Trajectory Prediction
Trajectory prediction involves forecasting the spatial coordinates of various dynamic entities on the road, such as cars, buses, pedestrians, tricycles, and animals, over short-term (1-3 seconds) and long-term (3-5 seconds) horizons. The task aims to provide accurate predictions by analyzing the distinct dynamic behaviors of these entities, such as aggressive or conservative driving styles, to enhance the safety and efficiency of autonomous driving systems and intelligent traffic management.
nuScenes
UniTraj (MTR)
Stanford Drone
Y-Net
ETH/UCY
NSP
ETH
Social-Implicit
JAAD
SGNet
PIE
SGNet
ActEV
Pishgu
INTERACTION Dataset - Validation
ITRA
PAID
DESIRE
TrajNet++
Social-NCE + Social-LSTM
ApolloScape
SpectralCows
HEV-I
SGNet
NGSIM
TraPHic
Apolloscape Trajectory
Trafficpredict
Argoverse
Argoverse2
ETH BIWI Walking Pedestrians dataset
Social Ways
ForkingPaths
Multiverse
GPS
GTA-IM Dataset
Skeleton-Graph
Hotel BIWI Walking Pedestrians dataset
Social Ways
Lyft Level 5
PROX
Skeleton-Graph
SDD
Multiclass-SGCN (ours)
STATS SportVu NBA [ATK]
DAG-Net
STATS SportVu NBA [DEF]
DAG-Net
TRAF
TraPHic
TrajAir: A General Aviation Trajectory Dataset
TrajAirNet
UCY
YJMob100K@B
YJMob100K@C
ST-MoE-BERT
YJMob100K@D