中文摘要：

为减少汽车起重机在吊装作业过程中吊臂与作业区障碍物发生碰撞的事故,建立障碍物分类模型,采用吊臂臂头自主探测学习方法,准确采集作业区障碍物的外形与位置信息。建立起重机圆柱坐标系,作业区按回转角度分成不同扇区,扇区按距回转中心距离分成不同扇格,以扇格精确存贮作业区环境数据,利用扇格数据生成吊装安全作业三维虚拟墙。以吊臂最近的20个时刻位置信息,用加权线性回归模型预测下一时刻吊臂的动作与位置参数。针对吊臂的预测动作及与虚拟墙的不同距离,采取不干预、限速、微动与禁止等吊臂控制策略。实际作业环境下的功能测试表明,该系统能有效地防止吊臂与作业区障碍物发生碰撞。

英文摘要：

To reduce collision accident occurred between the boom of truck crane and obstacles in work area during lifting operations. Based on obstacle classification model, the shape and location data of obstacles are collected accurately by adopting self-exploration way for boom head of crane. By establishing cylinder coordinate system of crane, the work area is divided into many sectors according to crane slewing angle, and then the sectors are divided into many fan grids based on the distance to the rotation center. Since obstacle data are stored precisely in the fan grids, the three-dimensional virtual walls of safe lifting operations are built by using these fan grids＇ data. The next boom＇s movement and position are predicted by the weighted linear regression model by using last 20 boom position data. Based on boom＇s movement trend and different distances to the virtual wall, the different boom control strategies of non-intervention, deceleration, micro-movement and prohibition are utilized. The function testing of actual operating environment demonstrate that the system could effectively prevent the collisions between the boom and obstacles of work area.