Autonomous navigation technology
Methods, systems, and apparatus, including computer programs encoded on computer -storage media, for swarming technology. In some implementations, the system can be modeled on the spatial reasoning process found in rodents and other mammals. A plurality of synaptic weights can be formed within a neural network. The neural network along with additional data of a system can be used to determine motion vectors for one or more devices within a group of devices or swarm. The motion vectors can help determine where the devices within a group of devices are located at a given time. Motion vectors can direct devices to areas of high activity in a similar manner to how spatial cells in a brain direct animals in accordance with high activity rates of particular cells within their brain.
1. A method for decentralized robot swarm navigation, the method comprising:
for each particular robot of the swarm:
determining a plurality of robot-based synaptic weights within a neural network, wherein each of the plurality of robot-based synaptic weights represents a distance based relationship between the particular robot and another robot of the swarm that is visible to the particular robot;
determining a plurality of objective-based synaptic weights within the neural network, wherein each of the objective-based synaptic weights represent a distance based relationship between the particular robot and an objective;
updating the plurality of robot-based synaptic weights and objective-based synaptic weights using the neural network and additional data, wherein the additional data is comprised of oscillating signals from at least one robot;
determining a motion vector that defines movement for the particular robot based on the updated weights; and
adjusting a current location of the particular robot to a subsequent location based on the determined motion vector.
2. The method of claim 1 , wherein the other robot is visible to the particular robot of the robot swarm if electronic signals can be exchanged between the other robot and the particular robot.
3. The method of claim 1 , wherein a robot-based synaptic weight of the plurality of robot -based synaptic weights that represents a connection between the other robot and the particular robot can be added in instances where the other robot moves from being non-visible to the particular robot to visible to the particular robot and a robot-based synaptic weight of the plurality of robot based synaptic weights can be deleted in instances where the other robot moves from being visible to the particular robot to non-visible to the particular robot.
4. The method of claim 1 , wherein the plurality of robot-based synaptic weights are stored within the particular robot of the swarm.
5. The method of claim 1 , wherein a location independent relationship is used in determining the plurality of synaptic weights.
6. The method of claim 1 , the method further comprising:
determining, by the particular robot and using the determined motion vector, a path towards a most highly activated visible particle in a model that is internal to or computed on behalf of the particular robot.
7. The method of claim 1 , wherein the particular robot receives data from an entity outside of the robot swarm and uses the data in determining motion.
8. A system for decentralized robot swarm navigation comprising:
one or more computers and one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to perform operations comprising:
for each particular robot of the swarm:
determining a plurality of robot-based synaptic weights within a neural network, wherein each of the plurality of robot-based synaptic weights represents a distance based relationship between the particular robot and another robot of the swarm that is visible to the particular robot;
determining a plurality of objective-based synaptic weights within the neural network, wherein each of the objective-based synaptic weights represent a distance based relationship between the particular robot and an objective;
updating the plurality of robot-based synaptic weights and objective-based synaptic weights using the neural network and additional data, wherein the additional data is comprised of oscillating signals from at least one robot;
determining a motion vector that defines movement for the particular robot based on the updated weights; and
adjusting a current location of the particular robot to a subsequent location based on the determined motion vector.
9. The system of claim 8 , wherein the other robot is visible to the particular robot of the robot swarm if electronic signals can be exchanged between the other robot and the particular robot.
10. The system of claim 8 , wherein a robot-based synaptic weight of the plurality of robot -based synaptic weights that represents a connection between the other robot and the particular robot can be added in instances where the other robot moves from being non-visible to the particular robot to visible to the particular robot and a robot-based synaptic weight of the plurality of robot based synaptic weights can be deleted in instances where the other robot moves from being visible to the particular robot to non-visible to the particular robot.
11. The system of claim 8 , wherein the plurality of robot-based synaptic weights are stored within the particular robot of the swarm.
12. The system of claim 8 , wherein the additional data contains oscillating signals that are location independent.
13. The system of claim 8 , wherein a location independent relationship is used in determining the plurality of synaptic weights.
14. The system of claim 8 , the operations further comprising:
determining, by the particular robot and using the determined motion vector, a path towards a most highly activated visible particle in a model that is internal to or computed on behalf of the particular robot.
15. The system of claim 8 , wherein the particular robot receives data from an entity outside of the robot swarm and uses the data in determining motion.
16. A non-transitory computer-readable medium storing software comprising instructions executable by one or more computers which, upon such execution, cause the one or more computers to perform operations comprising:
for each particular robot of the swarm:
determining a plurality of robot-based synaptic weights within a neural network, wherein each of the plurality of robot-based synaptic weights represents a distance based relationship between the particular robot and another robot of the swarm that is visible to the particular robot;
determining a plurality of objective-based synaptic weights within the neural network, wherein each of the objective-based synaptic weights represent a distance based relationship between the particular robot and an objective;
updating the plurality of robot-based synaptic weights and objective-based synaptic weights using the neural network and additional data, wherein the additional data is comprised of oscillating signals from at least one robot;
determining a motion vector that defines movement for the particular robot based on the updated weights; and
adjusting a current location of the particular robot to a subsequent location based on the determined motion vector.
17. The computer-readable medium of claim 16 , wherein a robot-based synaptic weight of the plurality of robot-based synaptic weights that represents a connection between the other robot and the particular robot can be added in instances where the other robot moves from being non -visible to the particular robot to visible to the particular robot and a robot-based synaptic weight of the plurality of robot based synaptic weights can be deleted in instances where the other robot moves from being visible to the particular robot to non-visible to the particular robot.
18. The computer-readable medium of claim 16 , wherein the plurality of robot-based synaptic weights are stored within the particular robot of the swarm.
19. The computer-readable medium of claim 16 , the operations further comprising:
determining, by the particular robot and using the determined motion vector, a path towards a most highly activated visible particle in a model that is internal to or computed on behalf of the particular robot.
20. The computer-readable medium of claim 16 , wherein the particular robot receives data from an entity outside of the robot swarm and uses the data in determining motion.