US 12,148,174 B2
Method for forecasting motion trajectory, storage medium, and computer device
Maosheng Ye, Shenzhen (CN); and Tongyi Cao, Shenzhen (CN)
Assigned to SHENZHEN DEEPROUTE.AI CO., LTD, Shenzhen (CN)
Filed by SHENZHEN DEEPROUTE.AI CO., LTD, Shenzhen (CN)
Filed on Nov. 19, 2021, as Appl. No. 17/530,465.
Prior Publication US 2023/0162374 A1, May 25, 2023
Int. Cl. G06T 7/246 (2017.01); G05D 1/00 (2006.01)
CPC G06T 7/246 (2017.01) [G05D 1/0253 (2013.01); G06T 2207/10028 (2013.01); G06T 2207/20081 (2013.01); G06T 2207/20084 (2013.01); G06T 2207/30241 (2013.01)] 14 Claims
OG exemplary drawing
 
1. A method for forecasting a motion trajectory, comprising:
obtaining an observed past trajectory of an object;
obtaining a spatial pointwise feature of each trajectory point in the observed past trajectory;
obtaining a temporal pointwise feature of the trajectory point according to the spatial pointwise feature of the trajectory points within a preset observation time interval; and
performing motion trajectory prediction on the object according to the spatial pointwise feature and the temporal pointwise feature of the trajectory points;
wherein obtaining the spatial pointwise feature of each trajectory point in the observed past trajectory comprises:
obtaining a first spatial pointwise feature of each trajectory point by performing feature extraction on each trajectory point;
transforming the first spatial pointwise feature of each trajectory point into a second spatial pointwise feature by point-voxel propagation; and
obtaining a spatial pointwise feature of the trajectory point by fusing the first spatial pointwise feature with the second spatial pointwise feature.