Accelerometer-based Chinese Traffic Police Gesture Recognition System

作者：Yuan Tao^*; Wang Ben

来源：Chinese Journal of Electronics, 2010, 19(2): 270-274.

• MEMS accelerometer
• Embedded system
• Chinese traffic police gesture recognition
• Human computer interaction

When an automatic traffic light system is not working e.g. due to traffic jam, it is necessary for traffic police to control the traffic by gestures. In this paper, the design of a system to synchronize traffic lights with traffic police gestures is presented. The gesture data are extracted by fixing a 3-axis accelerometer on the back of each hand. Signals obtained from the accelerometers are transformed into arm motion and palm orientation. A hierarchical classifier is designed to recognize the gestures, in which fuzzy rules are used to enable vague gesture recognition. An embedded micro-processor system has been developed for real-time gesture recognition. Wireless modules are used for communication and a digital compass is used to identify the direction, to which a traffic policeman faces. The integrate system has been tested, and the feasibility of the system to synchronize traffic lights has been demonstrated.

• 出版日期2010-4
• 单位清华大学