Determining transportation mode on mobile phones

As mobile phones advance in functionality and capability, they are increasingly being used as instruments for personal monitoring. Applications are being developed that take advantage of the sensing capabilities of mobile phones - many have accelerometers, location capabilities, imagers, and microphones - to infer contextual information. We focus on one type of context, the transportation mode of an individual, with the goal of creating a convenient (no requirement to place sensors externally or have specific position/orientation settings) classification system that uses a mobile phone with a GPS receiver and an accelerometer sensor to determine if an individual is stationary, walking, running, biking, or in motorized transport. The target application for this transportation mode inference involves assessing the hazard exposure and environmental impact of an individual's travel patterns. Our prototype classification system consisting of a decision tree followed by a first-order hidden Markov model achieves the application requirement of having accuracy level greater than 90% when testing with our dataset consisting of twenty hours of data collected across six individuals.

[1]  Ling Bao,et al.  Activity Recognition from User-Annotated Acceleration Data , 2004, Pervasive.

[2]  Blake Hannaford,et al.  A Hybrid Discriminative/Generative Approach for Modeling Human Activities , 2005, IJCAI.

[3]  Andreas Krause,et al.  Trading off prediction accuracy and power consumption for context-aware wearable computing , 2005, Ninth IEEE International Symposium on Wearable Computers (ISWC'05).

[4]  Mark A. Hall,et al.  Correlation-based Feature Selection for Machine Learning , 2003 .

[5]  C. Randell,et al.  Context awareness by analysing accelerometer data , 2000, Digest of Papers. Fourth International Symposium on Wearable Computers.

[6]  Min Y. Mun,et al.  Parsimonious Mobility Classification using GSM and WiFi Traces , 2008 .

[7]  Scott E Crouter,et al.  Validity of 10 electronic pedometers for measuring steps, distance, and energy cost. , 2003, Medicine and science in sports and exercise.

[8]  Albrecht Schmidt,et al.  Multi-sensor Activity Context Detection for Wearable Computing , 2003, EUSAI.

[9]  Henry A. Kautz,et al.  Learning and inferring transportation routines , 2004, Artif. Intell..

[10]  I. Anderson,et al.  Practical Activity Recognition using GSM Data ∗ , .

[11]  Gaetano Borriello,et al.  A Practical Approach to Recognizing Physical Activities , 2006, Pervasive.

[12]  Gaetano Borriello,et al.  Mobile Context Inference Using Low-Cost Sensors , 2005, LoCA.

[13]  William G. Griswold,et al.  Mobility Detection Using Everyday GSM Traces , 2006, UbiComp.

[14]  Mohammad H. Rahimi,et al.  Seeing our signals: combining location traces and web-based models for personal discovery , 2008, HotMobile '08.