The Somno Alert ® system is a software system that analyses existing data in the CAN bus of the vehicle and data provided by the Lane Recognition System in order to identify inadequate driving states related to driving quality. The system studies the driver’s driving quality during a predetermined time period and generates an alarm that is related to the driving quality. The system is designed for driving on motorways and main roads. The system is always engaged after the ignition is switched on and works at speeds above 70 km/h.

General inputs to the algorithm:

  • Vehicle speed

  • Blinkers

  • Yaw rate (steering wheel)

  • Fuel and brakes

  • Lane information


  • The thoracic effort signal is extracted from the video signal in real time using complex image processing algorithms.

  • Variability quantification algorithms analyse the thoracic effort signal extracted to identify drowsy or inattentive states while driving.

  • Basal thoracic effort signal memory when the driver is awake to analyse their degradation over time.

  • Contactless and car-independent

  • Can prevent accidents.

The system was developed in collaboration between Ficosa International S.A. and IBEC. My duties included the development of the algorithm that analyzed the real-time signals and predicted the drowsiness index, the integration of the algorithm in a Cortex M4 microcontroller, the algorithmic optimization to meet strict memory and CPU constraints, and the ultimate validation of the algorithm in a real vehicle.