15
TRADING OFF PREDICTION ACCURACY AND POWER CONSUMPTION FOR CONTEXT-AWARE WEARABLE COMPUTING Presented By: Jeff Khoshgozaran

TRADING OFF PREDICTION ACCURACY AND POWER CONSUMPTION FOR CONTEXT- AWARE WEARABLE COMPUTING Presented By: Jeff Khoshgozaran

  • View
    217

  • Download
    0

Embed Size (px)

Citation preview

Page 1: TRADING OFF PREDICTION ACCURACY AND POWER CONSUMPTION FOR CONTEXT- AWARE WEARABLE COMPUTING Presented By: Jeff Khoshgozaran

TRADING OFF PREDICTION ACCURACY AND POWER CONSUMPTION FOR CONTEXT-AWARE WEARABLE COMPUTING Presented By: Jeff Khoshgozaran

Page 2: TRADING OFF PREDICTION ACCURACY AND POWER CONSUMPTION FOR CONTEXT- AWARE WEARABLE COMPUTING Presented By: Jeff Khoshgozaran

Motivation

Wearable devices sensing user information Context-Aware Mobile Computing

Previous work Power consumption in full power mode

Quickly depletes a critically constrained resource High sampling rate to provide accuracy

Computational and space-intensive solutions Lack of scalability for knee and hip-worn sensors

Page 3: TRADING OFF PREDICTION ACCURACY AND POWER CONSUMPTION FOR CONTEXT- AWARE WEARABLE COMPUTING Presented By: Jeff Khoshgozaran

eWATCH

A context-aware wearable platform Several sensors including two-axis accelerometer Three power states for sensing and classifying data

Full Power Active CPU, active peripherals (~ 30 ms)

Idle State Core clock turned off, active peripherals Waiting for the next sample For SR=6 Hz, time interval~166ms

Low Power State is active most of the time Inactive CPU and peripherals Active real-time clock Scheduling next wake up Using selective sample algorithm

www.chronicle.pitt.edu, http://www.cmu.edu/

Page 4: TRADING OFF PREDICTION ACCURACY AND POWER CONSUMPTION FOR CONTEXT- AWARE WEARABLE COMPUTING Presented By: Jeff Khoshgozaran

Time/Frequency-Domain-Based Classification

5 second windows for computing features Time-based

Using means, variances, median, etc. Frequency-based

Using FFT on values of both accelerometer axes separately

http://www.seas.gwu.edu/~ayoussef/cs225/

Human movement is periodic

Frequency-based Approaches good for classifying accelerometer data

Less expressive for very low sampling rates

Page 5: TRADING OFF PREDICTION ACCURACY AND POWER CONSUMPTION FOR CONTEXT- AWARE WEARABLE COMPUTING Presented By: Jeff Khoshgozaran

Battery Lifetime Vs. Sampling Rate

8.17.6

More costly computation and more dimensions

While important, computation is not the dominant factor in reducing energy consumption

Page 6: TRADING OFF PREDICTION ACCURACY AND POWER CONSUMPTION FOR CONTEXT- AWARE WEARABLE COMPUTING Presented By: Jeff Khoshgozaran

Using SVM for Classification

A multi-class SVM used for actual classification Detect and exploit complex patterns in data Good for representing complex patterns Good for excluding unstable patterns (= overfitting) Computationally expensive training Very efficient classification (hardware friendly) Guassian Radial Basis Function used as kernel to

classify non-linear data The class of kernel methods implicitly defines the class of

possible patterns by introducing a notion of similarity between data

Implicit and non-linear embedding of data in high-dimensional spaces

Separated by a hyperplane in feature space

Page 7: TRADING OFF PREDICTION ACCURACY AND POWER CONSUMPTION FOR CONTEXT- AWARE WEARABLE COMPUTING Presented By: Jeff Khoshgozaran

Power-optimized Classification Experiments

Training data captured by 3 test participants

Each activity recorded for 10 minutes Data was split into different recorded activities Data was partitioned into blocks of 5 seconds Used to extract time/frequency domain features Labeled examples used for training multi class

SVM Prediction accuracy & power consumption

computed

Running/J oggingWalkingStandingSitting/WorkingClimbing/Descending Stairs

Page 8: TRADING OFF PREDICTION ACCURACY AND POWER CONSUMPTION FOR CONTEXT- AWARE WEARABLE COMPUTING Presented By: Jeff Khoshgozaran

Results

For all but extremely low frequency ranges, frequency based features perform superiorly.

