Method and system for radar-based odometry
OwnerMedhat Omr, Jacques Georgy +1inventors
InventorsMedhat Omr, Jacques Georgy, Billy Cheuk Wai Chan
Life of this patent
Estimating the term…
Abstract
An odometry solution for a device within a moving platform is provided using a deep neural network. Radar measurements may be obtained, such that static objects are detected based at least in part on the obtained radar measurements. Odometry information for the platform is estimated based at least in part on the detected static objects and the obtained radar measurements.
Claims19
A method for providing an odometry solution for a device within a moving platform using a deep neural network, the method comprising:
- a) obtaining radar measurements from at least one radar of the platform;
- b) detecting static objects based at least in part on the obtained radar measurements;
- and
- c) estimating odometry information for the platform based at least in part on the detected static objects and the obtained radar measurements,
- wherein at least one of step b) and c) employs the deep neural network and is based at least in part on signal strengths of the radar measurements and wherein signal strengths for the radar measurements comprise a strength image.
A system for providing an odometry solution for a device within a moving platform using a deep neural network, comprising at least one radar of the platform configured to output radar measurements and at least one processor, coupled to receive the radar measurements and operative to:
- i) detect static objects based at least in part on the radar measurements;
- and
- ii) estimate odometry information for the platform based at least in part on the detected static objects and the radar measurements,
- wherein at least one of i) and ii) employs the deep neural network and is based at least in part on signal strengths of the radar measurements and wherein signal strengths for the radar measurements comprise a strength image.
Dependencies are inferred from the claim text.