Detection of inappropriate driving states based on dynamic control degradation

Ariadna Bartra Cisa

Santiago Marco Colás

A main cause of vehicle accidents is the driver's condition. Several studies estimate that drowsiness and similar states cause around 16-30% of all traffic accidents. The existing drowsiness detection methods can be divided into four main groups: subjective, physiological, visual and driving behavioral. The first and second methods are not feasible for non professional divers due to their low reliability and/or need of devices attached to the driver. The third method can be already found in vehicles, but has the drawback of depending highly on lighting conditions. Finally, the fourth methods, based on the driving behavior analysis, have the advantage of being transparent and non-invasive to the user and also not depending highly on external conditions. Most of these detectors are developed in simulators, without considering the real environment's complexity. This thesis presents a system for real vehicles to estimate the driver's state based on an evaluation of his/her driving quality and interaction with the vehicle. We assume that this quality deteriorates with fatigue and drowsiness. The system analyses data available on the vehicle's bus and optionally data provided by a lane detection frontal camera. To validate our initial hypothesis simulator tests have been done with sleep-deprived drivers. The obtained data were analyzed to obtain the main drowsy driving characteristics. Afterwards, different real vehicle tests were carried out, where a significant increase of the environment's complexity appeared. Therefore, our development had a focus on false alarms prevention. Besides the major modules that generate penalties (based on the vehicle's position, steering movements and abrupt corrections), we presented other modules focused on identifying specific situations and conditions that modulate dynamically the system's sensitivity. The results show the feasibility of detecting inattentive states using the vehicle's dynamic control. The system is optimized to work in a vehicle's microcontroller, fulfilling automotive's standards and constraints.

Drowsiness is one of the main causes of traffic accidents, especially on long motorway journeys. Last year, 97,756 accidents with victims occurred in Spain, in which 1,700 people lost their lives and 134,500 were injured. It is estimated that between 16% and 30% of all accidents are caused by drowsiness. These are accidents with a high mortality rate, causing a great social and economic impact that leads key figures in the automotive industry to work together with the aim of reducing the number of current accidents. The automotive industry is devoting a lot of resources and efforts to the development of advanced driver assistance systems (ADAS, for its acronym in English Advanced Driver Assistance Systems); which aim to automate, adapt or improve vehicle safety systems, alerting the driver to possible problems or even taking control of the car if necessary. It is a highly innovative field that is in full expansion so that the next generations of vehicles are safer and lead to a reduction in accidents. The system developed in the presented thesis allows the identification of states unsuitable for driving in real environments from measurements based on the dynamic control of the vehicle. It only needs signals already existing in the car, although the addition of a front camera (common for other ADAS systems) improves the quality of the system's input information. The algorithms developed comply with the strict regulations of automotive software (MISRA) and are optimized to be able to run on microcontrollers common in automotive (ARM Cortex M4). 2 The results obtained from the thesis have a clear commercial and industrial orientation. A product of high interest to the automotive industry has been developed, which is looking for systems that add value and safety to its vehicles with the lowest possible cost and additional devices. The fact that the developed system complies with automotive standards and is highly optimized in terms of space and computational cost required means that it can be easily integrated into any vehicle microcontroller. In addition, it does not require additional material, thus limiting the additional cost of adding this new feature. Another point of great value is the fact that the system, unlike other detectors that have been developed in a simulator, has been built with real driving data and validated in a real environment by different automotive brands. Therefore, the detector presented in the thesis is a product clearly focused on being commercialized and incorporated into vehicles, both private and commercial. The company Ficosa, owner of the system, is currently working on its commercialization with different automotive companies. They present and describe the system as well as the results obtained in different cases and, when there is interest from the potential client, it is integrated into one of its demonstration vehicles so that they can test it in the conditions they wish and carry out their own tests. Last year, Ficosa reached an agreement with a major automotive company for the purchase of the system. As a result of this sale, the drowsiness detector can be found in one of the models released this year by one of the largest car manufacturers in the world, thus actively contributing to the improvement of global road safety. Ficosa is currently in an advanced stage of negotiations with a second car manufacturer to incorporate the system into its vehicles. The results of the thesis have also motivated Ficosa to register a worldwide patent, which came into force last year. This allows the company to obtain benefits not only from the sale of the product created but also from the knowledge and methods generated during its development, since the patent recognizes and ensures the exclusive commercial exploitation of the invention.

Vehicle Accidents; Driver's Condition; Drowsiness; Traffic Accidents; Drowsiness Detection Methods; Subjective Methods; Physiological Methods; Visual Methods; Driving Behavioral Methods; Non-Professional Drivers; Low Reliability; Attached Devices; Lighting Conditions; Driving Behavior Analysis; Transparent; Non-Invasive; External Conditions; Simulators; Real Environment Complexity; Real Vehicles; Driver's State Estimation; Driving Quality; Vehicle Interaction; Fatigue; Vehicle's Bus Data; Lane Detection Frontal Camera; Initial Hypothesis; Simulator Tests; Sleep-Deprived Drivers; Drowsy Driving Characteristics; Real Vehicle Tests; Environment's Complexity; False Alarms Prevention; Penalty Modules; Vehicle's Position; Steering Movements; Abrupt Corrections; Specific Situations; Modulating System's Sensitivity; Inattentive States; Vehicle's Dynamic Control; Vehicle's Microcontroller; Automotive Standards; Automotive Constraints