Particle filter to detect static and dynamic obstacles
Briefly, example methods, apparatuses, and/or articles of manufacture may be implemented to access, from at least one memory device, sensor measurements generated by one or more sensors mounted in a vehicle and to generate an occupancy grid having grid cells that track occupancy, by static obstacles and dynamic obstacles, in an environment of the vehicle based on the accessed sensor measurements. An apparatus may additionally allocate (i) a first set of particles to estimate occupancy of grid cells by the static obstacles and states of the static obstacles and (ii) a second set of particles to estimate occupancy of grid cells by the dynamic obstacles and states of the dynamic obstacles. An apparatus may additionally execute one or more operations to plan motion of the vehicle based on the estimated occupancy of grid cells by the static obstacles and dynamic obstacles.
1. An apparatus, comprising:
a processor coupled to at least one memory device, the processor configured to:
access, from the at least one memory device, sensor measurements generated by one or more sensors mounted in a vehicle;
generate an occupancy grid having grid cells that track occupancy, by static obstacles and dynamic obstacles, in an environment of the vehicle based on the accessed sensor measurements;
adjust an allocation of a first set of particles and a second set of particles based on the accessed sensor measurements to modify operation of a particle filter, which adjusts a movable division that is to separate the first set of particles from the second set of particles based on a ratio of a quantity of the first set of particles to a total quantity of particles, or a ratio of a quantity of the second set of particles, to the total quantity of particles;
allocate (i) the first set of particles to estimate occupancy of grid cells by the static obstacles and states of the static obstacles and (ii) the second set of particles to estimate occupancy of grid cells by the dynamic obstacles and states of the dynamic obstacles; and
execute one or more operations to plan motion of the vehicle based on the estimated occupancy of grid cells by the static obstacles and dynamic obstacles.
2. The apparatus of claim 1 , wherein the processor is further configured to:
allocate a third set of particles as newly spawned particles based on the accessed sensor measurements.
3. The apparatus of claim 1 , wherein the processor is further configured to:
reduce a magnitude of one or more velocity vectors of the first set of particles in relation to a magnitude of one or more velocity vectors of the second set of particles.
4. The apparatus of claim 3 , wherein the processor is further configured to:
predict a future position of at least one of the static obstacles relative to a static reference frame.
5. The apparatus of claim 1 , wherein the processor is further configured to:
resample one or more of the second set of allocated particles to update the estimated states of the dynamic obstacles; and
classify a cell of the occupancy grid as a dynamic cell based on of the second set of allocated particles.
6. The apparatus of claim 5 , wherein the processor is further configured to:
classify the cell of the occupancy grid as the dynamic cell further based on a confidence factor exceeding a threshold value.
7. A method of processing measurements to estimate states of one or more objects, the method being performed by one or more processors and comprising:
accessing, from a memory device, sensor measurements generated by one or more sensors mounted in a vehicle;
generating an occupancy grid having grid cells that track occupancy, by static obstacles and dynamic obstacles, in an environment of the vehicle based on the accessed sensor measurements;
adjusting an allocation of a first set of particles and a second set of particles based on the accessed sensor measurements to modify operation of a particle filter, which adjusts a movable division that is to separate the first set of particles from the second set of particles based on a ratio of a quantity of the first set of particles to a total quantity of particles, or a ratio of a quantity of the second set of particles, to the total quantity of particles;
allocating (i) the first set of particles to estimate occupancy of grid cells by the static obstacles and states of the static obstacles and (ii) the second set of particles to estimate occupancy of grid cells by the dynamic obstacles and states of the dynamic obstacles; and
performing motion planning of the vehicle based on the estimated occupancy of grid cells by the static obstacles and dynamic obstacles.
8. The method of claim 7 , further comprising:
allocating a third set of particles as newly spawned particles based on the accessed sensor measurements.
9. The method of claim 7 , further comprising:
reducing a magnitude of one or more velocity vectors of the first set of particles in relation to a magnitude of one or more velocity vectors of the second set of particles.
10. The method of claim 9 , further comprising:
predicting a future position of at least one of the static obstacles relative to a static reference frame.
11. The method of claim 7 , further comprising:
resampling one or more of the second set of allocated particles to update the estimated states of the dynamic obstacles; and
classifying a cell of the occupancy grid as a dynamic cell based on the resampled one or more of the second set of allocated particles.
12. The method of claim 11 , further comprising:
determining a confidence factor that the dynamic cell exceeds a threshold prior to classifying the cell of the occupancy grid as the dynamic cell.
13. An article, comprising:
a non-transitory storage media having instructions encoded thereon, which, when executed by a computer processor coupled to at least one memory device, cause the computer processor to:
access, from the memory device, sensor measurements generated by one or more sensors mounted in a vehicle;
generate an occupancy grid having grid cells that track occupancy, by static obstacles and dynamic obstacles, in an environment of the vehicle based on the accessed sensor measurements;
adjust an allocation of a first set of particles and a second set of particles based on the accessed sensor measurements to modify operation of a particle filter, which adjusts a movable division that is to separate the first set of particles from the second set of particles based on a ratio of a quantity of the first set of particles to a total quantity of particles, or a ratio of a quantity of the second set of particles, to the total quantity of particles;
allocate (i) the first set of particles to estimate occupancy of grid cells by the static obstacles and states of the static obstacles and (ii) the second set of particles to estimate occupancy of grid cells by the dynamic obstacles and states of the dynamic obstacles; and
execute one or more operations to plan motion of the vehicle based on the estimated occupancy of grid cells by the static obstacles and dynamic obstacles.
14. The article of claim 13 , wherein the encoded instructions are further to cause the computer processor to:
resample one or more of the second set of allocated particles to update the estimated states of the dynamic obstacles; and
classify a cell of the occupancy grid as a dynamic cell based on the resampled one or more of the second set of allocated particles.