关键词: 手势行为认知, 人机交互, 特装车辆, 指挥模拟训练, 动作特征, 运动控制

Abstract:

To overcome the practical shortcomings in the driving command training of special vehicles, a simulation training technology framework based on gesture behavior cognition and command interaction was proposed. A somatosensory interactive motion capture device and human skeleton model were used to conduct the real-time dynamic recognition and spatial coordinate transformation of the gesture actions of the command training personnel. The median filtering method was applied to eliminate the abrupt data and random noise. The spatial coordinates were processed consistently by using the human skeleton centralization and normalization method. Based on the command gesture action behavior model library and the spatial posture angle calculation of arm skeletons, the posture features of different command gesture actions were determined, and the gesture behavior cognition based on action features was realized. The intelligent vehicle motion command information and control parameters were formed through the association and mapping of gesture commands and sent to the motion control module through the wireless communication protocol. After parsing and processing, they were converted into voltage and current signals to drive the intelligent vehicle to move forward/backward and turn left/right, ultimately achieving real-time interactive control of its motion process by different command gestures. Test verification by multiple groups of command training personnel shows that the proposed technical means and method approaches have good versatility, consistency, and accuracy, which can provide an economical and efficient solution for the driving command training of special vehicles.

Key words: gesture behavior cognition, human-computer interaction, special vehicle, command simulation training, action feature, motion control