US 12,072,678 B2
Systems and methods for providing future object localization
Srikanth Malla, Sunnyvale, CA (US); and Chiho Choi, San Jose, CA (US)
Assigned to HONDA MOTOR CO., LTD., Tokyo (JP)
Filed by Honda Motor Co., Ltd., Tokyo (JP)
Filed on Mar. 24, 2020, as Appl. No. 16/828,343.
Claims priority of provisional application 62/901,216, filed on Sep. 16, 2019.
Prior Publication US 2021/0082283 A1, Mar. 18, 2021
Int. Cl. G05B 13/02 (2006.01); B60W 60/00 (2020.01); G06V 10/25 (2022.01); G06V 20/58 (2022.01); G08G 1/123 (2006.01)
CPC G05B 13/027 (2013.01) [G06V 10/25 (2022.01); G06V 20/58 (2022.01); G08G 1/123 (2013.01); B60W 60/0027 (2020.02)] 16 Claims
OG exemplary drawing
 
1. A system for future object localization, the system comprising:
one or more vehicle sensors for capturing host data associated with a host vehicle, wherein the host data includes a first series of image frames of an environment from the host vehicle and temporal information of past observations of the host vehicle based on the first series of image frames of the environment; and
a processor, the processor having:
a data receiving module, implemented via the processor, identifying one or more proximate vehicles within sensor range in the environment based on one or more of the host data and proximate data received from the one or more proximate vehicles, wherein the proximate data includes historical data of the one or more proximate vehicles as determined based on the first series of image frames of the environment;
a motion prediction module, implemented via the processor, configured to:
encode the host data in a host vehicle stage;
decode the encoded host data in the host vehicle stage to generate a first joint uncertainty distribution based on an initial joint uncertainty model and a host model distribution that includes the host data associated with the host vehicle; and
sample host kinematic predictions of future motions of the host vehicle in the host vehicle stage based on the first joint uncertainty distribution and the host data; and
an object localization module, implemented via the processor, configured to:
encode the proximate data in a proximate vehicle stage;
decode the encoded proximate data including proximate positions or proximate motions of the one or more proximate vehicles in the proximate vehicle stage to generate a second joint uncertainty distribution based on the initial joint uncertainty model and an object prediction model distribution of the host kinematic predictions of the host vehicle as a prior; and
sample proximate kinematic predictions in the proximate vehicle stage based on the second joint uncertainty distribution and the proximate data.