Adaptive sensor position determination for multiple mobile sensors
Techniques are provided for adaptive sensor position determination for multiple mobile sensors. One method comprises obtaining a spatio-temporal representation of sensor measurements, from multiple mobile sensors, wherein the spatio-temporal representation comprises multiple layers each corresponding to a different point in time, wherein a given layer comprises multiple positions, and wherein each position in the given layer corresponds to a possible location for at least one of the multiple mobile sensors in an environment; applying the spatio-temporal representation to an environment state prediction model that generates a prediction of at least one future sensor measurement value for multiple positions in the spatio-temporal representation; applying the predictions of the at least one future sensor measurement value to a sensor position determination model that determines a new position for each of one or more of the multiple mobile sensors; and initiating a movement of the one or more of the multiple mobile sensors to the new position.
1. A method, comprising:
obtaining a spatio-temporal representation of sensor measurements, from a plurality of mobile sensors, wherein the spatio-temporal representation comprises a plurality of layers each corresponding to a different point in time, wherein a given layer comprises a plurality of positions, and wherein each position in the given layer of the spatio-temporal representation corresponds to a possible location for at least one of the plurality of mobile sensors in an environment;
applying the spatio-temporal representation to an environment state prediction model that generates a prediction of at least one future sensor measurement value for a plurality of positions in the spatio-temporal representation;
applying the predictions of the at least one future sensor measurement value to a sensor position determination model that determines at least one new position for each of one or more of the plurality of mobile sensors using the predictions; and
initiating a movement of the one or more of the plurality of mobile sensors to the new position;
wherein the method is performed by at least one processing device comprising a processor coupled to a memory.
2. The method of claim 1 , wherein the environment state prediction model comprises a graph neural network.
3. The method of claim 1 , wherein the sensor position determination model comprises a reinforcement learning model.
4. The method of claim 3 , wherein the reinforcement learning model determines the new position for the at least some of the plurality of mobile sensors based at least in part on positions having a higher uncertainty value relative to other positions.
5. The method of claim 3 , wherein the reinforcement learning model determines the new position for a given mobile sensor based at least in part on a proximity of the given mobile sensor to the new position.
6. The method of claim 3 , wherein the reinforcement learning model determines the new position for a given mobile sensor based at least in part on an energy cost associated with moving the given mobile sensor to the new position.
7. The method of claim 1 , wherein a number of the positions in a plurality of the layers of the spatio-temporal representation is greater than a number of the plurality of mobile sensors.
8. The method of claim 1 , wherein the plurality of mobile sensors comprises a coordinated group of sensors.
9. The method of claim 1 , wherein the environment state prediction model further generates an uncertainty value for each prediction indicating a confidence of the environment state prediction model in a corresponding prediction and further comprising applying the uncertainty value for each prediction to the sensor position determination model that determines the new position for at least some of the plurality of mobile sensors.
10. An apparatus comprising:
at least one processing device comprising a processor coupled to a memory;
the at least one processing device being configured to implement the following steps:
obtaining a spatio-temporal representation of sensor measurements, from a plurality of mobile sensors, wherein the spatio-temporal representation comprises a plurality of layers each corresponding to a different point in time, wherein a given layer comprises a plurality of positions, and wherein each position in the given layer of the spatio-temporal representation corresponds to a possible location for at least one of the plurality of mobile sensors in an environment;
applying the spatio-temporal representation to an environment state prediction model that generates a prediction of at least one future sensor measurement value for a plurality of positions in the spatio-temporal representation;
applying the predictions of the at least one future sensor measurement value to a sensor position determination model that determines at least one new position for each of one or more of the plurality of mobile sensors using the predictions; and
initiating a movement of the one or more of the plurality of mobile sensors to the new position.
11. The apparatus of claim 10 , wherein the environment state prediction model comprises a graph neural network and the sensor position determination model comprises a reinforcement learning model.
12. The apparatus of claim 11 , wherein the reinforcement learning model determines the new position for the at least some of the plurality of mobile sensors based at least in part on positions having a higher uncertainty value relative to other positions.
13. The apparatus of claim 11 , wherein the reinforcement learning model determines the new position for a given mobile sensor based at least in part on one or more of a proximity of the given mobile sensor to the new position and an energy cost associated with moving the given mobile sensor to the new position.
14. The apparatus of claim 10 , wherein a number of the positions in a plurality of the layers of the spatio-temporal representation is greater than a number of the plurality of mobile sensors.
15. The apparatus of claim 10 , wherein the environment state prediction model further generates an uncertainty value for each prediction indicating a confidence of the environment state prediction model in a corresponding prediction and further comprising applying the uncertainty value for each prediction to the sensor position determination model that determines the new position for at least some of the plurality of mobile sensors.
16. A non-transitory processor-readable storage medium having stored therein program code of one or more software programs, wherein the program code when executed by at least one processing device causes the at least one processing device to perform the following steps:
obtaining a spatio-temporal representation of sensor measurements, from a plurality of mobile sensors, wherein the spatio-temporal representation comprises a plurality of layers each corresponding to a different point in time, wherein a given layer comprises a plurality of positions, and wherein each position in the given layer of the spatio-temporal representation corresponds to a possible location for at least one of the plurality of mobile sensors in an environment;
applying the spatio-temporal representation to an environment state prediction model that generates a prediction of at least one future sensor measurement value for a plurality of positions in the spatio-temporal representation;
applying the predictions of the at least one future sensor measurement value to a sensor position determination model that determines at least one new position for each of one or more of the plurality of mobile sensors using the predictions; and
initiating a movement of the one or more of the plurality of mobile sensors to the new position.
17. The non-transitory processor-readable storage medium of claim 16 , wherein the environment state prediction model comprises a graph neural network and the sensor position determination model comprises a reinforcement learning model.
18. The non-transitory processor-readable storage medium of claim 17 , wherein the reinforcement learning model determines the new position for the at least some of the plurality of mobile sensors based at least in part on positions having a higher uncertainty value relative to other positions.
19. The non-transitory processor-readable storage medium of claim 17 , wherein the reinforcement learning model determines the new position for a given mobile sensor based at least in part on one or more of a proximity of the given mobile sensor to the new position and an energy cost associated with moving the given mobile sensor to the new position.
20. The non-transitory processor-readable storage medium of claim 16 , wherein the environment state prediction model further generates an uncertainty value for each prediction indicating a confidence of the environment state prediction model in a corresponding prediction and further comprising applying the uncertainty value for each prediction to the sensor position determination model that determines the new position for at least some of the plurality of mobile sensors.