Multi-frequency radar array systems and sensor fusion for seeing around corners in autonomous driving
Disclosed herein are systems, devices, and methods that may be used for autonomous driving and/or in autonomous vehicles. Some embodiments use an integrated wide-aperture multi-band radar subsystem and leverage the unique propagation properties of multiple bands and/or multiple sensor technologies to significantly improve detection and understanding of the scenery and, in particular, to see around corners to identify non-line-of-sight targets. In some embodiments, at least one processor of the system is capable of jointly processing return (reflected) signals in multiple bands to provide high accuracy in a variety of conditions (e.g., weather). The disclosed radar subsystem can be used alone or in conjunction with another sensing technology, such as, for example, LiDAR and/or cameras.
1 . A method of identifying positions of targets in a scene, the method comprising:
transmitting a plurality of scanning signals;
collecting a plurality of reflected signals;
determining a respective classification for each signal of at least a subset of the plurality of reflected signals, wherein at least one signal of the subset is classified as a through-building return; and
based at least in part on (a) the at least the subset of the plurality of reflected signals, and (b) the respective classification for each signal of the at least the subset of the plurality of reflected signals, determining a projected position of a target in the scene.
2 . The method recited in claim 1 , wherein the respective classification each signal of the subset is one and only one of: line-of-sight, multipath, or through-building.
3 . The method recited in claim 1 , wherein determining the respective classification for each signal of the at least the subset of the plurality of reflected signals comprises determining, for a particular reflected signal of the at least the subset of the plurality of reflected signals, one or more of:
(i) a frequency of bands in which another reflected signal having a substantially similar time of flight and/or angle of arrival to the particular reflected signal is observed,
(ii) a number of bands in which the another reflected signal having the substantially similar time of flight and/or angle of arrival to the particular reflected signal is observed,
(iii) whether a strength of the particular reflected signal exceeds a threshold,
(iv) a rate of change over time of an angle of arrival and/or a time of flight of the particular reflected signal, or
(v) a polarization of the particular reflected signal.
4 . The method recited in claim 1 , wherein determining the projected position of the target is further based on information about at least one stationary object in the scene.
5 . The method recited in claim 4 , wherein the information about the at least one stationary object in the scene comprises at least one of: a location of the at least one stationary object, a position of the at least one stationary object, an orientation of the at least one stationary object, or a material of the at least one stationary object.
6 . The method recited in claim 4 , wherein the at least one stationary object includes a building.
7 . The method recited in claim 1 , wherein the plurality of scanning signals are radar signals.
8 . The method recited in claim 1 , wherein collecting the plurality of reflected signals comprises receiving the plurality of reflected signals.
9 . The method recited in claim 1 , wherein the projected position represents a range and an angle.
10 . The method recited in claim 1 , wherein determining the projected position of the target comprises:
determining a plurality of probabilities for a plurality of candidate positions of the target, each of the plurality of probabilities corresponding to a respective one of the plurality of candidate positions, each of the plurality of probabilities representing a likelihood that the target is in the respective one of the plurality of candidate positions,
identifying a largest one of the plurality of probabilities, and
choosing, as the projected position of the target, a particular one of the plurality of candidate positions that corresponds to the largest one of the plurality of probabilities.
11 . The method recited in claim 10 , wherein the at least the subset of the plurality of reflected signals is a first at least the subset of the plurality of reflected signals, wherein each of the first at least the subset of the plurality of reflected signals corresponds to a first time or a first time interval, and further comprising:
updating the plurality of probabilities based at least in part on a second at least the subset of the plurality of reflected signals, wherein each of the second at least the subset of the plurality of reflected signals corresponds to a second time or a second time interval.
12 . The method recited in claim 1 , wherein a first reflected signal of the plurality of reflected signals comprises a reflection of a first radar signal transmitted in a first frequency band, and a second reflected signal of the plurality of reflected signals comprises a reflection of a second radar signal transmitted in a second frequency band, wherein the first and second frequency bands are disjoint.
13 . The method recited in claim 12 , wherein the first frequency band is a L, S, W, X, C, K, Ka, Ku, or UHF band.
14 . The method recited in claim 12 , wherein the first frequency band is a C or K band, and the second frequency band is a UHF band.
15 . A system for identifying positions of targets in a scene, the system comprising:
a radar subsystem configured to transmit a plurality of scanning signals and collect a plurality of reflected signals; and
at least one processor coupled to the radar subsystem and configured to execute at least one machine-executable instruction that, when executed, causes the at least one processor to:
determine a respective classification for each signal of at least a subset of the plurality of reflected signals, wherein at least one signal of the subset is classified as a through-building return, and
determine a projected position of a target in the scene based at least in part on (i) the at least the subset of the plurality of reflected signals, and (ii) the respective classification for each signal of the at least the subset of the plurality of reflected signals.
16 . The system recited in claim 15 , wherein the radar subsystem is capable of transmitting a first subset of the plurality of scanning signals in a first radar band during a first time period and transmitting a second subset of the plurality of scanning signals in a second radar band during the first time period, wherein the first radar band differs from the second radar band.
17 . The system recited in claim 16 , wherein the radar subsystem comprises a plurality of transmitters and a plurality of receivers situated on and/or within a body of a vehicle.
18 . The system recited in claim 15 , wherein the radar subsystem comprises a sparse array, the sparse array comprising a plurality of transmit elements and a plurality of receive elements.
19 . The system recited in claim 18 , wherein the plurality of transmit elements and the plurality of receive elements are distributed over and/or within a body of a vehicle.
20 . The system recited in claim 18 , wherein the sparse array is three-dimensional.
21 . The system recited in claim 18 , wherein the sparse array is nonuniformly distributed.
22 . The system recited in claim 15 , wherein the plurality of reflected signals is a first plurality of reflected signals, and wherein the system further comprises:
a light detection and ranging (LiDAR) subsystem configured to transmit a plurality of optical signals and collect a second plurality of reflected signals,
and wherein, when executed, the at least one machine-executable instruction causes the at least one processor to determine the projected position of the target further based on the second plurality of reflected signals.
23 . A vehicle comprising the system recited in claim 15 .