Utilizing Gyroscope and Accelerometer for Self-Managed Crystal Methamphetamine Recovery and Private Monitoring

Abstract

Recovery from substance abuse is a continuous struggle as almost 60% of the addicts at any given instance in time during the recovery process experiences a relapse. The situation worsens during withdrawal, where the process, more so for those recovering from crystal methamphetamine addiction, often presents with intense symptoms that are really fatiguing. Mobile sensors offer a practical solution to continuously monitor these symptoms and provide near real-time insights into an individual's progress towards recovery. This paper explores two deep learning approaches, CNNs and LSTMs(Long Short Term Memory), to identify patterns related to withdrawal symptoms using accelerometer data. It will also address key limitations in current state-of-the-art methods, such as unbiased data collection, feature extraction, and mobile device implementation. Finally, the paper presents personalized reports designed to aid in therapy, helping individuals stay on track and reducing the likelihood of relapse.