Investigation of low-cost infrared sensing for intelligent deployment of occupant restraints

2018-11-13T12:52:51Z (GMT) by Amjad Juna
In automotive transport, airbags and seatbelts are effective at restraining the driver and passenger in the event of a crash, with statistics showing a dramatic reduction in the number of casualties from road crashes. However, statistics also show that a small number of these people have been injured or even killed from striking the airbag, and that the elderly and small children are especially at risk of airbag-related injury. This is the result of the fact that in-car restraint systems were designed for the average male at an average speed of 50 km/hr, and people outside these norms are at risk. Therefore one of the future safety goals of the car manufacturers is to deploy sensors that would gain more information about the driver or passenger of their cars in order to tailor the safety systems specifically for that person, and this is the goal of this project. This thesis describes a novel approach to occupant detection, position measurement and monitoring using a low-cost thermal imaging based system, which is a departure from traditional video camera-based systems, and at an affordable price. Experiments were carried out using a specially designed test rig and a car driving simulator with members of the public. Results have shown that the thermal imager can detect a human in a car cabin mock up and provide crucial real-time position data, which could be used to support intelligent restraint deployment. Other valuable information has been detected such as whether the driver is smoking, drinking a hot or cold drink, using a mobile phone, which can help to infer the level of driver attentiveness or engagement.