Optimum sampling frequency of 6 Hz85% Increase

Page 9: TRADING OFF PREDICTION ACCURACY AND POWER CONSUMPTION FOR CONTEXT- AWARE WEARABLE COMPUTING Presented By: Jeff Khoshgozaran

Selective Sampling vs. Prediction Accuracy

What: Further reduce energy consumption How: Selective Sampling Why: Human activity: a continuous process

Person more likely to continue an activity than to change to another at a point in time

Selective sampling schedules classification Reduces number of observations Saving energy from continuous monitoring to few points in

time Objective: keeping accuracy as high as possible

At is the user’s activity at time t

Page 10: TRADING OFF PREDICTION ACCURACY AND POWER CONSUMPTION FOR CONTEXT- AWARE WEARABLE COMPUTING Presented By: Jeff Khoshgozaran

Selective Sampling (cont.)

Select a set of observation times to maximize correct prediction of user’s activity for times when no sampling/classification is made

Minimize the expected loss:

Conditional plans

Maximum # of observations

Expected loss over all activity sequences a

Selected observation times for a

Minimize uncertainty

: Sequence of decisions: depending on observations so far, decides when next observation should be made

Entropy and dynamic programming used to find optimal

Page 11: TRADING OFF PREDICTION ACCURACY AND POWER CONSUMPTION FOR CONTEXT- AWARE WEARABLE COMPUTING Presented By: Jeff Khoshgozaran

4 Schemes to Select the Conditional Plan

Uniform Spacing Selects observation times at equally spaced intervals

Random Spacing B random length observation times selected at random

Exponential Backoff Maintains a maximum step size ∆max

If cur. act=last detected act., multiply ∆max by α else ∆max=1

Actual step size ∆ chosen uniformly at random from [1, ∆max]

Next observation made at t+ ∆ Entropy-based

Minimizing uncertainty using the entropy criterion Taking transition probabilities of states into account

More frequent sampling for activities with short durations

Page 12: TRADING OFF PREDICTION ACCURACY AND POWER CONSUMPTION FOR CONTEXT- AWARE WEARABLE COMPUTING Presented By: Jeff Khoshgozaran

Selective Sampling Experiments

Four new objects performing hour-long activities Subjects were indirectly asked to perform

representative tasks at random times User activities manually annotated by an

observer Resulting in <activity,duration> pairs sampled at

6Hz Data then partitioned into sequences of 5

seconds These blocks labeled with annotations and

classified using pre-trained classifiers in the frequency domain

Page 13: TRADING OFF PREDICTION ACCURACY AND POWER CONSUMPTION FOR CONTEXT- AWARE WEARABLE COMPUTING Presented By: Jeff Khoshgozaran

Results Continuous Sampling

Competitive for low

frequencies

Factor of 2 improvement

Using annotated data as “exact classification”No SVM (focusing on sampling)

Using classifier output instead of annotationsError=Sampling + Classification

Overall error dominated by classification from SVM and not by sampling

Classification accuracy lower than previous experiments due to 1.new subjects 2. noisy real-world environment

Roughly similar behavior to above experiment

@ 6Hz factor of 2.5 improvement

Page 14: TRADING OFF PREDICTION ACCURACY AND POWER CONSUMPTION FOR CONTEXT- AWARE WEARABLE COMPUTING Presented By: Jeff Khoshgozaran

Conclusion

High efficiency and accuracy for low range frequency of 1-10 Hz.

Competitive classification accuracy for the highly erratic and ambiguous (but convenient) wrist-based sensing

Four selective sampling strategies to further reduce the resource usage

Page 15: TRADING OFF PREDICTION ACCURACY AND POWER CONSUMPTION FOR CONTEXT- AWARE WEARABLE COMPUTING Presented By: Jeff Khoshgozaran

Comments

Using FFT for each dimension separately looses the correlation of among dimensions

Semi-controlled user behavior for test data generation

Authors assume continuous state change in a close set of predefined activities i.e., at any given time, one of these activities are taking place