System, method, and components providing compressive active range sampling
View Patent ↗A system, methodologies and components utilizing a random access scanning LiDAR sensor for object detection for use in autonomous and driver assisted transportation vehicles, wherein calculated expected potential range changes determined based on emission of range determining laser pulses are used to schedule subsequent emission of range determining laser pulses such that laser pulse generation and ray sample generation for individual rays with a larger expected potential range change is performed more frequently than laser pulse generation and ray sample generation for individual rays with a smaller expected potential change so as to provide compressive active range scanning.
1 . A system utilizing a random access scanning LiDAR sensor for object detection to implement assistive and/or autonomous driving functionality in a transportation vehicle, the system comprising:
at least one random access scanning LiDAR sensor; and
at least one data processor running a software algorithm stored in memory, the software algorithm being configured to control scheduling of individual rays for emission of range determining laser pulses generated by the at least one random access scanning LiDAR sensor to generate ray samples which are then analyzed to calculate expected potential range changes for the rays,
wherein the calculated expected potential range changes are then used to schedule subsequent emission of range determining laser pulses by the at least one random access scanning LiDAR sensor such that laser pulse generation and ray sample generation for individual rays with a larger expected potential range change is performed more frequently than laser pulse generation and ray sample generation for individual rays with a smaller expected potential change so as to provide compressive active range scanning,
wherein the compressive active range scanning maximizes information delivery by the at least one data processor running the software algorithm to generate ray request packets that request rays where the most change from a previous sampling of that ray is expected, whereby a given ray, as considered in a reference frame of the at least one random access scanning LiDAR sensor, is re-sampled more quickly based on an increased uncertainty of a likely result, thereby indicating an information-rich sample location.
2 . The system of claim 1 , wherein the potential range change from the previous sampling is determined based on radial velocity determined for that ray.
3 . The system of claim 1 , wherein the potential range change is determined based on uncertainty of a sampling having a specific result.
4 . The system of claim 1 , wherein the at least one random access scanning LiDAR sensor includes a micro-electro-mechanical system mirror to direct a laser pulse allow the beam to be steered in a complex path.
5 . The system of claim 1 , wherein the at least one random access scanning LiDAR sensor includes a laser emitter comprised of an array of micro-lasers.
6 . The system of claim 1 , wherein the at least one LiDAR sensor is configured to perform horizontal random-access scanning.
7 . The system of claim 1 , wherein the at least one LiDAR sensor is configured to perform vertical random-access scanning.
8 . The system of claim 1 , wherein the at least one LiDAR sensor is configured to provide dynamically configurable scan patterns.
9 . The system of claim 1 , wherein, on a periodic basis, the potential range change is computed for all rays in a grid, and the rays with the highest potential range change values are requested for the at least one LiDAR sensor to sample sooner than the other rays in the grid.
10 . The system of claim 1 , wherein range samples are returned to the at least one processor in User Datagram Protocol packets.
11 . The system of claim 1 , wherein the potential range change and x- and y-coordinates of the samples are saved in the memory in a sortable list, which is then used to schedule subsequent individual rays for emitting range determining laser pulses generated by the at least one random access scanning LiDAR sensor to generate ray samples.
12 . The system of claim 1 , wherein lateral motion across a frame of samples is detected by the at least one LiDAR by implementing velocity spreading such that a calculated radial velocity for a sampled point along a sampling ray is spread and saved for neighboring sampling rays such that potential range change in the neighboring sampling rays may be identified to indicate potential range change in a neighborhood of the sampled rays thereby triggering sampling of the neighborhood of the sampled rays to increase the probability of finding lateral movement across the frame of samples.
