Stress recognition supported by smartphones

A low-cost mobile electrocardiographic system for stress recognition

Constance, July 2014. According to the World Health Organisation (WHO), stress is recognized as a predominant disease with raising costs for treatment and rehabilitation. Each person perceives the stress in a subjective way; therefore, it is important to support an objective way detecting stress in mobile environments. While some people may have clear indications to recognize stress, others will not notice when they pass the threshold from being ‘busy’ to be ‘under stress’. Besides the traditional analysis of stress via the salivary cortisol level test, a growing availability of sensors enabled an alternative approach for non-invasive measurement of biological data relevant for stress detection. 

 IMG 9167

Wilhelm Daniel Scherz proposes a low-cost mobile electrocardiographic system for stress recognition. He developed the system in his Master Thesis at HTWG Konstanz that he presented on July 31, 2014.

The system collects the electronic characteristics of the heart and analyses the extracted characteristics. It is divided into four different hardware components. The first step made by the system is to extract the ECG signal from the user with electrodes, converting low electrical impulses into analogue voltage. The amplified signal will variegate according to the ECG signal. Then the signal is quantitated and the time needed for the operations is measured. After the data is calculated and the interval is measured, the data will be sent to the end device where the QRS data is extracted and the heartbeats are correlated. The correlated heartbeats represent the stress level. The stress level will be shown to the user as a visual feedback and later it can be used for informing and making recommendations to the user about his current stress status.

 Stress system architecture

System architecture

One of the biggest challenges in this work is the extraction and interpretation of the biological data and signals. Therefore, several specific algorithms have been developed like for the extraction of the heart rate from the ECG signal or the calculation of stress using these heart rates. For the prototype, a lightweight and resource friendly solution, capable of working on a small microcontroller have been set as a strong boundary condition.

The stress detection system was embedded into two parts: the first part is the hardware that is in charge of capturing the heartbeat and providing an analogue to digital conversion. The second part is the software, in charge of providing the analysis of the data (stress detection algorithm). Due to the fact that the system should be small, low-cost and light-weighted in terms of resources required, a small microcontroller should be used. The ECG data is extracted and pre-processed by the sensor shield. The Arduino Uno receives the data as analogue input and it forwards this to a PC or mobile device (smart phone) via serial communication like COM or with the help of a Bluetooth interface. The algorithm is implemented in software and executed only on the microcontroller. Currently, the data is logged via C# and analyzed with Matlab.

The presented work shows a possibility to detect stress with a reduced set of data detected by a light-weighted embedded system. The system uses a small and cheap microcontroller, enabling real-time stress detection. Different interfaces provide access for wired and wireless terminals like smartphones or desktop PCs. The algorithm developed is implemented with a small footprint and it is fitting into small prototyping boards. As an example, the system has been integrated into an Arduino Uno board, leaving space for further features, like data storage on a local smart card, more powerful Bluetooth bridges, WiFi-connection, GSM connectivity or embedded displays.

A more detailed discussion of the system and the results achieved will be provided during the 6th European Conference of the International Federation for Medical and Biological Engineering (MBEC 2014). The work of the Master Thesis has been accepted as full-paper after a peer-review process.


News Archive