Browse Prior Art Database

A Method and System for Automated Signal Tuning Based on Learned Usage of Chronological Patterns

IP.com Disclosure Number: IPCOM000010521D
Original Publication Date: 2002-Dec-10
Included in the Prior Art Database: 2002-Dec-10
Document File: 2 page(s) / 44K

Publishing Venue

IBM

Abstract

This paper describes a new broadcast tuning mechanism whereby tuner logic "learns" or is conditioned to, a listener or viewer's tuning habits, and based on these habits, preemptively tunes to a specific frequency or station when power is applied. Secondly, it describes a means to allow the user to preset such time-based preferences, removing all ambiguity from the selection process

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 50% of the total text.

Page 1 of 2

  A Method and System for Automated Signal Tuning Based on Learned Usage of Chronological Patterns

Advantages can be obtained by the adding logic to an audio tuner to allow for correlation of signal / station usage with the time of day, with learning of the user's listening habits. It could then act on the learned data by tuning to a specific signal when the tuner is power-enabled, based on the user preference at that time of day. For example, a user may listen to a station between the hours of 7 and 8 AM. The same user at lunch may listen to the CD player in the automobile, and during the evening drive between 5 and 6 PM the user listens to a different station. Upon reaching the "learning threshold" the tuner would store the preferred value(s) and upon next power-on during the specified time period, act upon those stored values. It should also be noted that the user would have the option of setting default stations for any given time rather than have them be learned, and could also disable the feature altogether. The learning behavior would follow the methodology described below. Firstly, definitions may be outlined: - IntervalBoundary == That interval at the chronological boundary condition, wherein the system is in a purely "learning" mode, and does not attempt to preset any stations for the user. This interval may be either measured in real time, e.g., calendar days, or it may be measured in user listening data points gathered, regardless of the number of days involved (e.g., if a user goes on vacation after initiating the learning mode, and comes home after two weeks away, the IntervalBoundary may or may not be satisfied depending on this choice). - TimeslotSampling == The unit of time over which station selections are sampled and for which a preference is stated and a correlation determined. The most likely interval for this is that of a half hour, wherein a tuning preference may be judged from 7:00AM to 7:30AM, from 7:30AM to 8:00AM, and so forth. However, the principles here may be extrapolated out to any desired time granularity. - ConsistencyCorrelation == Percentage of time that a radio is tuned to a given station within a given timeslot. - ConsistencyThreshold == Percentage of time that a user listens to a given station during a given timeslot, after which that station may be considered a "preference." A preferential move from an "unestablished correlation" to an "established correlation" occurs as the ConsistencyCorrelation passes ConsistencyThreshold. - Correlation Point == A station-tuning which may be correlated to a given timeslot, whether stated generically as either an "unestablished correlation" or an "established correlation." - Established Correlation == A station-tuning which has crossed the ConsistencyThreshold and which is now deemed to be a va...