Occupancy verification device and method
An occupancy verification including one or more processors, configured to receive occupancy grid data representing a plurality of cells of an occupancy grid and for each cell of the plurality of cells, a probability of whether the cell is occupied; receive object location data, representing one or more locations of one or more objects identified using a model; map the object location data to one or more cells of the plurality of cells based on the one or more locations of the one or more objects; determine comparison grid data, representing the plurality of cells of the occupancy grid and for each cell of the plurality of cells, a probability of whether the cell is occupied based on whether the object location data are mapped to the cell; and compare the occupancy grid data to the comparison grid data using at least one comparing rule.
1. An occupancy grid data verification device, comprising: an artificial neural network, configured to receive sensor data and to output object location data using the sensor data, wherein the object location data represent one or more locations of one or more objects as determined by the artificial neural network; one or more processors, configured to:
receive occupancy grid data, representing a plurality of cells of an occupancy grid for an autonomous driving system and for each cell of the plurality of cells, receive data representing a probability of whether the cell is occupied;
map the object location data to one or more cells of the plurality of cells of the occupancy grid based on the one or more locations of the one or more objects;
determine comparison grid data, wherein determining the comparison grid data comprises determining an occupancy status for each cell of the plurality of cells based on whether the object location data are mapped to the cell;
compare the occupancy grid data to the comparison grid data using at least one comparing rule;
output trust data representing a result of the comparison;
wherein the occupancy grid data further represent velocities associated with one or more cells of the plurality of cells; and wherein the object location data further represent velocities of the one or more objects; and wherein the one or more processors are further configured to:
determine velocity comparison grid data, wherein determining the velocity comparison grid data comprises determining a velocity associated with each cell of the plurality of cells based on a velocity of an object from the object location data that is mapped to the cell;
compare the velocities of the occupancy grid data with the velocity comparison grid data based on at least one velocity comparing rule; and
output velocity trust data representing a result of the comparison.
2. The occupancy verification device of claim 1 , wherein the at least one comparing rule comprises determining a distance between the occupancy grid data and the comparison grid data in accordance with a pre-defined distance metric.
3. The occupancy verification device of claim 2 , wherein the at least one comparing rule comprises determining a Bhattacharyya distance between the occupancy grid data and the comparison grid data.
4. The occupancy verification device of claim 1 , wherein determining the comparison grid data comprises determining the occupancy status as occupied if one or more objects represented by the object location data are mapped to the cell and determining the occupancy status as non-occupied if none of the one or more objects represented by the object location data are mapped to the cell.
5. The occupancy verification device of claim 1 , wherein the trust data comprise a tag representing an instruction to deactivate the object, an instruction to increase an occupancy probability of one or more cells to which the object location data were mapped, an instruction to define one or more cells to which the object location data were mapped as being occupied, or an instruction not to consider an object of the one or more objects in a driving decision if a result of the comparison is outside of a predetermined range.
6. The occupancy verification device of claim 1 , wherein the locations of the one or more objects are locations relative to an ego vehicle.
7. The occupancy verification device of claim 1 , wherein the one or more processors receive the object location data from a neural network.
8. The occupancy verification device of claim 1 , wherein the at least one velocity comparing rule comprises determining a Bhattacharyya distance between the velocities of the occupancy grid data and the velocity comparison grid data.
9. A non-transitory computer readable medium, comprising instructions which, if executed, cause one or more processors to:
receive occupancy grid data, representing a plurality of cells of an occupancy grid for an autonomous driving system and for each cell of the plurality of cells, receive data representing a probability of whether the cell is occupied;
implement an artificial neural network to generate object location data based on sensor data, wherein the object location data represent one or more locations of one or more objects identified using a model;
map the object location data to one or more cells of the plurality of cells of the occupancy grid based on the one or more locations of the one or more objects;
determine comparison grid data, wherein determining the comparison grid data comprises determining an occupancy status for each cell of the plurality of cells based on whether the object location data are mapped to the cell;
compare the occupancy grid data to the comparison grid data using at least one comparing rule; and
output trust data representing a result of the comparison;
wherein the occupancy grid data further represent velocities associated with one or more cells of the plurality of cells; and wherein the object location data further represent velocities of the one or more objects; further comprising:
determining velocity comparison grid data, wherein determining the velocity
comparison grid data comprises determining a velocity associated with each cell of the plurality of cells based on a velocity of an object from the object location data that is mapped to the cell;
comparing the velocities of the occupancy grid data with the velocity comparison grid data based on at least one velocity comparing rule; and
outputting velocity trust data representing a result of the comparison.
10. The non-transitory computer readable medium of claim 9 , wherein the at least one comparing rule comprises determining a Bhattacharyya distance between the occupancy grid data and the comparison grid data.
11. The non-transitory computer readable medium of claim 9 , wherein the trust data comprise a tag representing an instruction to deactivate the object, an instruction to increase an occupancy probability of one or more cells to which the object location data were mapped, an instruction to define one or more cells to which the object location data were mapped as being occupied, or an instruction not to consider an object of the one or more objects in a driving decision if a result of the comparison is outside of a predetermined range.
12. The occupancy grid data verification device of claim 1 , wherein at least one cell of the plurality of cells comprises a stored velocity value for an object corresponding to the cell; and
wherein the processor being configured to determine the velocity comparison grid data comprises receiving a velocity value associated with an object detected by the artificial neural network and comparing this velocity value with the stored velocity value in a corresponding cell of the occupancy grid.
13. The occupancy grid data verification device of claim 1 wherein the processor receiving data representing a probability of whether the cell is occupied comprises the processor receiving data about whether the cell is currently occupied.