Aggressive driving is the main cause of traffic accidents all over the world. The aggressive steering is the second leading cause of traffic accidents, just behind speeding. To detect the aggressive steering and warn the drivers to cultivate a good driving manners, it is necessary to automatically recognize the different vehicle steering modes, then to decide if they are aggressive or not. We investigate the various patterns of turning, changing lanes and U-turn, and then design and implement a system termed MultiWave that utilizes the gyroscope of a smartphone to automatically detect the vehicle steering patterns. MultiWave divides driving behaviors into different categories by analyzing the data collected by gyroscope sensors. By analyzing the gyroscopic sensor data of turning, changing lanes and U- turn, MultiWave is flexible and robust with most of the urban situations. The classification accuracy of MultiWave can reach an average of 92%, 76.5%, and 87% for vehicle turning, changing lanes and U-turn, respectively.