Preconditions for a Reliable & Robust Detection of Wrong-way Driving on Highways with GNSS and Autonomous Sensors

Abstract

Wrong-way drivers often cause serious accidents on highways. To prevent vehicle drivers to enter the highway from its exit, a GNSS-based navigation system is employed. It consists of a GNSS receiver, autonomous sensors (based on odometry and inertial navigation system) from the vehicle and a digital road map. A Kalman-filter is used to couple the GNSS position with the information provided by the autonomous sensors. The resulting GNSS-based position is projected onto the digital road map. This way, the GNSS-based position can be assigned to the corresponding road segment and the vehicle’s driving direction can be compared with the allowed driving direction given by the digital road map. If the vehicle goes in opposite direction than stipulated by the digital road map, wrong-way driving is detected. The focus of this work is to guarantee a reliable & robust detection of wrong-way driving. This particularly means that the probabilities of false alarm (when a driver is falsely informed to be a wrong-way driver) and missed detection (when wrong-way driving is not detected) are known and above all, can be regulated. To this end, statistical hypothesis tests are proposed that form the basis of the wrong-way detection algorithm. The detection algorithm is verified by several case studies. For that purpose, different GNSS technologies are taken into account. This way, preconditions can be formulated for enabling a reliable & robust detection of wrong-way driving. These preconditions cover accuracy, availability and continuity requirements on the GNSS-based navigation system and its single components. Furthermore, the preconditions reveal the key parameters that are crucial for ensuring a reliable & robust detection of wrong-way driving now and in future.