Systems and methods for generating and deforming a spatial template to identify landmarks
AssigneeToyota Jidosha Kabushiki Kaisha
InventorEitan Gilad Mendelowitz
Life of this patent
Estimating the term…
Abstract
Systems, methods, and other embodiments described herein relate to identifying landmarks by generating and deforming a spatial template using trace data from multiple vehicles. In one embodiment, a method includes forming a trace dataset for multiple vehicles from locations of landmarks and vehicles identified with sensor data by a learning model, the sensor data associated with the multiple vehicles. The method also includes generating a spatial template by a vehicle using the trace dataset having missing data. The method also includes aligning the landmarks for positioning by deforming the spatial template using a spring model that fills the missing data.
Claims20
An identification system comprising:
- a memory storing instructions that, when executed by a processor, cause the processor to:
- form a trace dataset for multiple vehicles from locations of landmarks and vehicles identified with sensor data by a learning model, the sensor data associated with the multiple vehicles;
- generate a spatial template by a vehicle using the trace dataset having missing data;
- and
- align the landmarks for positioning by deforming the spatial template using a spring model that fills the missing data.
A non-transitory computer-readable medium comprising:
- instructions that when executed by a processor cause the processor to:
- form a trace dataset for multiple vehicles from locations of landmarks and vehicles identified with sensor data by a learning model, the sensor data associated with the multiple vehicles;
- generate a spatial template by a vehicle using the trace dataset having missing data;
- and
- align the landmarks for positioning by deforming the spatial template using a spring model that fills the missing data.
A method comprising:
- forming a trace dataset for multiple vehicles from locations of landmarks and vehicles identified with sensor data by a learning model, the sensor data associated with the multiple vehicles;
- generating a spatial template by a vehicle using the trace dataset having missing data;
- and
- aligning the landmarks for positioning by deforming the spatial template using a spring model that fills the missing data.
Dependencies are inferred from the claim text.