According to a new study, a smartphone sensor similar to the GPS system could assist in determining if someone is drunk after taking marijuana. Intoxication caused by the consumption of cannabis has been linked to slowed response time, effects on work or school performance which can lead to accidents and fatalities.

The study, conducted by the Rutgers Institute for Health, Health Care Policy, and Aging Research, assesses the practicality of using smartphone sensor data to detect cannabis intoxication was published in the Drug and Alcohol Dependence journal. Surprisingly, the study was found to be 90% accurate, thanks to a combination of time features and sensor data. This is crucial for the existing methods including such as blood, urine, and saliva testing, to detect cannabis intoxication according to experts.

Travel patterns from GPS data and movement data from accelerometers were shown to be the most important phone sensors for detecting self-reported cannabis intoxication out of all the sensors available. Besides this, the researchers also monitored the time of day and day of the week to detect intoxication in the daily life of a cannabis consumer.

The data of responses of young adults between the ages of 18 and 25, who consume marijuana at least twice a week was used for the study. For a total of 30 days, the researchers tracked their responses. At the time of data collection, the participants had to self-report their cannabis use. At least three times a day, the data pointers comprised start/stop time and a subjective cannabis intoxication rating from 0 to 10, with 10 being very high. The researchers also gathered the data from the phone sensors of the participants.

Accelerometers, for example, were used to identify any changes in the participants' walking patterns when they said they were high. Based on routine, the researchers discovered that keeping track of the time and day had 60% accuracy in diagnosing cannabis intoxication in users. In order to detect such intoxication, researchers used a combination of time and data acquired from smartphone sensors, which had a 90% accuracy rate.