Adaptive scouting using multi-legged robots
Implementations are described herein for reducing the time and costs associated with the crop scouting in a crop field. In various implementations, a method is implemented using one or more processors, and the method include: operating, based on a type and arrangement of a crop field, a robot to travel along a trajectory through the crop field using a first gait. The robot includes one or more vision sensors. The first gait includes a first repeating cycle of poses of the robot. The method can further include: synchronizing operation of one or more of the vision sensors with one or more poses of the first repeating cycle of poses of the multi-legged robot to capture one or more initial sequences of images depicting one or more points-of-interest of crops growing in the crop field.
1. A method implemented using one or more processors, the method comprising:
operating, based on a type and arrangement of a crop field, a multi-legged robot to travel along a trajectory through the crop field using a first gait, wherein the multi-legged robot includes one or more vision sensors and the first gait includes a first repeating cycle of poses of the multi-legged robot;
synchronizing operation of one or more of the vision sensors with one or more poses of the first repeating cycle of poses of the multi-legged robot to capture one or more initial sequences of images depicting one or more points-of-interest of crops growing in the crop field;
processing the one or more initial sequence of images using one or more phenotypic machine learning models to infer one or more phenotypic traits of the crops;
selecting a second gait based on one or more of the inferred phenotypic traits of the crops, wherein the second gait includes a second repeating cycle of poses of the multi-legged robot that is different than the first repeating cycle of poses of the multi-legged robot; and
operating the multi-legged robot to travel along a portion of the trajectory using the second gait.
2. The method of claim 1 , wherein one or more of the inferred phenotypic traits of the crops includes sizes of the one or more crops.
3. The method of claim 1 , wherein the second gait is selected based on one or more terrain conditions of the crop field detected by one or more sensors of the multi-legged robot.
4. The method of claim 1 , further including synchronizing operation of one or more of the vision sensors with one or more poses of the second repeating cycle of poses of the multi-legged robot to capture one or more alternative sequences of images of points-of-interest of the crops.
5. The method of claim 4 , wherein images of the one or more alternative sequences images have greater spatial resolution than images of the one or more initial sequences of images.
6. The method of claim 4 , wherein images of the one or more alternative sequences images have greater temporal resolution than images of the one or more initial sequences of images.
7. The method of claim 1 , wherein the multi-legged robot travels slower using the second gait than the first gait, thereby increasing a spatial density of images captured using the one or more vision sensors.
8. The method of claim 1 , wherein the first and second gaits have different step heights.
9. The method of claim 1 , wherein the multi-legged robot includes a quadruped robot, the first gait includes a walk mode in which two feet of the quadruped robot contact a ground of the crop field at the same time, and the second gait includes a crawl mode in which three feet of the quadruped robot contact the ground at the same time.
10. The method of claim 1 , wherein the processing and selecting are performed onboard the multi-legged robot.
11. The method of claim 1 , wherein one or more of the vision sensors are coupled to the multi-legged robot via a first manipulator that is rotatable and extendable.
12. The method of claim 1 , wherein one or more of the vision sensors includes a stereo camera, and wherein the method further includes adjusting the stereo camera between a first depth and a second depth different from the first depth based on one or more of the inferred phenotypic traits of the crops.
13. The method of claim 1 , wherein the multi-legged robot includes a pulling manipulator and a cutting manipulator, and the method further includes:
identifying a portion of one or more of the crops to sample based on one or more of the inferred phenotypic traits;
operating a blade of the cutting manipulator to cut the portion of one or more of the crops; and
operating the pulling manipulator to transfer the cut portion to a storage compartment of the multi-legged robot.
14. The method of claim 1 , further including operating a manipulator to move foliage of one or more individual plants to expose previously occluded portions of the one or more individual plants.
15. A method implemented using one or more processors, the method comprising:
operating, based on a type and arrangement of a crop field, a multi-legged robot to travel along a trajectory through the crop field using a first gait, wherein the multi-legged robot includes one or more vision sensors, and wherein the first gait includes a first repeating cycle of poses of the multi-legged robot;
processing one or more initial sequence of images using one or more phenotypic machine learning models to infer one or more phenotypic traits of crops growing in the crop field;
based on one or more of the inferred phenotypic traits, selecting a second gait, wherein the second gait includes a second repeating cycle of poses of the multi-legged robot that is different than the first repeating cycle of poses of the multi-legged robot; and
operating the multi-legged robot to travel along a portion of the trajectory using the second gait.
16. A multi-legged robot comprising one or more processors and memory storing instructions that, in response to execution of the instructions, cause the one or more processors to:
operate, based on a type and arrangement of a crop field, the multi-legged robot to travel along a trajectory through the crop field using a first gait, wherein the multi-legged robot includes one or more vision sensors, and wherein the first gait includes a first repeating cycle of poses of the multi-legged robot;
synchronize operation of one or more of the vision sensors with one or more poses of the first repeating cycle of poses of the multi-legged robot to capture one or more initial sequences of images depicting one or more points-of-interest of crops growing in the crop field;
process the one or more initial sequence of images using one or more phenotypic machine learning models to infer one or more phenotypic traits of the crops;
select a second gait based on one or more of the inferred phenotypic traits of the crops, wherein the second gait includes a second repeating cycle of poses of the multi-legged robot that is different than the first repeating cycle of poses of the multi-legged robot; and
operate the multi-legged robot to travel along a portion of the trajectory using the second gait.
17. The multi-legged robot of claim 16 , wherein one or more of the inferred phenotypic traits of the crops includes sizes of the one or more crops.
18. The multi-legged robot of claim 16 , wherein the second gait is selected based on one or more terrain conditions of the crop field detected by one or more sensors of the multi-legged robot.
19. The multi-legged robot of claim 16 , wherein one of more of the processors are to synchronize operation of one or more of the vision sensors with one or more poses of the second repeating cycle of poses of the multi-legged robot to capture one or more alternative sequences of images of points-of-interest of the crops.
20. The multi-legged robot of claim 19 , wherein images of the one or more alternative sequences images have greater spatial resolution than images of the one or more initial sequences of images.