Collision avoidance support in roads with lateral and longitudinal maneuver prediction by fusing GPS/IMU and digital maps

Collision avoidance in roads can be addressed in several ways, being cooperative systems one of the most promising options. In cooperative collision avoidance support systems (CCASS) the vehicles which constitute a scene share by means of communication links information that can be useful to detect a potentially risky situation. Typically, this information describes the kinematic state of each vehicle and can be complemented with a prediction of its next state. Indeed, the timely prediction of the next maneuver of a vehicle results beneficial to estimate the risk factor of a scene. This article presents a solution to the problem of maneuver prediction which employs a reduced number of sensors: a Global Navigation Satellite System (GNSS) receiver, one gyro, one accelerometer and the odometry. Predictions are made by a bi-dimensional interactive multiple model (2D-IMM) filter in which longitudinal and lateral motions of the vehicle are distinguished and maneuvering states are described by different kinematic models. A number of experiments were carried out with two vehicle prototypes in several circuits. The results achieved prove the suitability of the proposed method for the problem under consideration.