Register sets of low-level features without data association

A system for vehicle localization using map data from multiple sources, the system comprising:

• a non-transitory memory configured to store instructions;
• at least one processor configured to execute the instructions to:
• determine a set of vehicle map data generated using vehicle sensors;
• project features of the vehicle map data into discretized feature representation;
• transform the discretized feature representation from a vehicle frame of reference to an external map frame of reference to generate a transformed discretized feature representation;
• register the transformed discretized feature representation with a map feature representation of a set of external map data and determine a cost function associated with the registration;
• repeat the registration for a plurality of transformations of the discretized feature representation and determine a cost function associated with each of the plurality of transformations;
• and
• use the transformation with the lowest cost function for vehicle localization.

A system for generating a cost surface for a discrete representation of three-dimensional (3D) point cloud data, the system comprising:

• a non-transitory memory configured to store instructions;
• at least one processor configured to execute the instructions to:
• determine one or more discrete representations of one or more images of the 3D point cloud data;
• determine normalized cross correlations between one or more feature points of the one or more images of the 3D point cloud data;
• generate one or more rigid body transforms corresponding to the one or more feature points based on the normalized cross correlations;
• and
• determine the cost surface for the one or more rigid body transformation based on elements of the one or more rigid body transformations.

A method for vehicle localization, comprising:

• determining a set of vehicle map data generated using vehicle sensors;
• projecting features of the vehicle map data into discretized feature representation;
• transforming the discretized feature representation from a vehicle frame of reference to an external map frame of reference to generate a transformed discretized feature representation;
• registering the transformed discretized feature representation with a map feature representation of a set of external map data and determine a cost function associated with the registration;
• repeating the registration for a plurality of transformations of the discretized feature representation and determine a cost function associated with each of the plurality of transformations;
• and
• using the transformation with the lowest cost function for vehicle localization.

A method of localization of a subject vehicle, comprising:

• receiving data regarding features in an environment of the subject vehicle;
• generating a feature map using data gathered from one or more of the plurality of sensors;
• generating a feature raster image from the feature map;
• registering a plurality of transformations of the feature raster image with a map feature raster image, to generate a cost function for the registrations of the plurality of transformations, and to determine the transformation with the lowest cost function;
• and
• using the transformation with the lowest cost function to determine localization of the subject vehicle.

A method for generating a cost surface for a discrete representation of three-dimensional (3D) point cloud data, the system comprising:

• determining one or more discrete representations of one or more images of the 3D point cloud data;
• determining normalized cross correlations between one or more feature points of the one or more images of the 3D point cloud data;
• generating one or more rigid body transforms corresponding to the one or more feature points based on the normalized cross correlations;
• and
• determining the cost surface for the one or more rigid body transformation based on elements of the one or more rigid body transformations.