. This technology poses significant challenges due to environmental factors and the influence of other pedestrians on pedestrian trajectories. The core of pedestrian trajectory prediction is a sequence prediction problem, which poses specific challenges in capturing inter-sequence dependencies and thus predicting multimodal distributions that conform to society. In recent years, neural network-based methods have been shown to outperform traditional methods in pedestrian trajectory prediction. In this paper, we will perform pedestrian detection and predict their future trajectories in first person view. This method has been evaluated on existing datasets and compared with traditional trajectory prediction methods, achieving advanced results.