Upsampling sensors to auto-detect a fitness activity
First Claim
1. A method for conserving power while reducing latency in collecting sufficient information for identifying particular fitness activities being performed by users of computing devices, the method comprising:
- determining, by a computing device while operating in a normal operational state of the computing device and based at least in part on applying a first Hidden Markov Model transformation to a sampling of a first set of a plurality of sensors of the computing device taken at a first rate that is a slowest rate from among three sampling rates so as to conserve power, whether a user of the computing device has transitioned from performing a non-fitness activity to initiating any fitness activity, wherein;
the computing device stores pre-defined identifiers of non-fitness and fitness activities in a set of pre-defined indications of activities;
determining whether the user of the computing device has transitioned from performing the non-fitness activity to initiating any fitness activity comprises identifying initial data from the first Hidden Markov Model transformation that precedes subsequent data from the first Hidden Markov Model transformation;
the initial data corresponds to one or more of the pre-defined identifiers of non-fitness activities in the set of predefined indications of activities; and
the subsequent data corresponds to one or more of the pre-defined identifiers of fitness activities in the set of predefined indications of activities;
responsive to determining that the user has transitioned from performing the non-fitness activity to initiating any fitness activity and until a probability that the user is engaged in a particular fitness activity satisfies a threshold, sampling, by the computing device while operating in an upsampling operational state of the computing device that uses more power than the normal operational state of the computing device, at a second rate that is a fastest rate from among the three sampling rates so as to more quickly obtain sufficient sensor data for identifying the particular fitness activities, a second set of the plurality of sensors to collect the sufficient sensor data to determine the probability that the user is engaged in a particular fitness activity, wherein the second set of the plurality of sensors includes a greater quantity of sensors than the first set of the plurality of sensors, wherein the probability that the user is engaged in the particular fitness activity is determined in part by applying a second Hidden Markov Model transformation to the sufficient sensor data;
responsive to determining that the probability satisfies the threshold, collecting, by the computing device in an active operational state of the computing device that uses less power than the upsampling operational state and more power than the normal operating state, at a third rate that is greater than the first rate and less than the second rate so as to efficiently collect fitness information, additional sensor data as fitness information associated with the particular fitness activity, the additional sensor data being collected from a particular set of the plurality of sensors that corresponds to a pre-defined identifier for the particular fitness activity in the set of pre-defined indications of activities; and
outputting, via a user interface of the computing device, the fitness information associated with the particular activity that was collected during the active operational state.
2 Assignments
0 Petitions
Accused Products
Abstract
In some examples, a method includes detecting, based at least in part on sampling a first set (122) of a plurality of sensors of a computing device (110) at a first rate, an indication that the user has initiated a fitness activity (112B), wherein the computing device (110) stores pre-defined identifiers of non-fitness and fitness activities in a set of pre-defined indications of activities; responsive to detecting the indication that the user has initiated the fitness activity, sampling, at a second rate that is greater than the first rate, a second set (120) of the plurality of sensors to determine a probability that the user is engaged in the fitness activity (112B); and responsive to determining that the probability satisfies a threshold, collecting, sensor data for the fitness activity using a particular set of the plurality of sensors that corresponds to a pre-defined identifier for the fitness activity (112B).
-
Citations
20 Claims
-
1. A method for conserving power while reducing latency in collecting sufficient information for identifying particular fitness activities being performed by users of computing devices, the method comprising:
-
determining, by a computing device while operating in a normal operational state of the computing device and based at least in part on applying a first Hidden Markov Model transformation to a sampling of a first set of a plurality of sensors of the computing device taken at a first rate that is a slowest rate from among three sampling rates so as to conserve power, whether a user of the computing device has transitioned from performing a non-fitness activity to initiating any fitness activity, wherein; the computing device stores pre-defined identifiers of non-fitness and fitness activities in a set of pre-defined indications of activities; determining whether the user of the computing device has transitioned from performing the non-fitness activity to initiating any fitness activity comprises identifying initial data from the first Hidden Markov Model transformation that precedes subsequent data from the first Hidden Markov Model transformation; the initial data corresponds to one or more of the pre-defined identifiers of non-fitness activities in the set of predefined indications of activities; and the subsequent data corresponds to one or more of the pre-defined identifiers of fitness activities in the set of predefined indications of activities; responsive to determining that the user has transitioned from performing the non-fitness activity to initiating any fitness activity and until a probability that the user is engaged in a particular fitness activity satisfies a threshold, sampling, by the computing device while operating in an upsampling operational state of the computing device that uses more power than the normal operational state of the computing device, at a second rate that is a fastest rate from among the three sampling rates so as to more quickly obtain sufficient sensor data for identifying the particular fitness activities, a second set of the plurality of sensors to collect the sufficient sensor data to determine the probability that the user is engaged in a particular fitness activity, wherein the second set of the plurality of sensors includes a greater quantity of sensors than the first set of the plurality of sensors, wherein the probability that the user is engaged in the particular fitness activity is determined in part by applying a second Hidden Markov Model transformation to the sufficient sensor data; responsive to determining that the probability satisfies the threshold, collecting, by the computing device in an active operational state of the computing device that uses less power than the upsampling operational state and more power than the normal operating state, at a third rate that is greater than the first rate and less than the second rate so as to efficiently collect fitness information, additional sensor data as fitness information associated with the particular fitness activity, the additional sensor data being collected from a particular set of the plurality of sensors that corresponds to a pre-defined identifier for the particular fitness activity in the set of pre-defined indications of activities; and outputting, via a user interface of the computing device, the fitness information associated with the particular activity that was collected during the active operational state. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16)