13 . A method for performing random access scanning using at least one LiDAR sensor for object detection to implement assistive and/or autonomous driving functionality in a transportation vehicle, the method comprising:
scheduling individual rays for emitting range determining laser pulses by at least one random access scanning LiDAR sensor by at least one data processor running a software algorithm stored in memory and generating ray request packets;
emitting the range determining laser pulses by the at least one random access scanning LiDAR;
analyzing data sensed in response to the emitted range determining laser pulses to generate ray samples;
analyzing the generated ray samples to determine expected potential range changes; and
using the expected potential range changes to schedule subsequent emission of range determining laser pulses by the at least one random access scanning LiDAR sensor such that laser pulse generation and ray sample generation for individual rays with a larger expected potential range change is performed more frequently than laser pulse generation and ray sample generation for individual rays with a smaller expected potential change so as to provide compressive active range scanning,
wherein the compressive active range scanning maximizes information delivery by at least one data processor running the software algorithm to generate ray request packets that request rays where the most change from a previous sampling of that ray is expected, whereby a given ray, as considered in a reference frame of the at least one random access scanning LiDAR sensor, is re-sampled more quickly based on an increased uncertainty of a likely result, thereby indicating an information-rich sample location.
14 . The method of claim 13 , wherein the potential range change from the previous sampling is determined based on radial velocity determined for that ray.
15 . The method of claim 13 , wherein the potential range change is determined based on uncertainty of a sampling having a specific result.
16 . The method of claim 13 , wherein the at least one LiDAR sensor is configured to perform horizontal random-access scanning.
17 . The method of claim 13 , wherein the at least one LiDAR sensor is configured to perform vertical random-access scanning.
18 . The method of claim 13 , wherein the at least one LiDAR sensor is configured to provide dynamically configurable scan patterns.
19 . The method of claim 13 , wherein, on a periodic basis, the potential range change is computed for all rays in a grid, and the rays with the highest potential range change values are requested for the at least one LiDAR sensor to sample sooner than the other rays in the grid.
20 . The method of claim 13 , wherein range samples are returned to the at least one processor in User Datagram Protocol packets.
21 . The method of claim 13 , wherein the potential range change and x- and y-coordinates of the samples are saved in the memory in a sortable list, which is then used to schedule subsequent individual rays for emitting range determining laser pulses generated by the at least one random access scanning LiDAR sensor to generate ray samples.
22 . The method of claim 13 , wherein lateral motion across a frame of samples is detected by the at least one LiDAR by implementing velocity spreading such that a calculated radial velocity for a sampled point along a sampling array is spread and saved for neighboring sampling rays such that potential range change in the neighboring sampling rays may be identified to indicate potential range change in a neighborhood of the sampled rays thereby triggering sampling of the neighborhood of the sampled rays to increase the probability of finding lateral movement across the frame of samples.
23 . A non-transitory computer readable medium including software instructions which, when performed on at least one computer processor, perform a method for performing random access scanning using at least one LiDAR sensor for object detection to implement assistive and/or autonomous driving functionality in a transportation vehicle, the method comprising:
scheduling individual rays for emitting range determining laser pulses by at least one random access scanning LiDAR sensor by at least one data processor running a software algorithm stored in memory and generating ray request packets;
emitting the range determining laser pulses by the at least one random access scanning LiDAR;
analyzing data sensed in response to the emitted range determining laser pulses to generate ray samples;
analyzing the generated ray samples to determine expected potential range changes; and
using the expected potential range changes to schedule subsequent emission of range determining laser pulses by the at least one random access scanning LiDAR sensor such that laser pulse generation and ray sample generation for individual rays with a larger expected potential range change is performed more frequently than laser pulse generation and ray sample generation for individual rays with a smaller expected potential change so as to provide compressive active range scanning,
wherein the compressive active range scanning maximizes information delivery by at least one data processor running the software algorithm to generate ray request packets that request rays where the most change from a previous sampling of that ray is expected, whereby a given ray, as considered in a reference frame of the at least one random access scanning LiDAR sensor, is re-sampled more quickly based on an increased uncertainty of a likely result, thereby indicating an information-rich sample location.