US 11,854,252 B1
Automated probabilistic axiom generation and incremental updates
Hyukseong Kwon, Thousand Oaks, CA (US); Amit Agarwal, San Francisco, CA (US); Amir M. Rahimi, Santa Monica, CA (US); Kevin Lee, Irvine, CA (US); Alexie Pogue, Los Angeles, CA (US); and Rajan Bhattacharyya, Sherman Oaks, CA (US)
Assigned to HRL LABORATORIES, LLC, Malibu, CA (US)
Filed by HRL Laboratories, LLC, Malibu, CA (US)
Filed on Mar. 22, 2022, as Appl. No. 17/700,802.
Application 17/700,802 is a continuation in part of application No. 17/133,345, filed on Dec. 23, 2020, granted, now 11,350,039.
Application 17/133,345 is a continuation in part of application No. 17/030,354, filed on Sep. 23, 2020, granted, now 11,334,767.
Application 17/133,345 is a continuation in part of application No. 17/700,802.
Application 17/700,802 is a continuation in part of application No. 17/030,354, filed on Sep. 23, 2020, granted, now 11,334,767.
Claims priority of provisional application 63/188,364, filed on May 13, 2021.
Claims priority of provisional application 62/984,713, filed on Mar. 3, 2020.
Claims priority of provisional application 62/905,059, filed on Sep. 24, 2019.
Int. Cl. G06K 9/62 (2022.01); G06V 10/98 (2022.01); G06V 20/56 (2022.01); G06V 10/762 (2022.01); G06V 10/764 (2022.01); G06V 10/10 (2022.01)
CPC G06V 10/993 (2022.01) [G06V 10/10 (2022.01); G06V 10/763 (2022.01); G06V 10/764 (2022.01); G06V 20/56 (2022.01)] 15 Claims
OG exemplary drawing
 
1. A system to evaluate and correct perception errors in object detection and recognition, the system comprising:
one or more processors and a non-transitory computer-readable medium having executable instructions encoded thereon such that when executed, the one or more processors perform operations of:
receiving, with a perception module, perception data from an environment proximate a mobile platform, the perception data reflecting one or more objects in the environment;
generating a plurality of perception probes from the perception data, wherein the plurality of perception probes describe perception characteristics of one or more object detections in the set of perception data;
for each perception probe, determining probabilistic distributions for true positive values and false positive values, resulting in true positive perception probes and false negative perception probes;
determining statistical characteristics of true positive perception probes and false positive perception probes;
based on the statistical characteristics, clustering true positive perception probes;
generating an axiom to determine statistical constraints for perception validity for each perception probe cluster;
evaluating the axiom to classify the plurality of perception probes as valid or erroneous;
generating optimal perception parameters by solving an optimization problem based on the axiom; and
adjusting the perception module based on the optimal perception parameters.