Time distribution scheme for wireless mesh networks
A wireless mesh network includes a mesh of continuously-powered devices (CPDs) and a mesh of battery-powered devices (BPDs). The BPDs are organized into hop layers based on hopping distance to the mesh of CPDs. The CPDs transmit time beacons to BPDs in a first hop layer during a first receive window associated with the first hop layer. The BPDs in the first hop layer then transmit time beacons to BPDs in a second hop layer during a second receive window. In this manner, the wireless mesh network propagates time values throughout the BPD mesh. Based on these time values, the BPDs power on during short time intervals to exchange data with neighboring BPDs, and then power off for longer time intervals, thereby conserving battery power. The techniques described herein for conserving battery power for BPDs may also be applied to conserve power consumption of CPDs.
1. A computer-implemented method for communicating time estimates between nodes in a wireless mesh network, the method comprising:
receiving, from a first uplink node, a first time estimate and a first uncertainty estimate corresponding to the first time estimate;
receiving, from a second uplink node that is different than the first uplink node, a second time estimate and a second uncertainty estimate corresponding to the second time estimate;
combining the first time estimate and the second time estimate based on the first uncertainty estimate and the second uncertainty estimate to generate a third time estimate;
generating a first time beacon based on the third time estimate; and
transmitting the first time beacon from a first node in the wireless mesh network to a second node in the wireless mesh network.
2. The computer-implemented method of claim 1 , wherein combining the first time estimate and the second time estimate comprises:
weighting the first time estimate based on the first uncertainty estimate to generate a first weighted time estimate;
weighting the second time estimate based on the second uncertainty estimate to generate a second weighted time estimate; and
averaging the first weighted time estimate and the second weighted time estimate to generate the third time estimate.
3. The computer-implemented method of claim 1 , further comprising combining the first uncertainty estimate with the second uncertainty estimate to generate a third uncertainty estimate that corresponds to the third time estimate.
4. The computer-implemented method of claim 3 , further comprising scaling an interval of time during which the second node is powered on and receiving data based on the third uncertainty estimate.
5. The computer-implemented method of claim 3 , wherein generating the first time beacon based on the third time estimate further comprises:
determining, based on the third time estimate, a fourth time estimate indicating a time at which the first time beacon is to be transmitted; and
determining, based on the third uncertainty estimate and a predicted clock drift of an harmonic oscillator, a fourth uncertainty estimate that corresponds to the fourth time estimate,
wherein the fourth time estimate and the fourth uncertainty estimate are included in the first time beacon.
6. The computer-implemented method of claim 1 , further comprising:
updating a first clock included in the first node based on the third time estimate; and
updating a second clock included in the second node based on the first time beacon.
7. The computer-implemented method of claim 1 , further comprising:
receiving a fourth time estimate and a fourth uncertainty estimate corresponding to the fourth time estimate;
determining that the fourth uncertainty estimate meets a threshold criterion; and
in response, discarding the fourth time estimate.
8. The computer-implemented method of claim 1 , wherein the first time estimate and the second time estimate are received from one or more battery-powered nodes that reside N hops away from a continuously powered device coupled to the wireless mesh network, wherein N comprises a positive integer.
9. The computer-implemented method of claim 8 , wherein the second node resides N+2 hops away from the continuously powered device and wherein the second node comprises a battery-powered node.
10. The computer-implemented method of claim 9 , wherein the one or more battery-powered nodes receive data during a first receive window but not during a second receive window, and the second node receives data during the second receive window but not the first receive window.
11. A first node, comprising:
a memory storing program instructions; and
a processor coupled to the memory and configured to:
receive, from a first uplink node, a first time estimate and a first uncertainty estimate corresponding to the first time estimate,
receive, from a second uplink node that is different than the first uplink node, a second time estimate and a second uncertainty estimate corresponding to the second time estimate,
combine the first time estimate and the second time estimate based on the first uncertainty estimate and the second uncertainty estimate to generate a third time estimate,
generate a first time beacon based on the third time estimate, and
transmit the first time beacon to a second node in the wireless mesh network.
12. The first node of claim 11 , wherein the processor, upon executing the program instructions, is configured to:
receive the first time estimate and the first uncertainty estimate;
receive the second time estimate and the second uncertainty estimate;
combine the first time estimate and the second time estimate;
generate the first time beacon; and
transmit the first time beacon to the second node.
13. The first node of claim 11 , wherein the processor is further configured to:
weight the first time estimate based on the first uncertainty estimate to generate a first weighted time estimate;
weight the second time estimate based on the second uncertainty estimate to generate a second weighted time estimate; and
average the first weighted time estimate and the second weighted time estimate to generate the third time estimate.
14. The first node of claim 11 , wherein the processor is further configured to combine the first uncertainty estimate with the second uncertainty estimate to generate a third uncertainty estimate that corresponds to the third time estimate.
15. The first node of claim 11 , wherein the processor is configured to generate the first time beacon based on the third time estimate by:
determining, based on the third time estimate, a fourth time estimate indicating a time at which the first time beacon is to be transmitted; and
determining, based on the third uncertainty estimate and a predicted clock drift of an harmonic oscillator, a fourth uncertainty estimate that corresponds to the fourth time estimate,
wherein the fourth time estimate and the fourth uncertainty estimate are included in the first time beacon.
16. The first node of claim 11 , wherein the processor is further configured to:
update a first clock included in the first node based on the third time estimate; and
update a second clock included in a second node based on the first time beacon.
17. The first node of claim 11 , wherein the processor is further configured to:
receive a fourth time estimate and a fourth uncertainty estimate corresponding to the fourth time estimate;
determine that the fourth uncertainty estimate meets a threshold criterion; and
in response, discard the fourth time estimate.
18. The first node of claim 11 , wherein the first node receives the first time estimate and the second time estimate from one or more battery-powered nodes that reside N hops away from a continuously powered device coupled to the wireless mesh network, wherein N comprises a positive integer, and wherein the second node resides N+2 hops away from the continuously powered device.
19. The first node of claim 18 , wherein the first node resides N+1 hops away from the continuously powered device and wherein the first node comprises a battery-powered node.
20. A non-transitory computer-readable medium storing program instructions that, when executed by a processor, cause the processor to communicate time estimates between nodes in a wireless mesh network, by performing the steps of:
receiving, from a first uplink node, a first time estimate and a first uncertainty estimate corresponding to the first time estimate;
receiving, from a second uplink node that is different than the first uplink node, a second time estimate and a second uncertainty estimate corresponding to the second time estimate;
combining the first time estimate and the second time estimate based on the first uncertainty estimate and the second uncertainty estimate to generate a third time estimate;
generating a first time beacon based on the third time estimate; and
transmitting the first time beacon from a first node in the wireless mesh network to a second node in the wireless mesh network.