-
-
17. A non-transitory computer-readable storage medium encoded with instructions that, when executed, cause at least one processor to:
-
determine, while a computing device is operating in a normal operational state of the computing device and based at least in part on applying a first Hidden Markov Model transformation to a sampling of a first set of a plurality of sensors of the computing device taken at a first rate that is a slowest rate from among three sampling rates so as to conserve power, whether a user of the computing device has transitioned from performing a non-fitness activity to initiating any fitness activity, wherein; the computing device stores pre-defined identifiers of non-fitness and fitness activities in a set of pre-defined indications of activities; the at least one processor determines whether the user of the computing device has transitioned from performing the non-fitness activity to initiating any fitness activity comprises identifying initial data from the first Hidden Markov Model transformation that precedes subsequent data from the first Hidden Markov Model transformation; the initial data corresponds to one or more of the pre-defined identifiers of non-fitness activities in the set of predefined indications of activities; and the subsequent data corresponds to one or more of the pre-defined identifiers of fitness activities in the set of predefined indications of activities; responsive to determining that the user has transitioned from performing the non-fitness activity to initiating any fitness activity and until a probability that the user is engaged in a particular fitness activity satisfies a threshold, sample, while the computing device is operating in an upsampling operational state of the computing device that uses more power than the normal operational state of the computing device, at a second rate that is a fastest rate from among the three sampling rates so as to more quickly obtain sufficient sensor data for identifying particular fitness activities, a second set of the plurality of sensors collect the sufficient sensor data to determine the probability that the user is engaged in a particular fitness activity, wherein the second set of the plurality of sensors includes a greater quantity of sensors than the first set of the plurality of sensors, wherein the probability that the user is engaged in the particular fitness activity is determined in part by applying a second Hidden Markov Model transformation to the sufficient sensor data; responsive to determining that the probability satisfies the threshold, collect, while the computing device is operating in an active operational state of the computing device that uses less power than the upsampling operational state and more power than the normal operating state, at a third rate that is greater than the first rate and less than the second rate so as to efficiently collect fitness information, additional sensor data as fitness information associated with the particular fitness activity, the additional sensor data being collected from a particular set of the plurality of sensors that corresponds to a pre-defined identifier for the fitness activity in the set of pre-defined indications of activities; and output, via a user interface of the computing device, the fitness information associated with the particular activity that was collected during the active operational state. - View Dependent Claims (18)
-
-
19. A computing device comprising:
-
one or more computer processors; a plurality of sensors operably coupled to the one or more computer processors; and a memory comprising instructions that when executed by the one or more computer processors cause the one or more computer processors to; determine, while the computing device is operating in a normal operational state of the computing device and based at least in part on applying a first Hidden Markov Model transformation to a sampling of a first set of a plurality of sensors of the computing device taken at a first rate that is a slowest rate from among three sampling rates so as to conserve power, whether a user of the computing device has transitioned from performing a non-fitness activity to initiating any fitness activity, wherein; the computing device stores pre-defined identifiers of non-fitness and fitness activities in a set of pre-defined indications of activities; the one or more processors determine whether the user of the computing device has transitioned from performing the non-fitness activity to initiating any fitness activity comprises identifying initial data from the first Hidden Markov Model transformation that precedes subsequent data from the first Hidden Markov Model transformation; the initial data corresponds to one or more of the pre-defined identifiers of non-fitness activities in the set of predefined indications of activities; and the subsequent data corresponds to one or more of the pre-defined identifiers of fitness activities in the set of predefined indications of activities; responsive to determining that the user has transitioned from performing the non-fitness activity to initiating any fitness activity and until a probability that the user is engaged in a particular fitness activity satisfies a threshold, sample, while the computing device is operating in an upsampling operational state of the computing device that uses more power than the normal operational state of the computing device, at a second rate that is a fastest rate from among the three sampling rates so as to more quickly obtain sufficient sensor data for identifying particular fitness activities, a second set of the plurality of sensors collect the sufficient sensor data to determine the probability that the user is engaged in a particular fitness activity, wherein the second set of the plurality of sensors includes a greater quantity of sensors than the first set of the plurality of sensors, wherein the probability that the user is engaged in the particular fitness activity is determined in part by applying a second Hidden Markov Model transformation to the sufficient sensor data; responsive to determining that the probability satisfies the threshold, collect, while the computing device is operating in an active operational state of the computing device that uses less power than the upsampling operational state and more power than the normal operating state, at a third rate that is greater than the first rate and less than the second rate so as to efficiently collect fitness information, additional sensor data as fitness information associated with the particular fitness activity, the additional sensor data being collected from a particular set of the plurality of sensors that corresponds to a pre-defined identifier for the fitness activity in the set of pre-defined indications of activities; and output, via a user interface of the computing device, the fitness information associated with the particular activity that was collected during the active operational state. - View Dependent Claims (20)
-
Specification