GAC: Energy-Efficient Hybrid GPS-Accelerometer-Compass GSM Localization

Moustafa Youssef, Mohamed Amir Yosef, Mohamed N. El-Derini

TL;DR

This paper introduces GAC, a hybrid localization system for Android phones that significantly reduces energy consumption by relying primarily on the accelerometer and compass, using GPS only for infrequent synchronization. This approach achieves exponential energy savings with only a linear loss in accuracy, making location-aware applications more feasible on battery-constrained mobile devices.

What Problem Does This Paper Address?

Location-aware applications are booming, but their widespread adoption on mobile devices is hampered by the significant battery drain caused by current localization technologies, particularly GPS. While GPS offers high accuracy, it can deplete a phone's battery in just a few hours. Existing alternatives like Wi-Fi or GSM-based localization often suffer from low accuracy or high energy consumption. This paper tackles the challenge of creating an energy-efficient localization system that leverages the increasingly common internal sensors (accelerometer and compass) found in modern smartphones, aiming to bridge the gap between accuracy and battery life.

Methods

The core of the GAC system is a hybrid approach that prioritizes low-power sensors. The authors first conducted an energy profiling study on an HTC Dream phone to understand the power consumption of GPS, accelerometer, and compass individually. They found that the accelerometer and compass, often integrated into a single chip, consume significantly less power than GPS.

GAC's strategy is to use the accelerometer and compass to estimate the phone's movement (direction and displacement) between GPS fixes. The compass provides the direction of motion, while the accelerometer's readings are integrated twice to calculate displacement. To mitigate the error accumulation inherent in inertial navigation systems, GPS is periodically used to re-synchronize the estimated position and velocity. The frequency of this GPS synchronization is a key parameter that balances energy savings and accuracy. The system also incorporates a mechanism to compensate for minor orientation changes due to road bumps or turns by using periods of constant speed to estimate the phone's orientation.

Datasets

The authors evaluated their system using data collected from an HTC Dream phone mounted on a car dashboard. The experiments were conducted in two distinct environments:

  • Intra-city driving: Representing a typical large city with slow traffic and numerous turns.
  • Highway driving: Considered an ideal environment for linear prediction-based localization.

No specific public datasets were mentioned; the data was collected by the authors for this study.

Key Insights and Ideas

  • Accelerometer/Compass as Primary Locators: The paper demonstrates that the accelerometer and compass, when used intelligently, can provide a viable and significantly more energy-efficient alternative to GPS for localization, especially when GPS is used sparingly.
  • Energy-Accuracy Tradeoff is Manageable: The GAC scheme shows that it's possible to achieve substantial energy savings (exponential) with only a modest, linear decrease in accuracy, making it a practical solution.
  • Normal Query Mode is Sufficient and Free: The "Normal" query mode for the accelerometer/compass (4Hz) is found to be sufficient for localization and, crucially, incurs no additional power cost because the phone's operating system already uses this mode continuously for features like screen auto-rotation.
  • Intra-city Performance Advantage: GAC shows a notable advantage over simpler linear prediction methods (like EnLoc) in intra-city driving, where the assumption of straight-line motion is frequently violated.

Key Results

  • Exponential Energy Savings: The GAC scheme achieves exponential energy savings compared to using GPS alone, with savings increasing significantly as the GPS synchronization period (TGsynch) increases.
  • Linear Accuracy Degradation: The accuracy loss is observed to be linear with respect to the increase in the GPS synchronization period.
  • Comparable to EnLoc on Highways: On highways, GAC's performance is comparable to the EnLoc scheme (a simple linear predictor).
  • Significant Intra-city Gain: Inside cities, GAC significantly outperforms EnLoc, especially for longer synchronization periods.
  • Power Consumption: GPS consumes about an order of magnitude more power than the combined accelerometer-compass. The GAC scheme, even without considering the "free" accelerometer usage, reduces power consumption considerably.

Limitations and Future Work

  • Sensor Noise and Accumulation: The paper acknowledges that noise in accelerometer and compass readings leads to error accumulation over time, which is why periodic GPS synchronization is necessary.
  • Assumption of Fixed Orientation: The initial model assumes the phone's orientation is fixed relative to the tracked object (e.g., car dashboard). While a compensation mechanism is introduced for minor changes, significant orientation shifts could still pose a challenge.
  • Future Work: The authors suggest extending GAC to indoor localization, incorporating Kalman filters for better error handling, dynamically adjusting the synchronization period based on error estimation, and further exploring the use of accelerometer data for orientation estimation during constant speed segments.

Summary updated

Abstract

Adding location to the available information enables a new category of applications. With the constrained battery on cell phones, energy-efficient localization becomes an important challenge. In this paper we introduce a low-energy calibration- free localization scheme based on the available internal sensors in many of today's phones. We start by energy profiling the different sensors that can be used for localization. Based on that, we propose GAC: a hybrid GPS/accelerometer/compass scheme that depends mainly on using the low-energy accelerometer and compass sensors and uses the GPS infrequently for synchronization. We implemented our system on Android-enabled cell phones and evaluated it in both highways and intra-city driving environments. Our results show that the proposed hybrid scheme has an exponential saving in energy, with a linear loss in accuracy compared to the GPS accuracy. We also evaluate the effect of the different parameters on the energy-accuracy tradeoff.
BibTeX
@inproceedings{youssef2010gac,
  author = {Moustafa Youssef and Mohamed Amir Yosef and Mohamed N. El-Derini},
  title = {{GAC: Energy-Efficient Hybrid GPS-Accelerometer-Compass GSM Localization}},
  booktitle = {GLOBECOM},
  year = {2010},
  url = {https://openreview.net/forum?id=Pepcf7QpR8},
}
Figure