Method, device and system for energy management

• US 9,471,127 B2
• Filed: 03/16/2011
• Issued: 10/18/2016
• Est. Priority Date: 03/16/2011
• Status: Active Grant

• Current Assignee: Nokia Technologies Oy (Nokia Corporation)
• Original Assignee: Nokia Technologies Oy (Nokia Corporation)
• Inventors: Saarinen, Jukka
• Primary Examiner(s): Prifti, Aurel
• Application Number:
  US14/004,374
• Publication Number:
  US 20140359325A1
• Time in Patent Office: 2,043 Days
• Field of Search: 706/25, 706/12, 718/1, 700/17, 382/103, 382/181, 709/238, 709/224, 707/723, 704/7
• US Class Current: 1/1
• CPC Class Codes: G06F 1/3206 Monitoring of events, devices or parameters that trigger a change in power modality
• CPC Class Codes: G06F 1/3234 Power saving characterised by the action undertaken

Abstract

The present application relates to carrying out energy management based on context information. In the method, apparatus and system according to the application, a context vector (820) is used to select a operational state plurality of operational state vectors (825, 835). The plurality of operational state vectors are indicative of the state and/or energy consumption of different operational units in the corresponding context. The selected operational state vectors are then used together in adapting an operational state of the operational units according to the needs in the context (855). There is also a method and devices for forming a data structure (810) that can be used in the selection of the operational state vectors. The method can be used in the energy management of mobile devices, where the operational units are components of these devices and the context can include data from sensors. The method may use an associative memory such as a sparse distributed memory in storing operational state vectors.

Claims

1. A method comprising:

providing, by an apparatus, a data structure comprising context vectors and operational state vectors of a plurality of operational units in a system, the data structure having an association between said context vectors and said operational state vectors, and the operational units including hardware, software applications, or a combination thereof;

receiving, by the apparatus, a context vector of the system;

using, by the apparatus, said context vector and a distance measure to select a plurality of operational state vectors, the plurality of operational state vectors being indicative of operational states and respective power consumption states of the operational units, and the respective power consumption states including one or more power-performance ratios and one or more transition costs; and

modifying, by the apparatus, the selected operational state vectors using a teaching operational state vector associated with said context vector by increasing a value of an element corresponding to the active operational unit in the teaching operational state vector according to said distance measure.

2. A method according to claim 1, wherein said selected operational state vectors are further modified by decreasing a value of an element corresponding to an inactive operational unit in said teaching operational state vector, and

wherein the operational units include a plurality of user devices.

3. A method according to claim 1,

wherein the system is a user device, and the hardware includes one or more processors, one or more memory units, one or more acceleration units, one or more communication units, or a combination thereof.

4. A method according to claim 1, further comprising thresholding said teaching operational state vector;

and

modifying said selected operational state vectors using said thresholded teaching operational state vector,

wherein context vector is indicative of environmental variables including one or more temperatures, one or more time points, one or more locations, one or more altitudes, one or more acceleration speeds, one or more lighting conditions, one or more noise conditions, one or more humidity conditions, or a combination thereof.

5. A method according to claim 1, further comprising:

forming said context vector by combining data from at least two time instances.

6. A method comprising:

receiving, by an apparatus, a context vector;

inputting, by the apparatus, said context vector as a single memory address input to an associative memory;

selecting, by the apparatus, a plurality of operational state vectors stored in the associative memory using an address distance between said inputted context vector, the plurality of operational state vectors being indicative of one or more power-performance ratios and one or more transition costs of a plurality of operational units of a system;

modifying, by the apparatus, the selected operational state vectors using a teaching operational state vector associated with said context vector by increasing a value of an element corresponding to the active operational unit in the teaching operational state vector according to said distance measure; and

managing, by the apparatus, power consumption of the operational units with said modified operational state vectors.

7. A method according to claim 6, wherein said associative memory is a sparse distributed memory, the method further comprising

masking the context vector with a mask before inputting said context vector into said associative memory at an input address,

wherein a plurality of data vectors of the operational state vectors are selected from the associative memory when the distance meets a threshold value.

8. A method according to claim 7, further comprising

choosing said associative memory among a group of associative memories based on said received context vector, before inputting said context vector into said associative memory,

wherein the address distance is a Hamming distance or a dot-product based distance.

9. A method according to claim 7, wherein a possible address space of context addresses comprises more entries than a total number of context addresses of the operational state vectors, and the addresses are assigned to the operational state vectors using a function.

10. A method according to claim 9, wherein the function is a random function.

11. A method according to claim 9, wherein the function corresponds to a statistical distribution of context vectors.

12. A method according to claim 6, further comprising:

adapting operational states of the operational units in the system based on the modified operational state vectors, wherein the adapted operational states change the power consumption.

13. An apparatus comprising:

at least one processor, and

at least one memory including computer program code for one or more programs, the at least one memory and said computer program code configured to, with the at least one processor, cause the apparatus to perform at least the following;

receive a context vector;

input said context vector as a single memory address input to an associative memory;

select a plurality of operational state vectors stored in the associative memory using an address distance between said inputted context vector and the operational state vectors, the plurality of operational state vectors being indicative of one or more power-performance ratios and one or more transition costs of a plurality of operational units of a system;

modify the selected operational state vectors using a teaching operational state vector associated with said context vector by increasing a value of an element corresponding to the active operational unit in the teaching operational state vector according to said distance measure; and

manage power consumption of the operational units with said modified operational state vectors.

14. An apparatus according to claim 13, wherein the associative memory is a sparse distributed memory, and the apparatus is further caused to:

mask said context vector with a mask before inputting said context vector into said associative memory at an input address;

wherein a plurality of data vectors of the operational state vectors are selected from the associative memory when the distance measure meets a threshold value.

15. An apparatus according to claim 13, further comprising a processor for producing said context vector.

16. An apparatus according to claim 13 further comprising different sub-units of said operational units, and a processor for enabling and disabling or setting operational states of said sub-units for managing the power consumption.

17. A non-transitory computer readable storage medium carrying one or more sequences of one or more instructions which, when executed by one or more processors, cause an apparatus to at least perform the following steps:

providing a data structure comprising context vectors and operational state vectors of a plurality of operational units in a system, the data structure having an association between said context vectors and said operational state vectors, and the operational units including hardware, software applications, or a combination thereof;

receiving a context vector of the system;

using said context vector and a distance measure to select a plurality of operational state vectors, the plurality of operational state vectors being indicative of operational states and respective power consumption states of the operational units, and the respective power consumption states including one or more power-performance ratios and one or more transition costs; and

modifying the selected operational state vectors using a teaching operational state vector associated with said context vector by increasing a value of an element corresponding to the active operational unit in the teaching operational state vector according to said distance measure.