Foreign object detection in tires by acoustic event detection

Nowadays cars are equipped with various different sensors, e.g., radar or ultrasonic sensors for distance measuring and parking assistance, cameras for light-control or lane departure alerting, traffic sign recognition etc. Acoustic sensors find their way into cars mainly for voice-control of in-car communication, entertainment and navigation systems. However, acoustical sensors are capable of recognizing many more car related events like sirens, lane-departure, road or tire-conditions, engine failures, broken or worn chassis parts, loose parts, broken tie rods, broken/cracked exhaust, worn clutch etc. In this paper, acoustically based detection of foreign objects in tires is investigated. This application scenario is important since objects in tires can cause punctures and, thus, lead to accidents in the worst case. The already existing pressure sensors for tires installed at the tires themselves can only detect an already existing critical air loss. However, microphones can detect the tire damage before a significant loss of air pressure is measurable. For this study, a database with acoustic sensors at different car positions has been recorded in real environments. Different algorithms to recognize objects in tires will be compared in this paper.