About AlcoGait & AlcoWear

Excessive alcohol use is the third leading lifestyle related cause of death in the United States. Smartphone sensing offers an opportunity to passively track alcohol usage and record associated drinking contexts. Drinkers can reflect on their drinking logs, detect patterns of abuse and self-correct or seek treatment.

AlcoGait: is a smartphone sensing app that passively detects a smartphone user's level of intoxication (how drunk) from their walk pattern (gait). AlcoGait extracts accelerometer features (time, frequency, statistical, wavelet and information theoretic domain) and gyroscope postural sway features (how much swaying) and classifies them using a Machine Learning approach.

AlcoWear: While gait sensor readings taken from a device attached to the user’s trunk (smartphone) are the most accurate, users often do not carry their phones (e.g. leave them on a table) while walking around during their day. Smartwatches are worn continuously but are less accurate due to noisier sensor readings (e.g. confounding hand gestures). AlcoWear extracts and classifies accelerometer and gyroscope features extracted from a smartwatch, and classifies them using a Machine Learning approach.

Faculty

• Professor Emmanuel Agu

Current Students

• Chai Nimkar (Masters student)
• Adonay Rezom (Undergraduate)
• Sam Huang (Undergraduate)
• Nicholas Cheung (Undergraduate)
• Joe Bremner (Undergraduate)
• Quoc Ho Lam (Undergraduate)

Past Students (Alumni)

• Vishal Rathi (Software developer at FactSet (First job))
• Anthony Topper
• Andrew McAfee
• Jacob Watson
• Ben Bianchi (Software Engineer, Wayfair)
• Rupak Lamsal
• Jules Voltaire (Software engineer, Constant Contact)
• Matthew Nguyen
• Linh Hoang
• Obatola Seward-Evans
• Muxi Qi
• Christina Aiello (Software engineer at VistaPrint/Cimpress (First job))
• Zach Arnold (MS CS Student Georgia Tech)
• Danielle LaRose (Senior application developer, Optum)

Medical Collaborators

• Dr Michael Stein
• Dr Ana Abranted

Press

Summary: AlcoGait has received over 120 media articles including Boston Globe, 2 NPR radio articles, BBC Radio, Worcester Telegram and Gazette, Boston TV channels 2,3,4 and 5 and 7

• [ Wired Magazine article ]
• [ Boston Globe Article: WPI develops an app to make drunk drivers toe the line]
• [ Worcester Telegram and Gazette Article: Walking the line: WPI students develop app as sobriety test]
• [ Communications of the ACM News Article ]

TV appearances:

• [ WCVB Boston Channel 5 News Article]
• [ WBZ Channel 4 Boston News article ]
• [ WGBH News article ]

Awards

• Honorable Mention, Best CS MQP in 2017-18
  Intoxigait Deep Learning
  Joseph Bremner, Nicholas Cheung, Quoc Ho Lam, Sam Huang
  
  
• Winner MS Level WPI Graduate Research Innovation Exchange (GRIE) Poster Competition, Arts and Science 2016
  Christina Aiello
  
  
• AProvost’s MQP award for the best MQP in the WPI CS department for the 2014-2015 academic session
  Smartphone Gait Inference MQP
  Zachary Arnold and Danielle Larose
  
  

Publications

Papers

• AlcoWear: Detecting Blood Alcohol Levels from Wearables,
  Andrew McAfee, Jacob Watson, Ben Bianchi, Christina Aiello, Emmanuel Agu,
  in Proceedings IEEE Conference on Ubiquitous Intelligence and Computing (UIC) 2017
  
  
• Investigating Postural Sway Features, Normalization and Personlization in Detecting Blood Alcohol Levels of Smartphone Users
  Christina Aiello and Emmanuel Agu,
  in Proc Wireless Health Conference 2016.
  
  
• A Factorial Experiment to Investigate Naturalistic Factors Affecting Smartphone Gait Analysis,
  Zach Arnold, Danielle LaRose, Emmanuel Agu,
  in Proc 17th International Conference on e-Health Networking, Applications and Services (Healthcom) 2015 (short paper)
  
  
• Smartphone Inference of Alcohol Consumption Levels from Gait
  Zachary Arnold, Danielle LaRose and Emmanuel Agu,
  in Proc. IEEE Conference on Healthcare Informatics 2015, Dallas, Texas
  
  

Masters Theses

• Investigating Gyroscope Sway Features, Normalization, and Personalization in Detecting Intoxication in Smartphone Users,
  Christina Aiello
  WPI Comp Science Dept., Advisor. Completed April 2016.
  
  
• A Comprehensive Comparative Performance Evaluation of Signal Processing Features for Detecting Intoxication using Smartphones,
  Muxi Qi,
  WPI Electrical & Computer Engr Dept., Advisor. Completed April 2016
  
  

Major Qualifying Projects (MQPs)

• Behavior Change Contextualizer
  Rupak Lamsal, Jules Voltaire and Matthew Nguyen,
  WPI CS Dept., Advisor. Completed March 2017
  
  
• AlcoWatch Intoxication Detection,
  Jacob Watson, Andrew McAfee, Benjamin Bianchi,
  WPI CS Dept., Advisor. Completed March 2017 (Top 5 out of 46 for Provost’s Award for best MQP in Computer Science Dept.)
  
  
• Smartphone Gait Inference
  Zachary Arnold, Danielle LaRose,
  WPI CS Dept., Advisor. Completed March 2015 (co-advised with Joseph Petrucelli, WPI Math Dept.) (Provost’s Award for best MQP in Computer Science Dept., 2014-15)
  
  

Funding

• NIH National Institute on Alcoholism R21 grant 1R21AA025193-01 (PA-14-188) Machine Learning Approach for Inferring Alcohol Intoxication Levels from Gait Data,
  PIs: Emmanuel Agu, Michael Stein, co-PI: Ana Abrantes
  Award Amount: $408,578, WPI sub-contract: $115,469. Award Dates 8/1/17-7/31/19
  
  

---

mailto:webmaster@cs.wpi.edu