OKCARS : Oklahoma Collision Analysis and Response System.

2012PDFTech ReportOTCREOS9.1-15-FDTRT06-G-0016Intelligent transportation systemsSensorsAutomatic crash notificationReal time data processingIntersectionsEmergency response timeTraffic safetyTraffic surveillanceWarning systemsAlgorithmsResearch HubTraffic safetyWarning systemsSensorsOklahomaUnited StatesOklahoma Transportation CenterCheng, QiChandler, DamonSheng, WeihuaOklahoma State University. School of Electrical and Computer EngineeringUniversity Transportation Centers Program (U.S.)NTL-INTELLIGENT TRANSPORTATION SYSTEMS-Collision Notification Systems (Vehicles)NTL-INTELLIGENT TRANSPORTATION SYSTEMS-Crash Prevention and SafetyNTL-SAFETY AND SECURITY-AccidentsNTL-SAFETY AND SECURITY-Highway SafetyUS Transportation CollectionBy continuously monitoring traffic intersections to automatically detect that a collision or nearcollision has occurred, automatically call for assistance, and automatically forewarn oncoming traffic, our OKCARS has the capability to effectively reduce emergency response time, and in turn potentially save thousands of lives and millions of dollars each year. We have designed and developed an affordable hardware platform consisting of four smart audio visual (SAV) nodes and a cellular modem. For networking of multiple nodes, we have also developed a software platform. To meet the critical and challenging system requirements, we have developed a near realtime vehicle detection and tracking algorithm requiring modest computing power. As an alternative and complement detection system, we have developed modules for efficient collision sound recognition and localization. We have shown that fusion of data from multiple microphone arrays and/or fusion of results from audio-video subsystems can significantly improve detection accuracy. We have developed a small-scale testbed for validating and verifying OKCARS and associated algorithms. OKCARS is non-intrusive, does not require specialized in-car equipment, operates using existing 3G communication technologies, and is relatively low-cost. It is a significant improvement from traffic monitoring systems currently available, where a human analyst has to make decisions by constantly monitoring several video stream inputs. Through improvement of service monitoring and emergency response preparedness, OKCARS has the potential to enhance roadway traffic safety and security